Microenvironmental regulation of tumour angiogenesis

Abstract

Tumours display considerable variation in the patterning and properties of angiogenic blood vessels, as well as in their responses to anti-angiogenic therapy. Angiogenic programming of neoplastic tissue is a multidimensional process regulated by cancer cells in concert with a variety of tumour-associated stromal cells and their bioactive products, which encompass cytokines and growth factors, the extracellular matrix and secreted microvesicles. In this Review, we discuss the extrinsic regulation of angiogenesis by the tumour microenvironment, highlighting potential vulnerabilities that could be targeted to improve the applicability and reach of anti-angiogenic cancer therapies.

Full text

Like normal organs, tumours need to establish a blood
supply to satisfy their demand for oxygen and nutri-
ents and accomplish other metabolic functions1. This
is achieved primarily through angiogenesis, the process
whereby new blood vessels develop from a pre-existing
vascular network. Hypoxia is a key driver of tumour angio-
genesis2. Hypoxic cancer cells secrete vascular endothe-
lial growth factor A (VEGFA), which initiates tumour
angiogenesis by engaging VEGF receptor 2 (VEGFR2)
expressed on the endothelial cells (ECs) of neighbouring
blood vessels1. Gradients of soluble VEGFA induce the for-
mation of motile ECs, called tip cells, which break down
the surrounding extracellular matrix (ECM) and lead the
growth of new vascular sprouts towards VEGFA. This
process requires the participation of additional signal-
ling molecules, including delta ligand-like 4 (DLL4) and
angio poietin 2 (ANGPT2), which, respectively, control
the tip-cell phenotype and destabilize ECjunctions1.
In pre-malignant stages of epithelial tumours (for
example, hyperplasia and carcinoma insitu), a basal
lamina separates the tumour from the vascularized peri-
tumoural tissues, so blood vessels rarely infiltrate these
early lesions3,4. In malignant tumours, cancer cells acquire
invasive behaviours and induce a stromal response
involving robust angiogenesis5. Therefore, tumour pro-
gression from a benign to a malignant stage is typically
associated with an angiogenic switch — the triggering
and development of a vascular network that is actively
growing and infiltrative5 (FIG.1). However, considerable
variation exists in the patterns of tumour vasculariza-
tion, which reflect differences in the tumour type, grade
and stage (for example, primary versus metastatic),
theanatomical site, the stromal cell composition and the
spatiotemporal expression of pro-angiogenic factors and
anti-angiogenic factors4,6–10.
Owing to excessive and sustained pro-angiogenic
signalling5, tumour-associated blood vessels (TABVs)
typically acquire an aberrant morphology, character-
ized by excessive branching, abundant and abnormal
bulges and blind ends, discontinuous EC lining, and
defective basement membrane and pericyte coverage.
These features are all indicative of — or conducive to —
impaired vascular maturation, poor vessel functionality
and incoherent tumour perfusion1,5,11. Furthermore, the
ECs of TABVs display structural and molecular traits
that distinguish them from their counterparts in normal
organs (BOX1).
Although the cancer cells can be an important
source of VEGFA and other pro-angiogenic medi-
ators12, recruited leukocytes increase VEGFA bioavail-
ability and signalling during the angiogenic switch13.
Furthermore, many signals that emanate from various
tumour-associated stromal cells (TASCs), and the ECM
in which they are embedded14, sustain angiogenesis
after the angiogenic switch through the subsequent
phases of tumour progression (TABLE1). In this Review,
we discuss the extrinsic regulation of angiogenesis by
the tumour microenvironment (TME), with the premise
that harnessing such regulation may be instrumental in
developing more effective anticancer therapies targeting
angiogenesis andbeyond.
Regulation of angiogenesis by TASCs
The abundance and composition of TASCs vary con-
siderably between tumours and in their diverse micro-
environments15–18. TASCs can be classified into two
main categories on the basis of their origin. Tumour-
infiltrating cells of haematopoietic origin are recruited
from the bone marrow to the tumour via the systemic
circulation and comprise diverse leukocyte types and
1The Swiss Institute for
Experimental Cancer
Research (ISREC), School
ofLife Sciences, École
Polytechnique Fédérale
deLausanne (EPFL), 1015
Lausanne, Switzerland.
2Department of Fundamental
Oncology, Ludwig Institute for
Cancer Research and Division
of Experimental Pathology,
University of Lausanne and
University of Lausanne
Hospital, 1066 Lausanne,
Switzerland.
Correspondence to M.D.P.
and T.V.P.
michele.depalma@epfl.ch;
tatiana.petrova@unil.ch
doi:10.1038/nrc.2017.51
Published online 14 Jul 2017
Hypoxia
The condition of low oxygen
availability. In tumours,
hypoxiais observed in cancer
cells that reside more than
70–150 μm away from a
perfused bloodvessel.
Pro-angiogenic factors
Biological molecules that
stimulate endothelial cell
proliferation and angiogenesis.
Anti-angiogenic factors
Biological molecules that
blockangiogenesis or promote
the regression of angiogenic
bloodvessels.
Microenvironmental regulation
oftumour angiogenesis
Michele De Palma1, Daniela Biziato1 and Tatiana V.Petrova2
Abstract | Tumours display considerable variation in the patterning and properties of angiogenic
blood vessels, as well as in their responses to anti-angiogenic therapy. Angiogenic programming
of neoplastic tissue is a multidimensional process regulated by cancer cells in concert with a
variety of tumour-associated stromal cells and their bioactive products, which encompass
cytokines and growth factors, the extracellular matrix and secreted microvesicles. In this Review,
we discuss the extrinsic regulation of angiogenesis by the tumour microenvironment,
highlighting potential vulnerabilities that could be targeted to improve the applicability and
reach of anti-angiogenic cancer therapies.
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
457
REVIEWS

Pericyte
Cell that enwraps and
promotes the survival of
endothelial cells, stabilizing
small blood vessels.
Tumour microenvironment
(TME). The complex and
dynamic ensemble of cancer
cells, tumour-associated
stromal cells (TASCs;
comprising primarily
leukocytes, fibroblasts and
vascular cells) and their
extracellular products.
Progenitors
Undifferentiated cells capable
of producing lineage-
committed cellular progeny.
subtypes, such as monocytes and macrophages, neutro-
phils, lymphocytes, as well as their immature precursors.
There are also reports of non-haematopoietic, bone
marrow- derived endothelial or mesenchymal progenitors
contributing to tumour angiogenesis19. Tissue-resident
cells are also recruited, including vascular cells (ECs
and pericytes), fibroblasts, adipocytes, but also some
tissue-resident leukocytes such as mast cells and macro-
phages. We discuss below the main TASC types involved
in the regulation of tumour angiogenesis.
Macrophages. In mouse cancer models, macro phages
largely derive from circulating monocytes that extra vasate to
tumours in response to various chemoattractants, including
Nature Reviews | Cancer
Pre-malignant Angiogenic
AD
AT
100 μm
ADCA
Blood vessels (CD31+)
TAMs (TIE2–GFP+)
Nuclei
TAM
TABV
Key angiostatic or
anti-angiogenic signals
Basal lamina, low
VEGFA gradients,
low MMP and
cathepsin activity,
ANGPT1, THBS1, PAI1,
angiostatin, IFNα,
CXCL9, CXCL10,
CXCL11 and ECM
MMTV–PyMT, breast
adenocarcinoma
model
RIP1–Tag2, pancreatic
neuroendocrine
tumour model
Blood vessels (lectin+)
Nuclei
PNET
ACPA
Islet
100 μm
VEGFA, FGF2,
CXCL8, PDGFs,
PlGF, ANGPT2,
IL-1β, IL-6, TNF,
BV8, MMPs,
cathepsins, EVs,
adipokines, lactate,
acidosis and ECM
Key pro-angiogenic
signals
a
b
c
Figure 1 | Angiogenesis during malignant progression. Early-stage (pre-malignant) tumours typically display scant or no
intra-tumoural vascularization, although a vascularized stroma surrounds the tumours and may adjoin parenchymal tumour
domains (part a, left panel). In malignant tumours, the cancer cells acquire invasive behaviours and induce a stromal response
involving robust intra-tumoural angiogenesis, along with leukocyte infiltration, fibroblast proliferation and extracellular
matrix (ECM) deposition (part a, right panel). In pre-malignant lesions, a basal lamina separates the tumour from the
surrounding tissues; this, together with angiostatic signals conveyed by some ECM components, and the relatively low levels
of pro-angiogenic factors, prevents intra-tumoural vascularization or constrains it into a quiescent state. In malignant lesions,
angiogenesis is largely controlled through the actions on vascular endothelial cells (ECs) of multiple pro-angiogenic
mediators, which include growth factors, cytokines, various ECM proteins, ECM-remodelling enzymes, as well as extracellular
vesicles (EVs) and by-products of deregulated tumour metabolism. Panels in b show early (left) or late (right) mammary
tumours of mouse mammary tumour virus–polyoma middle T antigen (MMTV–PyMT) transgenic mice. Avascular adenomas
(AD; left) develop from mammary glands embedded in vascularized adipose tissue (AT). In progressing adenocarcinomas
(ADCA; right), blood vessels (arrows) infiltrate the tumours, along with pro-angiogenic tumour-associated macrophages
(TAMs; green). In RIP1–Tag2 transgenic mice (part c), hyperplastic islets of Langerhans (islet; left) present a quiescent vascular
network, which becomes angiogenic and plethoric when the islets progress into pancreatic neuroendocrine tumour (PNET).
ACPA, acinar pancreas; ANGPT, angiopoietin; CXCL, CXC-chemokine ligand; EV, extracellular vesicle; FGF2, fibroblast
growth factor 2; GFP, green fluorescent protein; IFNα, interferon-α; IL, interleukin; MMP, matrix metalloproteinase;
PAI1,plasminogen activator inhibitor 1; PDGF, platelet-derived growth factor; PlGF, placental growth factor; TABV,tumour-
associated blood vessel; THBS1, thrombospondin 1; TNF, tumour necrosis factor; VEGFA, vascular endothelial growth
factorA. Images in panel b were adapted with permission from REF.215, Elsevier.
REVIEWS
458
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

Damage-associated
molecular patterns
(DAMPs). Biological
moleculesthat can initiate
aninflammatory response
independently of infection.
Clodronate liposomes
A formulation of small lipid
vesicles containing a
bisphosphonate that is capable
of inducing macrophage death
upon engulfment.
Angiostatic functions
Properties that promote
endothelial cell quiescence
andlimit angiogenesis.
chemokines, pro- inflammatory signalling mol ecules and
damage-associated molecular patterns (DAMPs)18,20. Upon
their extravasation, monocytes differentiate and mature
into tumour- associated macrophages (TAMs) under
the influence of colony-stimulating factor 1 (CSF1; also
known as M-CSF). Tcell cytokines and various tumour-
derived factors further sculpt the macrophage phenotype,
inducing TAMs to acquire substantial molecular and
functional heterogeneity, both within and across distinct
cancertypes18,20.
High macrophage numbers are frequently associated
with increased vascular density in human tumours21–23.
Accordingly, macrophages exert a pro-angiogenic role
in mouse cancer models. Mouse mammary tumour
virus–polyoma middle T antigen (MMTV–PyMT)
transgenic mice that had been rendered macrophage
deficient through Csf1 inactivation displayed decreased
vascularization in the mammary tumours24. Likewise,
the broad elimination of TAMs by clodronate liposomes
or CSF1 receptor (CSF1R) antibodies decreased angio-
genesis in various tumour models25–27. However, the
pro-angiogenic capacity of TAMs may depend on their
activation state, which is modulated by the cytokine
milieu to which they are exposed, and the specific TME
in which they reside18,28. In some developmental pro-
cesses, such as the remodelling of retinal blood vessels,
macrophages may acquire anti-angiogenic or angiostatic
functions29. However, there is currently little evidence for
TAMs having vascular-inhibitory roles in tumours.
TAMs secrete growth factors and inflammatory
cytokines that support angiogenesis by promoting EC
survival, activation and proliferation (FIG.2). TAMs are
an important source of VEGFA in both mouse and
human tumours26,30–32. The conditional elimination
of Ve g fa in myeloid cells (including TAMs) delays the
angiogenic switch and attenuates the abnormal features
of TABVs in mouse cancer models31. Furthermore,
Veg fa deficiency in TAMs limits their ability to restore
angiogenesis and to support the relapse of transplanted
tumours after chemotherapy32. TAM-derived VEGFA
also enhances vascular permeability, thereby facilitating
cancer cell intravasation and metastasis33. Additional
pro- angiogenic factors produced by TAMs include
two VEGF-family members, placental growth factor
(PlGF) and VEGFC, tumour necrosis factor (TNF),
interleukin-1β (IL-1β) and IL-6, CXC-chemokine
ligand8 (CXCL8; also known as IL-8) and fibroblast
growth factor 2 (FGF2); the angiogenic responses
evoked by these cytokines have been reviewed else-
where28,29,34. TAMs also express members of the WNT
family. The genetic deletion of Wnt7b in TAMs reduced
the expression of mitogenic WNT–β-catenin target
genes in tumour ECs and decreased the vascular density
in MMTV–PyMT mouse mammary carcinomas35.
TAMs often enwrap TABVs, and the intimate associ-
ation between perivascular TAMs and ECs creates an
instructive niche that supports tumour angiogenesis29,36.
TAMs secrete membrane-bound or soluble proteases
that, through ECM degradation, facilitate the infiltra-
tive growth of TABVs and mobilize pro-angiogenic
growth factors sequestered in the perivascular ECM5,14.
Macrophage-derived proteases include matrix metallo-
proteinases (MMPs; for example, MMP2, MMP9 and
MMP12) and serine or cysteine proteinases, such as
cathepsins and plasminogen activator14,37,38. Genetic or
bisphosphonate-mediated inhibition of MMP9 decreased
angiogenesis in human tumour xenografts39 and in a
mouse model of human papillomavirus 16 (HPV16)-
driven cervical cancer40. Likewise, the pharmacological
inhibition of cathepsin activity attenuated the vasculari-
zation of pancreatic neuroendocrine tumours (PNETs) in
RIP1–Tag2 transgenic mice41. However, MMPs and cathep-
sins are expressed by multiple cell types in tumours —
not only TAMs but also other leukocytes and cancer cells
— and exert broad pro-tumoural functions that can also
influence angiogenesis indirectly by regulating diverse
parameters of tumour progression14,37,38.
Conditional cell depletion studies have implicated
perivascular TAMs that express the ANGPT receptor
TIE2 (also known as TEK) in the promotion of tumour
Box 1 | Features of tumour endothelial cells
Morphology. Tumour endothelial cells (ECs) are structurally abnormal. They generally
present excessive fenestrations, uneven surfaces and intra-luminal projections, and
loosened intercellular junctions, and can also form multi-layered endothelia. These
features favour vascular leakage and may limit blood flow1,11,224.
Gene expression. The vascular ECs of different tissues and organs show distinct gene
expression profiles224. In analogy, tumour ECs may display considerable inter- and
intra-tumoural molecular heterogeneity224,225. Both gene expression profiling and the
use of phage-display peptide libraries identified several tumour-type or stage-specific
vascular markers (termed ‘tumour endothelial markers’ or ‘vascular zip codes’) in
mousemodels of cancer. The targeting of such tumour EC‑specific markers may
facilitate the selective delivery of therapeutic agents to tumour-associated blood
vessels (TABVs)226,227.
Proliferative signalling. Tumour ECs display increased proliferative, migratory and
tube-formation capabilities in response to growth factors and cytokines, compared
with non-tumour ECs1,224. Furthermore, they are resistant to senescence and can grow
exvivo in serum-free conditions. The upregulation of growth factor and cytokine
receptors (for example, vascular endothelial growth factor receptors (VEGFRs)) by
tumour ECs may account for such abilities224. Tumour ECs, but not normal ECs, may
express epidermal growth factor receptor (EGFR) and proliferate in response to EGF228.
They also show constitutive activation of PI3K–AKT signalling, which promotes cell
survival and resistance to apoptosis225.
Metabolism. Quiescent ECs display relatively high glycolysis rates1,160. However, tumour
ECs are hyper-glycolytic and largely use aerobic glycolysis to address their energy
requirements160.
Drug resistance. There is evidence for tumour ECs being more resistant than normal
ECs to various cytotoxic drugs229. For example, tumour ECs were shown to acquire
resistance to the cytotoxic agent paclitaxel through the upregulation of the
ATP-dependent efflux pump, P-glycoprotein 1 (PGY1; also known as ABCB1), which is
induced by VEGFA signalling230.
Genetic abnormalities. Gene and chromosomal abnormalities, including aneuploidy,
supernumerary centrosomes and translocations, have been documented in
subpopulations of tumour ECs of both mouse and human origin225,231. Tumour ECs may
accumulate genetic mutations through several routes. They produce substantial
amounts of reactive oxygen species (ROS) in response to cycles of anoxia–
reoxygenation (oxidative stress), and are directly exposed to ROS released by
tumour-infiltrating inflammatory cells and cancer cells. Furthermore, hypoxia represses
the cellular DNA repair machinery. Both processes, coupled to the high proliferation
rate of tumour ECs, can be directly mutagenic and also promote genetic instability in
tumour ECs232. Alternatively, genetic alterations in tumour ECs might result from the
direct trans-differentiation of cancer cells or, possibly, cancer stem cells into ECs233,234.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
459

Table 1 | Angiogenesis regulators in the TME
TME
component
Main angiogenesis regulators produced by
TME component
Effects of TME component on TABVs Refs
Macrophages VEGFA, FGF2, CXCL8, CXCL12, PlGF, VEGFC,
IL-1β, IL-6, TNF, WNT7B, MMPs and cathepsins
Pro-angiogenic; induce EC proliferation, migration and
survival, as well as ECM remodelling, to facilitate sprouting
angiogenesis
24,28–37,
40,41,67,
138
CXCL9, CXCL10, CXCL11 and TNF Potentially angiostatic under the influence of IFNγ and other
TH1 cytokines
83,85
Neutrophils and
MDSCs
VEGFA, FGF2, BV8 and MMP9 Pro-angiogenic; the role is well established during early
tumour stages or after therapeutic neutralization of VEGFA
13,54,55,57,
59,64,65
Mast cells FGF2, VEGFA, TNF, CXCL8, chymase, tryptase
and MMP9
Pro-angiogenic during the transition from non-angiogenic to
angiogenic tumours
72–74
Eosinophils VEGFA, FGF2, IL-6, CXCL8 and MMP9 Potentially pro-angiogenic, but relevance for tumour
angiogenesis unclear
77
TH2 cells IL-4 Potentially pro-angiogenic by stimulating the alternative
(M2-like) activation of TAMs
84
TH1 cells IFNγPotentially angiostatic through the induction of CXCL9,
CXCL10 and CXCL11 in TAMs or via direct angiostatic or
anti-angiogenic effects on ECs
80–83,85
TH17 cells IL-17 Pro-angiogenic by inducing CAFs to release CSF3, which
recruits pro-angiogenic neutrophils
219
Treg cells VEGFA Pro-angiogenic 86
B cells VEGFA, FGF2, MMP9 and IgG Potentially pro-angiogenic, either directly or via
IgG-dependent recruitment and activation of myeloid cells
78,79
Anti-VEGFA or anti-ANGPT2 IgG Potentially angiostatic through the production of
autoantibodies against pro-angiogenic cytokines in the
context of immunotherapy
222,223
NK cells VEGFA Potentially pro-angiogenic, but relevance for tumour
angiogenesis unclear
87,88
Platelets VEGFA, PDGFB, FGF2 and CXCL12 Pro-angiogenic 90,91,93,95–97
THBS1, PAI1, endostatin and ANGPT1 Potentially angiostatic 91,92,94
Pericytes VEGFA, ANGPT1 and ECM components Promote EC survival and, possibly, proliferation; they may
contriute to stabilization of TABVs
100,103
CAFs VEGFA, PDGFC, FGF2, CXCL12, osteopontin
and CSF3
Pro-angiogenic, both directly and indirectly by recruiting
myeloid cells and through ECM production
17,117–123
Adipocytes Adipokines and free fatty acids Pro-angiogenic and pro-inflammatory; stimulate
peri-tumoural angiogenesis
128,130,131,
134
ECM Periostin, tenascin C, fibronectin, osteopontin
and CCN-family proteins
Pro-angiogenic through the storage and concentration of
pro-angiogenic factors, and recruitment of pro-angiogenic
leukocytes
138,140–147
THBS1, osteonectin, decorin, proteolytic
fragments of typeIV and XVIII collagens
Potentially angiostatic 138,139,141
Hypoxia HIF-inducible genes: VEGFA, CXCL12 and
ANGPT2
Pro-angiogenic 2
Metabolites Lactate 153,154,157
H+Pro-angiogenic through increased expression and
stabilization of VEGFA mRNA
151,152
ROS Free radicals and non-radical ROS Potentially pro-angiogenic by enhancing HIF1 transcription
and the expression of pro-angiogenic and pro-inflammatory
factors; they also generate pro-angiogenic lipid oxidation
products
162,163
Tumour-
derived EVs
Various pro-angiogenic and inflammatory
mediators, ECM-remodelling enzymes and
mitogenic factors for ECs
Potential pro-angiogenic effects mediated via contacts
with, or transfer of their cargo to, ECs; relevance for tumour
angiogenesis unclear
176,178,182,
187
ANGPT, angiopoietin; CAF, cancer-associated fibroblast; CSF3, colony-stimulating factor 3; CXCL, CXC-chemokine ligand; EC, endothelial cell; ECM, extracellular
matrix; EV, extracellular vesicle; FGF2, fibroblast growth factor 2; HIF, hypoxia-inducible factor; IFNγ, interferon-γ; IgG, immunoglobulin G; IL, interleukin; MDSC,
myeloid-derived suppressor cell; MMP, matrix metalloproteinase; NK, natural killer; PAI1, plasminogen activator inhibitor 1; PDGF, platelet-derived growth factor;
PlGF, placental growth factor; ROS, reactive oxygen species; TABV, tumour-associated blood vessel; TAM, tumour-associated macrophage; TH, Thelper;
THBS1,thrombospondin 1; TME, tumour microenvironment; TNF, tumour necrosis factor; Treg cells, regulatory T cells; VEGF, vascular endothelial growth factor.
REVIEWS
460
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

RIP1–Tag2 transgenic mice
Expression of the SV40 large T
antigen (Tag) under the control
of the rat insulin promoter (RIP)
causes β-cell hyperplasia,
which progresses through a
series of rate-limiting stages
toinvasive pancreatic
neuroendocrine
tumour(PNET).
Vascular guidance
The guided, directional growth
of blood vessels.
angiogenesis36,42,43. Accordingly, the abundance of TIE2+
TAMs positively correlates with microvascular density
and/or distant metastasis in some types of human can-
cer33,44. Hypoxia-induced expression of CXCL12 (also
known as SDF1) and ANGPT2 stimulates the recruit-
ment and perivascular accumulation of TAMs that
express the respective cognate receptors CXC-chemokine
receptor 4 (CXCR4) and TIE2. These CXCR4+TIE2+
TAMs support angiogenesis in both treatment- naive43,45
and chemotherapy- or ionizing radiation-treated
tumours32,46–48. Ina transplant sarcoma model, TIE2
increased AKT activation in TAMs and protected them
from the pro- apoptotic effects of the chemotherapy drug
doxo rubicin49. Moreover, the genetic inactivation of Tie2
in TAMs impaired their ability to associate with imma-
ture blood vessels and sustain tumour angio genesis45 or
revascularization after chemotherapy49 in mouse tumour
models. Of note, genetic or pharmacological inhibition
of TIE2 in TAMs pheno copies some of the effects of
blocking EC-derived ANGPT2 in tumours, suggest-
ing that ANGPT2–TIE2 signalling regulates the pro-
angiogenic interactions between perivascular TAMs and
nascentTABVs36,45.
Additional cues may regulate TAM–EC interactions.
Notch signalling in macrophages has been implicated
in macrophage-assisted pathological angiogenesis50,
but currently its role in tumour angiogenesis is little
known. TAMs express vascular guidance molecules,
namely sema phorins, some of which modulate EC sur-
vival and migration51. The physical association between
TAMs and TABVs may, therefore, enhance EC survival,
activation and migration, to facilitate vascular growth
both in untreated tumours and during post-therapy
relapse36. Finally, macrophages were shown to perform
Nature Reviews | Cancer
CCL2
CSF1
VEGFA
VEGFR1 CCR2
VEGFR2
IL-4RαCSF1R
CSF1
IL-4
CSF1R
CSF3
BV8
IL-6
CSF3R
CSF3R
Blood vessel
anastomosis
TIE2
TIE2
CCL11 CSF2
CSF3
BV8
VEGFA
MMP9
Sprouting
angiogenesis
Cell differentiation
or recruitment by
tumour signals
Activation of
angiogenesis
Bone marrow
myeloid
precursors
Monocyte or M-MDSC
Granylocyte
or G-MDSC
TAM CCR3
TIE2-expressing TAM
TIE2-expressing
TAM
Neutrophil
Eosinophil
ECM
Angiogenesis regulators
VEGFA
FGF2
CXCL8
WNT7B
ANGPT2
IL-1β, IL-6
TNF
BV8
MMP2 or MMP9
Cathepsin
Figure 2 | Myeloid cell regulation of tumour angiogenesis. Various tumour-derived myeloid-cell chemoattractants — such
as CC-chemokine ligand 2 (CCL2), CCL11, colony-stimulating factor 1 (CSF1), CSF2, CSF3, vascular endothelial growth
factorA (VEGFA) and BV8 — recruit immature myeloid cells from the systemic circulation to the tumours. Upon their
extravasation, subsets of myeloid cells may differentiate into tumour-associated macrophages (TAMs), neutrophils and
eosinophils. Some myeloid cells maintain an immature phenotype in the tumour microenvironment (TME), and are referred to
as monocytic myeloid-derived suppressor cells (M-MDSCs) or granulocytic MDSCs (G-MDSCs). Myeloid cells promote tumour
angiogenesis by producing pro-angiogenic growth factors such as VEGFA, fibroblast growth factor 2 (FGF2), CXC-chemokine
ligand 8 (CXCL8), WNT7B and BV8. They also secrete various pro-inflammatory cytokines, namely interleukin-1β (IL-1β), IL-6
and tumour necrosis factor (TNF), and many proteases, including matrix metalloproteinases (MMPs) and cathepsins, which also
have pro-angiogenic roles. Myeloid-cell derived MMP9 mobilizes extracellular matrix (ECM)-bound VEGFA and enables its
binding to VEGF receptor 2 (VEGFR2), which is expressed on ECs, triggering angiogenesis. TIE2-expressing TAMs derive from
circulating monocytes; they associate with endothelial cells (ECs) and facilitate tumour angiogenesis by providing paracrine
pro-angiogenic and tissue-remodelling support to sprouting or anastomosing blood vessels. EC-derived angiopoietin 2
(ANGPT2) supports angiogenesis in an autocrine manner by binding to the TIE2 receptor but also promotes leukocyte
extravasation and mediates interactions between angiogenic ECs and TIE2-expressing TAMs. CSF1R, CSF1 receptor;
CCR,CC-chemokine receptor; IL-4Rα, IL-4 receptor-α.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
461

Vascular mimicry
The process whereby
vascular-like channels are
formed by non-endothelial
cells in certain tumours,
namely melanomas.
Non-thrombogenic EC-like
surfaces
Cellular surfaces capable of
preventing the formation of a
clot (or thrombus) when in
contact with blood.
TypeI interferon
A family of secreted proteins
with antiviral and
immunomodulatory functions,
which bind to a common
receptor.
Genetically engineered
mouse models (GEMMs)
ofcancer
Transgenic mice in which
cancer is initiated and driven
by defined genetic alterations,
such as the expression of
oncogene(s) and/or the
inactivation of tumour
suppressor gene(s), or both.
Immunocompetent mice
Mice that have an intact
immune system. They are
permissive to the growth of
transplanted tumours with
matched genetic background
(syngeneic tumours).
vascularmimicry52. Although macrophages have infiltrative
capa city and may transiently develop non- thrombogenic
EC-like surfaces53, it is currently unclear whether, and to
what extent, macrophage channels provide a scaffold for
subsequent endothelialization of bonafideTABVs.
Neutrophils. Neutrophils are the most abundant
granulo cytic population in the human blood and
generally account for a substantial proportion of the
haemato poietic cell infiltrate in experimental and
human cancers. CSF3 (also known as G-CSF) is a key
regulator of neutro phil production. CSF3 binds to its
receptor (CSF3R) expressed on neutrophil precursors
to activate the downstream Janus kinase (JAK)–signal
transducer and activator of transcription 3 (STAT3)
pathway, which promotes neutrophil proliferation and
expansion. Recruitment of neutrophils to tumours is
in part mediated by CXCL chemokines through the
cognate receptors CXCR1 and CXCR2 (REF.54).
Like macrophages, neutrophils are an important
source of pro-angiogenic factors and proteases in
theTME55 (FIG.2). In mice, STAT3 signalling controlsthe
pro-angiogenic functions of neutrophils and other mye-
loid cells by activating Veg f a, Fgf2 and Mmp9 transcrip-
tion56. Human neutrophils contain VEGFA-rich granules
that are rapidly deployed on stimulation with TNF57.
CSF3 induces neutrophils to upregulate the expression
of BV8 (also known as prokineticin 2; a hormone- like
protein) in a STAT3-dependent manner58; in turn,
BV8 promotes EC proliferation and angiogenesis in
tumours59. The pharmacological or genetic blockade of
CSF3, CSF3R or BV8 decreases intra-tumoural neutro-
phils and inhibits tumour angiogenesis and growth55.
Consistent with its angiostatic functions, typeI interferon
(IFN) signalling inhibits STAT3 activation and sup-
presses the production of VEGFA and MMP9 by neutro-
phils, hence limiting their pro-angiogenic capacity in
mouse models of cancer60.
The pro-angiogenic activity of neutrophils is cru-
cial during the early stages of tumour progression.
Neutrophil-derived MMP9 prompts the angiogenic
switch in RIP1–Tag2 mice13,55,61 by facilitating the mobi-
lization of ECM-bound VEGFA and its subsequent
binding to VEGFR2 on tumour ECs13. Accordingly,
neutro phil depletion by GR1 or LY6G antibodies delays
the angiogenic switch in both genetically engineered
mouse models (GEMMs) of cancer and tumours trans-
planted in immunocompetent mice60–62. Moreover, Mmp9
deficiency in myeloid cells impairs vascular matur ation
in transplant tumour models, suggesting that MMP9
also controls late events during tumour angio genesis63,64.
The absence of tissue inhibitor of metallo proteinases
(TIMPs) in complex with secreted pro-MMP9 is
required for the rapid activation and pro-angiogenic
capacity of secreted pro-MMP9 (REF.65). Neutrophils
are a key source of TIMP-free pro-MMP9, which
greatly exceeds the amount per cell produced by TAMs
in transplant tumour models64. However, considering
their abundance, TAMs may also provide a biologically
relevant source of MMP9 in the TME39,48,61. Proteases
released by activated neutrophils may also function as
negative regulators of angiogenesis. For example, pro-
teolysis of plasminogen by MMP9 and/or elastase lib-
erates angiostatin, which limits angiogenesis directly
by degrading VEGFA and FGF2, and indirectly by
preventing CXCL8-dependent neutrophil recruitment38.
Immature myeloid cells. In addition to mature macro-
phages and neutrophils, tumours contain abundant
infiltrates of immature myeloid cells, such as deactivated
dendritic cells (DCs) and myeloid-derived suppressor
cells (MDSCs). The latter comprise myeloid cells at vari-
ous stages of development and maturation, which can be
resolved into monocytic (M-MDSC) and granulocytic
(G-MDSC) cell populations66. Tumour-derived factors
such as CSF3, IL-1β and IL-6, fuel STAT3 activation
in MDSCs to promote their expansion, inhibit their
full maturation into macrophages or neutrophils and
enhance their pro-angiogenic functions in the TME66.
Although immature DCs and MDSCs display distinctive
metabolic properties and immunomodulatory capaci-
ties, their pro-angiogenic functions largely overlap with
those of mature macrophages and neutrophils67 (FIG.2).
M-MDSCs and macrophages have been thoroughly
characterized as both immunosuppressive and pro-
angiogenic in cancer66,68. Tcells extravasate to tumours
through a multi-step process that involves binding to cell
adhesion molecule (CAM)-family proteins expressed
on ECs. Under the influence of myeloid cell-derived
VEGFAand FGF2, the ECs of TABVs downregulate the
expression and abrogate the clustering of intercellular
adhesion molecule 1 (ICAM1) and vascular cell adhesion
molecule1 (VCAM1), hence limiting Tcell adhesion and
extra vasation68–70. These findings suggest that myeloid cells
may impair Tcell homing to tumours also through direct
effects of their pro-angiogenic products on TABVs68,70.
Mast cells and eosinophils. Mast cells are tissue-resident
granulocytes, the involvement of which in tumour angio-
genesis has long been postulated71. Mast cells release
pro-angiogenic factors, such as FGF2, VEGFA, TNF and
CXCL8, along with MMPs, including MMP9; they also
produce specific proteases (for example, chymase and
tryptase) that activate pro-MMPs72. The pro-angiogenic
functions of mast cells have been documented in GEMMs
of HPV16-driven skin cancer72, adenomatous polyposis
coli (Apc)Min intestinal adenoma73, and MYC-induced
PNET74. In these cancer models, mast cells surrounded
or infiltrated early pre-neoplastic lesions, and their inacti-
vation delayed the angiogenic switch and malignant pro-
gression. In PNETs, a mast cell inhibitor could also regress
established TABVs by inducing EC apoptosis74.
Eosinophils represent a minor granulocytic cell infil-
trate in experimental mouse tumours75. They are mainly
recruited to tumours by CC-chemokine ligand 11
(CCL11; also known as eotaxin) through CC-chemokine
receptor 3 (CCR3) and preferentially localize to hypoxic
areas in tumours76 (FIG.2). Eosinophils activated invitro
secrete various pro-angiogenic factors through degranu-
lation. They release VEGFA upon stimulation with
IL-5, whereas CCL11 and TNF prompt the secretion of
FGF2, IL-6, CXCL8 and MMP9, among others77. Further
REVIEWS
462
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

Thelper 1 (TH1) cells
T cells that can stimulate other
immune cells, such as
macrophages and cytotoxic T
lymphocytes, to kill infected or
cancer cells. They do so
primarily through the releaseof
the TH1 cytokine interferon-γ.
TH2 cells
T cells that stimulate B cells to
produce immunoglobulins.
They do so through the release
of TH2 cytokines, such as
interleukin-4.
Cytotoxic T lymphocytes
(CTLs). T cells capable of killing
other cells, including cancer
cells, generally through the
recognition of specific antigens.
studies are necessary to establish the importance of
eosinophil-derived pro-angiogenic factors for tumour
angiogenesis and revascularization after therapy.
Lym pho cytes . Lymphocytes are cells that accomplish
antigen-specific immune responses. By modulating
myeloid cell activation, B cells and Tcells may indi-
rectly control tumour angiogenesis. Furthermore, some
lymphocyte- derived cytokines directly influence EC
biology in tumours (FIG.3).
Bcells may facilitate angiogenesis in tumours by
expressing various pro-angiogenic mediators, includ-
ing VEGFA, FGF2 and MMP9, in a STAT3-dependent
manner78. They may also stimulate tumour angio genesis
indirectly through immunoglobulin G (IgG) and by
polarizing macrophages. For example, in a GEMM of
HPV16-driven skin cancer, deposition of Bcell-derived
IgG in the pre-malignant skin was shown to recruit and
activate pro-tumoural and pro-angiogenic TAMs, which
fostered skin carcinogenesis79. Pro-angiogenic TAM pro-
gramming was dependent on activating IgG receptors
(FcγRs) expressed on the macrophages. Indeed, mice
lacking FcγRs failed to mount a robust angiogenic
response and had delayed tumour progression and
reduced incidence of squamous cell carcinomas79.
There is also increasing evidence that T cells mod-
ulate tumour angiogenesis, both directly and indirectly.
Immunotherapy-elicited CD4+ Thelper 1 (TH1) cells can
directly inhibit tumour angiogenesis by enforcing the
maturation and/or quiescence of TABVs80. This process
may involve IFNγ, which restrains EC proliferation and,
when overexpressed experimentally, can cause the regres-
sion of immature blood vessels81,82. Tcells may also influ-
ence tumour angiogenesis indirectly. For example, CD4+
TH2 cells secrete IL-4 and stimulate the STAT6-dependent
alternative (or M2-like) activation of TAMs, which
entails immunosuppressive, tissue- remodelling and
pro-angiogenic functions83,84. Conversely, IFNγ secreted
by CD4+ TH1 cells or CD8+ cytotoxic T lymphocytes
(CTLs) may stimulate TAMs to upregulate the expres-
sion of the angiostatic cytokines CXCL9, CXCL10 and
CXCL11, in a STAT1-dependent manner83,85. There is,
Nature Reviews | Cancer
IL-4
M2-like
TAM
M1-like
TAM
IFNGR
IFNγ
IL-4Rα
IL-10
TGFβ
FcγR
IgG
IL-4Rα
Angiogenesis
B cell
TH1 cell
or CTL
TH2 cell
NK cell
IgG
Treg cell
Angiogenesis
Angiostasis
Angiogenesis regulators
VEGFA
FGF2
CXCL8
WNT7B
ANGPT2
IL-1β, IL-6
TNF
IFNγ
CXCL9 or CXCL10
MMP2 or MMP9
Cell activation through
cytokine release
Cell differentiation
Activation of angiogenesis
Inhibition of angiogenesis
IL-4
Figure 3 | Cross-talk between lymphocytes and myeloid cells regulates tumour angiogenesis. Tumour-infiltrating
Thelper 2 (TH2) cells secrete interleukin-4 (IL-4) and promote the differentiation of tumour-infiltrating monocytes and
macrophages into pro-angiogenic (M2-like) tumour-associated macrophages (TAMs). Conversely, TH1 cells and cytotoxic
Tlymphocytes (CTLs) secrete interferon-γ (IFNγ), which may stimulate monocytes and macrophages to exert angiostatic
(M1-like) functions through CXC-chemokine ligand 9 (CXCL9) and CXCL10 production. IFNγ can also inhibit tumour
angiogenesis directly by impairing endothelial cell (EC) proliferation. TH2 cells may activate humoral immunity, prompting
Bcells to secrete immunoglobulin G (IgG) that can stimulate pro-angiogenic macrophage programming via Fcγ receptor
(FcγR) engagement. TAM-derived immunosuppressive cytokines, such as IL-10 and transforming growth factor-β (TGFβ),
promote the expansion of regulatory T (Treg ) cells that sustain angiogenesis by releasing vascular endothelial growth factor A
(VEGFA). The tumour microenvironment (TME) may suppress natural killer (NK) cell cytotoxic activity and induce their
upregulation of VEGFA. In the context of cancer vaccines and immune checkpoint blockade, Bcells may secrete
autoantibodies against pro-angiogenic factors, namely angiopoietin2 (ANGPT2) and VEGFA, thereby neutralizing their
functions and disrupting tumour angiogenesis. FGF2, fibroblast growth factor 2; IFNGR, interferon γ receptor; IL-4Rα, IL-4
receptor-α; MMP, matrix metalloproteinase; TNF, tumour necrosis factor.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
463

Cancer immunosurveillance
The process whereby immune
cells, namely lymphocytes
andnatural killer cells,
recognize initiated cancer
cellsand eliminate them; it
may also lead to the selection
of less immunogenic cancer
cellclones.
however, little evidence for Tcells promoting angiostatic
or anti- angiogenic TAM reprogramming, at least in
treatment-naive, progressing tumours.
In contrast to TH1 cells and CTLs, immuno suppressive
regulatory T (Treg) cells seem to possess pro- angiogenic
capacities68. Treg cells may facilitate angiogenesis indi-
rectly by suppressing IFNγ-expressing effector TH1 cells.
Furthermore, hypoxia-induced CCL28 recruits Treg cells
that express VEGFA, and depletion of Treg cells abates
VEGFA levels and angiogenesis in the TME86. These obser-
vations add weight to the notion that immuno suppressive
cell networks involving myeloid cells and Tregcells not
only cause subsidence of anti tumour immunity, but also
function to stimulate tumour angiogenesis68.
Natural killer cells. Although natural killer (NK) cells
have important pro-angiogenic roles in the uterine
vasculature, their involvement in tumour angio genesis
is less well understood87. The genetic inactivation of
Stat5, which is required for NK cell-mediated cancer
immunosurveillance, upregulates VEGFA in NK cells and
enhances angiogenesis in mouse lymphoma models88.
Deactivated VEGFA-expressing NK cells have been
observed in various human cancer types (FIG.3), suggest-
ing potential associations between NK cell deactivation
and angiogenesis in progressing tumours87.
Platelets. The link between cancer progression and
thrombocytosis (increased platelet counts) is well estab-
lished89. Activated platelets are a rich source of pro-
angiogenic factors, including VEGFA, platelet- derived
growth factors (PDGFs) and FGF2 (FIG.4). They also
contain and deploy angiostatic molecules, such as throm-
bospondin 1 (THBS1), plasminogen activator inhibi-
tor1 (PAI1; also known as SERPINE1) and endostatin90.
Pro-angiogenic and angiostatic molecules are stored in
distinct α-granules, which may be selectively released
depending on the specific stimulus91. However, this
concept has been challenged by studies suggesting that
platelet secretion is, in fact, a stochastic process92.
In tumours, platelets are activated at sites of vascular
hyperpermeability and plasma leakage by contact with
collagen and cancer cells90,93. Tumours cause platelet
activation, aggregation and degranulation in their vas-
culature by expressing platelet-activating factors, such
as tissue factor (TF), thrombin and ADP. Tumour ECs
frequently overexpress TF, and positive correlations
between TF expression and microvessel density, or TF
and VEGFA expression, have been observed in several
cancer types90,93. Although disrupting platelet function
does not obviously impair tumour angiogenesis94, the
overall outcome of platelet activation and degranulation
in tumours appears to be pro-angiogenic90,93. In par-
ticular, platelet degranulation of VEGFA, CXCL12 and
PDGF, may initiate a ‘wound-healing’ response involv-
ing the recruitment and activation of myeloid cells and
cancer- associated fibroblasts (CAFs), and increased ECM
deposition, which in turn foster tumour angiogenesis14.
Interestingly, platelets can avidly sequester pro-
angiogenic factors in cancer-bearing hosts95, and plate-
lets isolated from cancer patients indeed contain higher
levels of pro-angiogenic factors compared with those
from healthy donors96,97. In one study, platelets were
shown to sequester pro-angiogenic factors from aggres-
sive mouse mammary tumours and to deploy them to
indolent tumours to induce angiogenesis and instigate
their progression98. Platelets can also promote angio-
genesis by stimulating the mobilization of myeloid cells
from the bone marrow and enhancing their homing to
tumours99. This may involve deployment to the bone
marrow niche of factors that had been sequestered at the
tumour site. Provocatively, the shuttling of sequestered
myeloid-cell chemoattractants (for example, CXCL12)
and pro-angiogenic mediators (for example, VEGFA)
by platelets99 might trigger the coordinate awaken-
ing of dormant disseminated cancer cells and thereby
induce metastatic outgrowth through the induction of
the angiogenic switch. Together, these findings illustrate
complex roles for platelets in the regulation of vascular
homeostasis and growth in tumours.
Pericytes. Pericytes are cells of mesenchymal origin that
enwrap and stabilize capillaries. They are embedded in
the basement membrane of small blood vessels and pro-
mote survival of ECs, while restraining their proliferation,
through the secretion of EC growth factors, MMP inhib-
itors and various ECM molecules. Pericytes also stabilize
EC junctions to limit vascular permeability100. At vari-
ance with quiescent capillaries, TABVs display uneven
and loose pericyte coverage11,101. The paucity of stable
pericyte–EC interactions in tumours enables sprouting
angiogenesis (FIG.4), but also generates a dysfunctional
vascular network characterized by EC hyperplasia,
defective cellular junctions and vascular leakiness11,100,101.
EC-derived PDGFB promotes the recruitment of
pericytes to the tumour vasculature100, whereas the
ANGPT–TIE2 system plays fundamental roles in regu-
lating subsequent pericyte–EC interactions102. The
binding of pericyte-derived ANGPT1 to TIE2 on ECs
inhibits EC proliferation, tightens EC junctions and
stabi lizes newly formed vessels103. Furthermore, peri-
cytes express neural cell adhesion molecule 1 (NCAM1)
and the NG2 proteoglycan, which contribute to vascu-
lar maturation by increasing pericyte recruitment104,105.
By contrast, angiogenic ECs produce ANGPT2, which
competes with ANGPT1 for binding to TIE2, disrupt-
ing pericyte–EC interactions and destabilizing the
TABVs to enable angiogenesis102. Accordingly, genetic or
pharmaco logical inhibition of ANGPT2 or TIE2 activa-
tion inhibits tumour angiogenesis and increases pericyte
coverage of the surviving blood vessels45,106–109.
Mounting data suggest that pericytes are hetero-
geneous cell subpopulations with different develop-
mental origins and diverse gene expression profiles100.
Two main pericyte subsets have been identified in mice,
termed type-1 (nestin−NG2+) and type-2 (nestin+NG2+)
pericytes110. Only type-2 pericytes were found in trans-
plant B16 melanoma and G26-H2 glioma tumour
models110, but it is currently unclear whether other
tumour models, such as GEMMs of cancer, also
lack type-1 pericytes. Tumour pericytes exert pro-
angiogenic functions and display an activated phenotype
REVIEWS
464
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

characterized by increased expression of α-smooth
muscle actin (αSMA; also known as ACTA2), regula-
tor of G-protein signalling-5 (RGS5) and endosialin,
and reduced levels of desmin and contractile proteins,
compared with normal-tissue pericytes70,100,110. Theacute
elimination of tumour pericytes, for instance, by PDGF
receptor (PDGFR) signalling inhibition or suicide
gene-based cell depletion approaches, disrupts angio-
genesis in both transplant and GEMMs of cancer111,112.
Furthermore, kinase inhibitors that concomitantly
block the VEGFRs and PDGFRs, such as sunitinib and
sorafenib, induce more pervasive and sustained TABV
regression than pure VEGFR inhibitors113,114. These find-
ings indicate that pericytes provide crucial pro-survival
cues to angiogenicTABVs.
Cancer-associated fibroblasts. CAFs have a key role in
producing a reactive stroma that frequently perpetuates
a tumour-promoting, tissue-repair response in solid
tumours17. CAFs largely derive from tissue-resident
fibroblasts that, under the influence of transforming
growth factor-β (TGFβ), acquire traits of functional
hyperactivation, including enhanced proliferation and
motility, along with robust ECM biosynthesis and depo-
sition capacity. Indeed, CAFs secrete enzymes, such as
lysyl oxidases (LOXs) and hydroxylases, which catalyse
the crosslinking of collagens to elastin and other ECM
molecules. By controlling the biomechanical properties
of the tumour stroma, including stiffness, elasticity and
interstitial fluid pressure, CAFs indirectly modulate
vascularization and blood flow in tumours115.
CAFs have well-established pro-angiogenic functions
in tumours (FIG.4). They often colocalize with TABVsin
human cancers, and co-implantation of CAFs and
cancer cells enhances angiogenesis, decreases cancer cell
dormancy and accelerates tumour growth in mice116,117.
CAFs are a major source of tumour VEGFA118,119, but
can also support tumour angiogenesis in a VEGFA-
independent manner120. CAF-derived PDGFC sus-
tains angiogenesis by further stimulating CAFs to
secrete pro-angiogenic growth factors, such as FGF2
and osteopontin121–123. The CAF secretome potentiates
Nature Reviews | Cancer
Cell differentiation or recruitment
by tumour signals
Activation of angiogenesis
Angiogenesis regulators
VEGFA
FGF2
CXCL12
TGFβ
PDGFB
PDGFC
ECM
Platelet
TGFβ
PDGFR
Basement
membrane
ANGPT1
Vascular
maturation
Vascular
maturation
CXCL12
VEGFA
ANGPT2 Angiogenesis
CAF
Pericyte
Vascular
leakage
Endothelial
cell
Fibroblast
Platelet
extravasation
and recirculation
Vascular
destabilization
VEGFR1
Granulocyte
or G-MDSC
Monocyte or
M-MDSC
CXCR4
Bone marrow
Myeloid
precursor cell
ECM
LOX
PDGFB
ANGPT1
Figure 4 | Chronic wound-healing response promotes tumour angiogenesis. Under the influence of transforming growth
factor-β (TGFβ) and other tumour-derived factors, peri-tumoural fibroblasts differentiate into cancer-associated fibroblasts
(CAFs), which secrete various components of the tumour extracellular matrix (ECM) and induce the crosslinking of collagen
fibres through lysyl oxidase (LOX) activity. Moreover, CAFs stimulate angiogenesis by secreting pro-angiogenic growth
factors, such as vascular endothelial growth factor A (VEGFA), fibroblast growth factor 2 (FGF2), CXC-chemokine ligand 12
(CXCL12) and platelet-derived growth factor C (PDGFC). The loose association of pericytes with tumour-associated blood
vessels (TABVs) favours chronic vascular leakage in tumours. This process is enhanced by autocrine angiopoietin 2 (ANGPT2)
signalling and is inhibited by ANGPT1 and PDGFB, which promote vascular maturation when VEGFA and ANGPT2 levels are
low. Platelet extravasation and degranulation at sites of vascular leakage liberates numerous pro-angiogenic mediators and
proteases, as well as cytokines and growth factors that support the proliferation and activation of CAFs, such as PDGFB
andTGFβ. Platelets may also sequester different tumour-derived factors, for example CXCL12 and VEGFA, in the tumour
microenvironment (TME) and deploy them to the bone marrow haematopoietic niche to enhance myelopoiesis and
myeloid-cell mobilization. CXCR4, CXC-chemokine receptor 4; G-MDSC, granulocytic myeloid-derived suppressor cell;
M-MDSC, monocytic myeloid-derived suppressor cell; VEGFR1, VEGF receptor 1.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
465

tumour angiogenesis also by attracting vascular ECs
and recruiting monocytes from the bone marrow, for
example, through the CXCL12–CXCR4 axis17,117. In mel-
anoma, aged CAFs secrete the WNT antagonist secreted
frizzled-related protein 2 (SFRP2), which exacerbates
the angiogenic and malignant behaviour of tumours in
old individuals124. Although CAFs also secrete angio-
genesis inhibitors, such as THBS1 (REF. 17), tumours
may overcome their angiostatic properties by adaptively
increasing the production of pro-angiogenic factors125.
Adipocytes. Adipose tissue may foster the growth of
initiated cancer cells through the promotion of angio-
genesis. Indeed, tumour growth was accelerated when
cancer cells were implanted in the white or brown adi-
pose tissue of mice, compared with the subcutaneous
space126,127. The tumours implanted in adipose tissue
displayed a more florid vascular network than those
implanted subcutaneously, suggesting a potential role for
adipocytes in accelerating angiogenesis127. Tumours that
arise in or in proximity to adipose tissue (for example,
breast, ovarian or colon cancers, as well as bone or lymph
node metastases) are exposed to a milieu of cytokines,
chemokines and hormones, collectively termed
adipo kines, some of which have well-established pro-
angiogenic functions (FIG.5). Pro-angiogenic adipokines
are secreted by adipocytes, infiltrating inflammatory
cells and other adipose tissue-associated stromal cells,
and may either target vascular ECs directly or recruit
vascular-modulatory inflammatory cells128.
Of note, obesity is associated with increased risk
of several cancer types129. The adipose tissue of obese
individuals is not only enlarged, but also chronically
inflamed and adipokine rich. Adipocytes isolated
from obese individuals enhanced EC proliferation
and migration invitro to a greater extent than adipo-
cytes from non-obese individuals130. Furthermore, in a
mouse mammary tumour model, pre-existing obesity
Nature Reviews | Cancer
Cell differentiation or
recruitment by tumour signals
Activation of angiogenesis
VEGFA
Adipokines
CCL2
CXCL12
MMPs
Angiogenesis regulators
CSF2
CSF3
CXCL8
VEGFA
Lactate
Lactate
Fatty acids
Adipocyte
reprogramming
Cancer-activated
adipocyte
Cancer
cells
GLUT1
MCT1
MCT1
Lipolysis
FABP4
Angiogenesis
Adipocytes
Oncogenes
• Hypoxia
• Low pH
• Low glucose
• High ROS
GLUT1
ROS
Angiogenesis
M2-like
TAM
TAM
Neutrophil
Figure 5 | Metabolic regulation of tumour angiogenesis. Under the influence of tumour-derived factors, peri-tumoural
adipocytes increase their lipolytic activity and liberate free fatty acids that, upon internalization by vascular endothelial cells
(ECs) through fatty acid-binding protein 4 (FABP4), increase the rate of β-oxidation in the tumour-associated blood vessels
(TABVs) to sustain angiogenesis. Cancer-activated adipocytes also liberate adipokines, a heterogeneous assortment of
growth factors, cytokines and hormones that promote tumour angiogenesis. Adipokines potentially relevant to tumour
angiogenesis include leptin, adiponectin, resistin, visfatin, oestrogens, tumour necrosis factor (TNF), interleukin-1β (IL-1β)
andIL-6, insulin growth factor 1 (IGF1), vascular endothelial growth factor A (VEGFA), fibroblast growth factor 2 (FGF2),
hepatocyte growth factor (HGF), angiopoietins (ANGPTs), CC-chemokine ligand 2 (CCL2) and colony-stimulating factor 2
(CSF2) (reviewed in REF.128). Chronic hypoxic conditions in the tumour microenvironment (TME) promote the expression of
hypoxia-inducible factor (HIF)-induced pro-angiogenic mediators, namely VEGFA and CXC-chemokine ligand 12 (CXCL12),
which directly stimulate tumour angiogenesis. Furthermore, sustained oncogenic signalling in the cancer cells is associated
with the upregulation of various myeloid-cell chemoattractants and activators, such as CSF2, CSF3, CXCL8 and VEGFA.
Metabolically active cancer cells secrete lactate, which is internalized by ECs and tumour-associated macrophages (TAMs)
through the lactate importer monocarboxylate transporter 1 (MCT1). Lactate stimulates tumour angiogenesis both by acting
directly on ECs and indirectly by promoting M2-like TAM programming. Tumour ECs respond to hypoxia and acidosis by
upregulating mediators of the glycolytic pathway, such as glucose transporter 1 (GLUT1). Finally, under the influence of
VEGFA, the ECs of TABVs produce various reactive oxygen species (ROS), which promote EC proliferation and angiogenesis
under conditions of metabolic stress. MMP, matrix metalloproteinase.
REVIEWS
466
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

Lipolysis
The process whereby
triglycerides are resolved into
glycerol and free fatty acids
through hydrolysis.
β-Oxidation
The process that occurs in the
mitochondrion and that uses
fatty acids to generate
acetyl-CoA, which is essential
for producing ATP through
oxidative phosphorylation.
facilitated tumour growth by inducing angiogenesis131.
Both human and mouse mammary adipocytes were
shown to recruit and activate macrophages through
a CCL2–IL-1β–CXCL12 signalling pathway. In turn,
the activated macrophages promoted stromal angio-
genesis before the appearance of cancer nodules131.
Consistent with these findings, leukaemic cells were
shown to preferentially thrive in so-called ‘milky spots’
— aggregates of immune cells embedded in highly vas-
cularized adipose tissue — in an experimental model of
peritonealmetastasis132.
Compared with normal adipose tissue, the peri-
tumoural adipose tissue is highly vascularized and
macro phage rich, and produces higher levels of pro-
teases, ECM proteins and various pro-angiogenic adi-
pokines. Moreover, cancer cells reprogramme adjacent
adipocytes to acquire an activated phenotype character-
ized by reduced cell size and sustained lipolysis126,133. Asa
result of increased lipolysis, cancer-associated adipo-
cytes may supply fatty acids to metastatic ovarian cancer
cells through the chaperone protein fatty acid-binding
protein 4 (FABP4), boosting β-oxidation of fatty acids
in cancer cells134. Of note, not only cancer cells but also
VEGFA-stimulated, angiogenic ECs express high levels
of FABP4 (REF.135). Angiogenic ECs can utilize fatty acid
β-oxidation for their growth136, so cancer-associated
adipo cytes may directly fuel peri-tumoural angiogenesis.
The extracellular matrix
The ECM is an intricate network of fibrous proteins,
glycosaminoglycans and matricellular proteins that pro-
vide structural support as well as biochemical and bio-
mechanical cues for cancer cell growth115,137,138. Vascular
ECs and mural cells produce a specialized ECM, the
basement membrane, which is crucial for blood vessel
integrity and function. Sprouting angiogenesis involves
the degradation of the basement membrane by MMPs
produced by activated ECs and recruited myeloid
cells, the ensuing formation of a provisional fibrin-
and fibronectin-rich ECM that supports EC prolifer-
ation andmigration, and the ultimate reassembly of a
mature basement membrane that, in the context of non-
pathological angiogenesis, contributes to EC quiescence
and vascular integrity1,138,139. Sustained pro-angiogenic
signalling in tumours impairs the subsequent steps of
vascular morpho genesis, namely the acquisition of a qui-
escent EC phenotype and the development of an intact
and selectively permeable vascular barrier1.
In tumours, the vascular basement membrane is fre-
quently discontinuous and loosely associated with ECs
and pericytes140, which contributes to increasing vas-
cular leakiness and facilitates cancer cell intra vasation
and metastasis1,115. Furthermore, the composition,
topo graphy and ligand density of both the vascular and
interstitial ECM are altered in tumours138. The ECM may
have both pro-angiogenic and vascular-stabilizing roles.
It serves as a depot for various pro- angiogenic growth
factors, notably VEGFA, FGFs, PDGFB and TGFβ,
which are released in their bioactive forms through
the proteolytic processing of the ECM by plasmin,
MMPs and other proteases5,138,139. The breakdown of
the ECM may also generate chemoattractants for pro-
angiogenic inflammatory cells, such as TAMs141. Direct
pro- angiogenic activities have been described for many
tumour ECM molecules, such as periostin, tenascins,
fibronectin, perlecan, osteopontin and CCN-family
proteins142–144. For example, in the RIP1–Tag2 PNET
model, tenascin C sustained angiogenesis by downregu-
lating Dickkopf-related protein 1 (DKK1) and increasing
WNT signalling145. Conversely, several ECM matricellu-
lar proteins, such as THBS1, osteonectin (also known as
SPARC) and the proteoglycan decorin, may exert angio-
static functions146,147. Sustained ECM remodel ling in
tumours may also generate biologically active fragments
of typeIV and XVIII collagens, which limit angiogenesis
by competing with intact collagen fibres for interaction
with EC integrins138,139.
The biophysical and mechanical properties of
the tumour ECM, such as the altered geometry and
increased density and crosslinking of collagen fibres,
influence tumour angiogenesis both directly and indi-
rectly 115,138. In experimental matrices, ECM stiffness
and contractility modulate the spatial organization of
VEGFA gradients and VEGFR2 expression by ECs148,149.
The abnormal arrangement of ECM fibres facilitates
tumour angiogenesis also by enhancing the migration
of ECs and pro-angiogenic TASCs, such as TAMs and
CAFs. Indeed, these cells migrate more rapidly on lin-
earized collagen fibres, which are enriched in tumours
compared with non-neoplastic tissues115,150.
Tumour metabolism
The key role of hypoxia in tumour angiogenesis is well
established2. The transcriptional activity of hypoxia-
inducible factor 1 (HIF1) induces the expression of sev-
eral pro-angiogenic genes, such as VEGFA, VEGFR2,
DLL4, CXCL12 and ANGPT2, in both cancer cells and
TASCs2. Under hypoxic conditions, cancer cells consume
glucose and secrete lactate, which generates an acidic
TME (FIG.5). Glucose deprivation and acidosis increase
VEGFA mRNA stability post-transcriptionally in the
cancer cells151,152. Also, ECs internalize cancer cell- derived
lactate through the lactate importer monocarboxy-
late transporter 1 (MCT1; also known as SLC16A1),
which enhances angiogenesis in a nuclear factor-κB
(NF-κB)- and HIF1-dependent manner153,154.
Similarly to cancer cells, TASCs also respond to
hypoxia2. Hypoxic conditioning of TASCs modulates the
tumour metabolic landscape and angiogenesis. Hypoxia
stimulates CAFs to secrete ECM-remodelling enzymes
and HIF-inducible pro-angiogenic factors (for exam-
ple, CXCL12), which facilitate tumour angiogenesis2.
Macrophages accumulate in hypoxic tumour regions155
and around nascent (non-perfused) TABVs29,36,45. These
hypoxic microenvironments fine-tune the activation of
TAMs and stimulate pro-angiogenic gene transcription156.
In analogy, lactate induces pro-angiogenic (M2-like)
TAM activation in a HIF1α-dependent manner157.
Hypoxic TAMs express several HIF-dependent glyco-
lytic genes, suggesting that they preferentially utilize a
glycolytic metabolism158. However, under the influence
of TH2 cytokines (for example, IL-4), TAMs may tune
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
467

Glycolysis
The metabolic process that
occurs in the cell cytoplasm
and that uses glucose to
generate pyruvate and the
high-energy molecules ATP and
NADH; in the presence of
oxygen, pyruvate may enter
the mitochondrion to sustain
oxidative metabolism.
Reactive oxygen species
(ROS). Chemically reactive
molecular species that contain
oxygen; by reacting with
biological molecules, ROS
canalter their structure
andfunction.
Oxidative metabolism
The metabolic processes
thatconverge on oxidative
phosphorylation to
produceATP.
Extracellular vesicles
(EVs). The heterogeneous
assortment of secreted vesicles
produced by virtually any cell
type through diverse
biogenesis processes.
Tetraspanins
A family of transmembrane
proteins that organize
microdomains enriched
inmembrane-bound
signallingproteins.
Orthotopic tumour
transplant
An experimental tumour that
results from the injection of
cancer cells into the tissue or
organ from which the cancer
cells were originally derived.
Ectopic tumour transplant
An experimental tumour that
results from the injection of
cancer cells into an anatomical
site that is different from the
one from which the cancer cells
were originally derived.
Generally, ectopic tumours
areinoculated in the
subcutaneous space.
down glycolysis and enhance oxidative phosphorylation,
a metabolic switch that is associated with the acquisition
of immunosuppressive and pro-angiogenic functions158.
Attenuated glycolysis in TAMs was also shown to facili-
tate glucose consumption by tumour ECs, leading to their
acquisition of robust angiogenic capacities159. Despite
their proximity to blood oxygen, tumour ECs mainly rely
on aerobic glycolysis (rather than oxidative phosphoryla-
tion) and fatty acid oxidation for their bioenergetics and
biosynthetic needs160. Such metabolic reprogramming
enables ECs to create new blood vessels while maximizing
oxygen transfer to surrounding tissues, limiting the pro-
duction of reactive oxygen species (ROS), and producing
ATP more rapidly than through oxidative metabolism160.
Cycles of hypoxia–reperfusion, high metabolic
activity and sustained oncogenic signalling, together
induce unbalanced ROS production in tumours161.
Depending on the exact species, concentration and cel-
lular source, ROS can either promote or inhibit tumour
angio genesis.ROS production in both cancer cells and
TASCs cell- autonomously induces VEGFA transcrip-
tion through HIF1, and also generates lipid oxidation
metabolites, such as end-products of docosahexaenoic
acid oxidation, which induce tumour angiogenesis in a
VEGFA-independent and Toll-like receptor 2 (TLR2)-
dependent manner162,163. Although ROS production
in ECs may sustain tumour angiogenesis164, excessive
ROS levels may blunt EC responsiveness to extracellular
VEGFA by increasing VEGFR2 recycling165, a process
that can be bypassed through the induction of anti-
oxidative responses166. It remains unclear whether ROS
generated by cancer cells and TASCs in highly hypoxic
and inflamed TMEs can directly penetrate ECs to
influence their angiogenic properties.
Finally, there is also experimental evidence for
oncogenic drivers to control tumour angiogene-
sis. Constitutively activated RAS and RAF proteins
directly induce the expression of pro-angiogenic fac-
tors, such as VEGFA and CXCL8, in cancer cells167–170.
Furthermore, mutant oncogenes may also elicit pro-
angiogenic responses indirectly (for instance, by induc-
ing the expression of myeloid cell chemoattractants171,172).
However, there is currently little clinical evidence that
specific oncogenes, such as mutant KRAS, confer higher
sensitivity to anti-angiogenic therapy, for example, in
colorectalcancer173,174.
Tumour-derived extracellular vesicles
Cancer cells and TASCs secrete various vesicles of dif-
ferent sizes, together referred to as extracellular vesicles
(EVs), which contain proteins, nucleic acids and lipids
that in part reflect the biomolecular composition of
the cell of origin175. The hypoxic and acidic TME may
enhance the production of tumour-derived EVs,and
increasing data suggest that tumour- derivedEVs
can influence vascular function, both locally in
tumours and remotely in distant organs through the
systemiccirculation29,175,176.
EVs secreted by cancer cells have been shown to
contain pro-angiogenic mediators, including VEGFA,
CXCL8, IL-6 and FGF2 (REFS177,178). Of note, the
acidic TME may facilitate EV disruption178, enabling
the interaction of pro-angiogenic molecules with
cognate receptors expressed on tumour ECs. Cancer
cell-derived EVs may also deploy pro-angiogenic ECM-
remodelling enzymes, such as urokinase plasminogen
activator (uPA), MMP2 and MMP9 (REFS 178,179).
Besides cancer cells, several TASC types produce EVs
with potential pro-angiogenic functions. These include
macrophages180, platelets181 and ECs182,183.
The EV surface displays several tetraspanins184. The
expression of certain tetraspanins on cancer cell- derived
EVs was shown to enhance EV internalization by ECs,
which in turn stimulated the transcription of angio-
genesis-related genes and promoted EC proliferation
and migration185. The fusion of tumour EVs with the
plasma membrane of ECs may also be conducive to
the horizontal transfer of mitogenic RNAs or proteins,
which may influence the biology of the recipient ECs
and stimulate tumour angiogenesis182,186–188. For example,
EVs secreted by cancer cells were reported to transfer
mutant epidermal growth factor receptor (EGFR) to
tumour ECs, inducing mitogenic MAPK and AKT sig-
nalling activation187. However, although an increasing
number of studies document the transfer of functional
macromolecules from tumour-derived EVs to TABVs
or vascular beds in pre-metastatic sites, the mechan-
istic underpinning and clinical implications of these
phenomena remain poorly understood175,184.
Distinct organ microenvironments
Heterogeneous vascular morphology and blood ves-
sel patterns are observed across distinct tumour types
and in different microenvironments of individual
tumours6–10. In mouse transplant cancer models, the site
of tumour inoculation (for example, orthotopic tumour
transplant versus ectopic tumour transplant) can markedly
influence angiogenesis along with tumour histopathol-
ogy, gene expression and several parameters of cancer
progression6,7. Also, the structure and density of metasta-
sis-associated blood vessels vary considerably according
to the location of the metastatic site after dissemination
from a primary human tumour9,10. Of note, the genetic
background of the mouse influences innate and adaptive
immune cell biology, which in turn may reverberate on
tumour angiogenesis. As a consequence of these many
variables, experimental tumours may substantially dif-
fer from human tumours in terms of vascular density,
functionality and phenotype, as well as responsiveness
to anti-angiogenic therapy.
Mouse cancer models typically display fast growth
kinetics and a ‘pushy’ (expansive) growth pattern, which
may exacerbate the requirement for angiogenesis and
the involvement of pro-angiogenic TASCs. Although
sprouting angiogenesis undeniably contributes to
human tumour vascularization5,189, non-angiogenic
modes have also been observed, especially in metastatic
human cancers. Vascular co-option — the infiltrative
growth of cancer cells along pre-existing host vessels —
has been documented in tumours that develop in highly
vascularized organs, such as the lung, liver, brain and
lymph nodes10,190–192. Remarkably, the analysis of 164
REVIEWS
468
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

human lung metastasis specimens derived from primary
cancers of the breast, colon or kidney found evidence for
vessel co-option in 80% of the cases190. Although mye-
loid cells have been implicated in metastasis-associated
angiogenesis in mouse models193,194, the contribution of
TASCs to vessel co-option in both primary and meta-
static tumours has been poorly studied, and further work
in this area isneeded.
Implications for anticancer therapies
As discussed above, the mechanisms involved in the
induction and maintenance of the tumour vasculature
are diverse and robust, and involve the action of mul-
tiple biochemical mediators and cell types with osten-
sibly redundant pro-angiogenic functions. Accordingly,
the pharmacological inhibition of VEGFA signalling
inhibits tumour angiogenesis in some but not all mouse
cancer models, and it typically does not block tumour
progression in mice and humans189. By striking con-
trast, the genetic inactivation of Veg f a impairs devel-
opmental angiogenesis and is embryonic lethal1. These
observations support the notion that the regulation
of tumour angiogenesis is a multidimensional pro-
cess that is less dependent on VEGFA signalling than
developmental angiogenesis. Furthermore, tumours
can rapidly adapt to the neutralization of individual
pro-angiogenic growth factors, including VEGFA189,195,
through routes that involve metabolic adaptation and
reprogramming196–201, the enforcement of compen-
satory pro-angiogenic signals108,202,203 or the acquisi-
tion of angiogenesis-independent modes of tumour
growth190,191,204,205 (FIG.6a–c).
Myeloid cells, macrophages and neutrophils in par-
ticular sustain both VEGFA-dependent and independent
angiogenesis in tumours. This is particularly relevant con-
sidering the notion that anti-angiogenic drugs provoke the
surge, in tumours, of hypoxia-inducible chemo attractants
for myeloid cells, which can rescue angio genesis through
VEGFA-independent pathways48,55,59,195,206,207. Strategies
that selectively impair pro-angiogenic macro phages,
for example, perivascular TIE2+ TAMs, may help to dis-
rupt compensatory pro-angiogenic cues36 while sparing
TAM subpopulations that have potential roles in antigen
presentation or the production of angiostatic factors in
response to TH1 cytokines18,83,85. Blocking ANGPT2 or
TIE2 signalling decreases TIE2+ TAMs208, impedes their
association with angiogenic blood vessels45 and increases
the proportion of TAMs that exhibit an M1-like (angio-
static) phenotype209,210. However, different myeloid-cell
types may contribute to limiting tumour responsiveness
to anti-angiogenic therapies62,67,211. For example, macro-
phage or neutrophil elimination in a mouse PNET
model did not impede the emergence of resistance to
sorafenib, an anti-angiogenic multi-kinase inhibitor,
but depleting both cell types improved the therapeutic
benefits62. These preclinical findings are consistent with
initial reports showing that anti-macrophage drugs, for
example, CSF1R inhibitors, have limited therapeutic
activity in patients with cancer212,213. Therefore, broadly
targeting myeloid cells may be required for effective abla-
tion of their pro-angiogenic capacity in the context of
cancer treatment. A promising targeted approach might
be inhibiting the γ-isoform of PI3K (PI3Kγ), which is
preferentially expressed in myeloid cells and sustains
their immunosuppressive and pro-angiogenic func-
tions62,214. Furthermore, TAMs can also be engineered to
express biologics that inhibit tumour angiogenesis and
reprogramme theTME215.
Pericytes have emerged as important regulators
of tumour angiogenesis and revascularization post-
therapy 70. Interestingly, regression of TABVs in response
to anti-VEGFA therapy leaves empty sleeves of basement
membrane and pericytes45,216, which provide a guiding
scaffold for rapid tumour revascularization after ther-
apy withdrawal216. Co-targeting pericytes and ECs
with inhibitors that potently block both PDGFRs and
VEGFRs, such as sunitinib, delays tumour revascular-
ization post-therapy compared with selective VEGFR
inhibitors113,114,195 and has clinical efficacy in PNETs217.
However, there is also evidence for pericytes limiting
cancer-cell intravasation and metastasis104. The latter
observation may explain the propensity of sunitinib
to increase metastasis from primary tumours in some
cancer models218. Because metastatic dissemination
from pericyte-depleted tumours may rely, at least in
part, on the acute release of ANGPT2 from sensitized
ECs, co-targeting ANGPT2 may serve to blunt the
pro- metastatic potential of pericyte elimination and to
improve the therapeutic benefits112.
There is increasing evidence that acute vascular
pruning by potent angiogenesis inhibitors may exacer-
bate or even instigate the pro-tumoural capabilities of
TASCs189,195,207. Notably, interception of VEGFA signal-
ling enhances M2-like TAM polarization108,206 and pro-
angiogenic CAF programming120. Moreover, TH17 cells
were shown to produce IL-17 in response to anti-VEGFA
therapy; in turn, IL-17 induced tumours to release CSF3,
which promoted VEGFA-independent tumour revascu-
larization and regrowth through neutrophil recruit-
ment219. Conversely, defined regimens of anti-angiogenic
drugs may normalize, rather than regress, TABVs in a
process involving the selective pruning of immature
capillaries and the concomitant stabilization of perfused
vessels220 (FIG.6d). Both suboptimal VEGFA neutralization
and ANGPT2 inhibition have been reported to normal-
ize TABVs, which may improve chemotherapy delivery,
enhance radiosensitivity and facilitate Tcell extravasation
in tumours70,209,210,220. The design of anti-angiogenic treat-
ments needs to incorporate these complexities in order
to maximize the therapeutic benefits in cancer patients.
This remains a challenging clinical task and open area of
preclinical research.
Concluding remarks
Most of the studies discussed in this Review employed
mouse cancer models as a platform for mechanistic
investigations of molecular or cellular players involved
in tumour angiogenesis. The many idiosyncratic details
inherent to each tumour model and its underlying biol-
ogy determine the experimental results and may limit
the applicability and relevance of the selected model to
human pathology. Interrogating the vascular-modulatory
TH17 cells
T cells that have roles in
protecting organ surfaces, in
particular the gut mucosa,
from pathogens. They produce
interleukin-17 and stimulate
Bcell-mediated humoral
immunity.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
469

Fig 6
Nature Reviews | Cancer
ANGPT1
Lactate
Re-vascularization
VEGFA inhibition
vascular regression Hypoxia
GLUT1
MCT1
Vascular regression
FABP3
and FABP7 Vascular
perfusion
Autophagy
Lipid
storage
Metabolic adaptation
Organ tissues
Tumour
a b
c d
CXCL9
CXCL10
Tumour necrosis
Pro-angiogenic TAM
Glucose
ANGPT2 Pericyte
CSF2
CSF3
CXCL12
CXCL12
FGF2
Angiogenesis Normalization Regression
Vessel co-option
VEGFA
(and others)
Activation of angiogenesis
Inhibition of angiogenesis
Cell differentiation or recruitment
by tumour signals
BV8
PlGF
CXCL8
M2-like
TAM
M2-like
TAM M1-like
TAM
Neutrophil Monocyte
Angiogenesis
re-oxygenation
TH1 cell
or CTL
Vascular regression and hypoxia
Figure 6 | Mechanisms of tumour escape from angiogenesis inhibition. Tumours can adapt to the acute neutralization
of key pro-angiogenic growth factors, including vascular endothelial growth factor A (VEGFA). a|By acutely disrupting
tumour angiogenesis and perfusion, anti-VEGFA therapy can activate metabolic or stress responses in the cancer cells,
which enable their survival under hostile conditions of oxygen and nutrient deprivation. Such mechanisms include
increased autophagy196 or the establishment of ‘metabolic symbiosis’ — the process whereby hypoxic cancer cells in
avascular tumour areas import glucose and export lactate, while normoxic cells in proximity to the surviving blood vessels
import and catabolize lactate198–200. Acute hypoxia may also precondition cancer cells for re-growth after withdrawal of
anti-angiogenic therapy by inducing fatty acid uptake and storage through the transporters fatty acid-binding protein 3
(FABP3) and FABP7 (REF.201 ). Accordingly, tumour re-growth after anti-angiogenic kinase inhibitors may rely on denovo
lipogenesis197. b|Alternative mechanisms of tumour adaptation to VEGFA deprivation include the induction of
compensatory pro-angiogenic growth factors, namely fibroblast growth factor 2 (FGF2), angiopoietin 2 (ANGPT2),
placental growth factor (PlGF) and BV8, which can rescue angiogenesis in VEGFA-depleted tumours. Furthermore, various
anticancer drug regimens provoke the surge, in tumours, of hypoxia-inducible chemoattractants for neutrophils and
macrophages, including colony-stimulating factor 2 (CSF2), CSF3 and CXC-chemokine ligand 12 (CXCL12), which recruit
angiogenesis-promoting myeloid cells. c|Cancer cells may circumvent dependence on angiogenesis by acquiring the
ability to hijack the pre-existing vasculature through an infiltrative growth mode called vascular co-option. Cancer-cell
growth along existing blood vessels has been implicated in tumour resistance to anti-angiogenic therapy in both
preclinical cancer models and patients with colon cancer liver metastases. d|Anti-angiogenic drugs may paradoxically
improve blood flow by normalizing the tumour-associated blood vessels (TABVs). Vascular normalization can be achieved
by attenuating pro-angiogenic signalling in tumours (for example, by chronically reducing VEGFA bioavailability or
blocking ANGPT2). Vascular normalisation may reshape the immune cell repertoire of tumours and facilitate antitumour
immunity, for example, by improving Tcell extravasation or by promoting the conversion of pro-angiogenic (M2-like) into
angiostatic (M1-like) tumour-associated macrophages (TAMs). CTL, cytotoxic T lymphocyte; GLUT1, glucose transporter 1;
MCT1, monocarboxylate transporter 1; TH1 cell, T helper 1 cell.
REVIEWS
470
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

functions of TASCs in patients with cancer is, therefore,
needed to understand the extrinsic regulation of human
tumour vascularization, in particular in the context of
metastatic disease and/or under therapeutic pressure.
The clinical testing of an expanding arsenal of drugs
targeting specific cellular components of the TME, such
as TAMs85,212,213, pericytes113,114,195,217 and Tcells210,221–223,
may provide clues about the roles played by these cells
in the regulation of human tumour angiogenesis. For
example, recent clinical studies have revealed unexpec-
ted roles for adaptive immune cells in the regulation of
human TABVs221–223. Tumour vascular destruction and
important humoral (IgG-mediated) reactions against
ANGPT2 and VEGFA were observed in long-term
responding patients who had received a cancer vaccine222.
Further studies illustrated that tumour regressions after
immune checkpoint blockade with antibodies targeting
cytotoxic T lymphocyte-associated antigen 4 (CTLA4)
or programmed cell-death protein 1 (PD1) were associ-
ated with heightened titres of anti-ANGPT2 serum
IgG, whereas therapy refractoriness or resistance were
associated with higher pre- or on-treatment ANGPT2
serum levels223. These provocative findings suggest that
tumour responses to immunotherapy may involve, or
even require, immune-mediated anti-angiogenic mech-
anisms. Therefore, if combined with the analysis of the
molecular, morphological, and functional properties
of TABVs, either on tumour biopsies or through non-
invasive imaging tools27,33,63,220, the clinical deployment
of TME-targeted drugs may help to shed new light
on the vascular- modulatory functions of TASCs in
humancancer.
1. Potente,M., Gerhardt,H. & Carmeliet,P. Basic
andtherapeutic aspects of angiogenesis. Cell 146,
873–887 (2011).
2. LaGory,E.L. & Giaccia,A.J. The ever-expanding role
of HIF in tumour and stromal biology. Nat. Cell Biol.
18, 356–365 (2016).
3. Bossi,P. etal. Angiogenesis in colorectal tumors:
microvessel quantitation in adenomas and carcinomas
with clinicopathological correlations. Cancer Res. 55,
5049–5053 (1995).
4. Bluff,J.E. etal. Angiogenesis is associated with the
onset of hyperplasia in human ductal breast disease.
Br. J.Cancer 101, 666–672 (2009).
5. Bergers,G. & Benjamin,L.E. Tumorigenesis and the
angiogenic switch. Nat. Rev. Cancer 3, 401–410
(2003).
6. Fukumura,D., Yuan,F., Monsky,W.L., Chen,Y. &
Jain,R.K. Effect of host microenvironment on the
microcirculation of human colon adenocarcinoma.
Am.J.Pathol. 151, 679–688 (1997).
7. Monsky,W.L. etal. Role of host microenvironment in
angiogenesis and microvascular functions in human
breast cancer xenografts: mammary fat pad versus
cranial tumors. Clin. Cancer Res. 8, 1008–1013
(2002).
8. Blouw,B. etal. The hypoxic response of tumors is
dependent on their microenvironment. Cancer Cell 4,
133–146 (2003).
9. Morrissey,C. etal. Differential expression of
angiogenesis associated genes in prostate cancer
bone, liver and lymph node metastases. Clin. Exp.
Metastasis 25, 377–388 (2008).
10. Jubb,A.M. etal. Vascular phenotypes in primary
non-small cell lung carcinomas and matched brain
metastases. Br. J.Cancer 104 , 1877–1881 (2011).
11. Morikawa,S. etal. Abnormalities in pericytes on
blood vessels and endothelial sprouts in tumors.
Am.J.Pathol. 160, 985–1000 (2002).
12. Inoue,M., Hager,J.H., Ferrara,N., Gerber,H.P. &
Hanahan,D. VEGF-A has a critical, nonredundant role
in angiogenic switching and pancreatic beta cell
carcinogenesis. Cancer Cell 1, 193–202 (2002).
13. Bergers,G. etal. Matrix metalloproteinase-9 triggers
the angiogenic switch during carcinogenesis.
Nat.CellBiol. 2, 737–744 (2000).
This study implicated myeloid-cell derived MMP9
as a key regulator of the angiogenic switch in a
GEMM of cancer.
14. Egeblad,M., Nakasone,E.S. & Werb,Z. Tumors as
organs: complex tissues that interface with the entire
organism. Dev. Cell 18, 884–901 (2010).
15. De Palma,M. & Hanahan,D. The biology of
personalized cancer medicine: facing
individualcomplexities underlying hallmark
capabilities. Mol. Oncol. 6, 111–127 (2012).
16. Cortez,E., Roswall,P. & Pietras,K. Functional subsets
of mesenchymal cell types in the tumor
microenvironment. Semin. Cancer Biol. 25, 3–9
(2014).
17. Kalluri,R. The biology and function of fibroblasts in
cancer. Nat. Rev. Cancer 16, 582–598 (2016).
18. Qian,B.Z. & Pollard,J.W. Macrophage diversity
enhances tumor progression and metastasis. Cell 141,
39–51 (2010).
19. Fang,S. & Salven,P. Stem cells in tumor angiogenesis.
J.Mol. Cell. Cardiol. 50, 290–295 (2011).
20. Lahmar,Q. etal. Tissue-resident versus monocyte-
derived macrophages in the tumor microenvironment.
Biochim. Biophys. Acta 1865, 23–34 (2016).
21. Leek,R.D. etal. Association of macrophage
infiltration with angiogenesis and prognosis in invasive
breast carcinoma. Cancer Res. 56, 4625–4629
(1996).
This report identified a positive correlation
between macrophage infiltration and a high
vascular grade in human breast cancers.
22. Koh,Y.W., Park,C.S., Yoon,D.H., Suh,C. & Huh,J.
CD163 expression was associated with angiogenesis
and shortened survival in patients with uniformly
treated classical Hodgkin lymphoma. PLoS ONE 9,
e870666 (2014).
23. Clear,A.J. etal. Increased angiogenic sprouting in
poor prognosis FL is associated with elevated
numbers of CD163+ macrophages within the
immediate sprouting microenvironment. Blood 115 ,
5053–5056 (2010).
24. Lin,E.Y. etal. Macrophages regulate the angiogenic
switch in a mouse model of breast cancer. Cancer Res.
66, 11238–11246 (2006).
25. Zeisberger,S.M. etal. Clodronate-liposome-mediated
depletion of tumour-associated macrophages: a new
and highly effective antiangiogenic therapy approach.
Br. J.Cancer 95, 272–281 (2006).
26. Priceman,S.J. etal. Targeting distinct tumor-
infiltrating myeloid cells by inhibiting CSF-1 receptor:
combating tumor evasion of antiangiogenic therapy.
Blood 115 , 1461–1471 (2010).
27. Lohela,M. etal. Intravital imaging reveals distinct
responses of depleting dynamic tumor-associated
macrophage and dendritic cell subpopulations.
Proc.Natl Acad. Sci. USA 111, E5086–E5095 (2014).
28. Squadrito,M.L. & De Palma,M. Macrophage
regulation of tumor angiogenesis: implications for
cancer therapy. Mol. Aspects Med. 32, 123–145
(2011).
29. Baer,C., Squadrito,M.L., Iruela-Arispe,M.L. &
DePalma,M. Reciprocal interactions between
endothelial cells and macrophages in angiogenic
vascular niches. Exp. Cell Res. 319, 1626–1634
(2013).
30. Lewis,J.S., Landers,R.J., Underwood,J.C.,
Harris,A.L. & Lewis,C.E. Expression of vascular
endothelial growth factor by macrophages is
up-regulated in poorly vascularized areas of breast
carcinomas. J.Pathol. 192, 150–158 (2000).
31. Stockmann,C. etal. Deletion of vascular endothelial
growth factor in myeloid cells accelerates
tumorigenesis. Nature 456, 814–818 (2008).
32. Hughes,R. etal. Perivascular M2 macrophages
stimulate tumor relapse after chemotherapy.
CancerRes. 75, 3479–3491 (2015).
33. Harney,A.S. etal. Real-time imaging reveals local,
transient vascular permeability, and tumor cell
intravasation stimulated by TIE2hi macrophage-
derived VEGFA. Cancer Discov. 5, 932–943 (2015).
34. Murdoch,C., Muthana,M., Coffelt,S.B. & Lewis,C.E.
The role of myeloid cells in the promotion of tumour
angiogenesis. Nat. Rev. Cancer 8, 618–631 (2008).
35. Yeo,E.J. etal. Myeloid WNT7b mediates the
angiogenic switch and metastasis in breast cancer.
Cancer Res. 74, 2962–2973 (2014).
36. Lewis,C.E., Harney,A.S. & Pollard,J.W.
Themultifaceted role of perivascular macrophages
intumors. Cancer Cell 30, 18–25 (2016).
37. Olson,O.C. & Joyce,J.A. Cysteine cathepsin
proteases: regulators of cancer progression and
therapeutic response. Nat. Rev. Cancer 15, 712–729
(2015).
38. Kessenbrock,K., Plaks,V. & Werb,Z. Matrix
metalloproteinases: regulators of the tumor
microenvironment. Cell 141, 52–67 (2010).
39. Huang,S. etal. Contributions of stromal
metalloproteinase-9 to angiogenesis and growth of
human ovarian carcinoma in mice. J.Natl Cancer Inst.
94, 1134–1142 (2002).
40. Giraudo,E., Inoue,M. & Hanahan,D. An amino-
bisphosphonate targets MMP-9-expressing
macrophages and angiogenesis to impair cervical
carcinogenesis. J.Clin. Invest. 114 , 623–633
(2004).
41. Joyce,J.A. etal. Cathepsin cysteine proteases are
effectors of invasive growth and angiogenesis during
multistage tumorigenesis. Cancer Cell 5, 443–453
(2004).
42. De Palma,M., Venneri,M.A., Roca,C. & Naldini,L.
Targeting exogenous genes to tumor angiogenesis by
transplantation of genetically modified hematopoietic
stem cells. Nat. Med. 9, 789–795 (2003).
43. De Palma,M. etal. Tie2 identifies a hematopoietic
lineage of proangiogenic monocytes required for
tumor vessel formation and a mesenchymal
population of pericyte progenitors. Cancer Cell 8,
211–226 (2005).
References 42 and 43 illustrated that the
conditional elimination of bone marrow-derived,
perivascular monocytes was sufficient to impair
angiogenesis and delay tumour growth in
transplant cancer models.
44. Matsubara,T. etal. TIE2-expressing monocytes as a
diagnostic marker for hepatocellular carcinoma
correlates with angiogenesis. Hepatology 57,
1416–1425 (2013).
45. Mazzieri,R. etal. Targeting the ANG2/TIE2 axis
inhibits tumor growth and metastasis by impairing
angiogenesis and disabling rebounds of proangiogenic
myeloid cells. Cancer Cell 19, 512–526 (2011).
46. Kioi,M. etal. Inhibition of vasculogenesis, but not
angiogenesis, prevents the recurrence of glioblastoma
after irradiation in mice. J.Clin. Invest. 120, 694–705
(2010).
47. Kozin,S.V. etal. Recruitment of myeloid but not
endothelial precursor cells facilitates tumor regrowth
after local irradiation. Cancer Res. 70, 5679–5685
(2010).
48. Welford,A.F. etal. TIE2-expressing macrophages limit
the therapeutic efficacy of the vascular-disrupting
agent combretastatin A4 phosphate in mice.
J.Clin.Invest. 121, 1969–1973 (2011).
49. Chen,L. etal. Tie2 expression on macrophages is
required for blood vessel reconstruction and tumor
relapse after chemotherapy. Cancer Res. 76,
6828–6838 (2016).
Immune checkpoint
A ligand–receptor pair that
either activates or inhibits a
stimulatory signal for
lymphocytes, resulting in the
activation or suppression of an
immune response.
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
471

50. Dou,G.R. etal. Myeloid-specific blockade of Notch
signaling attenuates choroidal neovascularization
through compromised macrophage infiltration and
polarization in mice. Sci. Rep. 6, 28617 (2016).
51. Tamagnone,L. Emerging role of semaphorins as major
regulatory signals and potential therapeutic targets in
cancer. Cancer Cell 22, 145–152 (2012).
52. Barnett,F.H. etal. Macrophages form functional
vascular mimicry channels invivo. Sci. Rep. 6, 36659
(2016).
53. Moldovan,N.I., Goldschmidt-Clermont,P.J.,
Parker-Thornburg,J., Shapiro,S.D. &
Kolattukudy,P.E. Contribution of monocytes/
macrophages to compensatory neovascularization:
the drilling of metalloelastase-positive tunnels in
ischemic myocardium. Circ. Res. 87, 378–384
(2000).
54. Coffelt,S.B., Wellenstein,M.D. & de Visser,K.E.
Neutrophils in cancer: neutral no more.
Nat.Rev.Cancer 16, 431–446 (2016).
55. Liang,W. & Ferrara,N. The complex role of
neutrophils in tumor angiogenesis and metastasis.
Cancer Immunol. Res. 4, 83–91 (2016).
56. Kujawski,M. etal. Stat3 mediates myeloid cell-
dependent tumor angiogenesis in mice. J.Clin. Invest.
118 , 3367–3377 (2008).
57. Gaudry,M. etal. Intracellular pool of vascular
endothelial growth factor in human neutrophils. Blood
90, 4153–4161 (1997).
58. Qu,X., Zhuang,G., Yu,L., Meng,G. & Ferrara,N.
Induction of Bv8 expression by granulocyte colony-
stimulating factor in CD11b+Gr 1+ cells: key role of
Stat3 signaling. J.Biol. Chem. 287, 19574–19584
(2012).
59. Shojaei,F. etal. Bv8 regulates myeloid-cell-dependent
tumour angiogenesis. Nature 450, 825–831 (2007).
60. Jablonska,J., Leschner,S., Westphal,K.,
Lienenklaus,S. & Weiss,S. Neutrophils responsive to
endogenous IFN-beta regulate tumor angiogenesis
and growth in a mouse tumor model. J.Clin. Invest.
120, 1151–1164 (2010).
61. Nozawa,H., Chiu,C. & Hanahan,D. Infiltrating
neutrophils mediate the initial angiogenic switch in a
mouse model of multistage carcinogenesis.
Proc. Natl Acad. Sci. USA 103, 12493–12498
(2006).
This study highlighted the contribution of
neutrophils to angiogenesis in the early stages of
tumour progression in a mouse model of PNET.
62. Rivera,L.B. etal. Intratumoral myeloid cells regulate
responsiveness and resistance to antiangiogenic
therapy. Cell Rep. 11, 577–591 (2015).
63. Nakasone,E.S. etal. Imaging tumor-stroma
interactions during chemotherapy reveals
contributions of the microenvironment to resistance.
Cancer Cell 21, 488–503 (2012).
64. Deryugina,E.I. etal. Tissue-infiltrating neutrophils
constitute the major invivo source of angiogenesis-
inducing MMP-9 in the tumor microenvironment.
Neoplasia 16, 771–788 (2014).
65. Ardi,V.C. etal. Neutrophil MMP-9 proenzyme,
unencumbered by TIMP-1, undergoes efficient
activation invivo and catalytically induces
angiogenesis via a basic fibroblast growth factor
(FGF-2)/FGFR-2 pathway. J.Biol. Chem. 284,
25854–25866 (2009).
66. Kumar,V., Patel,S., Tcyganov,E. & Gabrilovich,D.I.
The nature of myeloid-derived suppressor cells in the
tumor microenvironment. Trends Immunol. 37,
208–220 (2016).
67. Coffelt,S.B. etal. Elusive identities and overlapping
phenotypes of proangiogenic myeloid cells in tumors.
Am. J.Pathol. 176, 1564–1576 (2010).
68. Motz,G.T. & Coukos,G. The parallel lives of
angiogenesis and immunosuppression: cancer and
other tales. Nat. Rev. Immunol. 11 , 702–711 (2011).
69. Dirkx,A.E. etal. Tumor angiogenesis modulates
leukocyte-vessel wall interactions invivo by reducing
endothelial adhesion molecule expression. Cancer Res.
63, 2322–2329 (2003).
70. Johansson,A., Hamzah,J. & Ganss,R. More than a
scaffold: stromal modulation of tumor immunity.
Biochim. Biophys. Acta 1865, 3–13 (2016).
71. Kessler,D.A., Langer,R.S., Pless,N.A. & Folkman,J.
Mast cells and tumor angiogenesis. Int. J.Cancer 18,
703–709 (1976).
72. Coussens,L.M. etal. Inflammatory mast cells
up-regulate angiogenesis during squamous epithelial
carcinogenesis. Genes Dev. 13, 1382–1397 (1999).
73. Gounaris,E. etal. Mast cells are an essential
hematopoietic component for polyp development.
Proc. Natl Acad. Sci. USA 104, 19977–19982
(2007).
74. Soucek,L. etal. Mast cells are required for
angiogenesis and macroscopic expansion of Myc-
induced pancreatic islet tumors. Nat. Med. 13,
1211–1218 (2007).
References 72–74 revealed unexpected roles for
mast cells as pro-angiogenic cells in different
GEMMs of cancer.
75. Movahedi,K. etal. Different tumor microenvironments
contain functionally distinct subsets of macrophages
derived from Ly6C(high) monocytes. Cancer Res. 70,
5728–5739 (2010).
76. Cormier,S.A. etal. Pivotal advance: eosinophil
infiltration of solid tumors is an early and persistent
inflammatory host response. J.Leukoc. Biol. 79,
1131–1139 (2006).
77. Davis,B.P. & Rothenberg,M.E. Eosinophils and
cancer. Cancer Immunol. Res. 2, 1–8 (2014).
78. Yang,C. etal. Bcells promote tumor progression via
STAT3 regulated-angiogenesis. PLoS ONE 8, e64159
(2013).
79. Andreu,P. etal. FcRgamma activation regulates
inflammation-associated squamous carcinogenesis.
Cancer Cell 17, 121–134 (2010).
This study illustrated a role for B cells and secreted
IgG in promoting pro-angiogenic myeloid cell
programming in a mouse model of skin cancer.
80. Tian,L. etal. Mutual regulation of tumour vessel
normalization and immunostimulatory reprogramming.
Nature 544, 250–254 (2017).
81. Fathallah-Shaykh,H.M., Zhao,L.J., Kafrouni,A.I.,
Smith,G.M. & Forman,J. Gene transfer of IFN-gamma
into established brain tumors represses growth by
antiangiogenesis. J.Immunol. 164,
217–222 (2000).
82. Kammertoens,T. etal. Tumour ischaemia by interferon-
gamma resembles physiological blood vessel
regression. Nature 545, 98–102 (2017).
References 80–82 showed that T cells regulate the
tumour-associated vasculature by limiting
angiogenesis.
83. Biswas,S.K. & Mantovani,A. Macrophage plasticity
and interaction with lymphocyte subsets: cancer as a
paradigm. Nat. Immunol. 11, 889–896 (2010).
84. DeNardo,D.G. etal. CD4+ Tcells regulate pulmonary
metastasis of mammary carcinomas by enhancing
protumor properties of macrophages. Cancer Cell 16,
91–102 (2009).
85. Baer,C. etal. Suppression of microRNA activity
amplifies IFN-gamma-induced macrophage activation
and promotes anti-tumour immunity. Nat. Cell Biol.
18, 790–802 (2016).
86. Facciabene,A. etal. Tumour hypoxia promotes
tolerance and angiogenesis via CCL28 and Treg cells.
Nature 475, 226–230 (2011).
87. Bruno,A., Ferlazzo,G., Albini,A. & Noonan,D.M.
Athink tank of TINK/TANKs: tumor-infiltrating/tumor-
associated natural killer cells in tumor progression and
angiogenesis. J.Natl Cancer Inst. 106, dju200 (2014).
88. Gotthardt,D. etal. STAT5 Is a key regulator in NK cells
and acts as a molecular switch from tumor surveillance
to tumor promotion. Cancer Discov. 6, 414–429
(2016).
89. Sharma,D., Brummel-Ziedins,K.E., Bouchard,B.A.
&Holmes,C.E. Platelets in tumor progression: a host
factor that offers multiple potential targets in the
treatment of cancer. J.Cell. Physiol. 229, 1005–1015
(2014).
90. Sabrkhany,S., Griffioen,A.W. & Oude Egbrink,M.G.
The role of blood platelets in tumor angiogenesis.
Biochim. Biophys. Acta 1815, 189–196 (2011).
91. Chatterjee,M. etal. Distinct platelet packaging,
release, and surface expression of proangiogenic and
antiangiogenic factors on different platelet stimuli.
Blood 117 , 3907–3911 (2011).
92. Jonnalagadda,D., Izu,L.T. & Whiteheart,S.W.
Platelet secretion is kinetically heterogeneous in an
agonist-responsive manner. Blood 120, 5209–5216
(2012).
93. Yan,M., Lesyk,G., Radziwon-Balicka,A. & Jurasz,P.
Pharmacological regulation of platelet factors that
influence tumor angiogenesis. Semin. Oncol. 41,
370–377 (2014).
94. Ho-Tin-Noe,B., Goerge,T., Cifuni,S.M.,
Duerschmied,D. & Wagner,D.D. Platelet granule
secretion continuously prevents intratumor
hemorrhage. Cancer Res. 68, 6851–6858 (2008).
95. Klement,G.L. etal. Platelets actively sequester
angiogenesis regulators. Blood 113 , 2835–2842
(2009).
96. Peterson,J.E. etal. VEGF, PF4 and PDGF are elevated
in platelets of colorectal cancer patients. Angiogenesis
15, 265–273 (2012).
97. Salven,P., Orpana,A. & Joensuu,H. Leukocytes and
platelets of patients with cancer contain high levels of
vascular endothelial growth factor. Clin. Cancer Res. 5,
487–491 (1999).
References 95–97 showed that platelets could
sequester, and become a reservoir of,
pro-angiogenic mediators in patients with cancer.
98. Kuznetsov,H.S. etal. Identification of luminal breast
cancers that establish a tumor-supportive
macroenvironment defined by proangiogenic platelets
and bone marrow-derived cells. Cancer Discov. 2,
1150–1165 (2012).
99. Feng,W. etal. A novel role for platelet secretion in
angiogenesis: mediating bone marrow-derived cell
mobilization and homing. Blood 117 , 3893–3902
(2011).
100. Armulik,A., Genove,G. & Betsholtz,C. Pericytes:
developmental, physiological, and pathological
perspectives, problems, and promises. Dev. Cell 21 ,
193–215 (2011).
101. Eberhard,A. etal. Heterogeneity of angiogenesis and
blood vessel maturation in human tumors:
implications for antiangiogenic tumor therapies.
Cancer Res. 60, 1388–1393 (2000).
102. Saharinen,P., Eklund,L. & Alitalo,K. Therapeutic
targeting of the angiopoietin-TIE pathway. Nat. Rev.
Drug Discov. http://dx.doi.org/10.1038/nrd.2016.278
(2017).
103. Thurston,G. etal. Leakage-resistant blood vessels in
mice transgenically overexpressing angiopoietin-1.
Science 286, 2511–2514 (1999).
104. Xian,X. etal. Pericytes limit tumor cell metastasis.
J.Clin. Invest. 116 , 642–651 (2006).
105. You,W.K., Yotsumoto,F., Sakimura,K., Adams,R.H.
& Stallcup,W.B. NG2 proteoglycan promotes tumor
vascularization via integrin-dependent effects on
pericyte function. Angiogenesis 17, 61–76 (2014).
106. Nasarre,P. etal. Host-derived angiopoietin-2 affects
early stages of tumor development and vessel
maturation but is dispensable for later stages of tumor
growth. Cancer Res. 69, 1324–1333 (2009).
107. Holopainen,T. etal. Effects of angiopoietin-2-blocking
antibody on endothelial cell-cell junctions and lung
metastasis. J.Natl Cancer Inst. 104, 461–475
(2012).
108. Rigamonti,N. etal. Role of angiopoietin-2 in adaptive
tumor resistance to VEGF signaling blockade. Cell Rep.
8, 696–706 (2014).
109. Park,J.S. etal. Normalization of tumor vessels by
Tie2 activation and Ang2 inhibition enhances drug
delivery and produces a favorable tumor
microenvironment. Cancer Cell 30, 953–967 (2016).
110 . Birbrair,A. etal. Type-2 pericytes participate in
normal and tumoral angiogenesis. Am. J.Physiol. Cell
Physiol. 307, C25–C38 (2014).
111. Sennino,B. etal. Sequential loss of tumor vessel
pericytes and endothelial cells after inhibition of
platelet-derived growth factor B by selective aptamer
AX102. Cancer Res. 67, 7358–7367 (2007).
112 . Keskin,D. etal. Targeting vascular pericytes in hypoxic
tumors increases lung metastasis via angiopoietin-2.
Cell Rep. 10, 1066–1081 (2015).
113 . Erber,R. etal. Combined inhibition of VEGF and PDGF
signaling enforces tumor vessel regression by
interfering with pericyte-mediated endothelial cell
survival mechanisms. FASEB J. 18, 338–340 (2004).
114 . Bergers,G., Song,S., Meyer-Morse,N., Bergsland,E.
& Hanahan,D. Benefits of targeting both pericytes
and endothelial cells in the tumor vasculature with
kinase inhibitors. J.Clin. Invest. 111, 1287–1295
(2003).
References 113 and 114 highlighted the
anti-tumoural benefits of co-targeting pericytes
and ECs in mouse cancer models, paving the way to
the clinical testing of non-selective VEGFR and
PDGFR inhibitors in PNETs and other cancer types.
115 . Egeblad,M., Rasch,M.G. & Weaver,V.M. Dynamic
interplay between the collagen scaffold and tumor
evolution. Curr. Opin. Cell Biol. 22, 697–706 (2010).
116 . Tuxhorn,J.A., McAlhany,S.J., Dang,T.D.,
Ayala,G.E. & Rowley,D.R. Stromal cells promote
angiogenesis and growth of human prostate tumors in
a differential reactive stroma (DRS) xenograft model.
Cancer Res. 62, 3298–3307 (2002).
117 . Orimo,A. etal. Stromal fibroblasts present in invasive
human breast carcinomas promote tumor growth and
angiogenesis through elevated SDF-1/CXCL12
secretion. Cell 121, 335–348 (2005).
REVIEWS
472
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc

118 . Gerber,H.P., Kowalski,J., Sherman,D.,
Eberhard,D.A. & Ferrara,N. Complete inhibition of
rhabdomyosarcoma xenograft growth and
neovascularization requires blockade of both tumor
and host vascular endothelial growth factor.
CancerRes. 60, 6253–6258 (2000).
119 . Fukumura,D. etal. Tumor induction of VEGF promoter
activity in stromal cells. Cell 94, 715–725 (1998).
References 117–119 implicated CAFs as an
important source of VEGFA and myeloid-cell
chemoattractants in mouse tumour models.
120. Crawford,Y. etal. PDGF-C mediates the angiogenic
and tumorigenic properties of fibroblasts associated
with tumors refractory to anti-VEGF treatment.
CancerCell 15, 21–34 (2009).
121. Pietras,K., Pahler,J., Bergers,G. & Hanahan,D.
Functions of paracrine PDGF signaling in the
proangiogenic tumor stroma revealed by
pharmacological targeting. PLoS Med. 5, e19 (2008).
122. Anderberg,C. etal. Paracrine signaling by platelet-
derived growth factor-CC promotes tumor growth by
recruitment of cancer-associated fibroblasts.
CancerRes. 69, 369–378 (2009).
123. Hou,X. etal. PDGF-CC blockade inhibits pathological
angiogenesis by acting on multiple cellular and
molecular targets. Proc. Natl Acad. Sci. USA 107 ,
12216–12221 (2010).
124. Kaur,A. etal. sFRP2 in the aged microenvironment
drives melanoma metastasis and therapy resistance.
Nature 532, 250–254 (2016).
125. Filleur,S. etal. Invivo mechanisms by which tumors
producing thrombospondin 1 bypass its inhibitory
effects. Genes Dev. 15, 1373–1382 (2001).
126. Dirat,B. etal. Cancer-associated adipocytes exhibit an
activated phenotype and contribute to breast cancer
invasion. Cancer Res. 71, 2455–2465 (2011).
127. Lim,S., Hosaka,K., Nakamura,M. & Cao,Y. Co-option
of pre-existing vascular beds in adipose tissue controls
tumor growth rates and angiogenesis. Oncotarget 7,
38282–38291 (2016).
128. Cao,Y. Angiogenesis and vascular functions in
modulation of obesity, adipose metabolism, and insulin
sensitivity. Cell Metab. 18, 478–489 (2013).
129. Renehan,A.G., Zwahlen,M. & Egger,M. Adiposity
and cancer risk: new mechanistic insights from
epidemiology. Nat. Rev. Cancer 15, 484–498 (2015).
130. Bougaret,L. etal. Supernatants of adipocytes from
obese versus normal weight women and breast cancer
cells: in vitro impact on angiogenesis. J.Cell. Physiol.
232, 1808–1816 (2016).
131. Arendt,L.M. etal. Obesity promotes breast cancer
byCCL2-mediated macrophage recruitment and
angiogenesis. Cancer Res. 73, 6080–6093 (2013).
132. Hagiwara,A. etal. Milky spots as the implantation site
for malignant cells in peritoneal dissemination in mice.
Cancer Res. 53, 687–692 (1993).
133. Wagner,M. etal. Inflamed tumor-associated adipose
tissue is a depot for macrophages that stimulate tumor
growth and angiogenesis. Angiogenesis 15, 481–495
(2012).
134. Nieman,K.M. etal. Adipocytes promote ovarian
cancer metastasis and provide energy for rapid tumor
growth. Nat. Med. 17, 1498–1503 (2011).
References 133 and 134 showed that cancer cells
could reprogramme peri-tumoural adipocytes to
sustain tumour inflammation, angiogenesis and
growth.
135. Harjes,U., Bridges,E., McIntyre,A., Fielding,B.A. &
Harris,A.L. Fatty acid-binding protein 4, a point of
convergence for angiogenic and metabolic signaling
pathways in endothelial cells. J.Biol. Chem. 289,
23168–23176 (2014).
136. Schoors,S. etal. Fatty acid carbon is essential for
dNTP synthesis in endothelial cells. Nature 520,
192–197 (2015).
137. Dvorak,H.F. Tumor stroma, tumor bood vessels, and
antiangiogenesis therapy. Cancer J. 21, 237–243
(2015).
138. Bonnans,C., Chou,J. & Werb,Z. Remodelling the
extracellular matrix in development and disease.
Nat.Rev. Mol. Cell Biol. 15, 786–8014 (2014).
139. Arroyo,A.G. & Iruela-Arispe,M.L. Extracellular
matrix, inflammation, and the angiogenic response.
Cardiovasc. Res. 86, 226–235 (2010).
140. Baluk,P., Morikawa,S., Haskell,A., Mancuso,M.
&McDonald,D.M. Abnormalities of basement
membrane on blood vessels and endothelial sprouts
intumors. Am. J.Pathol. 163, 1801–1815 (2003).
141. Sorokin,L. The impact of the extracellular matrix on
inflammation. Nat. Rev. Immunol. 10 , 712–723
(2010).
142. Shao,R. etal. Acquired expression of periostin by
human breast cancers promotes tumor angiogenesis
through up-regulation of vascular endothelial growth
factor receptor 2 expression. Mol. Cell. Biol. 24,
3992–4003 (2004).
143. Van Obberghen-Schilling,E. etal. Fibronectin and
tenascin-C: accomplices in vascular morphogenesis
during development and tumor growth. Int. J.Dev. Biol.
55, 511–525 (2011).
144. Klenotic,P.A., Zhang,C. & Lin,Z. Emerging roles of
CCN proteins in vascular development and pathology.
J.Cell Commun. Signal. 10, 251–257 (2016).
145. Saupe,F. etal. Tenascin-C downregulates wnt inhibitor
dickkopf-1, promoting tumorigenesis in a
neuroendocrine tumor model. Cell Rep. 5, 482–492
(2013).
146. Lawler,P.R. & Lawler,J. Molecular basis for the
regulation of angiogenesis by thrombospondin-1 and
-2. Cold Spring Harb. Perspect. Med. 2, a006627
(2012).
147. Rivera,L.B., Bradshaw,A.D. & Brekken,R.A.
Theregulatory function of SPARC in vascular biology.
Cell. Mol. Life Sci. 68, 3165–3173 (2011).
148. Mammoto,A. etal. A mechanosensitive transcriptional
mechanism that controls angiogenesis. Nature 457,
1103–1108 (2009).
149. Rivron,N.C. etal. Tissue deformation spatially
modulates VEGF signaling and angiogenesis. Proc.Natl
Acad. Sci. USA 109, 6886–6891 (2012).
150. Mason,B.N., Starchenko,A., Williams,R.M.,
Bonassar,L.J. & Reinhart-King,C.A. Tuning three-
dimensional collagen matrix stiffness independently of
collagen concentration modulates endothelial cell
behavior. Acta Biomater. 9, 4635–4644 (2013).
151. Yun,H., Lee,M., Kim,S.S. & Ha,J. Glucose deprivation
increases mRNA stability of vascular endothelial growth
factor through activation of AMP-activated protein
kinase in DU145 prostate carcinoma. J.Biol. Chem.
280, 9963–9972 (2005).
152. Shi,Q. etal. Regulation of vascular endothelial growth
factor expression by acidosis in human cancer cells.
Oncogene 20, 3751–3756 (2001).
153. Vegran,F., Boidot,R., Michiels,C., Sonveaux,P. &
Feron,O. Lactate influx through the endothelial cell
monocarboxylate transporter MCT1 supports an
NF-kappaB/IL-8 pathway that drives tumor
angiogenesis. Cancer Res. 71, 2550–2560 (2011).
154. Sonveaux,P. etal. Targeting the lactate transporter
MCT1 in endothelial cells inhibits lactate-induced HIF-1
activation and tumor angiogenesis. PLoS ONE 7,
e33418 (2012).
155. Casazza,A. etal. Impeding macrophage entry into
hypoxic tumor areas by Sema3A/Nrp1 signaling
blockade inhibits angiogenesis and restores antitumor
immunity. Cancer Cell 24, 695–709 (2013).
156. Laoui,D. etal. Tumor hypoxia does not drive
differentiation of tumor-associated macrophages but
rather fine-tunes the M2-like macrophage population.
Cancer Res. 74, 24–30 (2014).
157. Colegio,O.R. etal. Functional polarization of tumour-
associated macrophages by tumour-derived lactic acid.
Nature 513, 559–563 (2014).
References 155–157 illustrated the effects of
tumour hypoxia on the pro-angiogenic macrophage
phenotype.
158. Biswas,S.K. Metabolic reprogramming of immune cells
in cancer progression. Immunity 43, 435–449 (2015).
159. Wenes,M. etal. Macrophage metabolism controls
tumor blood vessel morphogenesis and metastasis. Cell
Metab. 24, 701–715 (2016).
160. Potente,M. & Carmeliet,P. The link between
angiogenesis and endothelial metabolism. Annu. Rev.
Physiol. 79, 43–66 (2017).
161. Gorrini,C., Harris,I.S. & Mak,T.W. Modulation of
oxidative stress as an anticancer strategy. Nat. Rev.
Drug Discov. 12, 931–947 (2013).
162. Xia,C. etal. Reactive oxygen species regulate
angiogenesis and tumor growth through vascular
endothelial growth factor. Cancer Res. 67,
10823–10830 (2007).
163. West,X.Z. etal. Oxidative stress induces angiogenesis
by activating TLR2 with novel endogenous ligands.
Nature 467, 972–976 (2010).
164. Garrido-Urbani,S. etal. Targeting vascular NADPH
oxidase 1 blocks tumor angiogenesis through a
PPARalpha mediated mechanism. PLoS ONE 6,
e14665 (2011).
165. Warren,C.M., Ziyad,S., Briot,A., Der,A. &
Iruela-Arispe,M.L. A ligand-independent VEGFR2
signaling pathway limits angiogenic responses in
diabetes. Sci.Signal. 7, ra1 (2014).
166. Okuno,Y., Nakamura-Ishizu,A., Otsu,K., Suda,T. &
Kubota,Y. Pathological neoangiogenesis depends on
oxidative stress regulation by ATM. Nat. Med. 18,
1208–1216 (2012).
167. Rak,J. etal. Oncogenes and tumor angiogenesis:
differential modes of vascular endothelial growth
factor up-regulation in ras-transformed epithelial
cellsand fibroblasts. Cancer Res. 60, 490–498
(2000).
168. Sparmann,A. & Bar-Sagi,D. Ras-induced interleukin-8
expression plays a critical role in tumor growth and
angiogenesis. Cancer Cell 6, 447–458 (2004).
This study showed that a mutant RAS oncogene
fostered myeloid-cell infiltration and tumour
angiogenesis by activating the transcription of
Cxcl8 in cancer cells.
169. Bottos,A. etal. Targeting oncogenic serine/threonine-
protein kinase BRAF in cancer cells inhibits
angiogenesis and abrogates hypoxia. Proc. Natl Acad.
Sci. USA 109, E353–E359 (2012).
170. Sharma,A. etal. Mutant V599EB-Raf regulates
growth and vascular development of malignant
melanoma tumors. Cancer Res. 65, 2412–2421
(2005).
171. Phan,V.T. etal. Oncogenic RAS pathway activation
promotes resistance to anti-VEGF therapy through
G-CSF-induced neutrophil recruitment. Proc. Natl
Acad. Sci. USA 110, 6079–6084 (2013).
172. Pylayeva-Gupta,Y., Lee,K.E., Hajdu,C.H., Miller,G.
& Bar-Sagi,D. Oncogenic Kras-induced GM-CSF
production promotes the development of pancreatic
neoplasia. Cancer Cell 21, 836–847 (2012).
173. Cremolini,C. etal. FOLFOXIRI plus bevacizumab
versus FOLFIRI plus bevacizumab as first-line
treatment of patients with metastatic colorectal
cancer: updated overall survival and molecular
subgroup analyses of the open-label, phase 3 TRIBE
study. Lancet Oncol. 16, 1306–1315 (2015).
174. Tabernero,J. etal. Ramucirumab versus placebo in
combination with second-line FOLFIRI in patients with
metastatic colorectal carcinoma that progressed
during or after first-line therapy with bevacizumab,
oxaliplatin, and a fluoropyrimidine (RAISE):
arandomised, double-blind, multicentre, phase 3
study. Lancet Oncol. 16, 499–508 (2015).
175. Becker,A. etal. Extracellular vesicles in cancer:
cell-to-cell mediators of metastasis. Cancer Cell 30,
836–848 (2016).
176. Kholia,S. etal. Extracellular vesicles as new players in
angiogenesis. Vascul. Pharmacol. 86, 64–70 (2016).
177. Skog,J. etal. Glioblastoma microvesicles transport
RNA and proteins that promote tumour growth and
provide diagnostic biomarkers. Nat. Cell Biol. 10,
1470–1476 (2008).
178. Taraboletti,G. etal. Bioavailability of VEGF in tumor-
shed vesicles depends on vesicle burst induced by
acidic pH. Neoplasia 8, 96–103 (2006).
179. Graves,L.E. etal. Proinvasive properties of ovarian
cancer ascites-derived membrane vesicles. Cancer Res.
64, 7045–7049 (2004).
180. Squadrito,M.L. etal. Endogenous RNAs modulate
microRNA sorting to exosomes and transfer to
acceptor cells. Cell Rep. 8, 1432–1446 (2014).
181. Janowska-Wieczorek,A. etal. Microvesicles derived
from activated platelets induce metastasis and
angiogenesis in lung cancer. Int. J.Cancer 113,
752–760 (2005).
182. Sheldon,H. etal. New mechanism for Notch signaling
to endothelium at a distance by Delta-like 4
incorporation into exosomes. Blood 116 , 2385–2394
(2010).
183. Lombardo,G. etal. Activated Stat5 trafficking via
endothelial cell-derived extracellular vesicles controls
IL-3 pro-angiogenic paracrine action. Sci. Rep. 6,
25689 (2016).
184. Tkach,M. & Thery,C. Communication by extracellular
vesicles: where we are and where we need to go. Cell
164, 1226–1232 (2016).
185. Nazarenko,I. etal. Cell surface tetraspanin Tspan8
contributes to molecular pathways of exosome-
induced endothelial cell activation. Cancer Res. 70,
1668–1678 (2010).
186. Zhuang,G. etal. Tumour-secreted miR-9 promotes
endothelial cell migration and angiogenesis by
activating the JAK-STAT pathway. EMBO J. 31,
3513–3523 (2012).
187. Al-Nedawi,K., Meehan,B., Kerbel,R.S., Allison,A.C.
& Rak,J. Endothelial expression of autocrine VEGF
upon the uptake of tumor-derived microvesicles
containing oncogenic EGFR. Proc. Natl Acad. Sci. USA
106, 3794–3799 (2009).
REVIEWS
NATURE REVIEWS
|
CANCER VOLUME 17
|
AUGUST 2017
|
473

188. Umezu,T. etal. Exosomal miR-135b shed from
hypoxic multiple myeloma cells enhances angiogenesis
by targeting factor-inhibiting HIF-1. Blood 124,
3748–3757 (2014).
189. Ferrara,N. & Adamis,A.P. Ten years of anti-vascular
endothelial growth factor therapy. Nat. Rev. Drug
Discov. 15, 385–403 (2016).
190. Bridgeman,V.L. etal. Vessel co-option is common in
human lung metastases and mediates resistance to
anti-angiogenic therapy in preclinical lung metastasis
models. J.Pathol. 241, 362–374 (2017).
191. Frentzas,S. etal. Vessel co-option mediates resistance
to anti-angiogenic therapy in liver metastases.
Nat.Med. 22, 1294–1302 (2016).
References 190 and 191 highlighted the potential
clinical implications of vascular co-option for
resistance to anti-angiogenic therapy in metastatic
disease.
192. Pezzella,F. etal. Non-small-cell lung carcinoma tumor
growth without morphological evidence of neo-
angiogenesis. Am. J.Pathol. 151, 1417–1423
(1997).
193. Bonapace,L. etal. Cessation of CCL2 inhibition
accelerates breast cancer metastasis by promoting
angiogenesis. Nature 515, 130–133 (2014).
194. Gordon-Weeks,A.N. etal. Neutrophils promote
hepatic metastasis growth through fibroblast growth
factor 2-dependent angiogenesis in mice. Hepatology
65, 1920–1935 (2017).
195. Bergers,G. & Hanahan,D. Modes of resistance to
anti-angiogenic therapy. Nat. Rev. Cancer 8, 592–603
(2008).
196. Hu,Y.L. etal. Hypoxia-induced autophagy promotes
tumor cell survival and adaptation to antiangiogenic
treatment in glioblastoma. Cancer Res. 72,
1773–1783 (2012).
197. Sounni,N.E. etal. Blocking lipid synthesis overcomes
tumor regrowth and metastasis after antiangiogenic
therapy withdrawal. Cell Metab. 20, 280–294 (2014).
198. Allen,E. etal. Metabolic symbiosis enables adaptive
resistance to anti-angiogenic therapy that is
dependent on mTOR signaling. Cell Rep. 15,
1144–1160 (2016).
199. Jimenez-Valerio,G. etal. Resistance to antiangiogenic
therapies by metabolic symbiosis in renal cell
carcinoma PDX models and patients. Cell Rep. 15,
1134–1143 (2016).
200. Pisarsky,L. etal. Targeting metabolic symbiosis to
overcome resistance to anti-angiogenic therapy.
CellRep. 15, 1161–1174 (2016).
201. Bensaad,K. etal. Fatty acid uptake and lipid storage
induced by HIF-1alpha contribute to cell growth and
survival after hypoxia-reoxygenation. Cell Rep. 9,
349–365 (2014).
References 196–201 highlighted distinct
mechanisms whereby tumours escape the inhibitory
effects of anti-angiogenic therapy through
metabolic reprogramming.
202. Casanovas,O., Hicklin,D.J., Bergers,G. &
Hanahan,D. Drug resistance by evasion of
antiangiogenic targeting of VEGF signaling in late-
stage pancreatic islet tumors. Cancer Cell 8, 299–309
(2005).
203. Scholz,A. etal. Endothelial cell-derived angiopoietin-2
is a therapeutic target in treatment-naive and
bevacizumab-resistant glioblastoma. EMBO Mol. Med.
8, 39–57 (2016).
204. Paez-Ribes,M. etal. Antiangiogenic therapy elicits
malignant progression of tumors to increased local
invasion and distant metastasis. Cancer Cell 15,
220–231 (2009).
205. Rubenstein,J.L. etal. Anti-VEGF antibody treatment
of glioblastoma prolongs survival but results in
increased vascular cooption. Neoplasia 2, 306–314
(2000).
206. Gabrusiewicz,K. etal. Anti-vascular endothelial
growth factor therapy-induced glioma invasion is
associated with accumulation of Tie2-expressing
monocytes. Oncotarget 5, 2208–2220 (2014).
207. De Palma,M. & Lewis,C.E. Macrophage regulation of
tumor responses to anticancer therapies. Cancer Cell
23, 277–286 (2013).
208. Huang,H. etal. Specifically targeting angiopoietin-2
inhibits angiogenesis, Tie2-expressing monocyte
infiltration, and tumor growth. Clin. Cancer Res. 17,
1001–1011 (2011).
209. Kloepper,J. etal. Ang-2/VEGF bispecific antibody
reprograms macrophages and resident microglia to
anti-tumor phenotype and prolongs glioblastoma
survival. Proc. Natl Acad. Sci. USA 113 , 4476–4481
(2016).
210. Schmittnaegel,M. etal. Dual angiopoietin-2 and
VEGFA inhibition elicits antitumor immunity that is
enhanced by PD-1 checkpoint blockade. Sci. Transl
Med. 9, aak9670 (2017).
211 . Pahler,J.C. etal. Plasticity in tumor-promoting
inflammation: impairment of macrophage recruitment
evokes a compensatory neutrophil response.
Neoplasia 10, 329–340 (2008).
This study suggested that the genetic impairment
of macrophages could exacerbate the tumour-
promoting functions of neutrophils in a mouse
model of cervical cancer.
212. von Tresckow,B. etal. An open-label, multicenter,
phaseI/II study of JNJ-40346527, a CSF-1R inhibitor,
in patients with relapsed or refractory Hodgkin
lymphoma. Clin. Cancer Res. 21, 1843–1850 (2015).
213. Butowski,N. etal. Orally administered colony
stimulating factor 1 receptor inhibitor PLX3397 in
recurrent glioblastoma: an Ivy Foundation Early Phase
Clinical Trials Consortium phaseII study. Neuro Oncol.
18, 557–564 (2016).
214. Kaneda,M.M. etal. PI3Kgamma is a molecular
switch that controls immune suppression. Nature
539, 437–4424 (2016).
215. De Palma,M. etal. Tumor-targeted interferon-alpha
delivery by Tie2-expressing monocytes inhibits tumor
growth and metastasis. Cancer Cell 14, 299–311
(2008).
216. Mancuso,M.R. etal. Rapid vascular regrowth in
tumors after reversal of VEGF inhibition. J.Clin. Invest.
116 , 2610–2621 (2006).
217. Raymond,E. etal. Sunitinib malate for the treatment
of pancreatic neuroendocrine tumors. N.Engl. J.Med.
364, 501–513 (2011).
218. Ebos,J.M. etal. Neoadjuvant antiangiogenic therapy
reveals contrasts in primary and metastatic tumor
efficacy. EMBO Mol. Med. 6, 1561–1576 (2014).
219. Chung,A.S. etal. An interleukin-17-mediated
paracrine network promotes tumor resistance to
anti-angiogenic therapy. Nat. Med. 19, 1114–1123
(2013).
220. Carmeliet,P. & Jain,R.K. Principles and mechanisms
of vessel normalization for cancer and other
angiogenic diseases. Nat. Rev. Drug Discov. 10,
417–427 (2011).
221. De Palma,M. & Jain,R.K. CD4+ Tcell activation and
vascular normalization: two sides of the same coin?
Immunity 46, 773–775 (2017).
222. Schoenfeld,J. etal. Active immunotherapy induces
antibody responses that target tumor angiogenesis.
Cancer Res. 70, 10150–10160 (2010).
223. Wu,X. etal. Angiopoietin-2 as a biomarker and target
for immune checkpoint therapy. Cancer Immunol. Res.
5, 17–28 (2017).
References 222 and 223 revealed that the
anti-tumoural effects of certain cancer
immunotherapies may also involve immune-
mediated anti-angiogenic or angiostatic
mechanisms.
224. Aird,W.C. Endothelial cell heterogeneity. Cold Spring
Harb. Perspect. Med. 2, a006429 (2012).
225. Hida,K., Maishi,N., Torii,C. & Hida,Y. Tumor
angiogenesis — characteristics of tumor endothelial
cells. Int. J.Clin. Oncol. 21, 206–212 (2016).
226. Chaudhary,A. & St Croix,B. Selective blockade of
tumor angiogenesis. Cell Cycle 11, 2253–2259
(2012).
227. Pasqualini,R., Moeller,B.J. & Arap,W.
Leveragingmolecular heterogeneity of the vascular
endotheliumfor targeted drug delivery and imaging.
Semin. Thromb. Hemost. 36, 343–351 (2010).
228. Amin,D.N., Hida,K., Bielenberg,D.R. &
Klagsbrun,M. Tumor endothelial cells express
epidermal growth factor receptor (EGFR) but not
ErbB3 and are responsive to EGF and to EGFR
kinaseinhibitors. Cancer Res. 66, 2173–2180
(2006).
229. Xiong,Y.Q. etal. Human hepatocellular carcinoma
tumor-derived endothelial cells manifest increased
angiogenesis capability and drug resistance compared
with normal endothelial cells. Clin. Cancer Res. 15,
4838–4846 (2009).
230. Akiyama,K. etal. Tumor endothelial cells acquire drug
resistance by MDR1 up-regulation via VEGF signaling
intumor microenvironment. Am. J.Pathol. 180,
1283–1293 (2012).
231. Hida,K. etal. Tumor-associated endothelial cells
withcytogenetic abnormalities. Cancer Res. 64,
8249–8255 (2004).
232. Kim,Y.W. & Byzova,T.V. Oxidative stress in
angiogenesis and vascular disease. Blood 123,
625–631 (2014).
233. Soda,Y. etal. Transdifferentiation of glioblastoma cells
into vascular endothelial cells. Proc. Natl Acad. Sci.
USA 108 , 4274–4280 (2011).
234. Williamson,S.C. etal. Vasculogenic mimicry in small
cell lung cancer. Nat. Commun. 7, 13322 (2016).
Acknowledgements
Work in the authors’ laboratories is mainly supported by the
European Research Council (to M.D.P.), the Swiss National
Science Foundation (grants 31003A-165963 to M.D.P.;
31003A-156266, 31ER30-160674 and 316030-164119 to
T.V.P.), the Swiss Cancer League (grants KFS-3759-08-2015
to M.D.P.; KLS-3406-02-2014 to T.V.P.), the Leenaards
Foundation, the San Salvatore Foundation, and the Swiss
Bridge Foundation (to M.D.P. and T.V.P.).
Author contributions
M.D.P. conceived and wrote the article and display items.
T.V.P. contributed to discussions of the content and writing of
the article. M.D.P., D.B. and T.V.P. researched the data for the
article. M.D.P., D.B. and T.V.P. reviewed and edited the article
before submission.
Competing interests statement
The authors declare competing interests: see Web version
fordetails.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
REVIEWS
474
|
AUGUST 2017
|
VOLUME 17 www.nature.com/nrc