ROCK1 mediates leukocyte recruitment and neointima formation following vascular injury.
ABSTRACT Although Rho-associated kinase (ROCK) activity has been implicated in cardiovascular diseases, the tissue- and isoform-specific roles of ROCKs in the vascular response to injury are not known. To address the role of ROCKs in this process, we generated haploinsufficient Rock1 (Rock1(+/-)) and Rock2 (Rock2(+/-)) mice and performed carotid artery ligations. Following this intervention, we found reduced neointima formation in Rock1(+/-) mice compared with that of WT or Rock2(+/-) mice. This correlated with decreased vascular smooth muscle cell proliferation and survival, decreased levels proinflammatory adhesion molecule expression, and reduced leukocyte infiltration. In addition, thioglycollate-induced peritoneal leukocyte recruitment and accumulation were substantially reduced in Rock1(+/-) mice compared with those of WT and Rock2(+/-) mice. To determine the role of leukocyte-derived ROCK1 in neointima formation, we performed reciprocal bone marrow transplantation (BMT) in WT and Rock1(+/-) mice. Rock1(+/-) to WT BMT led to reduced neointima formation and leukocyte infiltration following carotid ligation compared with those of WT to WT BMT. In contrast, WT to Rock1(+/-) BMT resulted in increased neointima formation. These findings indicate that ROCK1 in BM-derived cells mediates neointima formation following vascular injury and suggest that ROCK1 may represent a promising therapeutic target in vascular inflammatory diseases.
- [show abstract] [hide abstract]
ABSTRACT: Remodeling of blood vessels underlies the pathogenesis of major cardiovascular disorders, including atherosclerosis, restenosis, and hypertension. Because remodeling of arteries is highly dependent on degradation of the extracellular matrix, which enables cells to migrate and proliferate, there is intense interest in the regulation and the roles of matrix metalloproteinases (MMPs) and the plasminogen activator-plasmin (PA-P) systems in vessel remodeling. Factors that promote vessel remodeling have been shown to be important in upregulating the activities of both proteolytic systems and include chronic changes in hemodynamics, vessel injury, cytokines involved in inflammation, and elevations in reactive oxygen species. The two proteolytic systems utilize common transcription factors to activate their respective genes and are frequently coexpressed in remodeling and atherosclerotic arteries. In this review, we discuss the effects of activating the MMP and PA-P systems on processes involved in vascular remodeling, factors regulating their expression and activation, their roles in restenosis, and the development and progression of atherosclerosis, as well as the ability of currently available inhibitors to prevent unfavorable remodeling and atherosclerosis.Current Hypertension Reports 01/2004; 5(6):466-72. · 3.74 Impact Factor
Article: Inflammation in atherosclerosis.[show abstract] [hide abstract]
ABSTRACT: Abundant data link hypercholesterolaemia to atherogenesis. However, only recently have we appreciated that inflammatory mechanisms couple dyslipidaemia to atheroma formation. Leukocyte recruitment and expression of pro-inflammatory cytokines characterize early atherogenesis, and malfunction of inflammatory mediators mutes atheroma formation in mice. Moreover, inflammatory pathways promote thrombosis, a late and dreaded complication of atherosclerosis responsible for myocardial infarctions and most strokes. The new appreciation of the role of inflammation in atherosclerosis provides a mechanistic framework for understanding the clinical benefits of lipid-lowering therapies. Identifying the triggers for inflammation and unravelling the details of inflammatory pathways may eventually furnish new therapeutic targets.Nature 420(6917):868-74. · 38.60 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: ROCKs, or Rho kinases, are serine/threonine kinases that are involved in many aspects of cell motility, from smooth-muscle contraction to cell migration and neurite outgrowth. Recent experiments have defined new functions of ROCKs in cells, including centrosome positioning and cell-size regulation, which might contribute to various physiological and pathological states.Nature Reviews Molecular Cell Biology 07/2003; 4(6):446-56. · 37.16 Impact Factor
1632?The?Journal?of?Clinical?Investigation? ? ? http://www.jci.org? ? ? Volume 118? ? ? Number 5? ? ? May 2008
ROCK1 mediates leukocyte recruitment and
neointima formation following vascular injury
Kensuke Noma,1 Yoshiyuki Rikitake,1 Naotsugu Oyama,1 Guijun Yan,2 Pilar Alcaide,3
Ping-Yen Liu,1 Hongwei Wang,1 Daniela Ahl,1 Naoki Sawada,1 Ryuji Okamoto,1 Yukio Hiroi,1
Koichi Shimizu,4 Francis W. Luscinskas,3 Jianxin Sun,2 and James K. Liao1
1Vascular Medicine Research Unit, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA.
2Department of Cell Biology and Molecular Medicine, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark,
New Jersey, USA. 3Center for Excellence in Vascular Biology, Department of Pathology, Brigham and Women’s Hospital and Harvard Medical School,
Boston, Massachusetts, USA. 4Donald W. Reynolds Cardiovascular Clinical Research Center, Department of Medicine,
Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA.
Vascular inflammation and smooth muscle proliferation contrib-
ute to vascular remodeling and obstructive vasculopathies such
as atherosclerosis and restenosis following percutaneous coronary
interventions (1, 2). The inflammatory process is characterized by
the activation of vascular wall cells and circulating leukocytes,
leading to the recruitment and infiltration of inflammatory cells
into the vessel wall (3, 4). The subsequent secretion of cytokines
and growth factors from these cells leads to increased migratory,
proliferative, and secretory responses of VSMCs, resulting in vas-
cular remodeling (5). Although recent studies have shed light on
some of the pathophysiological mechanisms involved in this pro-
cess, the intracellular signaling pathways that link these coordi-
nated responses in the vascular wall cells are not known.
Rho-associated kinases (ROCKs) are serine-threonine protein
kinases, which contribute to many downstream effects of the Rho
GTPases. Currently, there are 2 ROCK isoforms, namely ROCK1
(also referred to as ROKβ or p160ROCK) and ROCK2 (also
referred to as ROKα or Rho-kinase) (6). ROCKs regulate actin
cytoskeletal reorganization, focal adhesion complex formation,
smooth muscle contraction, cell migration, and gene expression
(7, 8). Phosphorylation of the myosin-binding subunit (MBS) of
myosin light chain phosphatase (MLCP) by ROCKs leads to inhi-
bition of MLCP activity and increase in MLC phosphorylation and
smooth muscle contraction (9).
Increased ROCK activity has been implicated in several cardio-
vascular diseases involving abnormal smooth muscle contraction
such as cerebral and coronary vasospasm (10, 11) and perhaps
hypertension (9). Furthermore, because ROCKs are also involved
in cellular proliferation, migration, and survival (6, 12), they may
also contribute to the development of atherosclerosis and vascu-
lar inflammation (13). Indeed, inhibition of ROCKs by the ROCK
inhibitors, fasudil and Y27632, has been shown to inhibit leuko-
cyte activation and infiltration (14), decrease tumor cell metastasis
and invasion (15, 16), and inhibit dissociation-induced apoptosis
of human embryonic stem cells (17).
Although pharmacological inhibition of ROCKs suggest that
they are important in promoting cardiovascular disease, the tissue-
and isoform-specific roles of ROCKs remain to be determined. For
example, previous studies with ROCK inhibitors are limited not
only by their nonselectivity for ROCK isoforms, but also, when
administered in vivo chronically and at higher concentrations,
ROCKs cannot be distinguished from other serine-threonine pro-
tein kinases such as protein kinase A and protein kinase C (18). In
addition, when given systemically, ROCK inhibitors cannot dis-
criminate among the tissue-specific roles of ROCKs in mediating
vascular disease process. Thus, a genetic approach with gene tar-
geting of specific ROCK isoforms offers the best strategy for dis-
secting and understanding the isoform-specific role of ROCKs.
We have previously generated mutant mice harboring deletion
of the Rock1 allele (19). Homozygous deletion of both Rock1 alleles
leads to embryonic and postnatal lethality (20–22). However, hap-
Nonstandard?abbreviations?used: BMT, bone marrow transplantation; MBS, myo-
sin-binding subunit; PCNA, proliferating cell nuclear antigen; ROCK, Rho-associated
Conflict?of?interest: The authors have declared that no conflict of interest exists.
Citation?for?this?article: J. Clin. Invest. 118:1632–1644 (2008). doi:10.1172/JCI29226.
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
loinsufficient ROCK1-knockout (Rock1+/–) mice are fertile and phe-
notypically normal with half of the protein level of ROCK1 (20).
In addition, we have generated mice with heterozygous deletion of
Rock2 allele (Rock2+/–). Although the homozygous ROCK2-knock-
out (Rock2–/–) mice die embryonically due to placental dysfunction
and intrauterine growth retardation, Rock2+/– mice similarly devel-
op normally, despite having half of the protein levels of ROCK2
(23, 24). Using these haploinsufficient ROCK mutant mice and
bone marrow transplantation (BMT) to isolate the effects of BM-
derived inflammatory cells, we investigated the role of ROCKs in
mediating neointima formation following vascular injury.
Generation of Rock1+/– and Rock2+/– mice. ROCK1-knockout mice on
C57BL/6 background were generated and analyzed as described
(19). Targeted deletion of a genomic fragment containing exon 3
of Rock2 gene was accomplished by homologous recombination
(see Supplemental Figure 1A; available online with this article;
doi:10.1172/JCI29226DS1). Homozygous deletion of Rock1 and
Rock2 results in embryonic and perinatal lethality (23, 24). How-
ever, heterozygous deletion of Rock1 and Rock2 leads to viable mice
with approximately half of the protein levels of the correspond-
ing ROCK isoform, with little if any compensatory changes in the
other ROCK isoform (see Supplemental Figure 1D).
Decreased neointima formation and ROCK activity after carotid artery
ligation in Rock1+/– mice. Basal and postligation systolic blood pres-
sures were not different between WT, Rock1+/–, and Rock2+/– mice
(n = 8 in each group; P = NS). Following ligation of the com-
mon carotid artery, the vessel typically undergoes inflammatory
changes, shrinkage, neointima formation, and narrowing of the
lumen (25). No neointima formation and lumenal narrowing
were observed in the unligated right common carotid arteries of
WT, Rock1+/–, and Rock2+/– mice (sham controls) (Table 1 and Fig-
ure 1A). In contrast, flow cessation caused by ligation of the left
common carotid artery led to substantial increase in neointima
formation in WT and Rock2+/– mice, but to a smaller increase in
Rock1+/– mice (P < 0.05 compared with WT and Rock2+/– mice). Most
of the cells in the neointima are VSMCs (data not shown, smMHC
staining). This was associated with decrease in intima to media
ratio in Rock1+/– mice (0.23 ± 0.11) compared with that of WT
(0.96 ± 0.13) or Rock2+/– mice (0.98 ± 0.21) and a reciprocal increase
in lumen size in vessels of Rock1+/– mice (P < 0.05 compared with
WT and Rock2+/– mice) (Table 1). The medial areas as well as the
areas surrounded by the external elastic lamina (EEL) and internal
EL (IEL) were similar in WT, Rock1+/–, and Rock2+/– vessels (P = NS).
These findings suggest that ROCK1, but not ROCK2, contributes
to neointima formation following vascular injury. Based upon
these findings, further investigations were focused on determin-
ing the mechanisms underlying the tissue-specific role of ROCK1
in neointima formation.
Using a specific phospho-MBS antibody that corresponds to
ROCK activity (19, 20), we found increased staining of the medial
and neointimal areas as well as adventitia following carotid liga-
tion of WT but not Rock1+/– mice (Figure 1B). This correlated with
decreased ROCK1 expression in the vascular wall of Rock1+/– mice
compared with that of WT mice. The level of staining for total
MBS in the media was similar between WT and Rock1+/–, mice
indicating that decreased ROCK activity in Rock1+/– mice was not
due to decreased substrate availability. Indeed, by western blot-
ting, overall ROCK activity was decreased in ligated carotid arter-
ies from Rock1+/– mice compared with that of WT or Rock2+/– mice
(see Supplemental Figure 2, A and B). There were no compensatory
changes in ROCK isoform expression in leukocytes of unligated
and ligated Rock1+/– and Rock2+/– mice (Figure 2A). Similarly, there
were no compensatory changes of ROCK isoform expression in
unligated and ligated carotid arteries from Rock1+/– and Rock2+/–
mice (Figure 2B). These findings indicate that changes in ROCK1,
but not ROCK2, in leukocytes and carotid arteries correlated with
reduced neointima formation.
Decreased leukocyte and macrophage recruitment to the vasculature in
Rock1+/– mice. Endothelial-leukocyte interaction and leukocyte
recruitment to the vascular wall contributes to the development
of neointima formation and vascular inflammation (5). Ex vivo
perfusion-fixed staining of the vessel wall following carotid artery
Lumenal and neointimal area of carotid arteries 4 weeks after ligation
WT to WT sham
WT to WT
WT to WT Rock1+/– sham
WT to Rock1+/–
Rock1+/– to WT sham
Rock1+/– to WT
60.6 ± 10.3
16.2 ± 4.3A,B
68.0 ± 6.2
38.1 ± 6.1D
62.7 ± 10.9
20.1 ± 4.4A,C
0.0 ± 0.0
26.8 ± 5.6A,C
0.0 ± 0.0
6.9 ± 3.8D
0.0 ± 0.0
31.2 ± 8.5A,B
12.6 ± 2.8
26.6 ± 2.6D
12.1 ± 1.3
26.6 ± 1.8A
12.4 ± 1.9
31.6 ± 3.3A
60.6 ± 10.3
43.0 ± 6.6
68.0 ± 6.2
45.1 ± 5.9
62.7 ± 10.9
51.2 ± 9.7
73.2 ± 8.7
69.6 ± 9.0
80.1 ± 7.2
71.7 ± 7.3
75.1 ± 9.5
82.8 ± 12.0
0.00 ± 0.00
0.96 ± 0.13A,B
0.00 ± 0.00
0.23 ± 0.11D
0.00 ± 0.00
0.98 ± 0.21A,B
61.0 ± 9.0
15.4 ± 5.7A,C
64.6 ± 12.4
17.6 ± 7.4D,C
65.0 ± 5.3
39.1 ± 6.4D
0.0 ± 0.0
25.6 ± 5.3A,B
0.0 ± 0.0
23.1 ± 5.7A,C
0.0 ± 0.0
8.6 ± 1.9A
14.2 ± 1.5
28.8 ± 1.5A
13.6 ± 0.9
32.0 ± 5.3A
13.9 ± 2.0
30.3 ± 3.4A
61.4 ± 9.0
40.9 ± 6.6
64.6 ± 12.4
40.7 ± 7.4
65.0 ± 5.3
47.6 ± 7.8
75.3 ± 7.7
69.7 ± 6.7
78.2 ± 11.9
72.7 ± 12.3
78.9 ± 4.2
77.9 ± 10.6
0.00 ± 0.0
0.93 ± 0.19A,B
0.00 ± 0.00
0.76 ± 0.14A,C
0.00 ± 0.00
0.27 ± 0.05
The ratio of intima to media was calculated as the intimal area/medial area. All results are presented as mean ± SEM. EEL, external elastic lamina; IEL,
internal EL. AP < 0.01 versus sham control in each group. BP < 0.01 versus Rock1+/– ligated vessel in non-BMT or Rock1+/– to WT ligated vessel in BMT
mice. CP < 0.05 versus Rock1+/– ligated vessel in non-BMT or Rock1+/– to WT ligated vessel in BMT mice. DP < 0.05 versus sham control in each group.
1634?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
ligation was performed in a blinded manner in WT and Rock1+/–
mice. Three days following carotid artery ligation, there was sub-
stantial increase in both leukocyte and macrophage accumulation
in ligated arteries of WT mice compared with unligated arteries in
WT and Rock1+/– mice (Figure 3). In unligated arteries, there was
minimal leukocyte recruitment in either WT or Rock1+/– mice. Com-
pared with ligated arteries of WT mice, the number of adherent leu-
kocytes was substantially less in Rock1+/– mice compared with that
of WT mice (4.3 ± 1.3 versus 14.1 ± 2.2 leukocytes/vessel section,
respectively; P < 0.05) (Figure 3, A and B). Using MOMA-2 stain-
ing to assess macrophage accumulation, we also found less mac-
rophage infiltration in Rock1+/– ligated vessels compared with that
of WT mice (4.2 ± 0.8 versus 8.3 ± 1.2 macrophages/vessel section;
P < 0.05) (Figure 3, C and D). These findings indicate that ROCK1
mediates leukocyte recruitment to the vessel wall after injury.
Decreased leukocyte chemotaxis, accumulation, and adherence in
Rock1+/– mice. Thioglycollate-induced peritonitis in WT, Rock1+/–,
and Rock2+/– mice was used to assess leukocyte homing response
and accumulation in vivo (26). Injection of thioglycollate increased
the numbers of neutrophils and macrophages in the perito-
neal cavity compared with injection of PBS (Figure 4, A and B).
There was no difference in the recruitment of neutrophils and
macrophages by PBS in WT, Rock1+/–, and Rock2+/– mice. In con-
trast, neutrophil accumulation was reduced in the peritoneal cav-
ity of Rock1+/– mice compared with that of WT and Rock2+/– mice
(7.04 ± 0.33 × 106 versus 11.96 ± 1.02 and 10.12 ± 1.09 × 106 neu-
trophils/ml, respectively; P < 0.01) (Figure 4A). Similarly, macro-
phage recruitment to the peritoneal cavity of Rock1+/– mice was
also reduced compared with that of WT and Rock2+/– mice (4.90 ±
0.23 × 106 versus 8.35 ± 0.76 and 8.17 ± 0.65 × 106 neutrophils/ml,
respectively; P < 0.01) (Figure 4B). These findings suggest that
neutrophil and macrophage recruitment and accumulation are
impaired in Rock1+/– but not Rock2+/– mice. Indeed, compared with
WT or Rock2+/– mice, VLA-4 expression is decreased by 80% in mac-
rophages from Rock1+/– mice (see Supplemental Figure 3C).
To determine whether endothelial ROCKs could also affect
monocyte adhesion, the interaction of U937 monocytes, sta-
bly expressed human L selectin (U937-LAM), with ECs from
WT, Rock1+/–, and Rock2+/– mice was studied under laminar
flow conditions. Compared with PBS, the adherence of U937-
LAM to thrombin-stimulated EC monolayers was substantially
increased in ECs from WT and Rock2+/– mice (P < 0.05 for both).
Decreased neointima formation after carotid
artery ligation in Rock1+/– mice. (A) Rep-
resentative cross sections of contralateral
unligated and ligated carotid arteries in WT,
Rock1+/–, and Rock2+/– mice at 28 days after
ligation. Scale bar: 100 μm. (B) Representa-
tive immunohistochemical analysis of ROCK
expression and activity in carotid arteries from
WT and Rock1+/– mice at 14 days after ligation.
Vessels were stained with nonspecific antibody
(IgG), and antibodies were directed at total MBS
(tMBS), phosphorylated MBS (pMBS), ROCK1,
and ROCK2. The I and M indicate the intima and
media, respectively. Scale bar: 50 μm.
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
In contrast, thrombin-stimulated ECs from Rock1+/– mice do not
exhibit increased U937-LAM adherence under laminar flow at
either 0.5 or 1.0 dynes/cm2 (P < 0.01 for both compared with
WT or Rock2+/–) (Figure 4, C and D). These findings suggest that
endothelial ROCK1, but not ROCK2, plays an important role in
mediating endothelial-leukocyte interaction in the vessel wall
under physiological flow conditions.
Decreased expression of endothelial adhesion molecules in Rock1+/– mice.
To determine whether ROCK1 in the vascular wall could also reg-
ulate leukocyte adhesion, the expression of adhesion molecules
such as ICAM-1 and VCAM-1 was evaluated following ligation in
WT and Rock1+/– mice (Figure 5, A and B). Semiqualitative analysis
of immunohistochemical staining by independent blinded observ-
ers was obtained using the following scale: 0, no staining; 1, weak
staining; 2, moderate staining; and 3, strong staining. Using this
qualitative scale to estimate the degree of endothelial expression
of adhesion molecules in WT and Rock1+/– mice compared with
expression in unligated vessel, we found that ICAM-1 and VCAM-1
expression were substantially increased in ligated vessels in WT
(ICAM-1, 1.3 ± 0.2; VCAM-1, 1.5 ± 0.2; n = 10; P < 0.05 for both com-
pared with unligated vessels) but less so in Rock1+/– mice (ICAM-1,
0.5 ± 0.2; VCAM-1, 0.7 ± 0.3; n = 10; P < 0.05 for both compared
with ligated WT vessels). Little or no expression of ICAM-1 or
VCAM-1 was observed in the unligated ves-
sels (n = 5 in each group). These findings
indicate that ROCK1 mediates ICAM-1
and VCAM-1 expression in the vessel wall
following vascular injury.
To confirm that ROCK1 mediates
endothelial expression of ICAM-1 and
VCAM-1, we isolated primary ECs from
WT and Rock1+/– mice and tested their
response to thrombin. In ECs from WT
mice, thrombin induced the mRNA
expression of Icam1 and Vcam1 (Figure
5C). In contrast, thrombin had little or no
effect on inducing the mRNA expression
of Icam1 or Vcam1 in ECs from Rock1+/–
mice. Western blot analysis showed that
the protein levels of ROCK1 in ECs from
Rock1+/– mice were approximately half of
that of ECs from WT mice (54.6% versus
WT; P < 0.01) (n = 4), and ROCK1 expres-
sion was not affected by thrombin stimu-
lation (Figure 5D). ROCK2 protein levels
in ECs from WT and Rock1+/– mice were
comparable, again suggesting no com-
pensatory changes in endothelial ROCK2
expression in Rock1+/– mice. Thrombin
increased ROCK activity within 5 minutes
in ECs from WT mice (Figure 5E). How-
ever, the effect of thrombin on ROCK
activity was substantially reduced in ECs
from Rock1+/– mice. In contrast, throm-
bin-induced extracellular signal-regulated
kinase (ERK) phosphorylation was rela-
tively similar, if not somewhat higher in
ECs from Rock1+/– mice compared with
ECs of WT mice (Figure 5E). These find-
ings suggest that the observed decrease in
the recruitment of leukocytes to the vascular wall following carotid
artery ligation in Rock1+/– mice may in part be due to the reduction
in endothelial adhesion molecule expression in Rock1+/– mice.
Decreased macrophage NF-κB activation and PDGF expression in
Rock1+/– mice. To determine the potential mechanisms underlying
ROCK1-mediated neointima formation, we assessed NF-κB acti-
vation and PDGF expression in thioglycollate-induced peritoneal
macrophages. Thioglycollate causes an inflammatory response
and has been shown to induce NF-κB activation and PDGF expres-
sion in peritoneal macrophages (27, 28). In thioglycollate-induced
Rock1+/– macrophages, higher steady-state level of IκB-α was
observed compared with that of WT and Rock2+/– macrophages
(Figure 6A). This corresponded to decreased NF-κB activation in
thioglycollate-induced Rock1+/– macrophages (Figure 6B). Similarly,
Pdgfa but not Pdgfa expression in thioglycollate-induced macro-
phages from Rock1+/– mice was reduced compared with those from
WT and Rock2+/– mice (P < 0.05) (Figure 6C). These findings sug-
gest that the regulation of NF-κB activation and PDGF-A expres-
sion by ROCK1 in macrophages may mediate some of the vascular
inflammatory response to injury.
Decreased VSMC proliferation in the neointima of Rock1+/– mice. To
determine the local vascular effects of ROCK1 on neointima for-
mation, we assessed the degree of cellular proliferation in the vas-
ROCK expression in leukocytes and carotids of Rock1+/– and Rock2+/– mice. (A) Expression
of ROCK isoforms in leukocytes from WT, Rock1+/–, and Rock2+/– mice with and without
carotid ligation. Representative western blot (upper panel). Quantification of ROCK1 expres-
sion (middle panel). Mean ± SEM; n = 3. Quantification of ROCK2 expression (lower panel).
Mean ± SEM; n = 3. (B) Expression of ROCK isoforms in unligated and ligated carotid arteries
from WT, Rock1+/–, and Rock2+/– mice. Representative western blot (upper panel). Quanti-
fication of ROCK1 expression (middle panel). Mean ± SEM; n = 4. Quantification of ROCK2
expression (lower panel). Mean ± SEM; n = 4. *P < 0.05 versus unligated WT mice; †P < 0.01
versus unligated Rock2+/– mice; §P < 0.01 versus ligated WT and Rock2+/– mice; **P < 0.01
versus unligated WT and Rock1+/– mice; ‡P < 0.01 versus ligated WT mice; #P < 0.05 versus
ligated Rock1+/– mice; ##P < 0.01 versus unligated WT and Rock2+/– mice; ***P < 0.01 versus
WT and Rock1+/– mice.
1636?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
cular wall by proliferating cell nuclear antigen (PCNA) staining.
PCNA staining in the vascular wall was minimal or not observed
prior to 14 days following carotid artery ligation. At day 14 after
carotid ligation, the ratio of PCNA-positive cells to total cells with-
in the arterial wall was substantially increased in WT mice com-
pared with that of Rock1+/– mice (16.7% ± 2.7% versus 5.0% ± 3.6%;
P < 0.01) (Figure 7A). Most, if not all, of the PCNA-positive cells
were VSMCs (data not shown, smMHC double staining). Similarly,
substantial increase in PCNA staining was observed in the media
in WT mice compared with that of Rock1+/– mice (9.1% ± 2.0% ver-
sus 3.4% ± 0.9%; P < 0.01).
To determine whether the decrease in proliferation is due to the
reduction in vascular inflammatory response or intrinsic decrease
in the ability of SMC to proliferate in Rock1+/– mice, we isolated pri-
mary VSMCs from aortas of WT and Rock1+/– mice and tested their
ability to proliferate and migrate in vitro. There was no difference
in the ability of WT or Rock1+/– VSMCs to adhere to the culture
dish or membranes (data not shown). Surprisingly, VSMC prolif-
eration as determined by cell number and thymidine incorpora-
tion in response to serum or PDGF was also not different between
VSMCs from WT and Rock1+/– mice (Figure 7, B and C). However,
the migration of VSMCs in response to PDGF as determined using
a modified Boyden chamber assay was substantially reduced in
Rock1+/– compared with that of WT mice (Figure 7D). Basal migra-
tion was not different between VSMCs from WT and Rock1+/– mice.
These findings suggest that the increase in VSMC proliferation
observed in the neointima and media following vascular injury is
probably due more to the indirect increase in the vascular inflam-
matory response elicited by leukocytes rather than by enhanced
intrinsic VSMC proliferation, since the proliferative response of
VSMCs to PDGF was similar between WT and Rock1+/– mice in
vitro. ROCK1, however, may contribute to increase VSMC migra-
tion and survival following vascular injury.
Similar to ECs, the expression of ROCK1 in VSMCs from Rock1+/–
mice was approximately half of that in WT mice (47.7% versus
WT; P < 0.01) (n = 4) (see Supplemental Figure 1E), with little or
no compensatory changes in ROCK2 expression. Furthermore,
thrombin stimulation did not affect ROCK1 or ROCK2 expres-
sion. In VSMCs from WT mice, PDGF increased ROCK activity
as measured by phosphorylation of MBS (Figure 7E). In VSMCs
from Rock1+/– mice, PDGF-induced ROCK activity was greatly
reduced. PDGF-stimulated ERK phosphorylation, however, was
comparable if not slightly higher in Rock1+/– mice compared with
that in WT mice. These findings suggest that the decreased VSMC
proliferation and neointima formation observed in Rock1+/– mice
are probably not mediated by ERK.
Decreased leukocyte recruitment to
ligated vessels of Rock1+/– mice. (A)
Representative histological sections
from carotid arteries in WT and Rock1+/–
mice stained with CD45 Ab for leuko-
cytes. Arrowheads indicate CD45-posi-
tive cells. Scale bar: 50 μm (left panel).
Quantitative analysis of the number of
CD45-positive leukocytes adherent to
the unligated (n = 3–4 in each group)
and ligated vessel walls (n = 14 in each
group) at 3 days after ligation (right
panel). (B) Representative histological
sections from carotid arteries in WT and
Rock1+/– mice stained with MOMA-2 Ab
for macrophage. Arrowheads indicate
MOMA-2–positive cells. Scale bar: 50
μm (left panel). Quantitative analysis of
the number of MOMA-2–positive mac-
rophages infiltrated into the unligated
(n = 3–4 in each group) and ligated ves-
sel (n = 12–13 in each group) at day 3.
*P < 0.01 versus unligated WT mice;
†P < 0.05 versus unligated Rock1+/– mice;
§P < 0.05 versus ligated WT mice.
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
Decreased neointima formation after carotid artery ligation by BM-
derived cells from Rock1+/– mice. In contrast to arterial wire injury and
transplant-associated arteriosclerosis, previous studies have shown
that BM-derived cells do not contribute to neointimal VSMCs in
the carotid artery ligation model (29). However, to determine the
contribution of ROCK1 from BM-derived inflammatory cells to
neointima formation, we performed reciprocal BMT using WT
and Rock1+/– mice, using both as donors and recipients of BMT.
Neither haploinsufficiency of ROCK1 nor BMT affected the total
number of peripheral blood erythrocyte and leukocyte counts or
leukocyte differential (see Supplemental Table 1). All mice that
did not receive BMT died within 3 weeks of the transplant. Four
weeks after BMT, peripheral blood counts returned to normal and
the mice underwent carotid artery ligation injury. Interestingly,
Rock1+/– to WT BMT mice exhibited less neointima formation and
decreased intima to media ratio compared with those of WT to
WT or WT to Rock1+/– BMT mice (Table 1 and Figure 8A). Indeed,
neointima formation in WT to Rock1+/– BMT mice was comparable
to that of WT to WT BMT mice. Circulating leukocytes isolated
from WT to WT and WT to Rock1+/– BMT mice showed similar
levels of ROCK1 and ROCK2 expressions compared with those of
leukocytes from untransplanted WT mice, indicating successful
BM replenishment (Figure 8B). Similarly, circulating leukocytes
isolated from Rock1+/– to WT BMT mice showed similar levels of
ROCK1 and ROCK2 expression compared with leukocytes from
untransplanted Rock1+/– mice. Furthermore, ROCK2 expression
in leukocytes was similar between WT to WT, WT to Rock1+/–,
and Rock1+/– to WT BMT mice. In contrast, ROCK1 and ROCK2
expression in ECs from WT to WT or Rock1+/– to WT BMT mice
was similar to that from WT mice (Figure 8C), indicating that
BMT did not affect the expression of ROCK1 and ROCK2 in the
vascular wall. These findings indicate that ROCK1 in circulating
leukocytes rather than in vascular wall cells is the predominant
mediator of neointima formation following carotid ligation.
Decreased leukocyte recruitment to the vessel wall in Rock1+/– BMT mice.
The role of leukocyte ROCK1 in mediating leukocyte recruitment
to the vascular wall was evaluated 3 days following carotid artery
ligation in BMT mice. The number of adherent leukocytes to the
vascular wall in Rock1+/– to WT BMT mice (4.8 ± 1.1 leukocytes/
vessel section) was substantially less compared with that of WT
to WT (11.1 ± 1.3 leukocytes/vessel section; P < 0.01) and WT to
Rock1+/– BMT mice (9.6 ± 1.8 leukocytes/vessel section; P < 0.05)
(Figure 9A). In addition, there was less macrophage infiltration in
vascular wall of Rock1+/– to WT BMT mice (3.4 ± 0.5 leukocytes/
vessel section) compared with that of WT to Rock1+/– BMT mice
(7.8 ± 1.8 leukocytes/vessel section) and WT to WT BMT mice
(8.9 ± 2.4 leukocytes/vessel section) (P < 0.05) (Figure 9B). However,
the expression of ICAM-1 and VCAM-1 in the vascular wall was
similar between WT to WT BMT mice (ICAM-1, 1.3 ± 0.2; VCAM-1,
1.6 ± 0.2; n = 9) and Rock1+/– to WT BMT mice (ICAM-1, 0.9 ± 0.2;
Decreased recruitment and adhesion of leukocytes in Rock1+/– mice. Neutrophil and macrophage recruitment was induced by i.p. injection of
thioglycollate. (A) Neutrophil recruitment 4 hours after i.p. injection of PBS (control) or 3% thioglycollate into WT, Rock1+/–, and Rock2+/– mice
(n = 10, 5, and 5, respectively). (B) Macrophage recruitment 4 days after i.p. injection of PBS (control) or 3% thioglycollate into WT, Rock1+/–, and
Rock2+/– mice (n = 9, 8, and 6, respectively). Confluent monolayers of ECs isolated from WT and Rock1+/– mice were incubated with PBS (control,
n = 4–6 in each group) or 10 units/ml of thrombin for 10 minutes (n = 9–10 in each group). They were then placed under laminar flow at 0.5 (C)
or 1.0 dynes/cm2 (D). U937 monocytes, stably expressed human L selectin (U937-LAM), were perfused over the EC monolayers, and adherent
cells were quantified after 3 minutes. *P < 0.01 versus each control group; †P < 0.05 versus each control group; ‡P < 0.01 versus Rock1+/– mice
with thioglycollate or Rock1+/– ECs with thrombin; §P < 0.05 versus Rock1+/– mice with thioglycollate or Rock1+/– ECs with thrombin.
1638?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
VCAM-1, 1.2 ± 0.2; n = 10) (P = NS) (see Supplemental Figure 3,
A and B). Adhesion molecule expression in the vascular wall, how-
ever, was decreased in WT to Rock1+/– mice (ICAM-1, 0.6 ± 0.2;
VCAM-1, 0.8 ± 0.2; n = 11; P < 0.05 compared with WT to WT and
WT to Rock1+/– BMT mice).
Following carotid artery ligation, VSMC proliferation was
evaluated in the vascular wall of these BMT mice. At day 14, the
ratio of PCNA-positive cells to total cells within the arterial wall
was substantially reduced in the intima and media of Rock1+/– to
WT BMT mice compared with those of WT to WT and WT to
Rock1+/– mice (P < 0.01 and P < 0.05, respectively) (Figure 9C).
These results suggest that leukocyte ROCK1 is a primary inducer
of neointima formation following vascular injury since in the
carotid artery ligation model, BM-derived cells do not differenti-
ate into intimal VSMCs (29).
Vascular remodeling is an adaptive process that occurs in response
to chronic hemodynamic changes, which includes compensatory
adjustment in vessel diameter and lumenal area (30). Although
the precise mechanism of arterial neointima formation following
vascular injury remains to be determined, growth factors, vasoac-
tive substances, inflammatory response, adhesion molecules, and
matrix modulators may all contribute to this process. Flow cessa-
tion models, such as the ligation of the common carotid artery, are
useful tools for investigating the mechanism of vascular remodel-
ing because they involve leukocyte infiltration, neointima forma-
tion, and luminal narrowing. Using this model, we have shown
that vascular inflammation and neointima formation after flow
cessation–induced vascular injury are substantially reduced in
Rock1+/– mice compared with those of WT or Rock2+/– mice. These
findings suggest a critical role for ROCK1 in the development of
the neointima. It should be noted that these effects were unrelated
to changes in blood pressure since systolic blood pressures were
comparable among WT, Rock1+/–, and Rock2+/– mice. Thus, target-
ing ROCK1, rather than ROCK2, may be therapeutically beneficial
in reducing neointima formation following vascular injury.
Several mechanisms contribute to the observed changes in
Rock1+/– ligated vessels, including inhibition of endothelial adhe-
sion molecule expression, impaired migratory response and sur-
vival of VSMCs, and decreased leukocyte recruitment and accumu-
lation in the vessel wall. Although ROCK2 expression was similar
in ECs and VSMCs of WT and Rock1+/– mice, total ROCK activity,
as defined by the phosphorylation of MBS in the vessel wall, was
reduced in ligated vessels from Rock1+/– mice compared with that of
WT mice. Thus, the decrease in vascular ROCK activity observed in
Decreased expression of endothelial
adhesion molecules in Rock1+/– mice.
Representative histological sections from
unligated and ligated carotid arteries at
7 days after ligation in WT and Rock1+/–
mice stained for ICAM-1 (A) and VCAM-1
(B). Scale bars: 50 μm. (C) Represen-
tative result of northern blot analysis of
Icam1 and Vcam1 mRNA expression in
ECs from WT and Rock1+/– mice. ECs
were stimulated with 5 U/ml of thrombin
for 6 hours. Gapdh mRNA expression
was used as an internal control. (D) Rep-
resentative result of western blot analy-
sis of ROCK1 and ROCK2 expressions
in ECs from WT and Rock1+/– mice with
or without thrombin stimulation for 5 min-
utes. Actin was used as an internal con-
trol. (E) Representative result of western
blot analysis of ROCK and ERK activities
in ECs from WT and Rock1+/– mice. ECs
were stimulated with 5 U/ml of thrombin
for the indicated time periods.
? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
Rock1+/– mice corresponded to a reduction in the vascular inflam-
matory response and neointima formation.
The adhesion and recruitment of leukocytes to the vessel wall are
the primary contributors to the inflammatory response following
vascular injury. Our finding of impaired leukocyte recruitment to
the vessel wall in Rock1+/– mice is consistent with previous stud-
ies showing that ROCK inhibitors decrease leukocyte recruitment
and adhesion following ischemia/reperfusion injury (31). Fur-
thermore, ROCK inhibitors have been shown to reduce neointima
formation after balloon injury in rats (32, 33). However, previous
studies with ROCK inhibitors have shown that ROCK inhibitors
cannot discriminate between the effects of ROCK1 and ROCK2,
and at higher concentrations, cannot distinguish the effects of
ROCKs from other serine-threonine protein kinases such pro-
tein kinase A or protein kinase C (18). The mechanism underly-
ing decreased leukocyte recruitment to the vessels of Rock1+/– mice
could be explained in part by decrease expression of endothelial
adhesion molecules. For example, administration of blocking anti-
bodies to ICAM-1 or VCAM-1 inhibits neointima hyperplasia after
arterial injury (34, 35). Furthermore, the ROCK inhibitor, Y27632,
inhibits thrombin-induced ICAM-1 expression through a NF-κB–
dependent mechanism (36). However, we observed no substantial
decrease in neointima formation in WT to Rock1+/– ligated vessels,
in which the expression levels of ICAM-1 and VCAM-1 were sub-
stantially lower compared with those of WT to WT BMT mice.
These findings suggest that the lower expression of ICAM-1 and
VCAM-1 in Rock1+/– mice may be sufficient to recruit leukocytes to
the vessel wall and that the decrease in leukocyte recruitment and
neointima formation observed in Rock1+/– mice is primarily due to
the loss of ROCK1 in leukocytes.
ROCK is also important in mediating the adhesion and trans-
migration of monocytes (14). Overexpression of the active form of
RhoA in monocytes leads to a dramatic increase in monocyte adhe-
sion to and migration across EC monolayers, both of which were
prevented by Y27632. Similarly, Y27632 has been shown to inhibit
monocyte chemoattractant protein-1–induced migration of leuko-
cytes (37). Indeed, we found that the recruitment of leukocytes such
as neutrophils and macrophages in Rock1+/– mice was substantially
reduced in vivo and under physiological flow conditions. The mech-
anism may in part be due to decrease activation of NF-κB and VLA-4
expression in Rock1+/– macrophages. Using BMT, we were able to sep-
arate the effects of ROCK1 in BM-derived cells versus local vascular
wall cells in mediating neointima formation following ligation. We
found that the recruitment of donor leukocytes from Rock1+/– mice
to sites of inflammation or vascular injury was impaired. This was
associated with decreased neointima formation in Rock1+/– to WT
BMT mice, despite comparable expression of endothelial adhesion
molecules as WT to WT BMT mice. These findings suggest that
ROCK1 in leukocytes may play a greater role than ROCK1 in vascu-
lar wall cells in mediating neointima formation.
Although the number of PCNA-positive cells in the intimal and
medial areas were substantially reduced in Rock1+/– mice compared
with that of WT mice, cellular proliferation and DNA synthesis
were not different between VSMCs isolated from WT and Rock1+/–
mice, although Rock1+/– VSMCs exhibited decreased migration and
survival. These findings suggest that the intrinsic ability of VSMCs
to proliferate in response to equal concentrations of serum or
PDGF is not altered in VSMCs from Rock1+/– mice compared with
that of WT mice. Indeed, the level of PDGF-induced ERK activa-
tion, which plays a critical role in cell growth, was comparable
between VSMCs of WT and Rock1+/– mice. Thus, the decrease in
neointimal VSMC proliferation observed in vivo in ROCK1 mice
may be due more to a reduction in overall inflammatory response
in the vessel wall (i.e., from decreased leukocyte recruitment) rath-
er than the intrinsic effects of ROCK1 in VSMCs. For example, the
smaller number of PCNA-positive cells observed in the vessels of
Rock1+/– mice as compared with that of WT mice may be due to
the decrease in leukocyte recruitment and the subsequent over-
all reduction in the release of inflammatory and growth stimuli
in the vessel wall of Rock1+/– mice. Indeed, growth factors such as
Decreased expression of PDGF-B, degradation of IκB-α, and activa-
tion of NF-κB in macrophages from Rock1+/– mice. (A) Representative
western blot analysis of IκB-α expression in peritoneal macrophages
from WT, Rock1+/–, and Rock2+/– mice. Actin was used as an internal
control. (B) Representative electrophoretic mobility shift assay of NF-κB
in WT, Rock1+/–, and Rock2+/– macrophages (left panel). Specificity
of NF-κB binding activity was analyzed by the addition of unlabeled
probe, by the pretreatment with excess unlabeled probe (cold) or
mutant probe (mutant cold), and by anti-p65 Ab (p65 Ab) supershift
gel assay (right panel). NF-κB binding band (NF-κB), nonspecific bind-
ing band (NS), free probe (free), and the raised bands supershifted
by Ab (supershift) are indicated on the right. (C) Expression of Pdgfa
and -b were analyzed by quantitative real-time PCR. Total RNA was
extracted from thioglycollate-induced peritoneal macrophages in WT,
Rock1+/–, and Rock2+/– mice (n = 6–8). *P < 0.05 versus WT and
1640?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
PDGF-B are secreted from macrophages only when they are acti-
vated (38). Thus, by preventing macrophage activation, the genetic
loss of ROCK1 in macrophages may lead to an overall decrease
in VSMC proliferation in the vessel wall. This is consistent with
our BMT studies showing that Rock1+/– to WT BMT mice have less
neointima formation compared with WT to WT BMT mice. As
mentioned, BM-derived cells do not differentiate into neointimal
VSMCs in the carotid ligation model (29).
In summary, we have shown that ROCK1 mediates leukocyte
recruitment and neointima formation following vascular injury.
ROCK1 in ECs contributes to the induction of endothelial adhe-
sion molecules, whereas ROCK1 in VSMCs mediates VSMC
migration and survival. More importantly, ROCK1 in circulating
leukocytes appears to be the primary determinant of leukocyte
recruitment to the vessel wall and is the critical mediator of neo-
intima proliferation. Thus, targeting ROCK1 in leukocytes may be
an effective strategy for treating vascular inflammatory and prolif-
erative diseases. Further studies are necessary to determine the pre-
cise mechanism by which ROCK1 regulates leukocyte function.
Generation of Rock1- or Rock2-knockout mice. ROCK1-knockout mice were
generated on C57BL/6 background as described previously (19). Deletion
of the Rock2 allele was accomplished by homologous recombination using
a targeting vector corresponding to a genomic fragment containing exon 3
of the Rock2 gene (see Supplemental Figure 1A). Successful targeting and
deletion of ROCK2 were confirmed by Southern blotting and PCR. Homo-
zygous deletion of ROCK2 on C57BL/6 background is embryonically
lethal. However, heterozygous Rock1+/– and Rock2+/– mice are viable and
have half the protein levels of ROCK1 and ROCK2, respectively, without
compensatory increase in the other ROCK isoform. All mice and their lit-
termates (i.e., all on C57BL/6 background) were maintained at the Harvard
Medical School animal facilities. The Standing Committee on Animals at
Harvard Medical School has approved all protocols pertaining to experi-
mentation with animals in this study.
Carotid artery ligation model. Carotid artery ligation was performed as
described previously (25, 39). Male Rock1+/–, Rock2+/–, and littermate WT
controls were used in this study. Carotid arteries were perfusion fixed
through the left ventricle with 4% paraformaldehyde and 10% sucrose in
PBS under physiological pressure as described previously (39).
Histology and immunostaining. Carotid arteries were embedded transversely
in OCT compound (Tissue-Tek). Sections (6 μm) from animals euthanized
3, 7, 14, and 28 days after ligation were obtained at 1 mm proximal to the
ligature. Elastica van Gieson staining was used for histochemical analysis
to evaluate neointima formation. Two or more researchers, who were blind-
ed to the experimental protocol, were enlisted to perform morphometric
analysis using the NIH image software, ImageJ (http://rsb.info.nih.gov/ij/).
Immunostaining with antiphosphospecific threonine 853 MBS (1:200)
Decreased cell proliferation in the neointima of Rock1+/– mice. (A) Representative histological sections from carotid arteries in WT and Rock1+/–
mice stained for PCNA at 14 days after ligation. Arrowheads indicate PCNA-positive cells. Scale bars: 50 μm (left panels). Quantitative analysis
of the ratio of PCNA-positive cells to total cell number in the intima and the media (n = 10–16) (right panel). *P < 0.01 versus WT mice. (B) Cell
proliferation in response to serum of VSMCs from WT and Rock1+/– mice. Experiments were performed 6 times in triplicates. (C) DNA synthesis
in response to PDGF of VSMCs from WT and Rock1+/– mice (n = 12). *P < 0.01 versus without PDGF (control). (D) Cell migration in response
to PDGF of VSMCs from WT and Rock1+/– mice (n = 8–9). *P < 0.01 versus without PDGF (control); †P < 0.01 versus control; §P < 0.01 versus
WT. (E) Representative western blot analysis of ROCK and ERK activities in VSMCs of WT and Rock1+/– mice. VSMCs were stimulated with 10
ng/ml of PDGF for the indicated time periods.
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
(19), anti-MBS (1:200; Covance Inc.) (19), anti-ROCK1 (1:100; Santa Cruz
Biotechnology Inc.), anti-ROCK2 (1:100; Santa Cruz Biotechnology Inc.),
anti–ICAM-1 (1:100; Santa Cruz Biotechnology Inc.), anti–VCAM-1 (1:100;
Southern Biotechnology Associates Inc.), anti-mouse CD45 (1:100; eBio-
science Inc.), and anti–MOMA-2 (1:100; Serotec Ltd.) Abs was performed.
After incubation with the biotinylated Abs (Vector Laboratories), antigen-
Ab complexes were visualized with horseradish peroxidase streptavidin
(Vector Laboratories), followed by 9-amino-3-ethylene-carbazole (AEC;
Dako Inc.) or 3,3′-diaminobenzidine (DAB; Vector laboratories). PCNA
staining was performed with the use of a PCNA Staining kit (Zymed labo-
ratories Inc.). The expression of ICAM-1 and VCAM-1 was semiquantita-
tively evaluated by blinded observers using the following scale: 0, negative;
1, variable or weak; 2, moderately or strongly positive.
Isolation and primary culture of mouse ECs and VSMCs. Isolation and culture
of ECs and VSMCs from WT or Rock1+/– mice were described previously (19,
40). At least 2 independent preparations for each experiment were used.
Thioglycollate-induced peritonitis. Each mouse was injected i.p. with 1 ml of
3% thioglycollate (Sigma-Aldrich) as previously described (26). Four hours
or 4 days after the injection, animals were killed by CO2 asphyxiation, and
9 ml of PBS containing 0.1% of BSA, 0.5 mM EDTA, and 10 U/ml of heparin
was injected i.p. Cells were recovered by peritoneal lavage. Total cell count
in the lavage was determined by a Coulter counter. Cytospin (Thermo Sci-
entific) preparations of the lavage were stained with Wright’s stain and dif-
ferentially counted to determine the percentage of neutrophils or macro-
phages. After centrifugation, harvested peritoneal cells were resuspended
in DMEM plus 10% FBS and plated into appropriate wells.
Cells were allowed to adhere for 2 hours and then washed free
of nonadherent cells (27, 28).
Northern blotting. Northern blotting was performed as
described previously (19). Primary cultures of ECs were stimu-
lated with 5 U/ml of thrombin (Sigma-Aldrich) in 1% serum
for 6 hours. Total RNA was isolated, separated on 1% agarose
gel, and transferred onto nitrocellulose membrane. The oli-
gonucleotide probes for Icam1 and Vcam1 were obtained
as a PCR product using specific primers for murine Icam1
(forward, 5′-CATCGGGGTGGTGAAGTCTGT-3′; reverse,
5′-TGTCGGGGGAAGTGTGGTC-3′) and Vcam1 (forward,
5′-CAGCTAAATAATGGGGAACTG-3′; reverse, 5′-GGGC-
GAAAAATAGTCCTTG-3′), respectively. Gapdh expression was
used as an internal control. Radiolabeling of probes was per-
formed using random hexamer priming, with [α-32P]CTP and
Klenow fragment of DNA polymerase I (Pharmacia Corp.).
Real-time quantitative RT-PCR. Total RNA was extracted from
the macrophages with TRIzol. The Quantitect SYBR Green RT-
PCR kit (QIAGEN) was carried out to perform amplifications
with the 1-step protocol as described by the manufacturer using
an ABI Prism 7900HT sequence detector (Applied Biosystems).
The following primers were used to amplify PDGF-A, PDGF-B,
and GAPDH partial cDNA: Pdgfa (forward, 5′-TGGCTC-
GAAGTCAGATCCACA-3′ and reverse, 5′-AGCCCCTACG-
GAGTCTATCTC-3′), Pdgfb (forward, 5′-CGAGCCAAGACGCCTCAAG-3′
and reverse, 5′-CATGGGTGTGCTTAAACTTTCG-3′), and Gapdh (forward,
5′-GCAGTGGCAAAGTGGAGATT-3′ and reverse, 5′-CACATTGGGGGTAG-
GAACAC-3′). The fluorescence curves were analyzed with software included
by Applied Biosystems. GAPDH was used as an endogenous control reference.
Fold change is shown as relative to that of WT control mice.
Western blotting. Western blotting was performed using phospho-specific
threonine 853 MBS polyclonal Ab (19), MBS polyclonal Ab (Covance Inc.),
ROCK1 monoclonal Ab (BD Transduction Laboratories), ROCK2 mono-
clonal Ab (BD Transduction Laboratories), IκB-α (Santa Cruz Biotechnol-
ogy Inc.), and actin polyclonal Ab (Sigma-Aldrich) as described previously
(19). Primary cultures of ECs were stimulated with 5 U/ml of thrombin
(Sigma-Aldrich) in 1% serum for the 5 indicated periods (see Methods and
Figure 5). Also, primary cultures of VSMCs were stimulated with 10 ng/ml
of human PDGF-BB in 0.4% serum-containing DMEM for the indicated
periods (see Methods and Figure 7).
Electrophoretic mobility shift assay. Nuclear proteins were isolated from mac-
rophage samples and NF-κB activity was examined by EMSA as described
previously (41). A 20-μl binding reaction mixture containing 5 μg of
nuclear proteins was incubated with 32P-labeled double-stranded NF-κB
consensus binding sequence (Santa Cruz Biotechnology Inc.) for 1 hour at
4°C. A supershift assay using antibodies to 2 μg of p65 Ab was performed
to confirm NF-κB binding specificity as described previously (41).
Leukocyte adhesion under defined laminar flow conditions. Leukocyte
interactions with ECs from either WT or Rock1+/– mice were examined
ROCK1 in BM-derived cells mediates to neointima forma-
tion. (A) Representative cross sections of contralateral
unligated and ligated carotid arteries in WT to WT, WT to
Rock1+/–, and Rock1+/– to WT BMT mice at 28 days after
ligation. Scale bars: 100 μm. Representative western blot
analysis of ROCK expression in leukocytes (B) and ECs
(C) from WT and Rock1+/– BMT mice. Actin was used as
an internal control.
1642?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
BM-derived cells contribute to leukocyte recruitment and cell proliferation in Rock1+/– to WT BMT mice. (A) Representative histological sections
from carotid arteries in WT and Rock1+/– BMT mice stained with CD45 Ab for leukocytes. Arrowheads indicate CD45-positive cells. Scale bar:
50 μm (left panel). Quantitative analysis of the number of CD45-positive leukocytes adherent to the unligated (n = 3–4) and ligated vessel wall
at 3 days after ligation (n = 8–12). (B) Representative histological sections from carotid arteries in WT and Rock1+/– BMT mice stained with
MOMA-2 Ab for macrophages. Arrowheads indicate MOMA-2–positive cells. Scale bar: 50 μm (left panel). Quantitative analysis of the number of
MOMA-2–positive macrophages infiltrated into the unligated (n = 3–4) and ligated vessel wall at days 3 (n = 7–8). (C) Representative histologi-
cal sections from carotid arteries in WT and Rock1+/– BMT mice stained for PCNA at 14 days after ligation. Arrowheads indicate PCNA-positive
cells. Scale bar: 50 μm (left panel). Quantitative analysis of the ratio of PCNA-positive cells to total cell number in the intima and the media
(n = 8–9). *P < 0.05 versus unligated mice; †P < 0.01, §P < 0.05 versus Rock1+/– to WT mice.
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
under conditions of fluid shear stress in a parallel plate flow chamber as
described previously (41, 42).
Cell proliferation assay. Cells were plated at a density of 40,000 cells per
well in 6-well dishes and cultured with or without 10 ng/ml of human
PDGF-BB in 0.4% serum-containing DMEM for the indicated periods (see
Methods and Figure 7). Cell numbers were quantified using an automated
cell counter and each value was derived from 6 independent experiments
performed in triplicate.
Thymidine incorporation. Cells were plated on to 24-well plates at a density of
20,000 cells/well and then serum-starved for 48 hours with 0.4% serum prior
to stimulation for 24 hours with or without PDGF. During the last 6 hours
of stimulation, cells were incubated with [3H] thymidine (Amersham).
Cell migration assays. Cell migration was measured using Transwell 24-well
chambers separated by a filter (8-μm pore size) coated with 0.1% gelatin. A total
of 2 × 104 cells were added to the upper chamber and serum-starved for 24
hours. PDGF-BB (10 ng/ml) was added to the lower chamber and incubated
for 4 hours. Cells that migrated on the filter were stained with the use of Proto-
col (Fisher), and cell numbers were counted using light microscopy (×200).
BMT protocol. Twenty-four male WT C57BL/6 (The Jackson Laboratory;
8 weeks old; n = 75) and Rock1+/– (8 weeks old; n = 37) mice underwent total
body lethal irradiation (9.5 Gy) to eliminate endogenous BM stem cells and
circulating leukocytes. BM cells used for repopulation were extracted from
the femur and tibia of 10 WT and 5 Rock1+/– mice. Irradiated mice were
injected intravenously with 107 BM cells from either WT or Rock1+/– mice.
At 4 weeks after BMT, blood and leukocyte counts returned to normal (see
Supplemental Table 1). Therefore, carotid artery ligation was performed
4 weeks after BMT. All irradiated mice that were not injected with trans-
planted BM cells die within 2–3 weeks after radiation.
Statistics. All data are expressed as the mean ± SEM. Statistical analysis
was performed by unpaired 2-tailed Student’s t test or ANOVA with a
Fisher’s exact post-test. A Mann-Whitney U test was used for comparison
between groups with noncontinuous parameters, i.e., histological grading
scores. A P value of less than 0.05 was considered statistically significant.
This work was supported by grants from the NIH (HL052233 and
HL080187 to J.K. Liao; HL053993 to F.W. Luscinskas; and GM-
67049, HL-67249, and HL-67283 to K. Shimizu), the American
Heart Association (Bugher Foundation Award to J.K. Liao, Sci-
entist Development Grant 0630047N to J. Sun and 0630010N to
K. Shimizu, and Postdoctoral Research Fellowship to D. Ahl), the
Japan Heart Foundation (Japan Heart Foundation/Bayer-Yakuhin
Research Grant Abroad to K. Noma, Y. Rikitake, and N. Oyama),
and the National Health Research Institute, Zhunan Town, Repub-
lic of China (to P.-Y. Liu).
Received for publication May 26, 2006, and accepted in revised
form February 27, 2008.
Address correspondence to: James K. Liao, Brigham and Women’s
Hospital, 65 Landsdowne Street, Room 275, Cambridge, Massa-
chusetts 02139, USA. Phone: (617) 768-8424; Fax: (617) 768-8425;
Kensuke Noma, Yoshiyuki Rikitake, and Naotsugu Oyama con-
tributed equally to this work.
1. Pasterkamp, G., Galis, Z.S., and de Kleijn, D.P.
2004. Expansive arterial remodeling: location,
location, location. Arterioscler. Thromb. Vasc. Biol.
2. Bobik, A., and Tkachuk, V. 2003. Metalloprotein-
ases and plasminogen activators in vessel remodel-
ing. Curr. Hypertens. Rep. 5:466–472.
3. Serrano, C.V., Jr., et al. 1997. Coronary angioplasty
results in leukocyte and platelet activation with
adhesion molecule expression. Evidence of inflam-
matory responses in coronary angioplasty. J. Am.
Coll. Cardiol. 29:1276–1283.
4. Tanaka, H., et al. 1993. Sustained activation of vas-
cular cells and leukocytes in the rabbit aorta after
balloon injury. Circulation. 88:1788–1803.
5. Libby, P. 2002. Inflammation in atherosclerosis.
6. Riento, K., and Ridley, A.J. 2003. Rocks: multifunc-
tional kinases in cell behaviour. Nat. Rev. Mol. Cell
7. Shimokawa, H., and Takeshita, A. 2005. Rho-
kinase is an important therapeutic target in car-
diovascular medicine. Arterioscler. Thromb. Vasc. Biol.
8. Noma, K., Oyama, N., and Liao, J.K. 2006. Physio-
logical role of ROCKs in the cardiovascular system.
Am. J. Physiol. Cell Physiol. 290:C661–C668.
9. Uehata, M., et al. 1997. Calcium sensitization of
smooth muscle mediated by a Rho-associated pro-
tein kinase in hypertension. Nature. 389:990–994.
10. Sato, M., Tani, E., Fujikawa, H., and Kaibuchi, K.
2000. Involvement of Rho-kinase-mediated phos-
phorylation of myosin light chain in enhancement
of cerebral vasospasm. Circ. Res. 87:195–200.
11. Masumoto, A., et al. 2002. Suppression of coronary
artery spasm by the Rho-kinase inhibitor fasudil
in patients with vasospastic angina. Circulation.
12. Rikitake, Y., and Liao, J.K. 2005. ROCKs as thera-
peutic targets in cardiovascular diseases. Expert Rev.
Cardiovasc. Ther. 3:441–451.
13. Mallat, Z., et al. 2003. Rho-associated protein
kinase contributes to early atherosclerotic lesion
formation in mice. Circ. Res. 93:884–888.
14. Honing, H., et al. 2004. RhoA activation promotes
transendothelial migration of monocytes via
ROCK. J. Leukoc. Biol. 75:523–528.
15. Lepley, D., Paik, J.H., Hla, T., and Ferrer, F. 2005.
The G protein-coupled receptor S1P2 regulates
Rho/Rho kinase pathway to inhibit tumor cell
migration. Cancer Res. 65:3788–3795.
16. Kamai, T., et al. 2003. Significant association of
Rho/ROCK pathway with invasion and metastasis
of bladder cancer. Clin. Cancer Res. 9:2632–2641.
17. Watanabe, K., et al. 2007. A ROCK inhibitor per-
mits survival of dissociated human embryonic
stem cells. Nat. Biotechnol. 25:681–686.
18. Rikitake, Y., et al. 2005. Inhibition of Rho kinase
(ROCK) leads to increased cerebral blood flow and
stroke protection. Stroke. 36:2251–2257.
19. Rikitake, Y., and Liao, J.K. 2005. Rho-kinase medi-
ates hyperglycemia-induced plasminogen activator
inhibitor-1 expression in vascular endothelial cells.
20. Rikitake, Y., et al. 2005. Decreased perivascular
fibrosis but not cardiac hypertrophy in ROCK1+/–
haploinsufficient mice. Circulation. 112:2959–2965.
21. Shimizu, Y., et al. 2005. ROCK-I regulates closure
of the eyelids and ventral body wall by induc-
ing assembly of actomyosin bundles. J. Cell Biol.
22. Zhang, Y.M., et al. 2006. Targeted deletion of
ROCK1 protects the heart against pressure over-
load by inhibiting reactive fibrosis. FASEB J.
23. Thumkeo, D., et al. 2003. Targeted disruption of
the mouse rho-associated kinase 2 gene results in
intrauterine growth retardation and fetal death.
Mol. Cell. Biol. 23:5043–5055.
24. Thumkeo, D., Shimizu, Y., Sakamoto, S., Yamada, S.,
and Narumiya, S. 2005. ROCK-I and ROCK-II coop-
eratively regulate closure of eyelid and ventral body
wall in mouse embryo. Genes Cells. 10:825–834.
25. Mukai, Y., et al. 2006. Decreased vascular lesion
formation in mice with inducible endothelial-spe-
cific expression of protein kinase Akt. J. Clin. Invest.
26. Mayadas, T.N., Johnson, R.C., Rayburn, H., Hynes,
R.O., and Wagner, D.D. 1993. Leukocyte rolling
and extravasation are severely compromised in P
selectin-deficient mice. Cell. 74:541–554.
27. Tavor, S., et al. 2002. Macrophage functional matu-
ration and cytokine production are impaired in C/
EBP epsilon-deficient mice. Blood. 99:1794–1801.
28. Bradley, M.N., et al. 2007. Ligand activation of LXR
beta reverses atherosclerosis and cellular cholester-
ol overload in mice lacking LXR alpha and apoE.
J. Clin. Invest. 117:2337–2346.
29. Tanaka, K., Sata, M., Hirata, Y., and Nagai, R. 2003.
Diverse contribution of bone marrow cells to neo-
intimal hyperplasia after mechanical vascular inju-
ries. Circ. Res. 93:783–790.
30. Geary, R.L., Kohler, T.R., Vergel, S., Kirkman,
T.R., and Clowes, A.W. 1994. Time course of flow-
induced smooth muscle cell proliferation and inti-
mal thickening in endothelialized baboon vascular
grafts. Circ. Res. 74:14–23.
31. Wolfrum, S., et al. 2004. Inhibition of Rho-kinase
leads to rapid activation of phosphatidylinositol
3-kinase/protein kinase Akt and cardiovascular pro-
tection. Arterioscler. Thromb. Vasc. Biol. 24:1842–1847.
32. Sawada, N., et al. 2000. Inhibition of rho-associ-
ated kinase results in suppression of neointimal
formation of balloon-injured arteries. Circulation.
33. Shibata, R., et al. 2001. Role of Rho-associated
kinase in neointima formation after vascular injury.
34. Oguchi, S., et al. 2000. Monoclonal antibody
against vascular cell adhesion molecule-1 inhibits
neointimal formation after periadventitial carotid
artery injury in genetically hypercholesterolemic
mice. Arterioscler. Thromb. Vasc. Biol. 20:1729–1736.
1644?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 118 Number 5 May 2008
35. Yasukawa, H., Imaizumi, T., Matsuoka, H.,
Nakashima, A., and Morimatsu, M. 1997. Inhibi-
tion of intimal hyperplasia after balloon injury by
antibodies to intercellular adhesion molecule-1
and lymphocyte function-associated antigen-1.
36. Anwar, K.N., Fazal, F., Malik, A.B., and Rahman, A.
2004. RhoA/Rho-associated kinase pathway selec-
tively regulates thrombin-induced intercellular adhe-
sion molecule-1 expression in endothelial cells via
activation of I kappa B kinase beta and phosphoryla-
tion of RelA/p65. J. Immunol. 173:6965–6972.
37. Worthylake, R.A., and Burridge, K. 2003. RhoA and
ROCK promote migration by limiting membrane
protrusions. J. Biol. Chem. 278:13578–13584.
38. Ross, R., et al. 1990. Localization of PDGF-B pro-
tein in macrophages in all phases of atherogenesis.
39. Kumar, A., and Lindner, V. 1997. Remodeling with
neointima formation in the mouse carotid artery
after cessation of blood flow. Arterioscler. Thromb.
Vasc. Biol. 17:2238–2244.
40. Sakata, Y., et al. 2004. Transcription factor CHF1/
Hey2 regulates neointimal formation in vivo
and vascular smooth muscle proliferation and
migration in vitro. Arterioscler. Thromb. Vasc. Biol.
41. Harari, O.A., Alcaide, P., Ahl, D., Luscinskas, F.W.,
and Liao, J.K. 2006. Absence of TRAM restricts Toll-
like receptor 4 signaling in vascular endothelial cells
to the MyD88 pathway. Circ. Res. 98:1134–1140.
42. Luscinskas, F.W., et al. 1994. Monocyte rolling,
arrest and spreading on IL-4-activated vascular
endothelium under flow is mediated via sequential
action of L-selectin, beta 1-integrins, and beta 2-
integrins. J. Cell Biol. 125:1417–1427.