ARTICLE
doi:10.1016/S1525-0016(03)00010-8
Localized Arteriole Formation Directly Adjacent to the
Site of VEGF-Induced Angiogenesis in Muscle
Matthew L. Springer, Clare R. Ozawa, Andrea Banfi, Peggy E. Kraft, Tze-Kin Ip,*
Timothy R. Brazelton, and Helen M. Blau†
Baxter Laboratory in Genetic Pharmacology, Stanford University School of Medicine, Stanford, California 94305-5175, USA
*Present address: Mission Internal Medical Group, 26732 Crown Valley Parkway, Suite 131, Mission Viejo, CA 92691.
†
To whom correspondence and reprint requests should be addressed at the Baxter Laboratory in Genetic Pharmacology, 269 Campus Drive, CCSR 4215,
Stanford University School of Medicine, Stanford, CA 94305-5175. Fax: (650) 736-0080. E-mail: [email protected].
We have shown previously that implantation of myoblasts constitutively expressing the VEGF-A
gene into nonischemic mouse skeletal muscle leads to overgrowth of capillary-like blood vessels
and hemangioma formation. These aberrant effects occurred directly at the implantation site. We
show here that these regions result from angiogenic capillary growth and involve a change in
capillary growth pattern and that smooth muscle-coated vessels similar to arterioles form directly
adjacent to the implantation site. Myoblasts genetically engineered to produce VEGF were
implanted into mouse leg muscles. Implantation sites were surrounded by a zone of dense
capillary-sized vessels, around which was a second zone of muscle containing larger, smoothmuscle-covered vessels but few capillaries, and an outer zone of muscle exhibiting normal
capillary density. The lack of capillaries in the middle region suggests that the preexisting
capillaries adjacent to the implantation site underwent enlargement and/or fusion and recruited
a smooth muscle coat. Capillaries at the implantation site were frequently wrapped around
VEGF-producing muscle fibers and were continuous with the circulation and were not observed to
include bone-marrow-derived endothelial cells. In contrast with the distant arteriogenesis resulting from VEGF delivery described in previous studies, we report here that highly localized
arterioles also form adjacent to the site of delivery.
Key Words: VEGF, myoblasts, angiogenesis, arteriogenesis, gene therapy
INTRODUCTION
Vascular endothelial growth factor (VEGF), also known as
vascular permeability factor, plays a major role in the
induction and maintenance of blood vessel formation
both during embryogenesis and throughout adult life
[1–3]. Due to its potency and specificity as an angiogenic
factor, VEGF is being used in the majority of therapeutic
angiogenesis clinical trials involving both protein delivery and gene therapy approaches [4]. Most clinical and
preclinical experiments involving VEGF-induced angiogenesis have focused on the endpoint with relatively little
characterization of the biological processes and intermediates involved. However, a variety of animal studies
[5–11] have made it clear that the physiological response
to VEGF is complex and can vary depending on the target
tissue, mode of delivery, and dosage. While clinical trials
using VEGF gene delivery approaches have yielded encouraging results [12–14], the interpretation of therapeu-
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
1525-0016/03 $30.00
tic angiogenesis trials continues to be controversial [15],
and many questions remain with regard to the nature of
vessels formed [16,17].
We previously demonstrated that the implantation of
myoblasts constitutively expressing VEGF-A164 from retroviral vectors into mouse skeletal muscle or myocardium
causes substantial, highly localized growth of endothelial
structures that ultimately give rise to hemangiomas, without affecting surrounding muscle tissue [5,18]. These effects occur even in nonischemic muscle, which is typically substantially less receptive to the actions of VEGF
than ischemic muscle [19,20]. Nonischemic muscle provides a model system that avoids the added complexity
caused by factors induced by preexisting hypoxia in ischemic muscle. Implantation of these myoblasts directly
into muscle and into subcutaneous and intraperitoneal
sites led to accumulation primarily of endothelial cells,
smooth muscle cells, and macrophages [5,6]. Despite
441
ARTICLE
these observations, there was very little known about the
mechanism by which these accumulations occurred and
whether the endothelial structures resulted from outgrowths of existing vessels (angiogenesis) or were formed
de novo by isolated endothelial cells or their precursors
(vasculogenesis). It was also unclear whether VEGF gene
introduction would result only in capillary growth or if
larger vessels with an intact smooth muscle layer could
form.
We show here that myoblast-mediated VEGF gene delivery results initially in a network of capillaries that are
continuous with the circulation and are associated with
neither isolated endothelial cells nor their bone marrow
precursors. This suggests that angiogenesis, the sprouting
from preexisting vessels, is the main mechanism responsible for the production of these vessels, rather than de
novo formation from precursor cells. Moreover, the host
muscle directly adjacent to the myoblast implantation
site supports the production of large, smooth musclecovered arterioles. We propose that these vessels may
form by a variation of the conventional arteriogenesis
process that occurs proximal to arterial occlusion and
tissue ischemia [21] and has been reported to increase
significantly after VEGF delivery to ischemic muscle [22–
24]. In the current case, these vessels form directly next to
the myoblast implantation site on a cellular level. The
proximity of these extra arterioles to the implantation site
raises the possibility that their formation is influenced by
a spatial gradient of VEGF or another VEGF-induced factor or by local hemodynamic effects triggered by the
nearby increase in capillaries. Notably, regions of muscle
receiving fewer myoblasts can still produce these arterioles. These data suggest that under appropriate conditions, cell-mediated constitutive VEGF gene delivery can
induce not only local angiogenesis and distant arteriogenesis but also local arteriole formation directly in the vicinity of the VEGF source in nonischemic muscle.
RESULTS
Constant VEGF Delivery at High Levels Induces Two
Distinct Morphologies of Endothelial Overgrowth
We showed previously that, within 1.5 months of implantation of 5 ⫻ 105 VEGF-producing primary myoblasts into
skeletal muscle, large hemangiomas that consisted of
blood pools and vascular channels formed, while control
myoblasts expressing only lacZ fused into muscle fibers
and did not cause any vascular abnormalities [5]. The
tissue that separated these channels possessed a tissue
architecture similar to that of an arterial wall, with a layer
of endothelial cells lining the lumen of the channels and
a deeper adjacent smooth muscle layer. To correlate the
histological methods used in this study with our previous
results, mouse legs that had been injected with myoblasts
expressing VEGF or lacZ were harvested at day 66 (late
time point) and tissue sections were examined using col-
442
doi:10.1016/S1525-0016(03)00010-8
orimetric immunostaining procedures that favor low
magnification. Implantation of control myoblasts did not
result in hemangioma formation. Examination of hemangiomas in legs implanted with VEGF-expressing cells
revealed that the endothelial structures within these hemangiomas fell into two categories (Fig. 1). Some regions of
hemangioma consisted mostly of fluid, surrounded by
tissue walls that were lined with endothelial cells and
identified by immunostaining for the endothelial protein
PECAM-1 (CD-31). Other regions of hemangioma were
mostly solid endothelial masses with relatively little fluid.
In either case, this standard number of myoblasts, i.e., this
standard dose of VEGF, caused a massive overgrowth of
endothelial structures. As in our previous study [5], control myoblasts expressing only lacZ did not cause any
such effects (not shown).
Distinct Physiological Effects Occur at Increasing
Distance from a Localized Source of VEGF
To understand how these structures develop from what is
initially a limited region of capillaries and cellular infiltrate, we isolated muscle from mice 14 –18 days after
implantation of VEGF-expressing myoblasts, a time
point preceding hemangioma development. We immunostained tissue sections from a sample myoblast implantation site with antibodies against VEGF and PECAM-1 (Fig.
2). Many of the muscle fibers and partially differentiated
small fibers at the myoblast implantation site stained
positive for VEGF. Some exhibited VEGF immunostaining throughout their cross-sectional area, while others
stained faintly or not at all within, but possessed a ring of
stained material around their periphery that was likely
due to VEGF’s binding to heparin. PECAM-1-immunostained capillaries were clearly observed growing through
the region and in many cases wrapping tightly around
individual muscle fibers, a growth pattern not normally
observed in skeletal muscle, which is typically characterized by long, straight capillaries running parallel to the
muscle fibers with occasional branches. Significantly, capillaries wrapped around not only fibers that stained entirely positive for VEGF but also those that stained only at
the periphery and even an occasional fiber in which VEGF
staining was not apparent (although it is unknown if
these fibers simply possessed an amount of VEGF below
the detection sensitivity of the immunostaining). Muscle
from animals implanted with control myoblasts exhibited
neither regions of excess VEGF immunostaining nor capillary growth (data not shown). Thus, the expression of
recombinant VEGF influenced not only the extent of angiogenesis but also the pattern of capillary growth.
The effects that typically characterize the implantation
site are quite localized and the majority of surrounding
muscle tissue remains essentially unchanged [5]. However, careful examination of the implantation sites of
VEGF-expressing myoblasts in several legs at days 14 –18
revealed an additional thin zone of host muscle directly
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00010-8
FIG. 1. Hemangiomas occurring 66 days after implantation of VEGF-expressing
myoblasts in two locations, stained immunohistochemically to visualize PECAM-1
(dark blue) with eosin counterstaining (pink) in (A–C) and with hematoxylin and
eosin (H&E) in (D). The two kinds of endothelial overgrowth are visible in A, with
dense endothelial staining identified by the arrow and more hemangioma-like
tissue composed of vascular channels inside the white box. B is an enlargement of
a different region of the section partially shown in A; C is an enlargement of the
white box in A. C and D are pictures of neighboring sections stained using the two
different methods. The PECAM staining reveals the thin endothelial layer that lines
the tissue surrounding the large vascular channels of the hemangioma. Asterisks in
C and D denote lumens of blood pools. Control myoblasts did not cause vascular
abnormalities (not shown). Bars, 1 mm in A and 500 ␮m in B–D.
ARTICLE
adjacent to the hyperendothelialized region, in which
larger caliber vessels were present with substantially
greater frequency than in surrounding normal muscle
(Fig. 3, labeled as zone 2) or in muscle implanted with
control myoblasts (not shown). Interestingly, capillaries
were not evident in this zone, despite the normal frequency of capillaries in the surrounding muscle on one
side (zone 3) and the masses of capillary-like vessels at the
implantation site on the other side (zone 1), suggesting
that these larger caliber vessels may have been derived
from the capillaries that originally inhabited this region
of the muscle. Double-immunolocalization of PECAM-1
and ␣-smooth muscle actin (SMA) revealed that these
larger caliber vessels possessed a thick smooth muscle coat
and therefore were classified as arterioles or arteriole-like
vessels (Figs. 3C–3E). The large number of SMA⫹ vessels
surrounding the implantation site was substantially
higher than the rare frequency in the rest of the muscle, as
immunolocalization of SMA normally detects only an
occasional arteriole. Notably, SMA⫹ vessels were present
only at low levels among the multitude of capillaries at
the implantation site itself. Additional amorphous SMA
staining at the implantation site was presumably due to
the transient expression of SMA by skeletal muscle myoblasts during differentiation [25]. Indeed, this amorphous
SMA staining was largely absent at later time points (data
not shown). SMA⫹ vessels were identified here and elsewhere by SMA⫹ cells wrapped around a smaller endothelial tube identified by PECAM-1 immunostaining, as
shown in detail by confocal microscopy in Fig. 3E. Thus,
three distinct zones were identified, the innermost being
the masses of capillary-like vessels containing very little
smooth muscle at the implantation site (zone 1) surrounded by a thin zone of host muscle containing numerous, large-caliber, SMA⫹ vessels and few if any capillaries
(zone 2) and further surrounded by a zone of seemingly
FIG. 2. Two serial sections taken from a VEGF-expressing myoblast implantation site after 2 weeks, one stained with H&E and the other immunostained to
visualize PECAM-1 (red) and VEGF (green). The merged image shows capillaries tightly wrapping around muscle fibers that stain positive for VEGF and also
around neighboring fibers that do not stain positive for VEGF, demonstrating that VEGF has substantially altered the growth pattern of the capillaries from their
typical muscle-fiber-parallel orientation. Bar, 100 ␮m.
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
443
ARTICLE
444
doi:10.1016/S1525-0016(03)00010-8
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00010-8
normal host muscle in which SMA immunostaining was
very rare (zone 3).
Another example of this tight localization of the newly
formed arterioles (shown in Figs. 3F–3I) was observed at
the needle track of an injection leading to a large implantation site after 66 days; that is, a small number of myoblasts had leaked back through the wound caused by the
needle during the injection, and these myoblasts fused
into fibers along that track. Despite the intense hypervascular response at the implantation site itself, this needle
track possessed only a few extra capillaries colocalizing
with the lacZ-positive myofibers, with no evidence of
hemangioma, chronic inflammation, or chronic edema,
based on hematoxylin and eosin (H&E) staining (not
shown). Once again, SMA⫹ arterioles were clustered
around this region in the adjacent muscle. Thus, the
arterioles that formed in response to implantation of
VEGF-expressing myoblasts were always directly adjacent
to the implantation site regardless of its size.
The VEGF-Induced Capillaries are Continuous with
the Circulation and Are Not Bone Marrow-Derived
While the vessels surrounding the implantation site were
easily identifiable, the nature of the endothelial staining
directly at the implantation site was harder to interpret in
these relatively thin 10-␮m tissue sections. Despite this,
we desired to ascertain whether this mass of endothelial
tissue was due entirely to angiogenic capillaries or if there
was a significant contribution from bone-marrow-derived
endothelial precursor cells [26,27] and/or individual endothelial cells. The following two approaches were taken
to address these questions: (1) use of laser scanning confocal microscopy to examine thicker tissue sections obtained from mice that had been perfused with a vascular
tracer, to assess vascular continuity, and (2) implantation
of VEGF-expressing myoblasts into mice that had received
bone marrow transplants (BMT) from donor mice with
genetically labeled bone marrow, to look for bone-marrow-derived cells.
To determine whether the PECAM-expressing endothelial structures at the myoblast implantation site consisted entirely of capillary-like vessels or also included
isolated endothelial cells/progenitors, mice that had received implantations of VEGF-expressing myoblasts in
both legs 14 days earlier were perfused through the left
ARTICLE
ventricle with a fluorescent vascular tracer and muscle
tissue was harvested immediately for thick cryosectioning
(50␮m). The tissue sections were then immunostained for
PECAM-1 or SMA. Because VEGF-induced vessels have the
potential to be leaky [2,7], the mice were perfused with a
suspension of 0.2-␮m fluorescent microspheres that are
large enough not to leak out of the vessels but small
enough that their suspension behaves as a fluid that
completely fills the vessel lumens [28] (Figs. 4A and 4B).
The sections were cut thick enough to allow threedimensional reconstruction by confocal microscopy to
yield high-resolution images. These reconstructions were
thoroughly examined for cells that stained positive for
PECAM-1 but were not associated with vascular tracer and
were not physically connected to any other PECAM-1⫹
structures. A painstaking search of many sections from 10
implanted hind limbs failed to turn up any such isolated
endothelial cells, further suggesting that at least the majority of the constant accumulation of capillaries directly
at the site of VEGF production is due to an angiogenic
mechanism as opposed to the in situ differentiation of
endothelial progenitor cells (Figs. 4E and 4F).
To assess the contribution of bone-marrow-derived endothelial precursor cells to the mass of endothelial tissue
at the myoblast implantation site, bone marrow from
mice that were transgenic for green fluorescent protein
(GFP) was transplanted into lethally irradiated normal
mice, their hematopoietic systems were allowed to be
reestablished for ⬎8 weeks, and then VEGF-expressing
myoblasts were implanted according to our standard procedure. The recipient mice, BMT donor mice, and myoblasts were all from the C57BL/6 genetic background to
avoid complications of an immune response. Because
macrophages that were already reported to be present at
such implantation sites [5,6] are themselves derived from
the bone marrow, a GFP-labeled cell would need to also
show positive immunostaining for PECAM-1 to be considered a bone-marrow-derived endothelial cell. Muscle
was sectioned lengthwise at 7 days (n ⫽ 2) and crosswise
at 14 days (n ⫽ 1) and examined with confocal microscopy using an optical section thickness of 1 ␮m to make
three-dimensional reconstructions, ensuring accuracy of
colocalization. While both PECAM-immunostained capillaries and GFP-labeled cells were easily detectable in
FIG. 3. Zone of arteriole formation surrounding a VEGF-expressing myoblast implantation site. (A, B) PECAM-1 immunostaining (dark blue) of endothelial
response to VEGF. B is an enlargement of the white box in A. (C, D) Immunofluorescent staining of PECAM-1 (red) and ␣-smooth muscle actin (SMA). Numbered
lines in B–D denote three sharply delineated zones of differing physiological effect. Zone 1 is the implantation site itself, zone 2 is the region of host muscle
possessing arterioles, and zone 3 is the remaining host muscle that is unchanged. Amorphous SMA staining in zone 1 of D (white arrowhead) is most likely
implanted myoblasts transiently expressing SMA during differentiation. (E) A confocal microscope-generated three-dimensional reconstruction of an arteriole
adjacent to an implantation site, stained as in previous images, showing green-stained smooth muscle cells wrapping around red-stained endothelial tubes (from
a 20␮m section). (F) A needle track leading to a large mass of capillaries in lower left. The white box in F is shown enlarged in neighboring sections showing
PECAM-1 (G), PECAM-1/SMA double-immunofluorescence of arteriole cluster (H), and X-gal staining of lacZ (I), which was also expressed by the myoblasts as
a marker gene. Arrows show corresponding locations. Bars, 500 ␮m in A, 100 ␮m in B–D, 10 ␮m in E, and 100 ␮m in F–I.
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
445
ARTICLE
doi:10.1016/S1525-0016(03)00010-8
FIG. 4. Vascular continuity of VEGF-induced vessel structures assessed by perfusion with fluorescent microspheres. (A, B) A 10-␮m section of a myoblast
implantation site from a mouse perfused with 0.2-␮m fluorescent blue-white microspheres and immunofluorescently stained for PECAM-1 (red). A shows
PECAM-1 staining alone; B shows both PECAM-1 staining and microsphere fluorescence in perfused capillaries indicating vascular continuity. (C–F) 30-␮m
three-dimensional reconstructions of thicker sections from mice perfused with “dark red” microspheres, pseudocolored as green, along with immunostaining for
PECAM-1 (red) and SMA (blue). E and F are high-magnification views from the front and side, respectively, of a reconstruction similar to that shown at low
magnification in C. Arrows in E and F point to two endothelial cells that appear to be isolated if viewed from the front, but are actually at the edge of the section
and probably connected to vessels in the neighboring tissue. (G) A confocal three-dimensional reconstruction of capillaries (red) at an implantation site in a
recipient of GFP-labeled bone marrow; marrow-derived inflammatory cells are thus green. (H) A 1-␮m-thick optical slice from the same region with nuclei in blue.
No marrow-derived endothelial cells are observed. Bars, 100 ␮m in A and B, 50 ␮m in C, 20 ␮m in D–H.
these sections, there was no evidence of GFP-labeled,
PECAM-immunostained endothelial cells, demonstrating
that bone-marrow-derived cells are at least not a major
contributor to the massive tangle of vessels at the source
of VEGF production (Fig. 4G). Some green cells appear to
be surrounded by red PECAM staining, but addition of
nuclei to the image (Fig. 4H) reveals that the red stain
comes from neighboring cells (for example, the green cell
at the upper right of Figs. 4G and 4H). As a positive
control, similar tissue sections were immunostained with
an antibody against the macrophage antigen F4/80. Cells
that contained both F4/80 staining and GFP fluorescence
were easily observed (data not shown), demonstrating
that positively immunostained GFP-expressing cells
would be detectable by these methods.
446
DISCUSSION
Despite its name, therapeutic angiogenesis presumably
requires more than just the angiogenic growth of capillaries to be truly beneficial to ischemic muscle. The new
capillary bed requires larger vessels to bring blood from
normal tissue to the ischemic region. Indeed, the primary
mechanism by which the body normally responds to arterial occlusion involves not only angiogenic capillary
growth but also the enlargement of preexisting collateral
arterioles into larger arteries by the process of arteriogenesis [21,29], which has been reported to be enhanced by
VEGF protein or gene delivery [22–24]. This process is
fundamentally different from angiogenesis in that the
latter involves sprouting of new vessels from preexisting
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
doi:10.1016/S1525-0016(03)00010-8
vessels, while arteriogenesis results not in a net increase of
vessels, but in enlargement of smooth-muscle-covered arterioles that were already present. Thus, the appearance of
new arterioles adjacent to the site of VEGF-expressing
myoblast implantation is an example of neither angiogenesis nor arteriogenesis in the strictest sense of the
term. Because the zone of muscle in which these arterioles
appear is devoid of most of the capillaries that must have
been present originally, it is possible that the arterioles
were derived from enlargement and/or fusion of local
capillaries in a manner similar to the enlargement of
distant arterioles during arteriogenesis.
What would drive such a process? In occlusive arterial
disease, arteriogenesis is driven by the pressure difference
experienced across preexisting collateral arterioles upstream of the starved capillaries. This leads to shear stress
and genetic and functional changes in the endothelium
that attract monocytes, which provide the factors necessary for arteriolar enlargement (reviewed by van Royen et
al. [21]). In our study, a vast new capillary bed formed at
the myoblast implantation site, potentially providing a
region of decreased resistance relative to the capillaries
located upstream. This would still result in high pressure
upstream relative to downstream and could trigger capillary arteriogenesis in the peripheral zone adjacent to the
implantation site. Thus, the formation of the arterioles
observed in this study may be only an indirect effect of
the delivery of VEGF, which would directly induce angiogenesis that itself triggers arteriole formation. This interpretation would be consistent with a report that the natural increase in VEGF expression in response to rabbit
hind-limb ischemia in one model was correlated temporally with angiogenesis but not arteriogenesis, which occurred subsequently [30]. Observations from a similar
model in which VEGF was not induced also showed that
arteriogenesis could still be induced in the absence of
endogenous VEGF expression and that exogenous VEGF
did not induce arteriogenesis in the absence of angiogenic
conditions [31]. However, the process reported here is
fundamentally distinct from the arteriogenic growth of
collaterals in that the arterioles form directly adjacent to
the site of VEGF delivery on a cellular level, in regions
where there were few, if any, preexisting muscular vessels.
If the above explanation is correct, it would suggest
that the peripheral, arteriole-containing zone was distinct
from the rest of the host muscle because of its proximity
to the new capillaries. An alternate explanation for the
existence of this peripheral zone invokes a potential steep
gradient of VEGF or another VEGF-induced factor produced at the implantation site. The VEGF-A isoform delivered, VEGF164, is partially heparin binding [32] and is
not expected to diffuse far from its source. Indeed, the fact
that capillaries wrapped around both the muscle fibers
that stained positive for VEGF and the neighboring fibers
that did not stain (Fig. 2) implies that the VEGF produced
by the muscle immediately bound to heparin on local
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
ARTICLE
fibers regardless of their genotype. Thus, a steep gradient
of VEGF concentration emanating from the implantation
site into neighboring host muscle may result in angiogenesis at high levels and arteriole formation at lower levels.
We were unable to visualize such a gradient using immunolocalization (Fig. 2), but that approach may not have
been sufficiently sensitive. Alternatively, the VEGF produced at the implantation site may have induced local
secretion of a different factor or factors that were also
capable of forming a spatial gradient. This alternative of a
second effector might be easier to reconcile with the presence of arterioles around some regions of lower myoblast
incorporation, since the size of a VEGF gradient would be
expected to vary as a function of the magnitude of the
VEGF source. There may be a threshold of VEGF concentration above which such a second effector is released,
possibly the same threshold that has been postulated for
the ability of VEGF to exert an effect on nonischemic
muscle [5]. It is interesting to note that VEGF has been
implicated before in arterial growth pattern formation
during embryonic development [33,34]; however, such a
role in adult muscle has not been previously described.
While it was initially believed that VEGF receptors
were specific to endothelial cells and monocytes, several
recent studies have demonstrated the existence of these
receptors on smooth muscle cells [35–38], suggesting that
VEGF may, in theory, play a direct role in accumulation of
the smooth muscle coat as well. It is notable that interaction of smooth muscle cells and monocytes in culture
leads to VEGF gene expression [39] and that VEGF-activated endothelium attracts platelets and induces them to
release more VEGF [40]. Since high-level, localized VEGF
expression ultimately induces hemangiomas possessing
all of the above cell types that continue to grow even if
the original source of VEGF becomes diminished [5], it is
very likely that such VEGF delivery leads to a feedback
loop that involves both angiogenic and arteriogenic
mechanisms.
The acquisition of a smooth muscle coat by vessels in
response to VEGF was also reported by Pettersson et al. [8]
to occur after adenoviral VEGF gene delivery in mice,
although this effect was observed in the ear and not in
skeletal muscle. An important difference between the injection of adenovirus and the implantation of myoblasts
is that the latter results in the production of VEGF directly
by muscle cells, while the adenovirus was shown to transduce interstitial cells in the muscle but not the muscle
fibers themselves. Thus, the identity of the cells producing VEGF may affect the nature of the response, potentially through the induced coexpression of other factors
(Ozawa et al., submitted for publication). In support of
this interpretation, Arsic et al. report [41] that arteriolelike vessels are also induced by VEGF gene transfer to
muscle using adeno-associated virus, which transduces
muscle fibers directly. Interestingly, local muscular arteriole formation has recently been reported to occur in
447
ARTICLE
doi:10.1016/S1525-0016(03)00010-8
response to delivery of genes encoding FGF2 and FGF6 in
a wound healing model [42], showing that this response is
not limited to VEGF gene delivery.
The rapidity and extent of endothelial proliferation in
response to VEGF-expressing myoblast implantation had
caused us previously to suspect that this effect was due in
part to recruitment of circulating bone-marrow-derived
endothelial progenitor cells [5], a mechanism that has
been reported to play a minor role in normal angiogenesis and a major role in pathological angiogenesis
[26,27,43,44]. Significantly, several attempts to detect endothelial cells that were either bone marrow derived or
not in contact with other capillaries failed to detect such
cells at the myoblast implantation site, implying that the
massive endothelial proliferation is due to a robust but
classical angiogenic response. It cannot be ruled out that
any differentiation of endothelial precursor cells has already occurred by the 14-day time point examined. However, because VEGF is still being produced at that time
(Fig. 2), and angiogenesis continues to occur beyond that
point, a process involving differentiation of precursor
cells would be expected to be ongoing and readily detectable. The formation of neo-arterioles that accompanies
this angiogenic response suggests that the increase in
large collateral vessels previously reported in preclinical
and clinical VEGF gene therapy experiments [22,45– 47]
may in fact reflect an increase in the number of arterioles,
in addition to the arteriogenic enlargement of preexisting
arterioles as previously assumed, and bodes well for therapeutic applications of VEGF gene delivery at appropriate
levels.
MATERIALS
AND
METHODS
Intramuscular implantation of myoblasts. Male C.B-17 SCID mice 8 –10
weeks of age were obtained from the Stanford University Department of
Comparative Medicine and from Taconic Farm (Germantown, NY) and
treated in accordance with the guidelines of the Stanford University Administrative Panel on Laboratory Animal Care. Primary myoblasts that
expressed the genes encoding murine VEGF164 and Escherichia coli lacZ or
lacZ alone [5] were cultured as described previously [48]. Myoblasts were
trypsinized and injected once into the tibialis anterior and once into the
lateral gastrocnemius of anesthetized mice as previously described [48,49].
Each injection consisted of 5 ⫻ 105 cells in 5 ␮l of PBS with 0.5% BSA.
Anesthesia consisted of inhaled methoxyflurane (Mallinckrodt Veterinary,
Mundelein, IL) throughout the procedure. The data are derived from a
total of 12 mice sacrificed after 14 –18 days and 6 mice sacrificed after 66
days in a total of three separate experiments, unless otherwise described.
Histological analyses. On the designated days, mice were sacrificed and
their muscle tissue was harvested, frozen, sectioned, and stained histologically with X-gal or H&E as described previously [48,49].
For immunofluorescent staining, the sections were fixed in 1.6% formaldehyde for 15 min and then were blocked with antibody buffer consisting of 2% normal goat serum and 0.3% Triton X-100 in PBS for ⬎45 min.
Sometimes 0.5% casein (“blocking reagent” from NEN Life Science Products, Boston, MA) was added to the blocking step to reduce background as
needed. The slides were incubated for 2 h at room temperature in antibody
buffer containing a rat monoclonal antibody against murine PECAM-1
(clone MEC 13.3; PharMingen, San Diego, CA; 1:100 dilution), a mouse
monoclonal antibody against smooth muscle actin (clone 1A4; ICN Bio-
448
medicals, Aurora, OH; 1:400 dilution), or a rabbit polyclonal antibody
against mouse VEGF (ab173; AbCam, Cambridge, UK; 1:100 dilution) or
combinations of the above. Negative controls lacking primary antibody
were always performed. Sections were rinsed in antibody buffer and then
incubated for 1–1.5 h in the appropriate secondary antibody conjugated to
Alexa488 or Texas Red (Molecular Probes, Eugene, OR). The slides were
then rinsed and mounted and viewed either with a Zeiss Axioplan fluorescence microscope or a Zeiss LSM510 laser scanning confocal microscope. Three-dimensional reconstructions of confocal microscope image
stacks were generated using Zeiss LSM software and presented as singleimage projections.
For colorimetric immunostaining for PECAM-1, 10-␮m sections were
fixed in ice-cold acetone for 10 min at ⫺20°C, allowed to dry for 5 min at
room temperature, and hydrated in PBS for 5 min. Sections were blocked
with 0.5% BSA in PBS for 30 min and then incubated in biotinylated rat
anti-mouse PECAM-1 monoclonal antibody (MEC 13.3; Pharmingen) at
1:250 dilution in blocking buffer. The sections were then stained with
ABC-AP reagent and BCIP/NBT substrate solution (Vector Laboratories,
Burlingame, CA) according to the manufacturer’s instructions. Levisamole
(Vector Laboratories) diluted to manufacturer’s specifications was added to
the substrate staining solution to inhibit endogenous alkaline phosphatase activity. Samples were rinsed and counterstained with eosin.
Bone marrow transplantation. Bone marrow was harvested from 8- to
10-week-old, male transgenic mice that ubiquitously expressed an enhanced version of GFP, a generous gift from M. Okabe (Osaka University)
[50], and transplanted into C57BL/6 mice as described previously [51].
Hematopoietic reconstitution was confirmed 8 weeks posttransplant by
counting the number of GFP⫹ circulating cells by FACS, and only mice in
which greater than 95% of circulating cells expressed GFP were evaluated
further. To harvest tissue, mice were deeply anesthetized with an intraperitoneal injection of 25 mg/ml Avertin (Aldrich) in PBS. The chest was
opened and the right atrium was removed. The left ventricle was entered
with a cannula, through which 1% formaldehyde in PBS was perfused at
120 mm Hg of pressure for 5 min [52]. Muscle that had been implanted
with myoblasts was harvested, frozen, and processed for immunofluorescence as described above without the fixation step. In some animals, legs
were harvested following perfusion, sectioned longitudinally using a
Vibratome, and processed for immunofluorescence.
Perfusion with fluorescent microsphere suspension. For vascular continuity assessments, mice that had previously been implanted intramuscularly
with VEGF-expressing myoblasts were perfused with a space-filling fluorescent microsphere suspension using a modification of our previously
published procedure [28]. Mice were preanesthetized with methoxyflurane
and then were deeply anesthetized with an intraperitoneal injection of 70
␮l of 50 mg/ml Nembutal (sodium pentobarbital; Abbott Laboratories,
North Chicago, IL) mixed with 50 ␮g/ml nitroglycerin (Abbott Laboratories) to maximize vasodilation. The chest was opened and an incision was
made in the right atrium to allow perfusate to drain. A small hole was
made in the apex of the left ventricle and the hole was entered with a
blunt-ended 18-gauge cannula connected to a syringe filled with PBS. The
animal was perfused with approximately 6 ml of PBS over approximately
1 min. The cannula was removed and a similar cannula was inserted,
through which was perfused 6 ml of a suspension of 0.2-␮m fluorescent
carboxylate-modified polystyrene beads (FluoSpheres; Molecular Probes),
diluted 1:6 with PBS (concentration as supplied was 2% solids in water).
Some experiments used “blue” FluoSpheres (365-nm excitation max,
415-nm emission max) and other experiments used “dark red” FluoSpheres (660-nm excitation max, 680-nm emission max). Leg muscle was
harvested immediately and frozen as described above. Ten- or 50-␮m
cryosections were cut and immunofluorescently stained as described
above except that for the 50-␮m sections, incubation times were extended
overnight and rinses were extended for several hours.
ACKNOWLEDGMENTS
We thank Georges von Degenfeld for invaluable discussion of the manuscript.
We also thank Donald McDonald, Gavin Thurston, and Nicole Glazer (Univer-
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
ARTICLE
doi:10.1016/S1525-0016(03)00010-8
sity of California, San Francisco) for procedural help in one experiment. This
work was supported by NIH Grant HL65572 to H.M.B. and M.L.S.; a grant from
the Muscular Dystrophy Association of America as well as NIH Grants
HD18179, AG09521, and CA59717 and funds from Aventis Corporation and
the Baxter Foundation to H.M.B.; a Howard Hughes Medical Institute predoctoral fellowship to C.R.O.; and a Life and Health Insurance Medical Research
Fund scholarship and an NIH predoctoral training grant to T.R.B.
RECEIVED FOR PUBLICATION SEPTEMBER 27, 2002; ACCEPTED JANUARY 8,
2003.
25.
26.
27.
28.
29.
REFERENCES
1. Ferrara, N., and Alitalo, K. (1999). Clinical applications of angiogenic growth factors
and their inhibitors. Nat. Med. 5: 1359 –1364.
2. Dvorak, H. F., Brown, L. F., Detmar, M., and Dvorak, A. M. (1995). Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and
angiogenesis. J. Pathol. 146: 1029 –1039.
3. Benjamin, L. E., Golijanin, D., Itin, A., Pode, D., and Keshet, E. (1999). Selective ablation
of immature blood vessels in established human tumors follows vascular endothelial
growth factor withdrawal. J. Clin. Invest. 103: 159 –165.
4. Isner, J. M., Vale, P. R., Symes, J. F., and Losordo, D. W. (2001). Assessment of risks
associated with cardiovascular gene therapy in human subjects. Circ. Res. 89: 389 – 400.
5. Springer, M. L., Chen, A. S., Kraft, P. E., Bednarski, M., and Blau, H. M. (1998). VEGF
gene delivery to muscle: potential role for vasculogenesis in adults. Mol. Cell 2: 549 –
558.
6. Springer, M. L., et al. (2000). Induction of angiogenesis by implantation of encapsulated primary myoblasts expressing vascular endothelial growth factor. J. Gene Med. 2:
279 –288.
7. Thurston, G., et al. (1999). Leakage-resistant blood vessels in mice transgenically
overexpressing angiopoietin-1. Science 286: 2511–2514.
8. Pettersson, A., et al. (2000). Heterogeneity of the angiogenic response induced in
different normal adult tissues by vascular permeability factor/vascular endothelial
growth factor. Lab. Invest. 80: 99 –115.
9. Sundberg, C., et al. (2001). Glomeruloid microvascular proliferation follows adenoviral
vascular permeability factor/vascular endothelial growth factor-164 gene delivery.
Am. J. Pathol. 158: 1145–1160.
10. Detmar, M., et al. (1998). Increased microvascular density and enhanced leukocyte
rolling and adhesion in the skin of VEGF transgenic mice. J. Invest. Dermatol. 111: 1– 6.
11. Larcher, F., Murillas, R., Bolontrade, M., Conti, C. J., and Jorcano, J. L. (1998). VEGF/VPF
overexpression in skin of transgenic mice induces angiogenesis, vascular hyperpermeability and accelerated tumor development. Oncogene 17: 303–311.
12. Vale, P. R., et al. (2001). Randomized, single-blind, placebo-controlled pilot study of
catheter-based myocardial gene transfer for therapeutic angiogenesis using left ventricular electromechanical mapping in patients with chronic myocardial ischemia.
Circulation 103: 2138 –2143.
13. Isner, J. M. (2002). Myocardial gene therapy. Nature 415: 234 –239.
14. Rajagopalan, S., Shah, M., Luciano, A., Crystal, R., and Nabel, E. G. (2001). Adenovirusmediated gene transfer of VEGF(121) improves lower-extremity endothelial function
and flow reserve. Circulation 104: 753–755.
15. Simons, M., et al. (2000). Clinical trials in coronary angiogenesis: an expert panel
summary. Circulation 102: E73– 86.
16. Blau, H. M., and Banfi, A. (2001). The well-tempered vessel. Nat. Med. 7: 532–534.
17. Springer, M. L. (2002). VEGF therapy in heart disease: for better or for worse? Drug
Discovery Today 7: 209 –211.
18. Lee, R. J., et al. (2000). VEGF gene delivery to myocardium: deleterious effects of
unregulated expression. Circulation 102: 898 –901.
19. Bauters, C., et al. (1995). Site-specific therapeutic angiogenesis after systemic administration of vascular endothelial growth factor. J. Vasc. Surg. 21: 314 –325.
20. Su, H., Lu, R., and Kan, Y. W. (2000). Adeno-associated viral vector-mediated vascular
endothelial growth factor gene transfer induces neovascular formation in ischemic
heart. Proc. Natl. Acad. Sci. USA 97: 13801–13806.
21. van Royen, N., et al. (2001). Stimulation of arteriogenesis: a new concept for the
treatment of arterial occlusive disease. Cardiovasc. Res. 49: 543–553.
22. Takeshita, S., et al. (1994). Intramuscular administration of vascular endothelial growth
factor induces dose-dependent collateral artery augmentation in a rabbit model of
chronic limb ischemia. Circulation 90: II228 –II234.
23. Tsurumi, Y., et al. (1996). Direct intramuscular gene transfer of naked DNA encoding
vascular endothelial growth factor augments collateral development and tissue perfusion. Circulation 94: 3281–3290.
24. Banai, S., et al. (1994). Angiogenic-induced enhancement of collateral blood flow to
MOLECULAR THERAPY Vol. 7, No. 4, April 2003
Copyright © The American Society of Gene Therapy
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
ischemic myocardium by vascular endothelial growth factor in dogs. Circulation 89:
2183–2189.
Springer, M. L., Ozawa, C. R., and Blau, H. M. (2002). Transient production of
alpha-smooth muscle actin by skeletal myoblasts during differentiation in culture and
following intramuscular implantation. Cell Motil. Cytoskeleton 51: 177–186.
Asahara, T., et al. (1997). Isolation of putative progenitor endothelial cells for angiogenesis. Science 275: 964 –967.
Shi, Q., et al. (1998). Evidence for circulating bone-marrow-derived endothelial cells.
Blood 92: 362–367.
Springer, M. L., Ip, T. K., and Blau, H. M. (2000). Angiogenesis monitored by perfusion
with a space-filling microbead suspension. Mol. Ther. 1: 82– 87.
Carmeliet, P. (2000). Mechanisms of angiogenesis and arteriogenesis. Nat. Med. 6:
389 –395.
Hershey, J. C., et al. (2001). Revascularization in the rabbit hindlimb: dissociation
between capillary sprouting and arteriogenesis. Cardiovasc. Res. 49: 618 – 625.
Deindl, E., et al. (2001). Role of ischemia and of hypoxia-inducible genes in arteriogenesis after femoral artery occlusion in the rabbit. Circ. Res. 89: 779 –786.
Houck, K. A., Leung, D. W., Rowland, A. M., Winer, J., and Ferrara, N. (1992). Dual
regulation of vascular endothelial growth factor bioavailability by genetic and proteolytic mechanisms. J. Biol. Chem. 267: 26031–26037.
Mukouyama, Y. S., Shin, D., Britsch, S., Taniguchi, M., and Anderson, D. J. (2002).
Sensory nerves determine the pattern of arterial differentiation and blood vessel
branching in the skin. Cell 109: 693–705.
Stalmans, I., et al. (2002). Arteriolar and venular patterning in retinas of mice selectively
expressing VEGF isoforms. J. Clin. Invest. 109: 327–336.
Ishida, A., et al. (2001). Expression of vascular endothelial growth factor receptors in
smooth muscle cells. J. Cell. Physiol. 188: 359 –368.
Grosskreutz, C. L., et al. (1999). Vascular endothelial growth factor-induced migration
of vascular smooth muscle cells in vitro. Microvasc. Res. 58: 128 –136.
Takagi, H., King, G. L., and Aiello, L. P. (1996). Identification and characterization of
vascular endothelial growth factor receptor (Flt) in bovine retinal pericytes. Diabetes 45:
1016 –1023.
Brown, L. F., Detmar, M., Tognazzi, K., Abu-Jawdeh, G., and Iruela-Arispe, M. L. (1997).
Uterine smooth muscle cells express functional receptors (flt-1 and KDR) for vascular
permeability factor/vascular endothelial growth factor. Lab. Invest. 76: 245–255.
Hojo, Y., et al. (2000). Interaction between human monocytes and vascular smooth
muscle cells induces vascular endothelial growth factor expression. Atherosclerosis 150:
63–70.
Verheul, H. M., et al. (2000). Vascular endothelial growth factor-stimulated endothelial
cells promote adhesion and activation of platelets. Blood 96: 4216 – 4221.
Arsic, N., et al. (2003). Induction of functional neovascularization by combined VEGF
and angiopoietin-1 gene transfer using AAV vectors. Mol. Ther. 7: 450 – 459.
Doukas, J., et al. (2002). Delivery of FGF genes to wound repair cells enhances
arteriogenesis and myogenesis in skeletal muscle. Mol. Ther. 5: 517–527.
Asahara, T., et al. (1999). VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. EMBO J. 18: 3964 –3972.
Kalka, C., et al. (2000). Vascular endothelial growth factor(165) gene transfer augments
circulating endothelial progenitor cells in human subjects. Circ. Res. 86: 1198 –1202.
Mack, C. A., et al. (1998). Biologic bypass with the use of adenovirus-mediated gene
transfer of the complementary deoxyribonucleic acid for vascular endothelial growth
factor 121 improves myocardial perfusion and function in the ischemic porcine heart.
J. Thoracic Cardiovasc. Surg. 115: 168 –177.
Takeshita, S., et al. (1996). Gene transfer of naked DNA encoding for three isoforms of
vascular endothelial growth factor stimulates collateral development in vivo. Lab. Invest.
75: 487–501.
Baumgartner, I., et al. (1998). Constitutive expression of phVEGF165 after intramuscular gene transfer promotes collateral vessel development in patients with critical limb
ischemia. Circulation 97: 1114 –1123.
Rando, T., and Blau, H. M. (1994). Primary mouse myoblast purification, characterization, and transplantation for cell-mediated gene therapy. J. Cell Biol. 125: 1275–1287.
Springer, M. L., Rando, T. A., and Blau, H. M. (1997). Gene delivery to muscle. In
Current Protocols in Human Genetics (A. L. Boyle, Ed.), pp. 13.4.1–13.4.19. Wiley, New
York.
Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., and Nishimune, Y. (1997). ‘Green
mice’ as a source of ubiquitous green cells. FEBS Lett. 407: 313–319.
Brazelton, T. R., Rossi, F. M., Keshet, G. I., and Blau, H. M. (2000). From marrow to
brain: expression of neuronal phenotypes in adult mice. Science 290: 1775–1779.
Thurston, G., Murphy, T. J., Baluk, P., Lindsey, J. R., and McDonald, D. M. (1998).
Angiogenesis in mice with chronic airway inflammation: strain-dependent differences.
Am. J. Pathol. 153: 1099 –1112.
449