p130Rb2and p27kip1cooperate to control mobilization
of angiogenic progenitors from the bone marrow
Anxo Vidala,b, Stergios Zacharoulisc,d,e,f,g, Wenjun Guoh, David Shaffera, Filippo Giancottih, Anna H. Bramleyd,e,f,g,
Carmen de la Hozi, Kristian K. Jensend,e,f,g, Daniel Katod,e,f,g, Daniel D. MacDonaldd,e,f,g, Joseph Knowlesd,e,f,g,
Nancy Yeha, Lawrence A. Frohmanj, Shahin Rafiik, David Lydend,e,f,g,l, and Andrew Koffa,l
Departments ofaMolecular Biology,cPediatrics, andhCellular Biochemistry, Memorial Sloan–Kettering Cancer Center, 1275 York Avenue,
New York, NY 10021;iDepartment of Cell Biology and Histology, School of Medicine and Dentistry, University of the Basque Country, Vizcaya,
48940 Leioa, Spain;dChildren’s Blood Foundation Laboratories and Departments ofePediatrics,fCell Biology,gGenetics, andkGenetic Medicine, Weill
Medical College at Cornell University, 515 East 71st Street, New York, NY 10021; andjSection of Endocrinology, University of Illinois, 1819
West Polk Street, MC 640, Chicago, IL 60612
Edited by Judah Folkman, Harvard Medical School, Boston, MA, and approved March 24, 2005 (received for review August 9, 2004)
elomonocytic and endothelial progenitor cells as well as endothe-
lial cells coopted from surrounding vessels. Cytokines induce these
cells to proliferate, migrate, and exit the cell cycle to establish the
vasculature; however, which cell cycle regulators play a role in
these processes is largely unknown. Here, we report that mice
lacking the cell cycle inhibitors p130 and p27 show defects in tumor
neoangiogenesis, both in xenografts and spontaneously arising
tumors. This defect is associated with impaired mobilization of
endothelial and myelomonocytic angiogenic progenitors from the
bone marrow. This article documents the role of these molecules in
angiogenesis and further suggests that cell expansion and mobi-
lization from the bone marrow of angiogenic precursors are
involves bothbone marrow-derivedmy-
angiogenesis ? cyclin-dependent kinase inhibitors ? stem cell
newly formed vessels (1). In addition to endothelial sprouting
from neighboring vessels, bone marrow-derived angiogenic cells
are recruited to facilitate formation and stability of the tumor
vasculature (2, 3). The angiogenic switch involves up-regulation
of cytokines such as VEGF from the tumor, which stimulates the
transfer of endothelial and hematopoietic stem cells from a
the mobilization of these cells into the circulation (4, 5). Circu-
lating endothelial precursors (CEPs), identified by expression of
VEGF receptor type 2 (VEGFR2), are considered to be em-
bryonic angioblasts that migrate, proliferate, and have the ability
to differentiate into mature endothelial cells within the tumor
bed. The angiogenic myeloid cells are identified by cell surface
expression of VEGF receptor type 1 (VEGFR1) and its sup-
porting function within the tumor bed is not fully understood.
Recent studies have shown the contribution of both VEGFR1?
and VEGFR2? cells to tumor neovascularization (4).
Despite the fact that angiogenic precursors have to undergo
proliferative changes to expand, mobilize, and differentiate, the
link between the cell cycle and the neoangiogenic process
remains poorly defined. Those molecules that have been iden-
tified are largely implicated to affect tumor cell properties, not
the vasculature. For example, several oncogenes and tumor
suppressors have been shown to affect angiogenesis by modu-
lating cytokine production (6–10). Indeed, even p130 has been
shown to suppress VEGF expression (11). However, the regu-
lation of cell expansion in the bone marrow and the need to
coordinate multiple cell types at the tumor site suggest the
possibility of further intersection between cell cycle regulators
and angiogenic signaling.
Here, we report that mice lacking the pocket protein p130 and
the cyclin-dependent kinase inhibitor p27 fail to properly form
new vessels, resulting in their inability to support the growth of
umors require activation of angiogenic pathways necessary
for recruitment of endothelial cells, the building blocks of
tumor xenografts. Defective neoangiogenesis is associated with
defects in the VEGF-induced expansion of VEGFR1? myeloid
cells and the VEGF-dependent mobilization of both VEGFR1?
and VEGFR2? endothelial precursors. Thus, these two cell
processes in tumor neoangiogenesis.
Materials and Methods
Animal and Xenograft Studies. Generation of p130?/?(12) and
mice were generated by interbreeding p130?/?p27?/?51animals.
Background controls were maintained in the same C57Bl6?SVJ
background in the same colony.
Tumor implantation experiments were performed as de-
scribed (4, 14). Sections of tissues were stained with hematox-
ylin-eosin or subjected to immunohistochemistry with antibodies
to CD31 (M-20, Santa Cruz Biotechnology).
For lumen measurements, images from randomly chosen
fields on at least six different sections from two independent
animals of each genotype were analyzed with METAMORPH 6.1
software (Universal Imaging, Downingtown, PA).
For transplants, donor bone marrow cells from 2-month-old
p130?/?p27?51/?51mice were injected i.v. into lethally irradiated
Rosa-26 (?-gal?) mice (14). Reciprocally, bone marrow from
Rosa-26 donors was injected into DKO recipients.
Analysis of Bone Marrow-Derived Cells. Mice were injected i.v. with
a single dose of 1.5 ? 108plaque-forming units of adenovirus
expressing VEGF (AdVEGF) or empty virus. Bone marrow and
mobilized cells were collected by flushing the femurs or retro-
orbital bleeding, respectively, and stained as described (15).
Antibodies against Sca-1 (E13.161.7 and D7) and CD11b (M1?
70) were from Pharmingen. Antibodies against mouse VEGFR1
and VEGFR2 (MF-1 and DC101, respectively) were from
ImClone Systems, New York.
Hematopoietic and endothelial colony assays were performed
as described (5, 15). 1,1?-Dioctadecyl-3,3,3?,3?-tetramethylindo-
carbocyanine-acetylated low-density lipoprotein? von Wille-
brand factor? colonies formed in the first 3 days after plating
were scored as early outgrowth endothelial colonies [colony-
This paper was submitted directly (Track II) to the PNAS office.
Freely available online through the PNAS open access option.
Abbreviations: CEP, circulating endothelial precursor; VEGFR1, VEGF receptor type 1;
VEGFR2, VEGF receptor type 2; DKO, double knockout; CFU, colony-forming unit; MLEC,
mouse lung endothelial cell; LLC, Lewis lung carcinoma; AdVEGF, adenovirus expressing
bPresent address: Department of Physiology, University of Santiago de Compostela, 15782
Santiago de Compostela, Spain.
lTo whom correspondence may be addressed. E-mail: email@example.com or dcl2001@
© 2005 by The National Academy of Sciences of the USA
May 10, 2005 ?
vol. 102 ?
forming unit (CFU)-EC early], whereas colonies formed after 2
weeks were considered late outgrowth endothelial colonies
Endothelial Cell Experiments. Mouse lung endothelial cells
(MLECs) were isolated by magnetic sorting using anti-rat IgG-
coated magnetic beads (M450, Dynal) coupled to anti-CD31
antibody (MEC13.3, Pharmingen) as described (16). Purity of
the isolated cells was consistently ?85% CD31? as assessed by
Proliferation was assessed by immunofluorescence with an
anti-BrdUrd mAb (Pharmingen).
Migration was measured by using a Boyden chamber assay,
and transmigrated cells were detected by staining with crystal
For cord formation assays, 50,000 cells were mixed with 5
mg?ml Matrigel (Becton Dickinson) and plated in complete
daily under the microscope. Cell extracts from MLECs were
resolved by SDS?PAGE, and proteins were transferred onto
poly(vinylidene difluoride) membranes and then incubated with
the appropriate antibodies for immunoblot. For kinase assay,
extracts were immunoprecipitated for cyclin E (M-20; Santa
Cruz Biotechnology) and kinase activity was measured by using
histone H1 as a substrate as described (17).
Statistical Analysis. Results are always expressed as the mean and
SD unless otherwise stated. Data were analyzed with unpaired
Student’s t test. P ? 0.05 was considered significant.
p130?/?p27?51/?51DKO Mice Exhibit Vascular Defects in the Pituitary.
In a C57?129 hybrid background, p130?/?mice undergo normal
development with no phenotypic abnormalities as described (12).
p27-deficient mice display organomegaly, infertility, and spontane-
These phenotypes were nearly identical to p130?/?p27?51/?51DKO
mice except that proliferation indices were increased in DKO
mice (Fig. 6, which is published as supporting information on the
PNAS web site). However, prominent vascular hemorrhages
resulting in widespread necrosis were seen in the pituitary
tumors arising in DKO mice compared with those in p27?51/?51
mice (Fig. 1). Well defined vessels, identified by staining for the
endothelial-specific marker CD31, were readily detected in
p27?51/?51but not in DKO tumors (Fig. 1). Similar results were
observed when staining for smooth muscle action (data not
shown). Thus, we investigated the possibility that p130 and p27
may be required for tumor angiogenesis.
Tumor Angiogenesis Is Impaired in Mice Lacking p130 and p27.Tumor
xenograft growth was measured in mice challenged with two
different tumor cell lines, B6RV2 lymphoma cells and Lewis
lung carcinoma (LLC) cells. After s.c. inoculation, WT mice
displayed a rapid increase in B6RV2 tumor size. Meanwhile, in
DKO mice, although there was an initial increase in growth,
tumors regressed (Fig. 2A). We next used corneal xenograft
assays. Normally, the immune system has no access to the cornea
and the area is avascular, allowing for easy detection of new
blood vessels. By day 7 postinoculation, implants of B6RV2 were
4?6; p27?51/?51, n ? 5?7; p130?/?, n ? 5?6), but no vascular
response or tumor growth was noted in DKO mice (n ? 0?6).
Tumors from LLC cells implanted s.c. grew at slower rates in
DKO mouse than in WT mice (data not shown). Widespread
to WT, although mature blood vessels were still present in the
viable peripheral rim of the tumor tissue. Smaller vessel size and
presence of scattered foci of CD31-positive endothelial cells in
the DKO were also noted (Fig. 2B). In fact, in two tumors grown
in DKO mice, the median vessel luminal area was 345 and 296
?m2(range from 44 to 6,405 ?m2), whereas in two tumors grown
in WT mice it was 456 and 497 ?m2(range from 92 to 18,125
?m2), a statistically significant difference (P ? 0.002) (Fig. 2C).
These findings suggest that p130 and p27 participated in either
VEGF-Induced VEGFR1? Cell Expansion Is Compromised in DKO Mice.
To determine whether the loss of p130 and p27 affected the
recruitment of VEGFR1? cells, we injected mice with adeno-
viral vectors expressing VEGF165[AdVEGF (15)] and measured
the number of VEGFR1? cells in the bone marrow and periph-
eral blood by two-color flow cytometry. In contrast to findings
with WT mice, the number of VEGFR1? bone marrow cells in
DKO mice did not increase in response to AdVEGF (Fig. 3E).
Similarly, in WT mice with AdVEGF injection there was a
marked increase in circulating VEGFR1?CD11b? cells,
whereas in DKO mice VEGFR1?CD11b? cells in the periph-
eral blood mononuclear cell (PBMC) pool increased to a lesser
extent (Fig. 3 A and B). Similar difference was observed when
PBMCs were cultured in vitro, and CFUs were counted (Table
1). These results suggested that the sustained response of
VEGFR1?CD11b? cells to VEGF required the presence of
p27, p130, or both.
Consistent with results from VEGF administration, when we
challenged animals with s.c. LLC xenografts, the proportion of
VEGFR1?CD11b? cells in the bone marrow increased in WT
mice, whereas cells in the marrow of DKO mice did not respond
to tumor stimulation (Fig. 7, which is published as supporting
information on the PNAS web site).
In unchallenged animals, VEGFR1?CD11b? cells represent
?0.01% of the PBMC population regardless of genotype. After
LLC inoculation in WT animals, the percentage of
VEGFR1?CD11b? myeloid cells increased in the PBMC from
day 2 to day 10, when the experiment was ended (Table 2, which
is published as supporting information on the PNAS web site).
However, in DKO animals, cells were detected only at day 10,
and even then it was only a weak response. This finding indicated
that in DKO mice the response of the VEGFR1? progenitors to
with hematoxylin-eosin (H&E, Top and Middle) or CD31 antibodies (Bottom).
Arrows show an area of necrosis. p27?51/?51, n ? 8; DKO, n ? 6. Genotypes are
shown above each column.
Vascular defects in p130?/?p27?51/?51pituitaries. Intermediate lobe
Vidal et al.PNAS ?
May 10, 2005 ?
vol. 102 ?
no. 19 ?
VEGF and to angiogenic signals initiated by tumor implantation
As in the WT, the proportion of VEGFR1?CD11b? cells
increased in the marrow of both single knockout mice (Fig. 7).
We also detected an increase of these cells in the peripheral
blood of both single knockout mice although these increases did
not parallel those seen in the bone marrow. We concluded that
p130 and p27 collaborated to ensure proper VEGF-induced
expansion of the VEGFR1? lineage in the bone marrow.
p27 and p130 Collaborate to Generate the Bone Marrow Niche for
VEGF-Induced Expansion of VEGFR1? Cells. To rule out a global
defect in bone marrow function in DKO mice, we performed
cellular blood counts in animals with or without VEGF stimu-
lation. Overall, we found no major differences in the white blood
cells, red blood cells, and platelet counts between WT and DKO
mice either unchallenged or after VEGF stimulation (Fig. 3C
and Fig. 8, which is published as supporting information on the
PNAS web site). However, in regards to the white blood cell
differential, the lymphocyte counts after VEGF administration
are lower in DKO mice than in WT (2.2 ? 1.1 vs. 4.0 ? 0.8 ?
103cells per ?l; P ? 0.04), a difference unlikely to explain the
defects seen in myeloid lineages in DKO mice. Additionally, we
found that the kinetics of recovery after 5-fluorouracil-induced
the colony-forming potential of bone marrow hematopoietic
precursors was similar (data not shown). Thus, the data suggest
that bone marrow function is not broadly impaired in DKO mice.
To determine whether the defect in expansion of VEGFR1?
cells was associated with the cellular component of the bone
marrow we transplanted donor DKO bone marrow into lethally
irradiated ROSA-26 WT animals and measured expansion of
VEGFR1? cells after VEGF stimulation as described above.
Myeloid VEGFR1? cells from DKO donors transplanted into
WT hosts were able to increase to the same extent as WT
controls in the bone marrow of reconstituted mice (Fig. 3E). In
contrast, the number of VEGFR1? cells from WT bone marrow
failed to increase in response to VEGF when transplanted into
DKO hosts (Fig. 3E). This finding indicated that the develop-
mental defect in DKO mice reflects the absence of an appro-
priate microenvironment or stem cell niche rather than a
cell-autonomous defect. Interestingly, even when the donor
DKO VEGFR1? cells were able to expand in the host WT bone
marrow, they failed to mobilize to the blood (Fig. 3F). This
finding indicated the existence of an additional defect in the
export of these cells to the peripheral blood. Thus, we concluded
that p27 and p130 were required both to generate the niche for
VEGFR1? hematopoietic cells to expand in response to VEGF
and for the mobilization of the VEGFR1? cells into the blood.
Defect in Mobilization of the Endothelial Precursors in DKO Mice.
Because myeloid VEGFR1? and endothelial VEGFR2? pre-
cursors are comobilized in response to the angiogenic switch, we
next determined whether the loss of p130 and p27 affected
VEGFR2? bone marrow-derived CEPs. After AdVEGF injec-
tion, VEGFR2? cells increased at similar extent in both WT and
DKO mice (Fig. 4A). Consistent with these results, no difference
was observed by staining bone marrow sections from WT and
DKO mice with antibodies to VEGFR2 (data not shown).
But in contrast, and similarly to what we found for VEGFR1?
cells, whereas elevation of plasma VEGF in WT mice increased
the numbers of VEGFR2? cells in the PBMC pool, in DKO
mice the response was drastically impaired (Fig. 4 B and C).
Comparable results were obtained after tumor inoculation,
where there was only a minimal response detectable by day 10
in DKO animals (Table 2). Accordingly, in vitro progenitor
assays showed a reduced number of CFU-EC colonies arising
from PBMC cultures (Table 1). Thus, the lack of VEGFR2?
cells in the blood of DKO mice suggests a mobilization defect
from the bone marrow, not a proliferative defect in the bone
marrow. Alternatively, decreased numbers of CEPs in the cir-
culation may reflect a shorter CEP half-life in peripheral blood.
To determine whether mobilization of VEGFR2? cells was
associated with exit from the cell cycle, we injected animals with
AdVEGF and examined the distribution of VEGFR2? cells in
the cell cycle by flow cytometry. In both WT and DKO mice
?35% of the VEGFR2? cells in the bone marrow were in S
phase, reflecting ongoing VEGF-dependent expansion. How-
ever, mobilized WT VEGFR2? cells accumulated mostly in G1
phase, suggesting that cell cycle exit accompanied or was a
prerequisite to mobilization from the bone marrow (data not
shown). The amount of cyclin E-associated kinase activity was
in the cell cycle may simply reflect the limited number of
export-competent cells in the VEGFR2? population at any
B6RV2 cells were inoculated into WT (n ? 6, gray bars) and DKO (n ? 3, red
bars) animals. Tumor size was measured with a caliper at the indicated time
points, and average tumor area with SD is shown. (B) Histology and vessel
cells, and tumors were removed 14 days later, fixed, sectioned, and stained
with either hematoxylin-eosin (H&E, Top and Middle) or antibodies for CD31
(Bottom) (WT, n ? 5; DKO, n ? 3). (C) METAMORPH analysis of smooth muscle
three intervals according to the luminal cross-sectional area, and the percent-
age in two independent tumors of each genotype is shown (WT, open bars;
at least four different sections of tumor.
Inhibition of xenograft growth and vascular defects in DKO mice. (A)
www.pnas.org?cgi?doi?10.1073?pnas.0405823102Vidal et al.
p130 and p27 Are Required for Endothelial Cell Differentiation. After
expansion and mobilization, recruited CEPs differentiate into
vessel-forming endothelial cells. The lack of detectable circulat-
ing CEPs in DKO mice precluded our study of their differenti-
ation, so we used primary endothelial cells. These cells are a
reasonable surrogate system as CEPs and mature endothelial
cells share many phenotypic features, such as expression of most
endothelial cell surface markers, migration, and the ability to
form cords in Matrigel (20–23). We isolated lung endothelial
cells from mice by using immunoaffinity adsorption to the
endothelial-specific marker CD31, and we found that p130 and
p27 accumulation correlated with Matrigel-induced endothelial
were necessary for differentiation of endothelial cell types, we
plated DKO MLEC in Matrigel and scored for cord formation.
WT, p130?/?, and p27?51/?51cells formed branched multicellular
vessel-like structures, whereas DKO cells were unable to assem-
ble into these structures and instead aggregated into groups of
extended cells (Fig. 5B), suggesting that p130 and p27 collabo-
rated during the morphological differentiation of endothelial
cells. Cyclin E-associated kinase activity was negligible in ex-
tracts from WT, p27?51/?51, and p130?/?differentiated cells, but
detectable in extracts from DKO cells (Fig. 5C). We were unable
to detect any difference in ability of WT and DKO cells to
migrate to growth factors (data not shown). Unlike WT cells,
DKO cells failed to exit the cell cycle under differentiation-
inducing conditions (Fig. 5F). This finding indicated that p130
and p27 are required to withdraw from the cell cycle in response
to differentiating signals.
In this article, we have established overlapping roles for cell cycle
regulators, p27 and p130, in multiple cellular responses in tumor
angiogenesis. These proteins are required for (i) the formation
of the bone marrow niche in which VEGFR1? cells develop in
response to angiogenic signaling, (ii) mobilization of both
VEGFR1? and VEGFR2? cells into the peripheral blood, and
(iii) proper endothelial differentiation. We provide evidence
Peripheral blood was collected after infection at days 3 and 5, and mobilized cells were identified by flow cytometry. (A) Representative profiles obtained for
mobilized precursors at day 3 are shown. The percent of VEGFR1?CD11b? cells is indicated in the boxes. (B) The number of VEGFR1? CD11b? cells detected
in the PBMC pool after adenoviral infection are shown. Data were compiled from two different experiments (six animals per group), and mean and SD is shown.
AdNull versus AdVEGF, P ? 0.05 at both times for WT; P ? 0.05 WT versus DKO at both times for AdVEGF. (C) White blood cell (WBC) counts (?103per ?l) of WT
and DKO mice 3 days after injection with AdNull or AdVEGF. (D) Hematopoietic recovery after 5-fluorouracil (5-FU) induced leucopenia. White blood cell counts
(WBC ?103per ?l) were determined at different days after injection of a single i.p. dose of 250 mg?kg 5-FU. (E and F) Bone marrow cells from mice were
transplanted into lethally irradiated mice as indicated below each column. Animals were then infected with AdVEGF, and the number of VEGFR1?CD11b?
myeloid cells was determined 3 days later in the bone marrow (E) and the peripheral blood (F). Results in E represent the fold increase compared with the basal
Defective expansion and mobilization of hematopoietic cells. (A and B) Mice were injected with AdVEGF or without transgene (AdNull) at day zero.
Table 1. No. of myelomonocytic and endothelial precursors in peripheral blood in response to
0.7 ? 0.6
1.0 ? 0.0
1.0 ? 1.0
1.0 ? 0.0
0.7 ? 0.6
1.7 ? 1.5
1.0 ? 0.0
1.0 ? 1.0
10.0 ? 0.0
11.0 ? 1.0
6.0 ? 1.7
12.0 ? 2.0
4.0 ? 1.0
2.0 ? 1.7
2.0 ? 0.0
3.0 ? 1.0
1.0 ? 1.0
1.0 ? 0.0
1.0 ? 0.0
1.0 ? 0.0
Peripheral blood was collected at day 0 (baseline) or day 3 after injection of the indicated adenoviruses.
Number of colony forming units per 100,000 plated PBMCs is shown (mean ? SD). Similar results were obtained
for day 5 after injection.
Vidal et al.PNAS ?
May 10, 2005 ?
vol. 102 ?
no. 19 ?
that VEGF-induced expansion of angiogenic progenitors in the
bone marrow is separable from the mobilization into the pe-
ripheral blood. As a consequence xenograft tumor growth is
The different behavior of VEGFR1? and VEGFR2? pro-
genitors regarding cell expansion suggests the existence of
different VEGF-responsive niches for different types of angio-
genic progenitors. Although we have not been able to identify a
particular cell type that defines this niche, we were able to rule
out a widespread defect in the establishment of the ‘‘bone
milieu’’ because VEGFR2? endothelial precursors expand in
DKO mice and hematopoiesis is mainly unaffected either in
unchallenged DKO mice or after myeloablation. Accordingly,
marrow. We have not noted any absence of SDF1, CXCR4, or
N-cadherin staining in DKO bone marrow (K.K.J. and D.L.,
unpublished data). Rather our results indicate that specific stem
cell populations could interact with particular niches, but we
recognize that other efforts to define a VEGFR1?-specific
niche must be undertaken.
To date, defects in recruitment of angiogenic precursors have
9?/?mice (4, 5). In those strains, defective expansion and
mobilization could not be distinguished from each other, and
bone marrow transplants rescued both phenotypes. Now, our
results in DKO mice allow us to uncouple expansion and
mobilization, suggesting that these are two separately regulated
processes. Transplanting DKO marrow into a WT recipient
rescues expansion of VEGFR1? cells, but these cells still fail to
be exported into the blood. On the other hand, VEGFR2?
blot of cell cycle regulators in extracts from MLECs grown in either gelatin-
coated dishes (AS) or after 4 days in Matrigel (M).*denotes a nonspecific
band cross-reactive to anti-p130 antibodies occasionally appearing in
mouse extracts. This experiment was done with three independent prep-
arations of cells. (B) MLECs purified by CD31 immunoaffinity were plated in
Matrigel, and cord formation was visualized by microscopy. The appear-
ance of representative cultures 4 days after plating is shown (WT, n ? 9;
p27?51/?51, n ? 6; p130?/?, n ? 6; DKO, n ? 8). (C) Cyclin E-associated histone
H1 kinase assay. Extracts from WT or DKO cells cultured in Matrigel for 4
days were immunoprecipitated with cyclin E-specific antiserum (?cyclin E)
or a rabbit anti-mouse (R?M) control antibody, and kinase activity was
measured by using histone H1 as a substrate. This experiment was repeated
three times with independent cultures. (D) MLEC from WT or DKO mice
were isolated and plated in Matrigel. After indicated times in culture (time
before labeling, TBL), cells were pulsed with 100 ?M BrdUrd for the
indicated time periods (labeling time, LT). After a total of 96 h in culture,
cells were recovered form the matrix by dispase digestion, and incorpora-
tion was scored by immunofluorescence as indicated in E (data from two
independent preparations; n ? 3 per preparation).
p130 and p27 are required for MLEC differentiation. (A) Immuno-
matched WT (n ? 6) and DKO mice (n ? 6) were injected with AdVEGF. Three
Mean and SD are shown for VEGF-injected mice (empty bars) and naı ¨ve
controls (filled bars). (B) Representative flow cytometric profiles obtained for
mobilized CEPs at day 3 after AdVEGF injection. (C) Number of
VEGFR1?CD11b? cells detected in the PBMC pool after adenoviral infection.
Data are compiled from two different experiments (six animals per group),
WT versus DKO for AdVEGF. (D) Effect of p130-p27 loss on cyclin E-cyclin-
dependent kinase 2 activity in VEGFR2? cells. Three days after infection with
AdVEGF, VEGFR2? cells were isolated from the bone marrow of WT and DKO
mice, and protein extract from an equal number of cells was subjected to
immunoprecipitation with antibodies to cyclin E. Immunoprecipitates were
twice with two pools of animals of each genotype.
Defective mobilization of endothelial precursors. (A) Sex- and age-
www.pnas.org?cgi?doi?10.1073?pnas.0405823102Vidal et al.
endothelial precursors expand in DKO marrow, but they fail to
efficiently exit into the peripheral circulation. Collectively, these
observations indicate that p27 and p130 act in a cell-autonomous
fashion to promote mobilization and that they can be used to
separate two VEGF-induced events, expansion and mobiliza-
tion, in angiogenic progenitor cell development.
However, the effects described here appear to be independent
of matrix metalloproteinase-9 because two of the defects ob-
served in matrix metalloproteinase-9-deficient mice were absent
in our DKO strain: first, the proliferative expansion of
VEGFR2? cells was unaffected; second, no differences in
hematopoietic recovery after chemically induced leukopenia
A common function of p27 and p130 is to promote cell cycle
exit, and we note that circulating VEGFR2? cells are enriched
in G0?G1phase of the cell cycle. Exit of cells from the cell cycle
is generally a prerequisite for differentiation, and there are many
examples of deregulated cyclin-dependent kinase activity inter-
fering with proper differentiation. Thus, we suggest that cell
cycle exit is contemporaneous with mobilization in angiogenic
progenitors and the increase in cyclin E-associated kinase ac-
tivity in the DKO cells may interfere with proper cell cycle arrest
and retard mobilization into the blood. We may not have
detected changes in cell proliferation by flow cytometry because
our ability to differentiate cells ready to mobilize from the bulk
of the VEGFR2? cells is limited by not having an appropriate
marker. Consequently, it remains possible that these proteins are
acting in a noncell cycle-related manner. Regardless, our data
indicate that accumulation of p27 and p130 uncouple prolifer-
ation within the bone marrow from mobilization into the pe-
ripheral blood and links mobilization to the down-regulation of
cyclin E–associated kinase activity by these proteins.
Furthermore, we found that p130 and p27 collaborate during
endothelial differentiation. As for VEGFR2? cells, the inability
of DKO cells to differentiate is associated with an increase in the
activity of cyclin E-associated kinase. This finding strongly
suggests that cyclin E is an important common target for p130
and p27 in angiogenesis.
Finally, one might wonder as to the significance of the
p130?p27-regulated events to tumor angiogenesis because spon-
taneous tumors still arose and grew in DKO animals. This result
contrasts with the dramatic effect on B6RV2 xenograft growth,
but is perhaps more similar to the findings with LLC xengorafts
where the nature of the tumor changed rather than the presence
or absence of tumor. Similar findings have been reported by
others (24, 25) and suggest that tumors, as well as xenografts,
have different degrees of dependency on angiogenic precursor
mobilization. In spontaneously arising pituitary tumors, we still
note a markedly different vasculature appearance, usually with
evidence of centralized necrosis, consistent with an angiogeni-
cally poor environment.
In conclusion, our observations support the notion that p27
and p130 cooperate to ensure angiogenic homeostasis at mul-
tiple levels. The significance of this overlap is well validated by
the observations that both spontaneous tumors and xenografts
display marked defects in angiogenesis. These proteins play an
important role in the control of stem?progenitor cell mobiliza-
tion in response to VEGF signaling, an event previously indis-
tinguishable from proliferation.
We thank J. Petrini (Memorial Sloan–Kettering Cancer Center), E.
Holland (Memorial Sloan–Kettering Cancer Center), K. Marians (Me-
morial Sloan–Kettering Cancer Center), R. Benezra (Memorial Sloan–
Kettering Cancer Center), R. Weiss (University of California, Davis), R.
Kalluri (Beth Israel Deaconess Medical Center, Boston), and K. Manova
(Memorial Sloan–Kettering Cancer Center) for discussions and critical
comments during the assembling of this work and S. Kerns (Weill
Medical College of Cornell University) and personnel at the Molecular
Cytology Core Facility (Memorial Sloan–Kettering Cancer Center) for
their assistance with graphics and illustration. This work was supported
by National Institutes of Health Grant CA89563, the Golfers Against
Cancer Foundation, and the Irma T. Hirschl Trust (to A.K.); the Bane
Foundation (L.A.F.); National Cancer Institute Grant CA98234-01, the
Children’s Blood Foundation, Emerald Foundation, Theodore A. Rapp
Foundation, Nicolaos Tzimas, and American Hellenic Educational
Progressive Association V District (to D.L.); and a National Cancer
Institute core grant to Memorial Sloan–Kettering Cancer Center. A.V.
was partially supported by the Ministerio de Educacion y Cultura, Spain.
D.S. acknowledges the support of CaPCURE.
1. Folkman, J. (2002) Semin. Oncol. 29, 15–18.
2. Rafii, S., Lyden, D., Benezra, R., Hattori, K. & Heissig, B. (2002) Nat. Rev.
Cancer 2, 826–835.
3. Rafii, S., Meeus, S., Dias, S., Hattori, K., Heissig, B., Shmelkov, S., Rafii, D.
& Lyden, D. (2002) Semin. Cell. Dev. Biol. 13, 61–67.
4. Lyden, D., Hattori, K., Dias, S., Costa, C., Blaikie, P., Butros, L., Chadburn,
A., Heissig, B., Marks, W., Witte, L., et al. (2001) Nat. Med. 7, 1194–
5. Heissig, B., Hattori, K., Dias, S., Friedrich, M., Ferris, B., Hackett, N. R.,
Crystal, R. G., Besmer, P., Lyden, D., Moore, M. A., et al. (2002) Cell 109,
6. Pal, S., Datta, K. & Mukhopadhyay, D. (2001) Cancer Res. 61, 6952–
7. Watnick, R. S., Cheng, Y. N., Rangarajan, A., Ince, T. A. & Weinberg, R. A.
(2003) Cancer Cell 3, 219–231.
8. Baudino, T. A., McKay, C., Pendeville-Samain, H., Nilsson, J. A., Maclean,
K. H., White, E. L., Davis, A. C., Ihle, J. N. & Cleveland, J. L. (2002) Genes
Dev. 16, 2530–2543.
9. Volpert, O. V., Pili, R., Sikder, H. A., Nelius, T., Zaichuk, T., Morris, C.,
Shiflett, C. B., Devlin, M. K., Conant, K. & Alani, R. M. (2002) Cancer Cell 2,
10. Lawler, J. (2002) J. Cell. Mol. Med. 6, 1–12.
11. Claudio, P. P., Stiegler, P., Howard, C. M., Bellan, C., Minimo, C., Tosi, G. M.,
Rak, J., Kovatich, A., De Fazio, P., Micheli, P., et al. (2001) Cancer Res. 61,
12. Cobrinik, D., Lee, M. H., Hannon, G., Mulligan, G., Bronson, R. T., Dyson, N.,
Harlow, E., Beach, D., Weinberg, R. A. & Jacks, T. (1996) Genes Dev. 10,
13. Kiyokawa, H., Kineman, R. D., Manova-Todorova, K. O., Soares, V. C.,
Hoffman, E. S., Ono, M., Khanam, D., Hayday, A. C., Frohman, L. A. & Koff,
A. (1996) Cell 85, 721–732.
14. Lyden, D., Young, A. Z., Zagzag, D., Yan, W., Gerald, W., O’Reilly, R., Bader,
B. L., Hynes, R. O., Zhuang, Y., Manova, K. & Benezra, R. (1999) Nature 401,
16. Dong, Q. G., Bernasconi, S., Lostaglio, S., De Calmanovici, R. W., Martin-
Padura, I., Breviario, F., Garlanda, C., Ramponi, S., Mantovani, A. & Vecchi,
A. (1997) Arterioscler. Thromb. Vasc. Biol. 17, 1599–1604.
17. Koff, A., Giordano, A., Desai, D., Yamashita, K., Harper, J. W., Elledge, S.,
Nishimoto, T., Morgan, D. O., Franza, B. R. & Roberts, J. M. (1992) Science
18. Fero, M. L., Rivkin, M., Tasch, M., Porter, P., Carow, C. E., Firpo, E., Polyak,
K., Tsai, L. H., Broudy, V., Perlmutter, R. M., et al. (1996) Cell 85, 733–744.
19. Nakayama, K., Ishida, N., Shirane, M., Inomata, A., Inoue, T., Shishido, N.,
Horii, I. & Loh, D. Y. (1996) Cell 85, 707–720.
20. Rafii, S. (2000) J. Clin. Invest. 105, 17–19.
21. Urbich, C. & Dimmeler, S. (2004) Trends Cardiovasc. Med. 14, 318–322.
22. Urbich, C. & Dimmeler, S. (2004) Circ. Res. 95, 343–353.
23. Bagley, R. G., Walter-Yohrling, J., Cao, X., Weber, W., Simons, B., Cook, B. P.,
Chartrand, S. D., Wang, C., Madden, S. L. & Teicher, B. A. (2003) Cancer Res.
24. Sikder, H., Huso, D. L., Zhang, H., Wang, B., Ryu, B., Hwang, S. T., Powell,
J. D. & Alani, R. M. (2003) Cancer Cell 4, 291–299.
25. Ruzinova, M. B., Schoer, R. A., Gerald, W., Egan, J. E., Pandolfi, P. P., Rafii,
S., Manova, K., Mittal, V. & Benezra, R. (2003) Cancer Cell 4, 277–289.
Vidal et al. PNAS ?
May 10, 2005 ?
vol. 102 ?
no. 19 ?