Control of endothelial cell proliferation and migration
by VEGF signaling to histone deacetylase 7
Shusheng Wang*, Xiumin Li*, Maribel Parra†, Eric Verdin†, Rhonda Bassel-Duby*, and Eric N. Olson*‡
*Department of Molecular Biology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9148; and†Gladstone
Institute of Virology and Immunology, University of California, San Francisco, CA 94158
Contributed by Eric N. Olson, March 24, 2008 (sent for review March 12, 2008)
and migration. However, the nuclear mediators of the actions of
VEGF in ECs have not been fully defined. We show that VEGF
induces the phosphorylation of three conserved serine residues in
histone deacetylase 7 (HDAC7) via protein kinase D, which pro-
motes nuclear export of HDAC7 and activation of VEGF-responsive
genes in ECs. Expression of a signal-resistant HDAC7 mutant
protein in ECs inhibits proliferation and migration in response to
VEGF. These results demonstrate that phosphorylation of HDAC7
serves as a molecular switch to mediate VEGF signaling and
angiogenesis ? PKD1 ? MEF2 ? class II HDAC
numerous vascular disorders (1). In response to angiogenic
to form primitive vascular labyrinths that undergo maturation
and remodeling, accompanied by recruitment of smooth muscle
cells to give rise to mature blood vessels. VEGF plays a key role
in angiogenesis by regulating the proliferation, migration, and
survival of ECs (2). The sensitivity of ECs to VEGF signaling is
exemplified by the lethal vascular abnormalities that result from
disruption of even a single VEGF allele in mice (3, 4). Other
peptide growth factors, such as fibroblast growth factor (FGF),
angiopoietin, transforming growth factors (TGFs), TNF-?, ape-
lin, and insulin-like growth factors (IGFs), also influence endo-
thelial proliferation and function (5, 6). However, most of these
factors act on other cell types and complement or coordinate
VEGF signaling rather than function as independent regulators
of angiogenesis and EC functions.
The binding of VEGF to its receptors induces receptor
dimerization and autophosphorylation, which activates several
downstream kinases, including protein kinases C and D (PKC
and PKD), phosphatidylinositol 3-kinase (PI3K), and MAPK
(2). VEGF signaling also leads to the activation of numerous
genes, such as those encoding regulator of calcineurin 2
(RCAN2) [formerly Down Syndrome Candidate Region 1
(DCSR1L1)] (7) and the nuclear receptor Nur77 (8), which
mediate the biological responses to VEGF. However, how
VEGF signals to the nucleus to regulate gene transcription and
angiogenesis is far from clear.
Histone acetyltransferases and histone deacetylases (HDACs)
are key regulators of chromatin structure and gene expression
(9). Mammalian class IIa HDACs (HDAC4, -5, -7, and -9)
contain an N-terminal extension that interacts with other tran-
scriptional cofactors and confers responsiveness to extracellular
signals (10). Phosphorylation of a series of conserved serine
residues in the N-terminal regulatory domain of class IIa
HDACs by calcium/calmodulin-dependent kinase (CaMK) and
PKD creates docking sites for 14-3-3 chaperone proteins, which
drives these HDACs from the nucleus to the cytoplasm and
derepresses gene expression (11–14). The MEF2 transcription
factor is a key downstream target for repression by class IIa
HDACs and for signal-dependent transcription in response to
he formation of blood vessels through the process of angio-
genesis is critical for normal vascular development and
signaling pathways that promote class IIa HDAC phosphoryla-
The functions of class IIa HDACs in vivo have been revealed
by gene knockout studies. HDAC5 and -9 function as stress-
responsive inhibitors of cardiac growth (16, 17), HDAC4 re-
presses chondrocyte growth and differentiation (18), and
HDAC7 maintains vascular integrity by repressing matrix me-
talloprotease 10 (MMP10) expression in ECs (19). A role for
HDAC7 in modulating angiogenesis has also been suggested
from in vitro assays (19, 20). However, whether VEGF acts
through HDAC7 to control angiogenesis and endothelial func-
tions is unknown.
Here, we show that HDAC7 and other class IIa HDACs are
phosphorylated and exported from the nucleus to the cytoplasm
in response to VEGF signaling. Moreover, blockade of HDAC7
phosphorylation with a signal-resistant HDAC7 mutant re-
presses EC proliferation and migration in response to VEGF.
Signaling by VEGF to HDAC7 regulates both MEF2-dependent
and independent target genes in ECs. Our results reveal a key
role for HDAC7 as a VEGF-dependent molecular switch that
governs EC functions.
Effects of Peptide Growth Factors on HDAC7 Phosphorylation in ECs.
To begin to explore the signaling pathways that regulate the
activity of class IIa HDACs in ECs, we compared the expression
of HDAC4, -5, -7, and -9 in human umbilical vein ECs
(HUVECs) and human aortic ECs (HAECs). HDAC4 and -7
mRNAs are abundant in these two cell types, whereas HDAC5
and -9 are expressed at relatively low levels [supporting infor-
mation (SI) Fig. S1]. Because HDAC7 is EC-restricted (19), we
tested a variety of growth factors known to modulate EC
functions for their effects on HDAC7 phosphorylation in ECs.
Three different phospho-HDAC7 antibodies, which specifically
recognize the conserved phosphoserines in HDAC7, were used
in the assay. These sites correspond to serines 178, 344, and 479
in mouse HDAC7 (Fig. 1A). The specificity of these antibodies
has been documented (21). HAEC cells were infected with
adenovirus-expressing mouse HDAC7 for 48 h, starved for 24 h
with low serum medium, and treated with different growth
factors. As shown in Fig. 1B, VEGF-A robustly induced HDAC7
phosphorylation at all three serine residues. Endothelin-1,
TGF-?, IGF, and FGF-2 also modestly induced HDAC7 phos-
patterns of phosphorylation. Apelin, another angiogenic factor
(5), failed to induce HDAC7 phosphorylation, possibly due to
the lack of its receptor in the HAEC cell line (5).
Author contributions: S.W. and E.N.O. designed research; S.W. and X.L. performed re-
analyzed data; and S.W., R.B.-D., and E.N.O. wrote the paper.
The authors declare no conflict of interest.
‡To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
June 3, 2008 ?
vol. 105 ?
Mutation of the three serine residues to alanine in HDAC7
(HDAC7-S/A) abolished inducible and basal phosphorylation
(Fig. 2D, lane 4). These results indicate that HDAC7 is phos-
phorylated in response to multiple growth factors in ECs,
whereas VEGF-A stands out as the most robust inducer of
VEGF rapidly induced HDAC7 phosphorylation with the
strongest response observed 5–10 min after VEGF treatment
(Fig. 1C). Phosphorylation of HDAC7 was paralleled by binding
to 14-3-3, as detected by coimmunoprecipitation. Phosphoryla-
tion of HDAC7 in response to VEGF-A occurred in a dose-
dependent manner with a maximum at 20 ng/ml (data not
shown). We conclude that VEGF-A dynamically regulates
HDAC7 phosphorylation in ECs.
The PKC/PKD Pathway Is Necessary and Sufficient for HDAC7 Phos-
phorylation by VEGF. To identify the kinases that phosphorylate
HDAC7 in response to VEGF, HAEC cells were infected with
HDAC7 virus and pretreated with kinase inhibitors before
VEGF-A treatment. The general serine/threonine kinase inhib-
itor staurosporine completely blocked HDAC7 phosphorylation
(data not shown). 1,2-bis(2-aminophenoxy)ethane-N,N,N?,N?-
tetraacetic acid-acetoxymethyl ester (BAPTA/AM) (a Ca2?
chelator), KN93 (a CaMK inhibitor), KN62 (a CaMKII-specific
inhibitor), and autocamtide inhibitory peptide II-2 (AIP II-2)
did not affect VEGF-induced HDAC7 phosphorylation (Fig. 2A
and data not shown), indicating that Ca2?and CaMKs are not
responsible for VEGF-induced HDAC7 phosphorylation. In
gram of HDAC7 showing the three signal-responsive serines in the N-terminal
regulatory domain. NLS, nuclear localization sequence. (B–D) HAEC cells were
infected with adenovirus expressing FLAG-HDAC7 for 48 h, transferred to
EBM-2 medium with 0.1%FBS for 24 h, and then treated with growth factors.
Antibodies recognizing HDAC7 phosphorylated at Ser-178, Ser-344, and Ser-
overexpressed HDAC7 was monitored by antibody against FLAG tag. The cells
were treated with: (B) 20 ng/ml IGF-1; 2.5 ng/ml TGF?; 10 ng/ml VEGF-A; 10?8
M endothelin-1; 1 ng/ml FGF-2; 10 ng/ml Apelin or no treatment (control) for
30 min, or (C) 10 ng/ml VEGF-A for the indicated times. A portion of the cell
lysate was immunoprecipitated with monoclonal anti-FLAG conjugated to
14–3-3 protein (IB).
Regulation of HDAC7 phosphorylation by VEGF. (A) Schematic dia-
(A and B) HAEC cells were infected with adenovirus expressing FLAG-HDAC7
for 48 h, transferred to EBM2 medium with 0.1% FBS for 24 h, treated with
kinase inhibitors, and then treated with 20 ng/ml of VEGF-A for 10 min.
Antibodies recognizing HDAC7 phosphorylated at Ser-178, Ser-344, and Ser-
479 were used to detect the phosphorylation state of HDAC7. The total
amount of overexpressed HDAC7 was monitored by antibody against FLAG
tag. The kinase inhibitors used were: (A) 30 ?M BAPTA/AM or 5 ?M KN93; (B)
500 nM Go ¨6983, 1.3 ?M Go ¨6976; DMSO is the control for the drug vehicle; (C)
HAEC cells were transfected with control siRNA or siRNA against PKD1 and
infected with adenovirus expressing FLAG-HDAC7, transferred to EBM-2 me-
HDAC7 phosphorylated at Ser-479 was used to detect the phosphorylation
state of HDAC7. The efficiency of si-PKD1 was examined by Western blot
analysis with PKD1 antibody. The total amount of overexpressed HDAC7 was
monitored by antibody against FLAG tag. Loading control is indicated by
antibody against GAPDH. (D) HAEC cells were infected with or without
adenoviruses expressing constitutively active PKD1 (Myc-tagged) and FLAG-
with VEGF-A. Western blot analysis was performed by using antibodies rec-
coimmunoprecipitation, the protein lysates were immunoprecipitaed with
anti-FLAG or anti-Myc antibodies, and immunoblotted with antibodies
against Myc, PKD or FLAG.
PKD is necessary and required for HDAC7 phosphorylation by VEGF.
Wang et al.
June 3, 2008 ?
vol. 105 ?
no. 22 ?
contrast, the PKC inhibitor Go ¨6983 and the Ca2?-dependent
PKC and PKD inhibitor Go ¨6976 (22) inhibited VEGF-induced
HDAC7 phosphorylation (Fig. 2B). Because the BAPTA/AM
studies suggested that calcium is not required for VEGF-induced
HDAC7 phosphorylation, these findings suggest that atypical
PKC, a known regulator of PKD, is required for VEGF-induced
HDAC7 phosphorylation, consistent with a recent report that
VEGF can induce PKC dependent PKD phosphorylation (23).
To directly test whether PKD is required for VEGF-induced
HDAC7 phosphorylation, a siRNA oligonucleotide targeted to
PKD1 or a control siRNA pool was transfected into HAEC cells.
The knockdown efficiency of si-PKD1 was confirmed by West-
ern blot analysis against endogenous PKD1 (Fig. 2C). Phosphor-
ylation of HDAC7 in response to VEGF was repressed by
levels were unaffected.
To further test whether PKD is sufficient to induce HDAC7
phosphorylation, HDAC cells were infected with adenovirus-
expressing Myc-tagged constitutively active PKD (PKD-CA),
adenovirus-encoding FLAG-tagged HDAC7 or the signal-
resistant HDAC7 mutant HDAC7-S/A, and HDAC7 phosphor-
ylation was examined. As shown in Fig. 2D, PKD-CA induced
HDAC7 phosphorylation to comparable levels as VEGF treat-
ment. In contrast, HDAC7-S/A was completely resistant to
phosphorylation by PKD-CA and VEGF. Coimmunoprecipita-
tion assays were performed to test whether PKD directly asso-
ciated with HDAC7 in ECs. HDAC7 interacted weakly with
endogenous PKD, without a detectable increase upon VEGF
treatment. Interestingly, the HDAC7-S/A mutant protein inter-
acted with PKD much more strongly than did the wild-type
HDAC7 protein (Fig. 2D). We interpret these findings to
HDAC7, whereas the nonphosphorylated HDAC7 mutant
forms a stable complex with PKD.
VEGF Induces HDAC7 Nuclear Export in ECs.Phosphorylation of class
IIa HDACs creates binding sites for the 14-3-3 chaperone
protein, which drives their export from the nucleus to the
cytoplasm (11–14, 16, 17, 24). As shown in Fig. 3A, HDAC7 is
exclusively localized in the nucleus in serum-starved HAEC cells
in the absence of VEGF. HDAC7 nuclear export was observed
beginning 30 min after VEGF treatment. The percentage of cells
with cytoplasmic HDAC7 continued to increase thereafter and
reached a maximum of ?70% after 4 h of VEGF treatment.
HDAC7 then gradually relocalized to the nucleus and was
primarily nuclear 24 h after VEGF treatment. Thus, ECs respond
to VEGF signaling by a wave of nucleocytoplasmic HDAC7
shuttling, reflecting the dynamic effects of VEGF on ECs.
Consistent with the requirement of the PKC/PKD pathway for
HDAC7 phosphorylation, Go ¨6976 inhibited HDAC7 nuclear
export in response to VEGF, whereas KN93 failed to block
VEGF-induced HDAC7 nuclear export (Fig. 3A). PKD-CA also
drove HDAC7 into the cytoplasm independent of VEGF treat-
ment (Fig. 3B). In contrast, HDAC7-S/A was localized to the
nucleus regardless of VEGF treatment and was insensitive to
PKD-CA. These results demonstrate that serines 178, 344, and
479 in HDAC7 are critical for VEGF-induced HDAC7 nucleo-
VEGF Induces Phosphorylation of HDAC4, -5, and -9. We further
tested the response of other class IIa HDACs to VEGF. An
antibody against phosphoserines 259 and 498 in HDAC5 also
recognizes the corresponding phosphoserines in HDAC4 and
MEF2-interacting transcription repressor (MITR) (a splice vari-
ant of HDAC9) (13). The specificity of the antibody was
demonstrated by the lack of signal with HDAC4-S/A and
MITR-S/A mutants (Fig. S2).
Serum-starved HAEC cells showed a basal level of phosphor-
ylation of HDAC4 and -5 and MITR, which was strongly
enhanced by VEGF (Fig. S2). Phosphorylation of HDAC4 and
-5 and MITR in response to VEGF was nearly abolished by the
PKC/PKD inhibitor Go ¨6976 but not the CaMK inhibitor KN93.
Moreover, PKD-CA alone potently induced HDAC4 and -5 and
HDAC4 and -5 and MITR were primarily located in the
nucleus in unstimulated HAEC cells (Fig. S3 A–C). HDAC4 and
-5 and MITR became significantly more cytosolic after 4 h of
-5 and MITR nuclear export independent of VEGF treatment.
Mutation to alanine of the key serine residues in HDAC4
(serines 246, 467, and 632), HDAC5 (serines 259 and 498), and
MITR (serines 218 and 448) rendered them irresponsive to
VEGF and PKD, demonstrating that these conserved residues in
HDAC4 and -5 and MITR are required for VEGF-induced
nuclear export in ECs.
HDAC7 Is Required for VEGF-Induced EC Proliferation and Migration.
The PKC/PKD pathway regulates EC proliferation in response
to VEGF (23). To study the functional significance of VEGF-
induced HDAC7 phosphorylation, we overexpressed HDAC7
and HDAC7-S/A in ECs and examined cell proliferation in
response to VEGF. The expression level of HDAC7 and
HDAC7-S/A was quantified by Western blot analysis (Fig. 4A).
As shown in Fig. 4B, overexpression of HDAC7-S/A, but not
HDAC7, blunted cell proliferation in response to VEGF as
assayed by [3H]-thymidine incorporation. These results demon-
strate that the VEGF-responsive phosphorylation sites in HDAC7
are critical for VEGF-induced endothelial proliferation.
presence or absence of Myc-PKD1-CA were treated with VEGF-A (10 ng/ml) as
indicated. Immunocytochemistry was performed by using antibody against
FLAG (red). Nuclei were stained with DAPI (blue). (C) Quantitative analysis of
the percentage of cytoplasmic HDAC7 after treatment with VEGF-A for the
indicated times (0–24 hr) without or with prior treatment of indicated inhib-
itor. KN93 (CaMK inhibitor, 5 ?M); Go ¨6976 (PKD inhibitor, 1.3 ?M).
VEGF induces HDAC7 nuclear export. HAEC cells infected with
www.pnas.org?cgi?doi?10.1073?pnas.0802857105 Wang et al.
The impact of HDAC7 on cell migration was examined by a
scratch-wound assay in which EC migration depends on VEGF.
Overexpression of HDAC7 in HAEC cells did not significantly
affect EC migration relative to a LacZ control (Fig. 4C).
However, overexpression of HDAC7-S/A significantly repressed
VEGF-induced EC migration. Twenty-four hours after wound-
ing, compared with the LacZ control, overexpression of HDAC7
and HDAC7-S/A in ECs resulted in 20% and 85% decreases in
the number of cells migrating to the scratched area in response
to VEGF, respectively. These results suggest that signaling from
VEGF to HDAC7 is required for VEGF-induced EC migration.
Importantly, ECs expressing HDAC7-S/A showed no signs of
toxicity. Failure of HDAC7 to be phosphorylated by VEGF
imposes a blockade to cell migration, suggesting that key genes
required for migration are irreversibly repressed by the signal-
resistant HDAC7 mutant.
HDAC7 Regulates VEGF-Induced Gene Expression in ECs. RCAN2 and
Nur77 are among the early VEGF response genes implicated in
angiogenesis (7, 25). To test whether these genes are controlled
by HDAC7 downstream of VEGF signaling, and whether the
interaction between HDAC7 and MEF2, a well defined target
for repression by HDAC7, confers VEGF responsiveness to
these genes, HAEC cells were infected with virus expressing
HDAC7, HDAC7-S/A, or HDAC7-S/A-?MEF, which lacks the
MEF2-binding domain. LacZ was used as a control. Real-time
PCR was performed to examine the level of candidate gene
expression after 1 h of VEGF treatment. GAPDH and cyclo-
philin A, which were not significantly affected by VEGF treat-
ment, were used as controls.
RCAN2 and Nur77 were induced ?8- and 40-fold, respec-
tively, upon 1 h of VEGF treatment (Fig. 5). Although overex-
pression of wild-type HDAC7 did not significantly affect the
expression of RCAN2 and Nur77, presumably because it is
phosphorylated and rendered inactive by VEGF, overexpression
of HDAC7-S/A blunted VEGF-induced RCAN2 and Nur77
expression. These results indicate that phosphorylation of
HDAC7 is required for expression of RCAN2 and Nur77 in
response to VEGF. HDAC7-S/A-?MEF failed to effectively
were infected with adenovirus expressing (A) FLAG-HDAC7 or FLAG-HDAC7-
with adenovirus expressing lacZ (Ad-LacZ), FLAG-HDAC7 (Ad-HDAC7), or
FLAG-HDAC7-S/A (Ad-HDAC7-S/A) and subjected to cell proliferation and (C)
migration assays in response to VEGF treatment. (P value ?0.007 between
Ad-LacZ (or Ad-HDAC7) and Ad-HDAC7-S/A). Of note, there are slightly less
cells in Ad-HDAC7-S/A infected samples compared with the other two con-
trols. The results in B and C are representative of three independent
Regulation of EC proliferation and migration by HDAC7. HAEC cells
HAEC cells infected with adenovirus expressing lacZ (Ad-LacZ), FLAG-HDAC7
(Ad-HDAC7), FLAG-HDAC7-S/A (Ad-HDAC7-S/A), or HDAC7-S/A-?MEF (Ad-
HDAC7-S/A-?MEF) were treated with 10 ng/ml VEGF for 1 h. Real-time quan-
tative RT-PCR was performed to examine the expression level of (A) RCAN2
and (B) Nur77. (C–E) HAEC cells transfected with control siRNA pool or siRNA
Expression of (C) HDAC7, (D) RCAN2, and (E) Nur77 was determined by
real-time RT-PCR. The results are representative of three independent exper-
iments. The P value shown was calculated by Student’s t test.
Inhibition of VEGF-dependent gene expression by HDAC7. (A and B)
Wang et al.
June 3, 2008 ?
vol. 105 ?
no. 22 ?
block expression of Nur77 but was as effective as HDAC7-S/A
in blocking VEGF-induced expression of RCAN2, suggesting
that VEGF-induced Nur77 expression depends on MEF2,
These results indicate that HDAC regulates VEGF-induced
gene expression in both a MEF2-dependent and independent
We further tested the effect of HDAC7 knockdown on
VEGF-induced gene expression. siRNA was used to knockdown
HDAC7 in HAEC cells (Fig. 5C), causing expression of RCAN2
and Nur77 to increase by 2-fold in the absence of VEGF
treatment (Fig. 5 D and E). Moreover, knockdown of HDAC7
decreased the magnitude of induction of RCAN2 and Nur77
expression in response to VEGF compared with cells transfected
with control siRNA (Fig. 5 D and E). These data indicate that
the presence of HDAC7 is required for repressing VEGF-
regulated RCAN2 and Nur77 expression. The reduced effect of
siHDAC7 on gene expression might reflect partial HDAC7
knockdown efficiency and redundant roles of other class IIa
HDACs in regulating RCAN2 and Nur77 expression by VEGF.
In addition, the reduction of HDAC7 might up-regulate other
repressors, such as HADC9 (27), which might compromise the
VEGF effect on gene expression.
The results of this study demonstrate that VEGF activates the
PKD signaling pathway, which modulates HDAC7 phosphory-
lation and nuclear export and in turn the migration and prolif-
eration of ECs by regulating VEGF-responsive gene expression
Regulation of VEGF Signaling by HDAC7 in ECs.HowVEGFsignaling
is dynamically regulated in the nucleus is a subject of great
interest. Our results show that VEGF-A rapidly induces HDAC7
phosphorylation and nuclear export, which promotes EC pro-
liferation and migration by relieving HDAC7-dependent repres-
sion of VEGF responsive genes. Maximal HDAC7 phosphory-
lation occurs ?5–10 min after VEGF-A stimulation, and
HDAC7 nuclear export is maximized ?4 h later. Both HDAC7
phosphorylation and nuclear export then decrease with time,
with HDAC7 becoming relocalized to the nucleus after 24 h of
VEGF treatment. These findings indicate that HDAC7 acts as a
molecular switch to dynamically regulate VEGF signaling. The
ability of HDAC7 to mediate VEGF signaling is consistent with
recent studies showing that HDAC7 is critical for angiogenesis
that other peptide growth factors, which complement or coor-
dinate the actions of VEGF in regulating endothelial functions,
also induce HDAC7 phosphorylation, albeit to a reduced level
for multiple signaling pathways. Consistent with this notion, the
T cell receptor has been shown to regulate apoptosis in thymo-
cytes via phosphorylation and nuclear export of HDAC7
Activation of the PKC/PKD pathway is necessary and suffi-
cient to mediate VEGF-A-induced HDAC7 phosphorylation
and nucleocytoplasmic shuttling, consistent with a prior report
that VEGF can induce PKC-dependent PKD phosphorylation in
founding member of a the PKD family, which also includes
PKD2 and PKD3/PKC? (30). PKD1 has been identified as class
IIa HDAC export kinase in cardiomyocytes and thymocytes (29,
31, 32). PKD has been implicated in diverse processes, such as,
vesicular transport, metastasis, immune responses, apoptosis,
and cell proliferation (30). Our data and those of others showed
migration (23, 33), indicating an important role for PKD in EC
Role of Class II HDACs in VEGF-Induced Angiogenesis. Our results
indicate that HDAC7 controls EC functions by repressing the
expression of MEF2-dependent and independent genes. MEF2
likely plays a role in the regulation of VEGF-induced EC
proliferation and migration downstream of the PKC/PKD path-
way. Prior studies have implicated MEF2 in the control of EC
survival downstream of the MAPK pathway (34, 35).
MMP10, a gene regulated by HDAC7 in ECs, was repressed
by HDAC7-S/A but not significantly regulated by 1 h of VEGF
treatment (data not shown). Similarly, calcineurin-binding pro-
tein (calsarcin-1) and growth arrest-specific 1 (GAS1), which
have been shown to be up-regulated in HUVEC cells after
siRNA-mediated knockdown of HDAC7 (19), were not signif-
icantly regulated by 1 h of VEGF treatment (data not shown).
These findings suggest that the regulation of these genes by
HDAC7 may reflect a late effect of VEGF on HDAC7. Alter-
natively, HDAC7 might also regulate VEGF-independent gene
expression in ECs.
Although we have focused on the regulation of HDAC7
phosphorylation by VEGF because of the specificity and abun-
dant expression of this HDAC in ECs, our results indicate that
all class IIa HDACs are phosphorylated in response to VEGF.
Thus, it is likely that the responses of ECs to VEGF represent
the combined effects of phosphorylation of all class IIa HDACs.
However, in vivo functional studies by gene knockouts show that
only HDAC7 is required for maintaining endothelial function,
whereas the absence of the other HDACs does not evoke a
vascular phenotype under normal conditions (16–19). It is likely
that different Class IIa HDACs might also have distinct targets
and functions, although they are regulated similarly in response
role in regulating EC function.
Therapeutic Implications. Our finding that VEGF acts through
PKD to modulate HDAC7 phosphorylation and EC functions
raises interesting prospects for pharmacological manipulation of
the VEGF-PKD-HDAC7 axis in the settings of vascular disor-
ders. Up-regulation of VEGF-A underlies angiogenesis associ-
ated with diabetic retinopathy, tumor growth, metastasis, arthri-
tis, and atherosclerosis. Antiangiogenic therapy has shown
promise in suppression of tumor growth and other diseases
related to abnormal angiogenesis (2). Our findings suggest that
pharmacological inhibition of PKD should prevent HDAC7
HDAC7. In this model, VEGF induces phosphorylation and nuclear export of
HDAC7 via a PKC/PKD1 pathway. HDAC7 controls VEGF-induced EC prolifer-
ation and migration by regulating VEGF responsive gene expression.
A model for the control of EC gene expression by VEGF signaling to
www.pnas.org?cgi?doi?10.1073?pnas.0802857105Wang et al.
phosphorylation, thereby repressing EC proliferation and mi- Download full-text
gration. In this regard, development of specific PKD inhibitors
or inhibitors of HDAC phosphorylation may hold potential in
Materials and Methods
EC Culture and Immunocytochemistry. HAEC (Clonetics) and HUVEC (American
Type Culture Collection) cells were grown in EC growth medium (Clonetics/
Cambrex) supplemented with 2% FBS. Kinase inhibitor and/or growth factor
treatments of adenovirus infected cells are described in SI Text. For immuno-
cytochemistry, cells were plated on gelatin-coated glass coverslips and fixed
and stained as described (13).
Adenovirus Generation and Infection. Recombinant adenoviruses were gener-
ated as described in SI Text or in refs. 13, 14, and 17.
Western Blot Analysis and Immunoprecipitation. Western blot analysis was
performed according to standard procedures by using the ECL detection kit
(Amersham Pharmacia Biotech). Immunoprecipitation was performed as de-
scribed (36) and outlined in SI Text.
siRNA Transfection and Real-Time Quantitative RT-PCR. Details of transfection
with siRNA and real-time PCR are described in SI Text. Primers are listed in SI
EC Proliferation Assay and Scratch-Wound Assay. EC proliferation assays were
vitro scratch assay’’ (37) and as described in SI Text.
Note Added in Proof. During the preparation of this manuscript, Ha et al.
published a report (38), consistent with our data, showing that HDAC5 reg-
ulates VEGF signaling and angiogenesis.
ACKNOWLEDGMENTS. We thank S. Chang, M. Avkiran, and T. McKinsey for
scientific input and advice. We thank Jose Cabrera for graphics and Jennifer
Institutes of Health, the Donald W. Reynolds Cardiovascular Clinical Research
Center, and the Robert A. Welch Foundation. S.W. was supported by a
fellowship grant from the American Heart Association.
1. Risau W (1997) Mechanisms of angiogenesis. Nature 386:671–674.
2. Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L (2006) VEGF receptor signalling -
in control of vascular function. Nat Rev Mol Cell Biol 7:359–371.
lacking a single VEGF allele. Nature 380:435–439.
4. Ferrara N, et al. (1996) Heterozygous embryonic lethality induced by targeted inacti-
vation of the VEGF gene. Nature 380:439–442.
5. Cox CM, D’Agostino SL, Miller MK, Heimark RL, Krieg PA (2006) Apelin, the ligand for
the endothelial G-protein-coupled receptor, APJ, is a potent angiogenic factor re-
quired for normal vascular development of the frog embryo. Dev Biol 296:177–189.
6. Distler JH, et al. (2003) Angiogenic and angiostatic factors in the molecular control of
angiogenesis. Q J Nucl Med 47:149–161.
7. Qin L, et al. (2006) Down syndrome candidate region 1 isoform 1 mediates angiogen-
esis through the calcineurin-NFAT pathway. Mol Cancer Res 4:811–820.
8. Liu D, Jia H, Holmes DI, Stannard A, Zachary I (2003) Vascular endothelial growth
and the related nuclear receptors Nur77, Nurr1, and Nor1. Arterioscler Thromb Vasc
9. Jenuwein T, Allis CD (2001) Translating the histone code. Science 293:1074–1080.
10. Verdin E, Dequiedt F, Kasler HG (2003) Class II histone deacetylases: Versatile regula-
tors. Trends Genet 19:286–293.
11. Backs J, Song K, Bezprozvannaya S, Chang S, Olson EN (2006) CaM kinase II selectively
signals to histone deacetylase 4 during cardiomyocyte hypertrophy. J Clin Invest
12. Grozinger CM, Schreiber SL (2000) Regulation of histone deacetylase 4 and 5 and
transcriptional activity by 14–3-3-dependent cellular localization. Proc Natl Acad Sci
13. McKinsey TA, Zhang CL, Lu J, Olson EN (2000) Signal-dependent nuclear export of a
histone deacetylase regulates muscle differentiation. Nature 408:106–111.
14. Vega RB, et al. (2004) Protein kinases C and D mediate agonist-dependent cardiac
hypertrophy through nuclear export of histone deacetylase 5. Mol Cell Biol 24:8374–
15. Potthoff MJ, Olson EN (2007) MEF2: A central regulator of diverse developmental
programs. Development 134:4131–4140.
16. Chang S, et al. (2004) Histone deacetylases 5 and 9 govern responsiveness of the heart
to a subset of stress signals and play redundant roles in heart development. Mol Cell
17. Zhang CL, et al. (2002) Class II histone deacetylases act as signal-responsive repressors
of cardiac hypertrophy. Cell 110:479–488.
18. Vega RB, et al. (2004) Histone deacetylase 4 controls chondrocyte hypertrophy during
skeletogenesis. Cell 119:555–566.
19. Chang S, et al. (2006) Histone deacetylase 7 maintains vascular integrity by repressing
matrix metalloproteinase 10. Cell 126:321–334.
a key step in angiogenesis. Circ Res 101:1237–1246.
controls its nucleocytoplasmic shuttling, and inhibits apoptosis in thymocytes. Genes
22. Gschwendt M, et al. (1996) Inhibition of protein kinase C mu by various inhibitors.
Differentiation from protein kinase C isoenzymes. FEBS Lett 392:77–80.
23. Wong C, Jin ZG (2005) Protein kinase C-dependent protein kinase D activation mod-
ulates ERK signal pathway and endothelial cell proliferation by vascular endothelial
growth factor. J Biol Chem 280:33262–33269.
24. Kao HY, et al. (2001) Mechanism for nucleocytoplasmic shuttling of histone deacety-
lase 7. J Biol Chem 276:47496–47507.
25. Zeng H, et al. (2006) Orphan nuclear receptor TR3/Nur77 regulates VEGF-A-induced
angiogenesis through its transcriptional activity. J Exp Med 203:719–729.
26. Youn HD, Liu JO (2000) Cabin1 represses MEF2-dependent Nur77 expression and T cell
27. Haberland M, et al. (2007) Regulation of HDAC9 gene expression by MEF2 establishes
Cell Biol 27:518–525.
28. Dequiedt F, et al. (2003) HDAC7, a thymus-specific Class II histone deacetylase, regu-
lates Nur77 transcription and TCR-mediated apoptosis. Immunity 18:687–698.
29. Parra M, Kasler H, McKinsey TA, Olson EN, Verdin E (2005) Protein kinase D1 phos-
phorylates HDAC7 and induces its nuclear export after T-cell receptor activation. J Biol
30. Rykx A, et al. (2003) Protein kinase D: A family affair. FEBS Lett 546:81–86.
31. Harrison BC, et al. (2006) Regulation of cardiac stress signaling by protein kinase D1.
Mol Cell Biol 26:3875–3888.
32. Huynh QK, McKinsey TA (2006) Protein kinase D directly phosphorylates histone
deacetylase 5 via a random sequential kinetic mechanism. Arch Biochem Biophys
for VEGF-A165-induced angiogenesis through its interaction and regulation of phos-
pholipase Cgamma phosphorylation. J Biol Chem 281:32550–32558.
34. Olson EN (2004) Undermining the endothelium by ablation of MAPK-MEF2 signaling.
J Clin Invest 113:1110–1112.
35. Hayashi M, et al. (2004) Targeted deletion of BMK1/ERK5 in adult mice perturbs
vascular integrity and leads to endothelial failure. J Clin Invest 113:1138–
36. Wang S, et al. (2004) Chick Pcl2 regulates the left-right asymmetry by repressing Shh
expression in Hensen’s node. Development 131:4381–4391.
37. Liang CC, Park AY, Guan JL (2007) In vitro scratch assay: A convenient and inexpensive
method for analysis of cell migration in vitro. Nat Protoc 2:329–333.
38. Ha CH, et al. (2008) Protein kinase D-dependent phosphorylation and nuclear export
of histone deacetylase 5. J Biol Chem 283:14590–14599.
Wang et al.
June 3, 2008 ?
vol. 105 ?
no. 22 ?