reviews

anticancer strategies involving the vasculature

Victoria L. Heath and Roy Bicknell
abstract | The growth and metastasis of solid tumors critically depends on their ability to develop their own
blood supply, a process known as tumor angiogenesis. Over the past decade much work has been performed to
understand this process, and modifying this process provides a key point of therapeutic intervention in the fight
against cancer. This review explores the development of anti-veGF-based antiangiogenic therapies, of which
there are currently three licensed for clinical use worldwide. Although originally anticipated to inhibit the growth
of tumor vessels, the induction of vascular normalization caused by these approved agents has provided a
novel means of effective delivery of known chemotherapeutic agents. The development of small molecules that
target veGF receptors has resulted in the generation of inhibitors with not only vascular activity but antitumor
activity in certain cancers. This review will address the current status of vascular-disrupting strategies, such as
therapies designed to induce tumor collapse by selectively destroying existing tumor vessels. These therapies
can be broadly divided into small-molecular-weight vascular-disrupting agents and ligand-directed approaches.
we discuss the current status of development, drug mechanisms of actions, combination with conventional
chemotherapy and radiotherapy, and potential future targets for therapeutic intervention.
Heath, v. L. & Bicknell, r. Nat. Rev. Clin. Oncol. 6, 395–404 (2009); published online 5 May 2009; doi:10.1038/nrclinonc.2009.52

Introduction
tumor vessels are structurally abnormal; they are leaky, the process of tumor angiogenesis is promoted by
tortuous, deficient in pericyte coverage, and lack the proangiogenic factors such as veGF, fibroblast growth
normal hierarchical arrangement of arterioles, capillaries factor (FGF) and interleukin 8, which are produced by
and venules (Figure 1).1 Consequently, tumors tend to tumor cells and cells of the tumor stroma, with the most
be poorly perfused and also have a high interstitial pres- critical being veGF. the veGF family of growth factors
sure because of a lack of lymphatics. Both of these factors comprises veGF a–e and placental growth factors
compromise the delivery of chemotherapeutics.2 this poor (PlGF) 1 and 2. veGF a is the predominant mediator of
perfusion results in a highly hypoxic and acidic tumor angiogenesis in tumors.8 veGF induces angiogenesis via
microenvironment,3 and this combination of features is binding and activation of the tyrosine kinase veGF recep-
known to affect the physiology and gene expression of tors 1–3 (veGFr1–3). veGF a can bind and activate both
tumor endothelial cells. tumors can use a variety of strat- veGFr1 and veGFr2, but signaling through veGFr2 is
egies to enhance their blood supply, including sprouting more potent than signaling through veGFr1 at inducing
angiogenesis, vessel co-option, intussusception of exist- an angiogenic response in endothelial cells.9 upregulation
ing vessels, and the recruitment of bone-marrow-derived of veGF production can be caused by hypoxia,8 and via
endothelial progenitor cells into growing vessels.4,5 vessel growth factors such as epidermal growth factor (eGF),10
co-option involves the tumor, particularly in the case and platelet-derived growth factor (PDGF).11
of metastases, growing in proximity to the pre-existing Data from animal models in the field of antiangiogenic
vasculature and deriving nutrients from it.6 vessel intus- agents has not always translated to the clinic. Despite this,
susception is a process whereby vessels split longitudinally blockade of veGF signaling in a number of different
and thereby increase vascular density. unlike sprouting experimental xenograft models was shown to reduce tumor
angiogenesis, this process is fast and provides a rapid growth from 25–95% in human xenograft models using institute for Biomedical
research, Birmingham
means of increasing vascular density without requiring the different tumor cell types grafted either orthotopically or University Medical
proliferation of endothelial cells.7 By contrast, sprouting ectopically, and these results paved the way for the clinical school, Birmingham,
angiogenesis is a slower, invasive process that enables the development of veGF signaling antagonists.12,13 UK (V. l. Heath,
R. Bicknell).
formation of vessels from pre-exisiting vessels, which are
able to bridge avascular areas. sprouting endothelial cells Bevacizumab Correspondence:
r. Bicknell, Cancer
produce proteases to degrade extracellular matrix proteins; the first antiangiogenic drug to be licensed in 2004 by research UK
the cells subsequently migrate, proliferate and join other the FDa was bevacizumab (a humanized anti-veGF Angiogenesis Group,
existing vessels. Finally, the endothelial cells form a lumen antibody)—as a first-line treatment for metastatic colo- institute for Biomedical
research, Birmingham
and recruit pericytes to stabilize the newly formed vessel.4 rectal cancer in combination with 5-fluorouracil (5-Fu)- University Medical
based chemotherapy. subsequently in 2006, it was also school, edgbaston,
Birmingham B15 2TT,
competing interests licensed for second-line treatment of colorectal cancer UK
The authors declare no competing interests. in combination with 5-Fu, and also in the treatment of r.bicknell@bham.ac.uk

nature reviews | clinical oncology volume 6 | JulY 2009 | 395

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

Key points in optimizing the scheduling of antiangiogenic therapies

combined with other treatment modalities.16 this issue
■ Blockade of veGF signaling results in vascular normalization, and the use
of bevacizumab, an anti-veGF agent, is effective only when combined with of timing of antiangiogenic therapies in conjunction with
chemotherapy other treatment modalities was explored in a phase i/ii
■ small-molecular-weight veGF receptor inhibitors such as sorafenib and trial of neoadjuvant bevacizumab in combination with
sunitinib inhibit a range of receptor tyrosine kinases chemoradiation in 11 patients with locally advanced
■ sorafenib and sunitinib cause vascular normalization and additionally target rectal cancer.2 the initial treatment cycle consisted of
stromal cells and certain tumor-cell types; they are currently being used as bevacizumab at either 5 or 10 mg/kg administered alone,
monotherapies while in the following three cycles it was administered
■ Microtubule-disrupting agents selectively destroy the tumor vasculature, in combination with chemoradiation (5-Fu 225 mg/m2
leading to hemorrhagic necrosis of the tumor infused during 24 h throughout each treatment week of
■ The distinctive tumor microenvironment results in the differential expression cycles 2 to 4, and external-beam irradiation delivered
of cell-surface and extracellular matrix molecules; these distinctive expression during cycles 2 to 4 for a total dose of 50.4 Gy delivered in
profiles can be used for imaging and targeting therapeutics specifically to 28 fractions during 5.5 weeks). ten days after treatment
the tumor with bevacizumab alone, vascular density, tumor blood
flow and tumor interstitial pressure were decreased
and tumor-cell apoptosis increased.2 Despite this, there
patients with unresectable, locally advanced, recurrent was no difference in the appearance of the tumors
or metastatic non-squamous non-small-cell lung cancer. macroscopically and there was no reduction in tumor
Bevacizumab is currently being assessed in at least 300 metabolism. after three treatment cycles that included
clinical trials in 20 different tumor types, including early- chemoradiation, there was decreased tumor meta-
stage cancers. Despite this success, the combination of bolism and significant tumor regression, which in two
bevacizumab and chemotherapy has not been successful cases included a full response. the results of this study
in the treatment of all cancers, and it failed to confer any support the hypothesis that bevacizumab caused vascular
benefit in terms of overall survival when combined with normalization.2 a decrease in circulating endothelial cells
chemotherapeutics in the treatment of previously treated and endothelial progenitor cells has also been observed
metastatic breast cancer.14 to optimize treatment it would with bevacizumab.2,17 although it is clear that such cells
be helpful to identify biomarkers that could predict the contribute to tumor angiogenesis,5 the extent to which
effectiveness of bevacizumab and other veGF-targeted reductions in the numbers of these circulating cells con-
therapies. the reason for the varied success of bevaci- tributes to the vascular normalization, or other clinical
zumab when combined with chemotherapy is probably benefit afforded by this therapy, is currently unclear.
because of the production of other proangiogenic factors Bevacizumab is a generally well-tolerated and safe
by the resistant tumors.15 therapy, although there are a number of adverse effects
at least three mechanisms of action have been pro- that might be exacerbated by concomitant chemotherapy.
posed for bevacizumab; firstly, an antiangiogenic low-grade and manageable adverse effects include
mechanism based on in vitro mouse xenograft models, hypertension, proteinuria, nosebleed, upper respiratory
secondly the inhibition of circulating endothelial cells infection, gastrointestinal symptoms and headache.
and endothelial progenitor cells colonizing the tumor rarely, are serious adverse effects such as gastrointestinal
vasculature, and thirdly vascular normalization. 2,5,13 perforations, hemorrhage and thrombolytic events
Clinical trials have demonstrated that in general, beva- observed.18 in general, all the adverse effects have been
cizumab is not effective as a monotherapy and its major attributed to blocking the normal role of veGF in the
activity is not in blocking angiogenesis, but rather via vasculature, and it has been recognized that veGF has a
the normalization of the tumor vasculature (Figure 1). role in maintaining the normal vasculature.18
this normalization involves the pruning and remod- studies in mice show that blockage of veGF signaling
eling of the tumor vasculature so that it more closely results in regressions of normal capillaries, leaving sleeves
resembles normal vasculature.16 this leads to enhanced of pericytes that persist for several weeks and promote
blood flow, reduced interstitial pressure and improved revascularization upon cessation of the therapy.18 levels
oxygenation of the tumor, all of which enhance the of endothelial fenestrations have also shown to be depen-
delivery of chemotherapeutics, and the increased oxy- dent on levels of veGF signaling, and this might explain
genation renders the tumor cells more radiosensitive and the proteinuria seen in patients treated with bevaci-
in some cases more chemosensitive.2 thus, the normal- zumab. likewise, hypertension might occur because
izing effects of veGF-signaling blockade are likely to blockade of veGF results in a reduction of nitric oxide,
improve delivery of chemotherapy, and it will be critical which is a known vasodilator.18,19 it will be especially
to improve the timing of the two treatments to maxi- important to study the effects of veGF blockade on the
mize their combined effectiveness. as the normalizing normal vasculature as antiangiogenic therapies are now
effects of veGF blockade are transient, lasting 6 days in being used for longer durations and in younger patients.
one experimental model, the determination of both the other veGF signaling blockers currently being tested in
onset and termination of this process will be essential clinical trials include veGF trap—an engineered soluble

396 | JULY 2009 | voLUme 6 www.nature.com/nrclinonc

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

veGF receptor (aflibercept, regeneron Pharmaceuticals, a b Abnormal tumor vessels

tarrytown, nY)20 and antibodies against veGFr-1 or Hypoxia,
EGFR signaling
veGFr-2 (imClone, new York, nY).one potential pitfall
of monotherapies is the eventual development of resis-
tance; the genetic instability of tumor cells enables them Tumor
cells
to mutate and develop mechanisms of resisting anti-
cancer therapies, and this is more likely to occur when a
single protumorigenic pathway is targeted than when VEGF
multiple pathways are targeted simultaneously. a study Pericyte
recruitment
has shown that extended veGFr blockade resulted in
eventual tumor revascularization, which was dependent Poor perfusion
Hypoxia
on other proangiogenic factors including FGF.21 another High interstitial
PDGF
possible mechanism of resistance is vascular co-option, pressure
in which the tumor grows around existing vessels, deriv-
ing its nutrients from it, and thus obviating the need for c VEGF blockade d Vascular disruption
Normalization
the development of new vessels.22 Necrotic
core
Small-molecule inhibitors of VEGF signaling Efficient delivery of
an alternative strategy of blocking veGF signaling is chemotherapeutics
Increased oxygenation
via inhibition of the veGFr by small-molecule tyrosine and radiosensitivity
kinases inhibitors. such small-molecule inhibitors have
the advantage of being orally available, and will ultimately
be more cost-effective as they are less expensive to both Tumor
manufacture and administer to patients than biological rim
Chemotherapeutics
therapeutics. two such inhibitors have been licensed delivered to target
by the FDa; sorafenib for the treatment of unresect- the rim of surviving
tumor cells
able hepatocellular carcinoma and advanced renal-cell
carcinoma, and sunitinib for first-line and second-line Figure 1 | The development of the tumor vasculature, and the effect of therapies that
target it. a | Hypoxia and eGFr signaling can lead to the production of veGF by tumor
treatment of metastatic renal-cell carcinoma and for
cells. This induces the sprouting of endothelial cells and eventually the development
the treatment of gastrointestinal stromal tumors. a of new vessels. PDGF produced by endothelial cells is involved in the recruitment of
number of other inhibitors with similar activities are listed pericytes, which surround and stabilize new vessels. b | The tumor vasculature is
in table 1. veGFr 1–3 belong to the split-kinase domain highly abnormal, tortuous, and lacks the normal architecture of arterioles, capillaries
family of receptor tyrosine kinases, which also contains and venules. Consequently the tumor is poorly perfused and hypoxic. A lack of
PDGFr-α and β and c-Kit.23 the majority of these veGF lymphatic drainage causes high interstitial pressure. c | veGF blockade causes
inhibitors have activity against veGFr 1–3, PDGFr-α vascular normalization, pruning and stabilization of the vasculature, resulting in good
and β, and c-Kit, which is probably because of their struc- perfusion and the efficient delivery of chemotherapeutics. The increased
oxygenation results in increased radiosensitivity of tumor cells. d | vascular-
tural similarity. additionally, they variously have activity
disrupting agents cause the collapse of tumor vessels, resulting in tumor-cell death
against more distantly related tyrosine kinases including and hemorrhagic necrosis. A surviving rim of tumor cells can be targeted by
eGFr and FGFr (table 1), and this promiscuity in target chemotherapeutics. Abbreviation: PDGF, platelet-derived growth factor.
inhibition gives them the potential to extend their activity
against other cell types such as tumor cells and pericytes.
this multikinase inhibition might enable these small antagonists have been approved for the treatment of
molecules to be used as monotherapies, something that metastatic pancreatic cancer, non-small-cell lung cancer,
has not proved effective with the pure veGF antagonists; squamous-cell carcinoma of the head and neck, and
however, further clinical data are needed to confirm if this colorectal cancer.29 thus, the antitumor and antivascular
is the case. aberrant receptor tyrosine kinase signaling is properties of vandetanib, a dual veGFr and eGFr inhibi-
found in a variety of tumor-cell types including gastro- tor, could be especially effective in the treatment of these
intestinal stromal tumors (which overexpress PDGFr and types of cancers.30 Clinical trials have shown little effect of
c-Kit), hematologic malignancies (which overexpress vandetanib in the treatment of colorectal cancer, although
Flt3), and thyroid cancer (ret) and small-cell lung more promising results have been shown in non-small-
carcinoma (which overexpress c-Kit).24–26 sorafenib also cell lung cancer. Phase ii clinical data of vandetanib, used
inhibits raf kinase, part of the raf/meK/erK signaling as either a single agent or in combination with the chemo-
cascade involved in cell growth and survival.27 raf kinases therapeutic agent docetaxel, was found to significantly
are found to be overactivated in a number of solid tumors improve progression-free survival by several weeks com-
including hepatocellular carcinoma, renal-cell carcinoma, pared with the control arms of gefitinib (an eGFr antago-
melanoma and papillary thyroid carcinoma.28 nist) or docetaxel alone, respectively.31 these data were
aberrant eGFr signaling is involved in the patho- sufficient to warrant progression to phase iii trials. as well
genesis of many epithelial-derived cancers, and eGFr as its involvement in tumor-cell signaling, PDGF signaling

nature reviews | clinical oncology volume 6 | JulY 2009 | 397

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

Table 1 | Multikinase inhibitors in clinical use and development

inhibitor alternative names company Drug targets indications and clinical trial
testing phase
sunitinib sutent®, sU11248 Pfizer veGFr2, PDGFrα First-line and second-line
and β, c-KiT, Flt3, treatment of metastatic rCC
reT90,91 Treatment of GisTs
Phase iii: breast, colorectal
and lung cancer
sorafenib Nexavar®, Bayer HealthCare veGFr2 and veGFr3, Unresectable hepatocellular
BAY 43-9006 Pharmaceuticals and raf, PDGFrβ, Flt3 and carcinoma, advanced rCC
Onyx Pharmaceuticals c-KiT27 Phase iii: NsCLC and melanoma
Pazopanib Armala™, Gw786034 GlaxosmithKline veGFr1–3, PDGFrα Phase i: colorectal cancer
and β, c-KiT92 Phase ii: NsCLC, ovarian cancer,
sarcoma
Phase iii: rCC
Axitinib AG-013736 Pfizer veGFr1–3, PDGFrβ, Phase ii: lung, GisTs, thryroid,
c-KiT93 breast and rCC
Phase iii: pancreatic
Cediranib recentin™, AZD2171 Astrazeneca veGFr1–3, PDGFrα Phase iii: recurrent glioblastoma
and β, c-KiT94 Phase ii/iii: NsCLC, colorectal
cancer
vatalanib PTK787, ZK222584 Novartis veGFr1–3, PDGFrα Phase ii: metastatic
and β, c-KiT34 neuroendocrine tumors, brain and
central-nervous-system tumors,
sarcoma
vandetanib Zactima™, ZD6474 Astrazeneca veGFr1–3, PDGFrβ, Phase iii: NsCLC
eGFr and reT30,95 Phase ii: medullary thyroid cancer
Brivanib alaninate BMs-582664 Bristol–Myers squibb veGFr1–3, FGFr1–396 Phase iii: colorectal cancer
Phase ii: hepatocellular
carcinoma, multiple tumor types
BiBF 1120 vargatef™ Boehringer ingelheim veGFr1–3, PDGFrα Phase iii: NsCLC
and PDGFrβ, FGFr1–3,
Flt3, src, Fyn, Lck97
Abbreviations: FGFr, fibroblast growth factor receptor; GisTs, gastrointestinal stromal tumors; NsCLC, non-small-cell lung cancer; PDGFr, platelet-derived
growth factor receptor; rCC, renal-cell carcinoma.

is required for pericyte recruitment and vessel stabiliza- in the treatment of glioblastoma found increased levels
tion. the combined inhibition of PDGF and veGF might, of plasma basic FGF and stromal-cell-derived factor 1α,
therefore, have more-profound effects against the tumor and increased viable circulating endothelial cells in blood
vasculature than veGF inhibition alone.32 when tumors escaped and ceased to respond to treat-
similar to bevacizumab, there is evidence to show that ment.17 additionally, progression of the tumor during
small-molecule inhibitors cause vascular normaliza- treatment interruption was accompanied by an increase
tion.17,33 interestingly, vatalanib—unlike bevacizumab— in circulating endothelial progenitor cells.17 Finally the
was found not to confer benefit when combined with question arises as to the effect of increasing the level of
chemotherapy in the treatment of metastatic colorectal the veGF blockade. one study that combined sorafenib
cancer.34 one possible explanation could be that vata- with bevacizumab demonstrated enhanced antitumor
lanib treatment did not block veGFr signaling as effi- activity in ovarian cancer but the combination also caused
ciently as bevacizumab. alternatively, blocking PDGF enhanced renal toxicity. the toxic effects included pro-
signaling might have had a detrimental effect on vascu- teinurea, thrombocytopenia and hypertension, and were
lar normalization because of its potential inhibition of sufficiently severe to warrant reductions in the dose of
pericyte recruitment.34 Currently, sunitinib and sorafenib sorafenib, but in most cases treatment was continued.35
are licensed for use as monotherapies, but on the basis
of their vascular normalizing properties they might also Vascular targeting agents
prove effective when combined with chemotherapy or the induction of tumor collapse via vascular target-
radiotherapy. several clinical trials are being undertaken ing of the tumor endothelium was shown in mouse
to investigate this. experiments.36 vascular targeting is particularly good
the identification of blood-borne biomarkers, which for destroying cells deep within a tumor that cannot
can indicate blockade of veGF signaling and success- be accessed by other forms of chemotherapy; however,
ful vascular normalization, would greatly facilitate the tumor cells can survive and eventually give rise to often
use of these cancer treatments. a study using cediranib fast regrowth of the tumor (Figure 1). these surviving

398 | JULY 2009 | voLUme 6 www.nature.com/nrclinonc

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

cells are accessible to chemotherapeutic agents, and it Fibronectin ED-B ED-A III-CS
is envisioned that vascular targeting would be used in
combination with such agents to be maximally effective.
there are two approaches to vascular targeting; the first Tenascin-C
type is ligand-directed, where antibodies or peptides 1 2 3 4 5 A1 A2 A3 A4 B AD C D 6 7 8
target therapeutic molecules either on the endothelial-
cell surface or basement membrane of tumor vessels. the Fibronectin type I homology domain EGF-like domain
second approach uses small molecules designed to selec-
Fibronectin type II homology domain Tenascin assembly domain
tively inhibit biological functions of tumor endothelial
cells. these include microtubule-disrupting agents and Fibronectin type III homology domain Fibrinogen-related domain

DmXaa, a molecule that induces endothelial-cell death Figure 2 | structure and alternative splicing of fibronectin and tenascin-C. The
and cytokine production.37 alternatively spliced domains are showing in pale shading. The fibronectin in fetal
For ligand-directed vascular targeting, it is necessary tissue and tumors contains the eD-B, the iii-Cs region is also subject to alternative
to identify cell-surface molecules that are found only on splicing. The large isoforms of tenascin-C containing the domains A1–D are found in
the surface of tumor but not normal endothelial cells. a a range of tumor types.15,49,55 Abbreviations: eD; extra domain; iii-Cs, type iii
number of approaches have been taken to identify such connecting segment.
molecules. Bioinformatic approaches using publicly
available, expressed sequence tags and serial analysis of silica to exclusively isolate membranes from endothelial
gene-expression data have been successful in identifying cells resulted in the identification of annexin a as a lung
endothelial specific genes, which, upon further investi- tumor endothelial marker.47
gation, have in some cases given tumor-endothelial-
specific expression patterns. one example is robo4, a Extracellular matrix molecules
cell-surface molecule involved in angiogenesis in vitro extracellular matrix molecules are suitable targets for
and in vivo, which has been found to be selectively ligand-directed therapy because of their proximity
expressed on tumor endothelium in adults, and which is to the tumor vasculature and their accumulation to a
a suitable ligand for molecular targeting.38–41 endothelial much higher level than would be found for a marker
gene expression has been examined using techniques expressed on the plasma membrane of endothelial
such as microarray, serial analysis of gene-expression cells. Fibronectin is a large glycoprotein found in body
libraries and subtractive hybridization. these techniques fluids, extracellular matrices and in most basement
have used both cells cultured in vitro under conditions membranes, and has a role in many different processes
that resemble those of the tumor microenvironment, including cell adhesion, migration and differentiation.48
and in vivo using endothelial cells isolated from differ- alternative mrna splicing has given rise to tumor-
ent types of solid tumors and comparing them with those specific isoforms of fibronectin and tenascin-C, which
from adjacent normal tissue.42 a study performed in have been used for vascular targeting. a splice variant of
mice aimed to compare endothelial-cell expression from fibronectin, which contained an extra type iii homology
a number of different normal tissues, different tumor domain whose expression was found to be restricted to
types and from the regenerating liver, where extensive sites of active angiogenesis, was identified 20 years ago
angiogenesis occurs but in a non-tumor context.43 the in tumors and fetal tissues (fibronectin extra domain
results of these different studies highlight two important [eD]-B; Figure 2).49,50 this form is secreted from tumor
features of the endothelial transcriptome. Firstly, endo- cells and can be found abundantly in the tumor stroma
thelial cells from different tissues have unique molecular and the abluminal side of tumor vessels. Phage display
signatures,44 and secondly, endothelial cells undergoing technology has been used successfully to generate a
different forms of angiogenesis express different sets high-affinity single-chain antibody (l19) to this domain,
of genes. the aim would be to target those molecules which has been used to characterize the distribution of
expressed only in the context of tumor angiogenesis to eD-B within tumors.51 this reagent has been invaluable
avoid disrupting physiological angiogenesis occurring in in the development of vascular targeting in experimental
wound healing and the female reproductive cycle. systems and has enabled the testing of a wide range of
innovative proteomic approaches have also been therapeutic agents in this setting, such as tissue factor,
undertaken to identify tumor-specific proteins acces- cytokines, radioisotopes and cytotoxic molecules.52 the
sible from the vasculature. Perfusion of normal kidney role of eD-B as a target in the clinic might be somewhat
tissue and kidney tumors with biotin resulted in the limited, however, as it is expressed only in aggressively
identification of a number of tumor-specific molecules growing tumors, and was found to be expressed in a
on the endothelial-cell surface and extracellular matrix minority (<25%) of certain cancers—such as epithelial-
including periostin and versican.45 this methodology derived cancers, including breast cancers. By contrast, it
has subsequently been adapted for use in tissue that is has been successfully used for imaging of head and neck
difficult to perfuse such as breast cancer tissue, where tumors in patients.53,54
versican was identified as a target tumor marker. 46 an tenascin-C is a large extracellular matrix protein,
alternative methodology using perfusion of colloidal which undergoes alternative splicing of its mrna

nature reviews | clinical oncology volume 6 | JulY 2009 | 399

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

transcript to include variable numbers of type iii clinical trials, other clinical studies have also shown that
fibronectin domain homology repeats a1–D (Figure 2). a Psma-specific monoclonal antibody effectively targets
expression of the large isoform of tenascin-C has been the vasculature of metastases of different tumor types,
shown to occur in different neoplasms, often in types of and it is hoped that this will pave the way for its use in
cancer where expression of fibronectin eD-B is low, such targeting therapeutics to tumor vessels.70
as breast cancer. the splicing patterns of tenascin-C are
complex and the development of antibodies to specific Effector moieties used in vascular targeting
domains is crucial for elucidating which isoforms are there are a range of effector moieties that can be used in
expressed in which cancers. For example, the C domain vascular targeting, such as the cytotoxic agents ricin a
is highly expressed in adult glioblastoma but rarely and neocarzinostatin, as well as radioisotopes (131i).54
in other cancers, whereas the a1 domain is expressed in Biological agents have also been used with some success;
head and neck, breast, and lung cancers.55,56 in a phase ii tissue factor has been used to promote coagulation of
trial, administration of radiolabeled anti-tenascin-C blood within tumor vessels, and cytokines such as inter-
antibodies to resection cavities in combination with leukin 2, interleukin 12 and tumor necrosis factor-α
conventional chemotherapy and radiotherapy increased have been targeted to the fibronectin eD-B to promote
the median survival time of patients with glioblastoma, immune antitumor responses. 71 liposomes, which
and a phase i trial of this agent has also shown promise in have the advantage of carrying larger amounts of toxin,
the treatment of non-Hodgkin lymphoma.57,58 Cilengitide have also been used to target cytotoxic agents.71 only
(merck serono, Geneva, switzerland) is a cyclic peptide radioisotopes have been used for vascular targeting in
inhibitor of the integrins αvβ3 and αvβ5, which are patients, with some effective pretargeting strategies being
expressed in tumor vessels and have a key role in tumor developed to enhance the delivery of radioisotopes to the
angiogenesis.59,60 this inhibitor showed antiangiogenic tumors and minimize exposure of the rest of the vascu-
effects in an orthotopic model of brain cancer, while it lature.70,72,73 one such strategy involves administering a
was ineffective for the same tumor types when implanted biotinylated targeting antibody, followed by avidin and
subcutaneously, which indicates that it might be a useful streptavidin, and the biotinylated radionucleotide.
antiangiogenic agent in brain cancer.61 indeed, phase i
trials of cilengitide in patients with recurrent malignant Microtubule-disrupting agents
glioma showed that the drug was well tolerated with some antimitotic microtubule-disrupting agents have long
antitumor responses. in newly diagnosed patients with been known to be able to damage the tumor vasculature,
glioblastoma multiforme, single-arm studies combining causing immediate and dramatic reduction in tumor blood
cilengitide with radiotherapy and temozolomide have flow and, ultimately, necrosis of the tumor. tumor vessels
shown encouraging activity and have led to a planned are more susceptible to microtubule-disrupting agents
randomized phase iii trial.62 compared with normal vessels because of their abnor-
there are a number of tumor endothelial molecules that mal structure, leakiness, and the presence of proliferat-
have been effectively used for vascular targeting in experi- ing endothelial cells. in order to use such agents in the
mental systems. these include a CD44 isoform recog- clinic it was necessary to identify molecules that exhib-
nized by the monoclonal antibody tes23, annexin a1, ited this activity at doses well below their toxic limit,
endoglin, prostate surface membrane antigen (Psma) and combretastatins fulfilled these requirements.37,74,75
and phosphatidylserine.47,63–66 the latter two endothelial subsequently, a number of other chemically distinct
markers have been further developed for the clinic and are microtubule-disrupting molecules have been identified
the subject of ongoing clinical trials. Phosphatidylserine with similar activities. table 2 lists different microtubule
is an anionic phospholipid that is normally found on the inhibitors used as vascular-disrupting agents in clinical
inner leaflet of the plasma membrane; however, it is also trials. the most studied of these is disodium combreta-
found on the outer leaflet of tumor endothelium and it is, statin a-4 3-o-phosphate (Ca4P). within 6 h of treat-
therefore, a potential vascular target for the treatment of ment, Ca4P causes a rapid shutdown of the tumor
cancer.67 radiotherapy increases phosphatidylserine levels vasculature, which persists for 12 h and is accompanied by
on the tumor endothelial surface and treatment with a hemorrhagic necrosis resulting from tumor damage.74
an anti-phosphatidylserine monoclonal antibody pro- as well as causing microtubule depolymerization, Ca4P
moted endothelial-cell death, most likely by promoting causes a change in endothelial shape, increased actino-
antibody-dependent cell-mediated cytotoxicity in an Fc'- myosin contractility, assembly of actin stress fibers and
dependant manner. Bavituximab, an anti-phosphatidyl- formation of focal adhesions.76 oXi4503 (combretastatin
serine monoclonal antibody (Peregrine Pharmaceuticals, a-1 disodium phosphate) represents a second genera-
tustin, Ca), is now in clinical trial testing for the treat- tion class of combretastatin, which shows more potent
ment of solid cancer.63 Psma is a type ii integral mem- vascular shutdown and antitumor activity compared
brane protein that is upregulated by prostate tumor cells. with Ca4P in mouse experimental tumor models.77 it
it has also been found to be present on the vasculature of is hoped that this can be developed as a monotherapy
many different solid tumors.68,69 while it has been used to induce hemorrhagic necrosis and target the resistant
as a target in a number of prostate cancer therapies in tumor cells at the tumor rim. more-recently identified

400 | JULY 2009 | voLUme 6 www.nature.com/nrclinonc

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

Table 2 | small-molecule microtubule-disrupting agents in clinical development

Drug ongoing chemical company Tumor types
phase i–iii trials
Zybrestat™ ii/iii Combretastatin 4-A phosphate Oxigene Anaplastic thyroid cancer,
NsCLC, ovarian cancer, head
and neck cancer
OXi4503 i Combretastatin A-1 phosphate Oxigene Advanced solid tumors
Ave8062 iii Combretastatin A-4 serine sanofi-Aventis sarcoma, NsCLC
prodrug (AC7700)
ABT-751 i/ii Orally active sulfonamide Abbott Laboratories ALL, neuroblastoma
ZD6126 i Phosphate prodrug of AstraZeneca NA
N‑acetylcolchinol (discontinued 2005)
Abbreviations: ALL, acute lymphoblastic leukemia; NA, not applicable; NsCLC, non-small-cell lung carcinoma.

vascular-disrupting agents are mn-029, nPi-2358 and high, thus providing a new therapeutic target—PlGF is
ePC2407, which are in the early stages of clinical develop- not widely expressed in human cancers.85,86
ment.78 Published clinical trial data for Ca4P, aBt-751 over the past decade it has become apparent that
and tZt-1027 have indicated that these agents are gen- receptors involved in axonal guidance also have a role
erally well tolerated with responses seen occasionally in in angiogenesis, and perturbing the signaling of these
some patients. all these agents are being investigated receptors in endothelial cells was shown to inhibit angio-
in combination with other chemotherapeutic agents.75 genesis. examples of these newly identified candidates
these responses included a partial response lasting more include Delta-like ligand 4 (Dll4), unc 5b, robo4 and the
than 57 weeks in a patient with metastatic liposarcoma ephrin ligand/ephrin receptor molecules.
treated with tZt-1027,75 and a partial response lasting Dll4 is an endothelial specific membrane anchored
for more than 81 weeks in a patient with non-small-cell ligand for the notch receptor, and inactivating even one
lung cancer treated with aBt-751.79 Preclinical studies allele of the gene encoding this protein results in embry-
have shown that treatment of vascular-disrupting agents onic lethality.87 in Dll4-deficient mice, the vasculature is
can be successfully combined with traditional chemo- excessively branched and nonfunctional, and it is thought
therapy and radiotherapy, with most success observed that ligation of Dll4 to notch delivers an inhibitory signal
when the vascular-disrupting agent is administered a few to endothelial cells restraining their exploratory pro-
hours after the chemotherapeutic agent.80 in the clinic, angiogenic activity.87 Dll4 is found to be upregulated in
the choice of chemotherapy will depend on the type of a number of tumors including renal-cell carcinoma and
cancer. this combination strategy might prove difficult bladder cancer. Blockade of Dll4 using either anti-Dll4
in the case of cancers that are difficult to treat or that antibodies, or soluble Dll4 has been used successfully
have become refractory to conventional chemotherapies. in experimental tumor models to reduce tumor size.87
an alternative strategy that has proved successful in examination of these tumors revealed an increased
experimental models is to combine antiangiogenics with vascular density but decreased perfusion with vascu-
vascular-disrupting agents.81 in phase i clinical trials, the lar tracers, suggesting that uncontrolled angiogenesis
use of Ca4P and ZD6126 has been associated with some resulted in the formation of nonfunctional vessels.87 the
cardiovascular toxic effects and Ca4P was suggested not observed effect is in direct contrast to that seen with the
to be suitable for the treatment of patients with known vascular normalization after veGF blockade.
heart disease.82,83 endothelial sprouting has also been shown to be nega-
tively regulated by signaling of netrin-1 through the
Novel antiangiogenic targets endothelial unc5B receptor. unc5B is expressed at sites of
as our knowledge of the angiogenic process increases active angiogenesis, and ectopic expression of its ligand
and new molecules involved in this process are identified, netrin-1 by transplanted tumor cells delayed tumor
more targets of potential novel antiangiogenic therapies angiogenesis in an experimental model.88 robo4 is an
emerge.84 after veGF, the most studied angiogenic factors endothelial specific member of the roundabout family
with regard to antiangiogenic strategies are probably of axonal guidance molecules discovered in 2000.38 it is
members of the FGF family. highly expressed on endothelial cells of tumor vessels but
neutralizing PlGF antibodies were found to be effec- not on those of the normal vasculature, and a soluble
tive in blocking tumor growth in an experimental mouse robo4 molecule can block angiogenesis in both in vivo
model, and, interestingly, were effective in tumors that and in vitro angiogenesis models.40 ephrin receptors are
were resistant to veGFr2 blockade. 85 PlGF binds receptor tyrosine kinases often overexpressed on tumor
veGFr1 and part of its proangiogenic activity is thought cells. Blockade of ligation of either epha2 or ephB4 to
to be due to its recruitment of monocytes. with the excep- their cognate ephrin ligands was shown to reduce tumor
tion of breast and gastric cancer—where its expression is growth and angiogenesis in experimental models.89

nature reviews | clinical oncology volume 6 | JulY 2009 | 401

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

Knowing the extent to which each of these different microtubule-disrupting agents being further in clinical
signaling pathways contributes to angiogenesis in human development. ongoing efforts to identify tumor endo-
cancer will be a crucial aspect of research required to thelial markers and tumor-specific extracellular matrix
translate these experimental antiangiogenics into the proteins will be useful for targeting of therapeutics and
clinical setting. imaging of the tumor vasculature. Promising markers,
including robo4, fibronectin eD-B and the large isoform
Conclusions of tenascin-C, have been identified and many more are
the development of therapies that target the tumor being investigated. it seems likely that many markers will
vasculature is an intense and exciting area of academic be restricted to a range of tumor types. vascular disrup-
and commercial research. mechanistically, these strat- tion is effective at destroying central areas of tumors, and
egies can be divided into those that cause vascular it is possible that agents with this mode of action will need
normalization and those that destroy the tumor vascu- to be combined with therapies that target cancer cells at
lature. the licensing of bevacizumab was a landmark in the tumor rim. over the next decade it is anticipated that
the development of antiangiogenics, and as a result of these different strategies of targeting the tumor vascu-
a large number of clinical trials, its use is likely to be lature will make an increasingly valuable contribution
extended to more types of cancer and in the treatment to the arsenal of anticancer therapies.
of early stage disease. it will be critical to determine how
the use of this vascular normalizing agent can be most Review criteria
effectively combined with other treatment modalities. The information for this review has been derived from
Furthermore, it will be important to elucidate its effects both primary and secondary literature. The NiH Clinical
on the normal vasculature and how tumors eventually Trials website was searched, along with the websites
evade veGF blockade. the use of small molecules to of pharmaceutical companies including Pfizer, Bayer,
target veGFr signaling has generated compounds with GlaxosmithKline, AstraZeneca and Boehringer ingelheim.
activities against the endothelium, tumor cells and cells The PubMed database was searched for articles
published before 1 June 2008. The search criteria were
of the tumor stroma, and these agents are successfully
the anti-angiogenic drugs and vascular targeting agents
being used as monotherapies. the clinical development described in this review.
of vascular-disrupting agents is less advanced, with the

1. Baluk, P., Hashizume, H. & McDonald, D. M. 10. Petit, A. M. et al. Neutralizing antibodies against tumor vasculature and alleviates edema in
Cellular abnormalities of blood vessels as epidermal growth factor and erbB-2/neu glioblastoma patients. Cancer Cell 11, 83–95
targets in cancer. Curr. Opin. Genet. Dev. 15, receptor tyrosine kinases downregulate vascular (2007).
102–111 (2005). endothelial growth factor production by tumor 18. Kamba, T. & McDonald, D. M. Mechanisms of
2. Duda, D. G., Jain, r. K. & willett, C. G. cells in vitro and in vivo: angiogenic implications adverse effects of anti-veGF therapy for
Antiangiogenics: the potential role of for signal transduction therapy of solid tumors. cancer. Br. J. Cancer 96, 1788–1795 (2007).
integrating this novel treatment modality with Am. J. Pathol. 151, 1523–1530 (1997). 19. Horowitz, J. r. et al. vascular endothelial
chemoradiation for solid cancers. J. Clin. Oncol. 11. reinmuth, N. et al. induction of veGF in growth factor/vascular permeability factor
25, 4033–4042 (2007). perivascular cells defines a potential paracrine produces nitric oxide-dependent hypotension.
3. Helmlinger, G., Yuan, F., Dellian, M. & Jain, r. K. mechanism for endothelial cell survival. evidence for a maintenance role in quiescent
interstitial pH and pO2 gradients in solid FASEB J. 15, 1239–1241 (2001). adult endothelium. Arterioscler. Thromb. Vasc.
tumors in vivo: high-resolution measurements 12. Brekken, r. A. et al. selective inhibition of Biol. 17, 2793–2800 (1997).
reveal a lack of correlation. Nat. Med. 3, vascular endothelial growth factor (veGF) 20. riely, G. J. & Miller, v. A. vascular endothelial
177–182 (1997). receptor 2 (KDr/Flk-1) activity by a monoclonal growth factor trap in non small cell lung cancer.
4. Adams, r. H. & Alitalo, K. Molecular regulation anti-veGF antibody blocks tumor growth in Clin. Cancer Res. 13, s4623–s4627 (2007).
of angiogenesis and lymphangiogenesis. mice. Cancer Res. 60, 5117–5124 (2000). 21. Casanovas, O., Hicklin, D. J., Bergers, G. &
Nat. Rev. Mol. Cell Biol. 8, 464–478 (2007). 13. Gerber, H. P. & Ferrara, N. Pharmacology and Hanahan, D. Drug resistance by evasion of
5. rafii, s., Lyden, D., Benezra, r., Hattori, K. pharmacodynamics of bevacizumab as antiangiogenic targeting of veGF signaling in
& Heissig, B. vascular and hematopoietic monotherapy or in combination with cytotoxic late-stage pancreatic islet tumors. Cancer Cell
stem cells: novel targets for anti-angiogenesis therapy in preclinical studies. Cancer Res. 65, 8, 299–309 (2005).
therapy? Nat. Rev. Cancer 2, 826–835 (2002). 671–680 (2005). 22. Auguste, P., Lemiere, s., Larrieu-Laharque, F.
6. Holash, J. et al. vessel cooption, regression, 14. Miller, K. D. et al. randomized phase iii trial & Bikfalvi, A. Molecular mechanisms of tumor
and growth in tumors mediated by of capecitabine compared with bevacizumab vascularization. Crit. Rev. Oncol. Hematol. 54,
angiopoietins and veGF. Science 284, plus capecitabine in patients with previously 53–61 (2005).
1994–1998 (1999). treated metastatic breast cancer. J. Clin. Oncol. 23. Hubbard, s. r. & Till, J. H. Protein tyrosine
7. Burri, P. H., Hlushchuk, r. & Djonov, v. 23, 792–799 (2005). kinase structure and function. Annu. Rev.
intussusceptive angiogenesis: its emergence, 15. relf, M. et al. expression of the angiogenic Biochem. 69, 373–398 (2000).
its characteristics, and its significance. factors vascular endothelial cell growth factor, 24. Arora, A. & scholar, e. M. role of tyrosine
Dev. Dyn. 231, 474–488 (2004). acidic and basic fibroblast growth factor, tumor kinase inhibitors in cancer therapy.
8. Olsson, A. K., Dimberg, A., Kreuger, J. growth factor beta-1, platelet-derived J. Pharmacol. Exp. Ther. 315, 971–979 (2005).
& Claesson-welsh, L. veGF receptor endothelial cell growth factor, placenta growth 25. Lodish, M. B. & stratakis, C.A. reT oncogene
signaling—in control of vascular function. Nat. factor, and pleiotrophin in human primary in MeN2, MeN2B, MTC and other forms of
Rev. Mol. Cell Biol. 7, 359–371 (2006). breast cancer and its relation to angiogenesis. thyroid cancer. Expert Rev. Anticancer Ther. 8,
9. Gille, H. et al. Analysis of biological effects and Cancer Res. 57, 963–969 (1997). 625–632 (2008).
signaling properties of Flt-1 (veGFr-1) and 16. Jain, r. K. Normalization of tumor vasculature: 26. Gilliland, D. G. & Griffin, J. D. role of FLT3 in
KDr (veGFr-2). A reassessment using novel an emerging concept in antiangiogenic therapy. leukemia. Curr. Opin. Hematol. 9, 274–281
receptor-specific vascular endothelial growth Science 307, 58–62 (2005). (2002).
factor mutants. J. Biol. Chem. 276, 3222–3230 17. Batchelor, T. T. et al. AZD2171, a pan-veGF 27. wilhelm, s. M. et al. BAY 43-9006 exhibits
(2001). receptor tyrosine kinase inhibitor, normalizes broad spectrum oral antitumor activity and

402 | JULY 2009 | voLUme 6 www.nature.com/nrclinonc

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

targets the rAF/MeK/erK pathway and 44. Chi, J. T. et al. endothelial cell diversity 62. reardon, D. A., Nabors, L. B., stupp, r. &
receptor tyrosine kinases involved in tumor revealed by global expression profiling. Proc. Mikkelsen, T. Cilengitide: an integrin-targeting
progression and angiogenesis. Cancer Res. 64, Natl Acad. Sci. USA 100, 10623–10628 arginine-glycine-aspartic acid peptide with
7099–7109 (2004). (2003). promising activity for glioblastoma multiforme.
28. Gollob, J. A., wilhelm, s., Carter, C. & 45. Castronovo, v. et al. A chemical proteomics Expert Opin. Investig. Drugs 17, 1225–1235
Kelley, s. L. role of raf kinase in cancer: approach for the identification of accessible (2008).
therapeutic potential of targeting the raf/ antigens expressed in human kidney cancer. 63. He, J., Luster, T. A. & Thorpe, P. e. radiation-
MeK/erK signal transduction pathway. Semin. Mol. Cell. Proteomics 5, 2083–2091 (2006). enhanced vascular targeting of human lung
Oncol. 33, 392–406 (2006). 46. Castronovo, v. et al. identification of specific cancers in mice with a monoclonal antibody
29. Ciardiello, F. & Tortora, G. eGFr antagonists reachable molecular targets in human breast that binds anionic phospholipids. Clin. Cancer
in cancer treatment. N. Engl. J. Med. 358, cancer using a versatile ex vivo proteomic Res. 13, 5211–5218 (2007).
1160–1174 (2008). method. Proteomics 7, 1188–1196 (2007). 64. Huang, X., Bennett, M. & Thorpe, P. e. Anti-
30. Herbst, r. s., Heymach, J. v., O’reilly, M. s., 47. Oh, P. et al. subtractive proteomic mapping of tumor effects and lack of side effects in mice
Onn, A. & ryan, A. J. vandetanib (ZD6474): an the endothelial surface in lung and solid of an immunotoxin directed against human and
orally available receptor tyrosine kinase inhibitor tumors for tissue-specific therapy. Nature 429, mouse prostate-specific membrane antigen.
that selectively targets pathways critical for 629–635 (2004). Prostate 61, 1–11 (2004).
tumor growth and angiogenesis. Expert Opin. 48. Hynes, r. O. & Yamada, K. M. Fibronectins: 65. seon, B. K., Matsuno, F., Haruta, Y., Kondo, M.
Investig. Drugs 16, 239–249 (2007). multifunctional modular glycoproteins. J. Cell & Barcos, M. Long-lasting complete inhibition
31. Hanrahan, e. O. & Heymach, J. v. vascular Biol. 95, 369–377 (1982). of human solid tumors in sCiD mice by
endothelial growth factor receptor tyrosine 49. Zardi, L. et al. Transformed human cells targeting endothelial cells of tumor vasculature
kinase inhibitors vandetanib (ZD6474) and produce a new fibronectin isoform by with antihuman endoglin immunotoxin. Clin.
AZD2171 in lung cancer. Clin. Cancer Res. 13, preferential alternative splicing of a previously Cancer Res. 3, 1031–1044 (1997).
s4617–e4622 (2007). unobserved exon. EMBO J. 6, 2337–2342 66. wakai, Y. et al. effective cancer targeting using
32. Abramsson, A., Lindblom, P. & Betsholtz, C. (1987). an anti-tumor tissue vascular endothelium-
endothelial and nonendothelial sources of 50. Castellani, P. et al. The fibronectin isoform specific monoclonal antibody (Tes-23). Jpn J.
PDGF-B regulate pericyte recruitment and containing the eD-B oncofetal domain: Cancer Res. 91, 1319–1325 (2000).
influence vascular pattern formation in tumors. a marker of angiogenesis. Int. J. Cancer 59, 67. ran, s., Downes, A. & Thorpe, P. e. increased
J. Clin. Invest. 112, 1142–1151 (2003). 612–618 (1994). exposure of anionic phospholipids on the
33. Taguchi, e., Nakamura, K., Miura, T., 51. Pini, A. et al. Design and use of a phage display surface of tumor blood vessels. Cancer Res.
shibuya, M. & isoe, T. Anti-tumor activity and library. Human antibodies with subnanomolar 62, 6132–6140 (2002).
tumor vessel normalization by the vascular affinity against a marker of angiogenesis 68. Chang, s. s. et al. Five different
endothelial growth factor receptor tyrosine eluted from a two-dimensional gel. J. Biol. anti-prostate-specific membrane antigen
kinase inhibitor KrN951 in a rat peritoneal Chem. 273, 21769–21776 (1998). (PsMA) antibodies confirm PsMA expression in
disseminated tumor model. Cancer Sci. 99, 52. Neri, D. & Bicknell, r. Tumor vascular targeting. tumor-associated neovasculature. Cancer Res.
623–630 (2008). Nat. Rev. Cancer 5, 436–446 (2005). 59, 3192–3198 (1999).
34. Prat, A., Casado, e. & Cortés, J. New approaches 53. Carnemolla, B. et al. A tumor-associated 69. Ghosh, A. & Heston, w. D. Tumor target
in angiogenic targeting for colorectal cancer. fibronectin isoform generated by alternative prostate specific membrane antigen (PsMA)
World J. Gastroenterol. 13, 5857–5866 (2007). splicing of messenger rNA precursors. J. Cell and its regulation in prostate cancer. J. Cell
35. Azad, N. s. et al. Combination targeted therapy Biol. 108, 1139–1148 (1989). Biochem. 91, 528–539 (2004).
with sorafenib and bevacizumab results in 54. santimaria, M. et al. immunoscintigraphic 70. Milowsky, M. i. et al. vascular targeted therapy
enhanced toxicity and antitumor activity. J. Clin. detection of the eD-B domain of fibronectin, with anti-prostate-specific membrane antigen
Oncol. 26, 3709–3714 (2008). a marker of angiogenesis, in patients with monoclonal antibody J591 in advanced solid
36. Burrows, F. J. & Thorpe, P. e. eradication of cancer. Clin. Cancer Res. 9, 571–579 (2003). tumors. J. Clin. Oncol. 25, 540–547 (2007).
large solid tumors in mice with an immunotoxin 55. Brack, s. s., silacci, M., Birchler, M. & Neri, D. 71. Thorpe, P. e. vascular targeting agents as
directed against tumor vasculature. Proc. Natl Tumor-targeting properties of novel antibodies cancer therapeutics. Clin. Cancer Res. 10,
Acad. Sci. USA 90, 8996–9000 (1993). specific to the large isoform of tenascin-C. Clin. 415–427 (2004).
37. Tozer, G. M., Kanthou, C. & Baguley, B. C. Cancer Res. 12, 3200–3208 (2006). 72. Goldenberg, D. M., sharkey, r. M., Paganelli, G.,
Disrupting tumor blood vessels. Nat. Rev. 56. Carnemolla, B. et al. identification of a Barbet, J. & Chatal, J. F. Antibody pretargeting
Cancer 5, 423–435 (2005). glioblastoma-associated tenascin-C isoform advances cancer radioimmunodetection
38. Huminiecki, L. & Bicknell, r. In silico cloning of by a high affinity recombinant antibody. Am. J. and radioimmunotherapy. J. Clin. Oncol. 24,
novel endothelial-specific genes. Genome Res. Pathol. 154, 1345–1352 (1999). 823–834 (2006).
10, 1796–1806 (2000). 57. Akabani, G. et al. Dosimetry and radiographic 73. reardon, D. A., Zalutsky, M. r. & Bigner, D. D.
39. Huminiecki, L., Gorn, M., suchting, s., analysis of 131i-labeled anti-tenascin 81C6 Antitenascin-C monoclonal antibody
Poulsom, r. & Bicknell, r. Magic roundabout is murine monoclonal antibody in newly radioimmunotherapy for malignant glioma
a new member of the roundabout receptor diagnosed patients with malignant gliomas: patients. Expert Rev. Anticancer Ther. 7,
family that is endothelial specific and a phase ii study. J. Nucl. Med. 46, 1042–1051 675–687 (2007).
expressed at sites of active angiogenesis. (2005). 74. Dark, G. G. et al. Combretastatin A-4, an agent
Genomics 79, 547–552 (2002). 58. rizzieri, D. A. et al. Phase 1 trial study of that displays potent and selective toxicity
40. suchting, s., Heal, P., Tahtis, K., stewart, L. M. 131
i-labeled chimeric 81C6 monoclonal toward tumor vasculature. Cancer Res. 57,
& Bicknell, r. soluble robo4 receptor inhibits antibody for the treatment of patients with 1829–1834 (1997).
in vivo angiogenesis and endothelial cell non-Hodgkin lymphoma. Blood 104, 642–648 75. Lippert, J. w. 3rd. vascular disrupting agents.
migration. FASEB J. 19, 121–123 (2005). (2004). Bioorg. Med. Chem. 15, 605–615 (2007).
41. Herbert, J. M., stekel, D., sanderson, s., 59. Dechantsreiter, M. A. et al. N-Methylated cyclic 76. Kanthou, C. & Tozer, G. M. The tumor vascular
Heath, v. L. & Bicknell, r. A novel method of rGD peptides as highly active and selective targeting agent combretastatin A-4-phosphate
differential gene expression analysis using alpha(v)beta(3) integrin antagonists. J. Med. induces reorganization of the actin
multiple cDNA libraries applied to the Chem. 42, 3033–3040 (1999). cytoskeleton and early membrane blebbing
identification of tumor endothelial genes. 60. eliceiri, B. P. & Cheresh, D. A. The role of alphav in human endothelial cells. Blood 99,
BMC Genomics 9, 153 (2008). integrins during angiogenesis: insights into 2060–2069 (2002).
42. Li, J. L. & Harris, A. L. The potential of new potential mechanisms of action and clinical 77. Hua, J. et al. Oxi4503, a novel vascular
tumor endothelium-specific markers for the development. J. Clin. Invest. 103, 1227–1230 targeting agent: effects on blood flow and
development of antivascular therapy. Cancer (1999). antitumor activity in comparison to
Cell 11, 478–481 (2007). 61. MacDonald, T. J. et al. Preferential susceptibility combretastatin A-4 phosphate. Anticancer Res.
43. seaman, s. et al. Genes that distinguish of brain tumors to the antiangiogenic effects of 23, 1433–1440 (2003).
physiological and pathological angiogenesis. an alpha(v) integrin antagonist. Neurosurgery 78. Cai, s. X. small molecule vascular disrupting
Cancer Cell 11, 539–554 (2007). 48, 151–157 (2001). agents: potential new drugs for cancer

nature reviews | clinical oncology volume 6 | JulY 2009 | 403

© 2009 Macmillan Publishers Limited. All rights reserved

 reviews

treatment. Recent Patents Anticancer Drug 85. Fischer, C. et al. Anti-PlGF inhibits growth of molecule multitargeted tyrosine kinase
Discov. 2, 79–101 (2007). veGF(r)-inhibitor-resistant tumors without inhibitor for renal cell carcinoma. Expert Opin.
79. Mauer, A. M. et al. A phase ii study of affecting healthy vessels. Cell 131, 463–475 Investig. Drugs 17, 253–261 (2008).
ABT-751 in patients with advanced non-small (2007). 93. rugo, H. s. et al. Phase i trial of the oral
cell lung cancer. J. Thorac. Oncol. 3, 631–636 86. Jain, r. K. & Xu, L. alphaPlGF: a new kid on the antiangiogenesis agent AG-013736 in patients
(2008). antiangiogenesis block. Cell 131, 443–445 with advanced solid tumors: pharmacokinetic
80. siemann, D. w., Chaplin, D. J. & (2007). and clinical results. J. Clin. Oncol. 23,
Horsmann, M. r. vascular-targeting therapies 87. Yan, M. & Plowman, G. D. Delta-like 4/Notch 5474–5483 (2005).
for treatment of malignant disease. Cancer signaling and its therapeutic implications. Clin. 94. wedge, s. r. et al. AZD2171: a highly potent,
100, 2491–2499 (2004). Cancer Res. 13, 7243–7246 (2007). orally bioavailable, vascular endothelial growth
81. siemann, D. w. & shi, w. Dual targeting of 88. Larrivee, B. et al. Activation of the UNC5B receptor factor receptor-2 tyrosine kinase inhibitor for
tumor vasculature: combining Avastin and by Netrin-1 inhibits sprouting angiogenesis. Genes the treatment of cancer. Cancer Res. 65,
vascular disrupting agents (CA4P or OXi4503). Dev. 21, 2433–2447 (2007). 4389–4400 (2005).
Anticancer Res. 28, 2027–2031 (2008). 89. Heroult, M., schaffner, F. & Augustin, H. G. eph 95. wedge, s. r. et al. ZD6474 inhibits vascular
82. Beerepoot, L. v. et al. Phase i clinical receptor and ephrin ligand-mediated interactions endothelial growth factor signaling,
evaluation of weekly administration of the during angiogenesis and tumor progression. Exp. angiogenesis, and tumor growth following oral
novel vascular-targeting agent, ZD6126, in Cell Res. 312, 642–650 (2006). administration. Cancer Res. 62, 4645–4655
patients with solid tumors. J. Clin. Oncol. 24, 90. Mendel, D. B. et al. In vivo antitumor activity of (2002).
1491–1498 (2006). sU11248, a novel tyrosine kinase inhibitor 96. Huynh, H. et al. Brivanib alaninate, a dual
83. Cooney, M. M. et al. Cardiovascular safety targeting vascular endothelial growth factor and inhibitor of vascular endothelial growth factor
profile of combretastatin a4 phosphate in a platelet-derived growth factor receptors: receptor and fibroblast growth factor receptor
single-dose phase i study in patients with determination of a pharmacokinetic/ tyrosine kinases, induces growth inhibition in
advanced cancer. Clin. Cancer Res. 10, 96–100 pharmacodynamic relationship. Clin. Cancer Res. mouse models of human hepatocellular
(2004). 9, 327–337 (2003). carcinoma. Clin. Cancer Res. 14, 6146–6153
84. rusnati, M. & Presta, M. Fibroblast growth 91. Chow, L. Q. & eckhardt, s. G. sunitinib: from (2008).
factors/fibroblast growth factor receptors as rational design to clinical efficacy. J. Clin. Oncol. 97. Hilberg, F. et al. BiBF 1120: triple angiokinase
targets for the development of anti- 25, 884–896 (2007). inhibitor with sustained receptor blockade
angiogenesis strategies. Curr. Pharm. Des. 13, 92. sonpavde, G., Hutson, T. e. & sternberg, C. N. and good antitumor efficacy. Cancer Res. 68,
2025–2044 (2007). Pazopanib, a potent orally administered small- 4774–4782 (2008).

404 | JULY 2009 | voLUme 6 www.nature.com/nrclinonc

© 2009 Macmillan Publishers Limited. All rights reserved