acuseeseceseerecccessccnncesaesereneeereeennsssnssesessaassunacsnnnesosssesossseeeesenneatons REVIEW
`
`
`
`
`
`
`Clinical applications of angiogenic growth factors
`and their inhibitors
`
`|
`
`NAPOLEONE FERRARA! &
`Kari ALITALO?
`
`patients with nonhealing ischemic ulcers and/orrest pain due
`
`In embryos, blood vessels form through
`two distinct processes, vasculogenesis
`and angiogenesis. Vasculogenesis
`in-
`volves the de novodifferentiation of en-
`in
`dothelial cells
`from mesodermal precursors, whereas
`angiogenesis new vessels are generated from pre-existing ones’.
`Vasculogenesis takes place only during embryonic develop-
`ment andleads to the formation of a primary vascular plexus.
`‘Later these rather uniformly sized endothelial channels are re-
`modeled into a mature system consisting ofa tree-like hierar-
`chy of large and small vessels. New capillaries then form
`through angiogenesis, either by sprouting or by splitting (in-
`tussusception)from their vessels of origin. In adults, angiogen-
`esis is essential for the female reproductive cycle, and for repair,
`remodeling and regeneration of tissues, for example during
`wound healing*. Neovascularization is also importantin patho-
`logical processes such as tumor growth and metastasis’.
`The known endothelial cell specific growth factors and their
`receptors can beclassified into vascular endothelial growth fac-
`| tor (VEGF) and angiopoietin (Ang) families(Fig. 1).Among the
`various angiogenic factors, VEGF is probably the most essen-
`tial for the developmentand differentiation of the vascular
`system*. Loss of a single VEGF allele results in embryonic
`lethality®® (Fig. 2) . Even selective inactivation of the heparin-
`binding isoforms of VEGF,
`leaving one functional isoform
`(VEGF,),
`is insufficient for the proper developmentof the
`cardiovascular system andresults in myocardial ischemia and
`perinatal or early postnatal lethality’. Also, other angiogenic
`factors, such as FGFs may work moreindirectly, some of them
`through the VEGFs and their receptors*, so that a thorough
`knowledge of the signal transduction pathways of VEGFs and
`angiopoietinsis essential for their use in therapeutic settings.
`
`|
`
`Therapeutic angiogenesis andinhibition of arterial restenosis
`An exciting frontier of cardiovascular medicineis therapeutic
`angiogenesis. Promoting the formation of new collateral ves-
`sels on the ischemic myocardium,leg muscles and othertissues
`would have an important effect on the treatment of disorders
`for which pharmacologicalinterventionhasbeenineffective in
`controlled trials and for which therapy is now limited to surgi-
`cal revascularization or endovascularinterventional therapy”.
`Several angiogenic molecules have been tested in animal
`models, including bFGF, aFGF, FGF-S, VEGF isoforms, VEGF-
`C, HGF/SF and Ang-1/Ang-2. The factors tested most exten-
`sively are VEGF and bFGF. In somecases, the recombinant
`protein wastested. In others, gene transfer using naked DNA
`or adenoviral vectors was used. A single intra-arterial adminis-
`tration of 500-1000 ugof rhVEGF,.s augmented perfusion and
`development of collateral vessels
`in a
`rabbit model of
`hindlimb ischemia in which the femoral artery wassurgically
`| removed", Similar results were obtained in the same model
`
`NATURE MEDICINE + VOLUMES * NUMBER 12 ¢ DECEMBER 1999
`
`with intramuscular or intra-arterial ad-
`ministration of aFGF, bFGF, HGF/SF and
`VEGF-C (refs. 11-14). VEGF administra-
`tion after removal of the femoral artery
`not only resulted in increased vascularization butalso led to
`recovery of the normal endothelial reactivity to various media-
`tors'®. Arterial gene transfer with cDNA encoding VEGFiso-
`formsalso led to revascularization to an extent comparable to
`that achieved with the recombinantprotein'®. Moreover, ad-
`ministration of a VEGF,,; adenovirus vectorshortly after com-
`moniliac artery ligation in the rat was capable of stimulating
`an angiogenic response that protects against subsequent oc-
`clusion of the femoralartery, indicating that gene transfer of
`VEGF might be useful in the prophylaxis of advancing arterial
`occlusive disease”. As little as 2 ug rhVEGF delivered over 4
`weeks periadventitially, distal to the occlusion, resulted in a
`significant increase in coronary blood flow and functional im-
`provementin a pig model of chronic myocardial ischemia’.
`Very similar results were obtained using bFGF (ref. 19).
`Unexpectedly, even a single intracoronary administration of
`VEGE (or bFGF) was efficacious in this model to an extent
`comparable to that of 4-week infusion, despite the fact that
`only a small fraction of protein localizes to the ischemic area”’.
`Given suchresults, it is conceivable that young and otherwise
`healthy animalsare very responsive to exogenous growth fac-
`tors in the context of ischemia. At least some of this respon-
`siveness may be due to the upregulation of VEGFreceptors in
`the endothelia of ischemic tissues?!, Adenovirus-mediated
`gene transfer of VEGF). (ref. 22) or FGF-S (ref. 23) also resulted
`in collateral vessel growth and functional improvement in
`porcine models of myocardial ischemia.
`These encouraging animalstudiesled to clinical trials using
`recombinant VEGF,,;, aFGF, bFGF or gene therapy with plas-
`mid or with adenoviral vectors. There is considerable debate
`whether gene therapy or administration of recombinantpro-
`tein would be preferable. Delivery of angiogenic proteins by
`gene therapy might not only minimizetheir systemic side ef-
`fects, such as hypotension (VEGF) or nephrotoxicity (bFGF),
`but also provide a slow release of the encoded factor for 1-2
`weeks,
`leading to a more lasting angiogenic response.
`However, slow release of the recombinantprotein, using mi-
`crospheres or heparin-alginate formulations, might achieve
`the sameresults, without the potential risks associated with
`the use ofviral vectors.
`Arterial gene transfer of naked plasmid DNA encoding
`VEGF,,; ina patient with severe limb ischemia produced angio-
`graphic andhistologic evidence of angiogenesis in the knee,
`mid-tibial and ankle levels 4 weeks after the transfer™. In a sub-
`sequent study, the VEGFys plasmid cDNA was injected intra-
`muscularly”®. Gene transfer was done in ten limbs of nine
`
`1359
`
`CELLTRION - EXHIBIT 1045
`0001
`
`0001
`
`CELLTRION - EXHIBIT 1045
`
`

`

`...................................
`..............
`.....................
`..
`......................
`......
`.........
`. ..............
`.............
`............
`REVIEW···--· .. ·····
`
`. ...... .
`
`0
`•
`
`2•
`
`29
`
`�c,o
`
`VECFR-2
`
`VECFR-3
`
`such as VF.GF
`proteins
`3•, but secreted
`wall was not very efficient
`in the ankle­
`lmpfOvement
`disease.
`arterial
`to peripheral
`using infu.
`trials
`gene transfer
`could be used for therapeutic
`were reported
`flow in eight limbs
`brachial index and distal
`21•
`or histamine-induced
`increase of en.
`catheters'•
`sion-perfusion
`also shown that
`by the same group have
`small trials
`Additional
`VEGF and VEGF-C share one
`". Because
`permeability
`dothelial
`in clinical
`DNA resulted
`plasmid
`of the VEGf,6S
`local injection
`VEGF-C
`in the other receptor,
`but differ
`(VEGFR-2)
`receptor
`ischemia
` or
`by myocardial
`affected
`in patients
`improvement
`in the
`effects
`but distinct
`might have overlapping
`and VEGF,-s
`obliterans)
`(thromboangiitis
`disease
`Burger's
`none
`17• However,
`inhibits
`intimal
`VEGF-C gene transfer
`wall. However,
`vessel
`placebo-controlled.
`using
`trials
`Clinical
`of these studies were
`is at least equ;il to
`effect
`and the protective
`early,
`thickening
`gene transfer
`of
`mediated
`DNA or adenovirus
`VEGF-C naked
`8•
`that seen with VEGF,6s gene transfer"
`are now in phase r.
`patients
`ischemia
`VEGF,21 in myocardial
`before
`with limb ischcmia,
`from a patient
`angiograms
`Femoral
`growth factor
`endothelial
`vascular
`inhibition of
`Therapeutic
`of a VEGF,6, plasmid/liposome
`and 3 months after transfection
`VilScular density
`after the
`show increased
`vector,
`expression
`Tumors
`chamben in
`in transparent
`The gwwth of tumor xenografts
`and some cau­
`is ongoing
`the trial
`(Fig. 3). However,
`treatment
`indicating
`density,
`in vascular
`by an increase
`mice is preceded
`such data, until more pa­
`tion should be used in interpreting
`on the development
`that the rapid growth of tumors depends
`evaluated.
`are more extensively
`of placebo
`and the effect
`tients
`of angiogenesi
`In 1971, inhibition
`supply".
`of a neovascular
`VEGF,65 and bFGF are also
`using recombinant
`trials
`Clinical
`s
`for the treatment
`of solid tu­
`as a valid strategy
`was proposed
`ischemia
`with coronary
`patients
`In a phase l study in
`ongoing.
`of tumor angiogenesis
`for the mediator(s)
`mors and the search
`by intracoronary
`infu­
`was administered
`in which rllVEGF,.s
`at all doses tested".
`tolera:ed
`was safely
`sion, the molecule
`was begu11'
`41) or angiopoietin/Tie2
`of bFGF (ref.
`inhibition
`Although
`in seven of
`in perfusion
`of improvement
`There was evidence
`so far VEGJ: and its re­
`tumor growth,
`may inhibit
`42,43)
`(refs.
`in five of seven
`and improved col lateralization
`subjects
`fifteen
`system in
`investigated
`the most extensively
`coronary mgiography. However,
`constitute
`ceptors
`a
`follow-up
`who underwent
`of anti-cancer
`and are now a main target
`tumor angiogenesis
`phase II study, in which
`placebo-controlled
`subsequent
`upregulated
`in most
`VEGF mRNA is substantially
`strategies.
`infusion,
`fol­
`intracoronary
`as a single
`rhVEGF was delivered
`has not demonstrated
`infusions,
`lowed by three intravenous
`the lllain
`tumor cells represent
`Although
`human tumors'.
`stroma is also an important
`source of VEGF, tumor-as�ociated
`than placebo
`in
`was not better
`• The treatment
`bencfit
`clinical
`at 60 days:?Y. Brief expo­
`There is a correlation between
`at least
`time and pain relief,
`VEGF
`treadmill
`site ofVEGF production".
`cancer
`breast
`in primary
`sel density
`and microves
`expression
`in this t rial, may be
`such as those achieved
`sures to rhVEGF,.,.
`in several
`has been described
`correl.ation
`A similar
`sections••.
`a therapeutically
`meaning­
`and maintain
`to trigger
`insufficient
`of extensive
`in the context
`especially
`response,
`ful angiogenic
`gastric carcinoma••.
`including
`other malignancies,
`of VEGF in
`administration
`of
`Also, systemic
`disease.
`atherosclerotic
`plasma levels
`there are increases in
`Furthermore,
`individuals,
`and
`with tumor-free
`compared
`tumor patients
`rhVEGF,.,1 or other factor
`an appropriate
`may fail to generate
`with a
`from bchemic to non-is•
`gradient
`concentration
`angiogenic
`are associated
`before chemotherapy
`high VEGF levels
`poor outcome".
`in a variety
`of
`of angiogenesis
`aspect
`a requisite
`chemic areas,
`of VEGJ'.' in tumorigenesis
`for involvement
`evidence
`Direct
`Moreover,
`circumstances'.
`the
`and pathological
`physiological
`suspected,
`and
`than initially
`greater
`is probably
`placebo effect
`against
`antibodies
`using monoclonal
`was first demonstrated
`myocar-
`with ve1y compromised
`even patients
`improve­
`a substantial
`may show
`dial function
`A phase II study With bFGF
`placebo.
`ment with
`is now ongoing.
`ischemia
`for coronary
`into the vascular
`wall of­
`Local gene transfer
`for the treatment
`alternative
`fers a promising
`sVECFR-1 VEGFR.J
`after percuta­
`of restenosis
`of the complication
`(PTCA) and
`angioplasty
`neous transcoronary
`Restenosis occurs
`in many
`stenting.
`coronary
`in 6 months, leading
`to ob-
`patients
`treated
`in 20-35% of the patients"'.
`The
`struction
`depends on en-
`of r estenosis
`pathogenesis
`which also predisposes
`ar•
`damage,
`dothelial
`conditions,
`such as
`to other pathological
`teries
`of resteno-
`Prophylaxis
`spasms or thrombosis.
`be based on strategies
`for
`sis could therefore
`of en-
`or enhancement
`protection
`endothelial
`ial growth factors
`dothelial repair
`and endothel
`or vascular
`could be used for
`gene transfer
`Fig. 1 VEGFs,
`in cells and tissues.
`effects
`endothelial
`and some of their
`their receptors
`in balloon-injured
`this". Re-endothelization
`auto/transphosphorylation,
`and subsequent
`dimerization
`binding induces
`receptor
`Ligand
`rat carotid
`by a single
`was accelerated
`artery
`responses.
`cellular
`and leads to differential
`pathways
`transduction
`signal
`various
`activates
`into the
`VEGF injected
`dose of recombinant
`1; a,b,
`NP-1, neuropilin-
`proteoglycan;
`sulphate
`HSPG, heparan
`VECFR-1;
`soluble
`sVECFR-1,
`2•33• Vessel
`or locally'
`bloodstream
`status
`was
`ref. 95). VE·
`integrin
`VEGFR-2;
`with activated
`complex
`to make a molecular
`a,b, (reported
`of VEGF plasmid
`also improved by injection
`for VEGF-deperdent
`a requirement
`with VEGFR-2,
`is also able to form a complex
`cadherin
`carotid arter­
`of rabbit
`surface
`into adventitial
`point mutation of
`P1114l,
`the P13-kinase/Akt
`pathway'.
`involving
`anti-apoptotic
`signals
`in the arterial
`gene transfer
`ies". lntravascular
`with lymphoedema••.
`in a family
`patients
`affecting
`VECFR-3
`
`VEGF-C
`VEGF
`�
`"".
`�V:t/�i1/
`
`av6l V£.
`
`Cadherin
`
`T
`
`Pll l◄L
`
`survival
`ty, DNA synthesis,
`permeabili
`Migration,
`
`angiogenesis Lymphangiogenesis
`
`1360
`
`• VOLUME 5 • NUMBER 12 • DECEMBER 199'
`NATURE MEDICINE
`
`0002
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`························ .. ·········· .. ···· .. ····· .. ····· ............................................................................................................................................... REVIEW
`
`
`
`
`
`
`
`
`
`sexually immature animals that have not completed statural
`
`
`
`
`
`growth55• rhuMAb VEGF is now in phase II clinical trials for the
`
`
`
`
`
`treatment of non-small cell lung carcinoma and colorectal car­
`
`
`
`cinoma in conjunction with standard chemotherapy and for
`
`
`
`
`breast and renal cell carcinoma as a single agent. In addition,
`
`
`
`
`small molecules that inhibit VEGFR·2 signal transduction are
`
`
`undergoing phase II clinical trials in cancer patients
`16•
`
`
`
`
`l'urthermore, monoclonal antibodies against VEGFR-2 are en­
`
`
`tering clinical trials.
`
`
`Retinal ischemia and other conditions
`Diabetes mellitus, occlusion of the central retinal vein or pre­
`
`
`
`
`
`
`
`
`matu1ity with subsequent exposure to oxygen can all be associ­
`
`
`ated with intraocular neovascularization". A common
`
`
`
`
`denominator among these conditions is retinal ischemia';. The
`
`Fig. 2 Yolk sac of E10.5 VEGF•1• and VEGF .,. mouse embryos'. There is
`
`
`
`
`new blood vessels may lead to vitreous hemorrhage, retinal de­
`
`
`in the yolk sac of the heterozygous, an apparent absence of vasculature
`
`
`
`
`
`tachment, neovascular glaucoma, and eventual blindness.
`
`
`
`which die around E11. This is probably the only example among verte­
`
`
`
`Diabetic retinopathy is the leading cause of blindness in the
`
`
`
`
`
`brates of lethality after inactivation of a single allele of a gene that is not
`
`
`working population. The hypothesis that ischemia-induced
`
`maternally imprinted.
`
`
`VEGF may be pathogenic in these conditions was initially
`
`
`
`
`
`In of patients. tested by measuring VEGF levels in the eye fluids
`
`
`
`VEGF in human xenografts in nude mice••. These initial studies
`
`
`
`
`a large series with 165 patients, a strong correlation was found
`
`showed that several tumor cell lines can be substantially
`
`
`
`between concentrations of VE.CF in both aqueous and vitreous
`
`8• These findings were ex­
`growth-inhibited by this treatment'
`
`
`
`
`
`and active proliferative retinopathy associated with diabetes,
`
`
`
`
`tended to a broad variety of tumor cell lines, including carcino­
`
`
`
`
`
`occlusion of central retinal vein or prematurity58• Direct evi­
`
`
`
`mas, sarcomas and gliomas•. Intravital videomicroscopy
`
`
`dence fo, the role of VF.GF as a mediator of intraocular neovas­
`
`
`
`techniques have augmented our understanding of VEGF in tu­
`
`
`
`cularization has been generated in several animal models,
`morigenesis
`
`
`...,·,.,. Non-invasive imaging of the vasculature
`
`
`
`including a primate model of Iris neovascuiarization and a
`
`
`
`demonstrated a nearly complete suppression of tumor-associ­
`
`
`
`mouse model of retinopathy of prematurity. in the former, in­
`
`
`
`ated angiogenesis in animals treated with monoclonal antibod­
`
`
`
`
`traocular administration of monoclonal antibodies against
`
`
`
`ies against VEGF compared with controls. providing a direct
`
`
`
`VEGF substantially inhibits the neovascularization that follows
`
`
`
`verification that inhibition of angiogenesis is the mechanism
`
`
`
`the occlusion of central retinal veins$9
`
`• Likewise, soluble
`9• lntravital
`
`VEGFR-1 or VEGFR-2 extracellular domains fused to the im­
`
`
`of tumor suppression after anti•VEGF treatment'
`
`
`microscopy techniques have also been used to investigate the
`
`
`
`
`munoglobulin y Fe domain suppress retinal angiogenesls in the
`
`mouse model(,<). There is also evidence that growth
`
`
`
`effects of VEGF on the permeability and other properties of
`tumor vessels
`
`
`
`50• Treatment with antibodies against VEGF re­
`
`
`
`hormone/insulin-like growth factor-I is involved in ischemia­
`
`
`
`
`sulted in time-dependent reductions in vascular permeability,
`
`
`induced retinal neovascularization
`61•
`
`
`
`in the diameter and tortuosity and eventually to a regression of
`
`
`
`Neovascularization is a principal cause of visual loss also in
`
`
`
`survival fac­tumor blood vessels; thus, V£G r is also an essential
`
`
`
`the wet form of age-related macular degeneration (AMO), the
`
`
`
`
`overall leading cause of blindness"'. Several studies have docu­
`
`
`
`
`tor for tumor endothelial cells'°. Further evidence that VEGF
`
`
`
`mented the immunohistochemical localization of VEGF in sur­
`
`
`
`
`action Is required for tumor angiogenesis has been provided by
`
`
`
`the finding that retrovirus-mediated expression of a dominant
`
`
`
`
`
`gically resected choroidal neovascular membranes from AMO
`
`
`
`
`
`negative VE.GFR-2 mutant, which inhibits ,ignal transduction
`
`
`
`63• These findings suggest involvement of V£GF in the
`patients
`
`
`
`
`progression of AMD-related choroidal neovascularization.
`
`
`
`
`through wild-type VEGFR-2 receptor, suppresses the growth of
`
`glioblastoma multiforme as well as other tumor cell lines in
`
`
`
`Anti-VEGF strategies for AMD are now being explored in clini•
`
`
`
`
`cal trials. One approach. consists in the intravitreal administra­
`
`
`vi,·os'. Furthermore, high local expression of the soluble extra­
`
`
`tion of a recombinant humanized anli-VEGF Fab antibody
`
`
`
`
`cellular domain of VEGFR-1 or VEGFR-2, achieved by aclminis­
`
`
`
`
`
`fragment. Another strategy involves the injection of 2'-fluo­
`
`
`
`
`tration of the recombinant proteins, adenoviral-mediated gene
`
`
`
`
`transfer or by stable transfection of tumor cells, may signifi­
`
`
`
`ropyrimidine RNA oligonucleotide ligands (aptamerst
`'.
`
`
`VEGF inhibition may also have therapeutic value for the
`
`
`
`
`cantly inhibit tumor growth, metastasis and mortality rate in
`nude mices2•13•
`
`
`
`treatment of ischemic•reperfusion related brain edema and in­
`
`
`
`jury. VEGF antagonism has shown beneficial effects in a mouse
`
`
`
`Several strategies have been used to generate VEGF inhibitors
`
`model of cortical ischemia
`
`65; reducing acutely the volume of
`trials. One approach involves the 'human­
`
`
`sunable for clinical
`
`
`
`
`
`edematous tissue and resulting in a significant sparing of corti­
`
`
`
`
`ization' of mouse monoclonal antibodies. A chief advantage of
`cal tissue.
`
`
`
`'humanized' antibodies is a high degree of specificity, com­
`in the female reproduc­VEGF is important in angiogenesis
`
`
`
`
`
`bined with a long half-life and little or no immunogenicity. A
`
`
`
`
`tive tract. VEGF inhibition results in suppression of corpus lu­
`
`
`
`
`'humanized' high-affinity monoclonal antibody against VEGF
`
`
`
`teum angiogenesis in rodents� and primates". VEGF inhibitors
`
`
`(rhuMAb VEGF) with the same affinity and biological proper­
`
`
`ties as the original murine antibody has been described".
`
`
`
`might be used to treat conditions characterized by ovarian hy•
`
`
`
`
`Toxicological studies in primates have shown that the effects of
`
`
`perplasia and hypervascularity, such as the polycystic ovary
`
`
`
`
`rhuMAb VEGF are limited to inhibition of angiogenesis in the
`
`
`66. VEGF-dependent angiogenesis may also be impor­
`syndrome
`
`
`
`female reproductive tract and in the epiphyseal growth plate in
`
`
`tant pathogenically in endometriosis. Furthermore, VEGF is a
`
`
`
`NATURE MEDICINE • VOLUMES • NUMBER 12 • DECEMBER 1999
`
`
`
`1361
`
`0003
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`REVIEW ...........................................................................................................................................................................................................
`
`
`.................. ..
`
`mediator or the ovarian growth and increased vascular perme­
`
`
`
`
`Also, it is unknown whether an angiogenic treatment may be
`
`
`
`
`
`
`
`ability of ovarian hyperstimulation syndrome, a potentially
`
`
`
`sufficient to induce functional blood vessels for prolonged
`pe­
`
`
`
`fatal condition characterized by massive ovarian enlargement
`
`
`riods or wi II need to be re-administered periodically in order to
`
`
`
`that may follow medical induction of ovulation with go­
`
`maintain such vessels.
`nadotropins
`6'.
`A K14-driven VEGF-C transgene induced lymphangiogenesis
`
`
`
`VEGF-c but no angiogenesis in mouse skinS<, and recombinant
`
`
`Perspectives
`
`
`
`
`also stimulated lymphatic vessel hyperplasia in mature chick
`VEGF,65 binds to neuropilin-1, which functions as a ligand
`
`
`
`
`
`
`chorioallantoic membrane••. Thus, besides anglogenesls, it may
`
`
`
`
`binding subunit of putative transmembrane receptors mediat­
`
`
`
`also become possible to direct therapeutic lymphangiogene
`sis
`
`
`
`
`ing specific signals for different semaphorins, the molecules
`
`
`
`in patients, such as after evacuation of axillary lymph nodes in
`
`
`
`is mediating the collapse of axonal growth cones&1. Neuropilin
`
`
`breast carcinoma surgery.
`
`
`
`
`expressed in endothelial cells and enhances the mitogenic ef­
`Despite the potential redundancy of tumor angiogenesis fac­
`
`
`
`
`
`fects of VEGFR-2 upon VEGF,65 stimulation. Thus, there may be
`
`
`
`
`tors, inhibition ofVEGF alone seems sufficient to achieve con­
`
`
`
`
`signals between cross-regulation of cellular an as-yet ill-defined
`
`
`
`
`
`siderable tumor growth suppression in a wide variety of
`
`
`
`these two families of factors. These findings lead to the intrigu­
`
`
`
`
`
`models. However, it remains to be established whether tumors
`
`
`
`ing conclusion that the processes of axon guidance and devel­
`
`
`
`
`are able to activate, after prolonged therapy, alternative angio­
`
`
`opment of a network of capillary tubes share at least some
`
`
`
`genic pathways that might confer resistance to the treatmen
`t.
`
`
`
`common molecular mechanisms. In addition, the angiopoietin
`
`
`These issues should be addressed in the current clinical
`trials
`
`
`
`
`
`receptorffie and ephrin families of endothelial tyrosine kinases
`
`
`
`with various VEGF inhibitors. A challenge now in anti-VEGF
`
`
`
`have important functions in the formation and maintenance
`
`
`
`
`(and anti-angiogenic) therapy is devising appropriate and reli­
`
`
`
`of the vascular system""·". Endothelial cell-specific members of
`
`
`
`able markers to monitor tumor progression. There is consider­
`
`
`the TGF-13 receptor and Notch families have also been de­
`able debate whether blood vessel count in biopsy
`
`
`
`
`5·'6 may provide a reliable indicator of respome to
`
`
`
`scribed12·73. Given this complexity of vascular endothelial sig­
`specimens'
`
`
`naling, therapies using VEGF alone or any other single
`
`
`
`
`the treatment. There are also efforts to identify surrogate end­
`
`
`
`
`
`angiogenic factor may produce incompletely functioning or
`
`
`
`
`points, applying non-invasive approaches, such as magnetic
`
`
`
`
`unstable endothelial channels with defective arteriovenous
`
`resonance imaging84.
`
`
`
`and pericellular differentiation, characteristic of many tu­
`VEGF is not only a mitogen but also a potential survival fac­
`
`
`
`
`
`
`mors". Combinations of growth factors may be preferable In
`
`
`
`
`tor for endothelial cells•. Such a 'maintenance' function seems
`
`
`
`
`
`future therapies directed to neovascularization of tissues, with
`
`
`
`
`to be developmentally regulated, as it is very dependent on the
`
`
`
`an adequate investment of the formed vessels with perien­
`
`age of the animal87. VEGF inactivation during early postnatal
`
`
`
`dothelial matrix and pericyte/smooth muscle cells. In fact, a
`
`
`
`
`
`
`of gene targeting inducible life, achieved by Cre-loxf>-mediated
`
`
`
`more heterogenous set of genes coordinating angiogenic func­
`re­
`
`
`
`by administration of a soluble VEGFR-1 chimeric protein,
`
`
`
`
`tions may be provided by active ongoing research of hypoxia­
`
`
`
`sults in regression of the vasculature, kidney failure and lethal­
`
`
`
`
`regulated gene expression in mammalian cells". Also, some
`
`
`
`
`itf'. However, in adult animals a similar treatment has no
`
`
`proteins, such as the VEGFR-2 activating HIV
`virus-encoded
`
`
`
`
`on the existing vasculature. Therefore, a process of mat­
`effects
`
`
`
`G-protein­, Kaposi sarcoma herpesvirus-associated
`Tat protein'
`
`
`
`uration occurs in endothelial cells such that VEGF eventually is
`6
`
`
`not essential for survival. This switch seems to take place in the
`
`
`
`
`coupled receptor" or Orf virus encoded VEGF-E'S-30 may offer
`
`
`
`
`new insights into the mechanism of regulation of angiogenesis.
`
`mouse around the fourth postnatal week. Absence
`of pericyte
`
`
`
`
`
`Although recent research has focused on the combination of
`
`
`
`VEGF and Ang-1 as being especially promising, it is not known
`
`
`now which growth factor combinations will prove to be the
`
`
`most effective therapeutically. VEGF and bFGF have a very
`
`
`
`
`synergistic effect in the induction of angiogenesis, both in vitro
`
`
`and in vivo'. The interaction between VEGF and HGF/SF is also
`
`
`
`
`
`being actively investigated. Although transgenic expression of
`
`
`Ang-1 in the skin epidennis under the keratin (K)14 promoter
`
`has been associated with neovascularization
`8', other studies,
`
`
`
`
`using defined amounts of the recombinant protein in a model
`
`
`
`
`of adult neovascularization, have failed to demonstrate strong
`
`
`angiogenic responses to Ang-1, unless it is used in combina­
`
`
`
`tion with VEGF (refs. 71,82). This discrepancy may be ex­
`
`
`plained by the fact that the expression ot the K14 promoter is
`
`
`
`
`initiated already at midgestation, and thus the results may re­
`
`
`flect persistence of the fetal neovascularization. It is possible,
`
`
`
`however, that Ang-1 may provide a co-factor for combination
`
`
`
`therapies. A further unresolved issue is the correct dosage of
`
`
`
`growth factor(s). This seems particularly important for a mole­
`Fig. 3 Angiography of the lower extremity of a patient with limb is·
`
`
`
`
`
`cule like VEGF, which has several isoforms and such a tight
`
`
`
`chemia before (PRE) and 3 months after (3 MO) the transfection or a
`
`
`
`
`dose-response effect that a 50% reduction in expression re­
`
`
`VEG Fl 65 plasrnid/liposomc expression vector, showing strongly increasi!d
`
`
`
`
`
`
`sults in lethality during embryonic life5•6. Conversely, continu­
`
`
`vascular density after the treatment. Courtesy H. Manninen, P. Matsi, K.
`
`
`
`
`A. I.
`
`
`ous local overexpression of VEGF may result in a
`uala,
`
`
`
`Makinen, M. Hilpelainen, M. Laitinen, E. Alhava and S. Yla-Hertt
`
`
`hemangioma-llke vasculature and thus can be deleterious".
`
`
`
`
`
`
`Virtanen Institute and Kuopio University Hospital (Kuopio, Finland).
`
`1362
`
`1919
`, , .. nr:rn..AA(R
`NATURE MEDICINE • VOLUMES • ,-,11u,AAl!O
`
`0004
`
`

`

`,......
`
`
`
`
`
`························································ ... · ................................................................... REVIEW
`
`
`
`
`
`---------------
`------ --�
`- - -----------------------------
`
`
`
`
`
`their determining vessels may be a factor coverage in immature
`
`in t,eatment of cardio­and vasculogenesis
`
`
`9. Rivard, A. &: Isner, J.M. Angiogenesis
`
`vascular disease. Mo/. Med. 4, 429-440 (1998).
`
`
`
`
`
`
`sug­other evidence dependence on VEGF (ref. 88). However,
`10. Takeshita, S. et al. Therapeutic angiogenesis. A single inlra-arterial bolus of vas­
`
`
`
`
`
`
`gests that the molecular/intracelullar nature of thls switch may
`
`
`
`
`cular endothelial growth factor augments revascularization in a rabbit ischemic
`
`
`be more complex and mostly still to be determined"'. In juve­
`
`
`
`hind Hmb model./. Oin. /1,ve,r. 93, 662-670 (1994).
`
`
`
`11. Pu, L.Q. et al. Enhanced revascularization of the ischemic limb by angiogenic
`
`
`
`bone forma­for endochondral nile animals, VEGF is essential
`
`
`therapy. Circulation 88, 208-215 (1993).
`growth89·s1. In the fully developed
`tion and longitudinal
`
`
`
`
`
`12. Asahara, T. et al. Synergistic effe<t of vascular endothelial growth factor and
`
`
`angiogenic anlmal, VEGF may be required mainly for active
`
`
`
`
`
`basic fibroblast growth factor on angiogenesis in vivo. Cirwlotion 92, 11365-11371
`(1995).
`
`
`wound heal­processes such as corpus luteum development or
`13. Van Belle, E. et al. Potentiated angiogenic effect of scatter factor hepatocyte
`
`
`
`
`
`
`
`
`for endothelial ing. Neverthless, VEGF may be important
`
`
`
`
`
`growth factor via induction ol vascular endothelial growth factor: the case for
`
`
`
`ptmu:::rine c:,mplifi�ation of angiogenesis, 97, 381-390 (1998).
`
`
`
`homeostasis In the adult In certain circumstances; for example,
`Circvlotion
`
`
`
`
`14 Witzenbichler, 8. et al. Vascular endothelial growth factor-C (VEGF-CNEGF-2)
`
`
`
`
`
`
`during disease states. Indeed, prolonged VEGF inhibition failed
`
`
`
`
`
`
`promotes angiogenesis in the setting of tissue ischemia. Am. /. Pothol. 1S3,
`si or ro­
`
`to induce glomerular damage in normal primates
`
`381-394 (1998).
`15. Bauters, C. et al. Physiological assessment of augmented vascularity induced by
`
`
`
`
`
`
`dents8'·90, despite
`
`
`the strong constitutive expression of the
`
`
`
`
`VEGF in ischemicrabbit hindlimb. Am./. PhySiol. 267, H1263-Hl 271 (1994).
`
`VEGF mRNA in podocytes and other cell types in the adult kid­
`
`
`
`16. Takeshita, S. et al. Gene transfer of naked DNA encoding for three isoforms of
`
`
`
`ney'. However, administration of VEGF inhibitors to rats with
`
`
`
`
`
`vascular endothelial growth factor stimulates collateral development in vivo.
`
`
`Lab. Invest, 75, 487-501 (1996),
`
`
`(!lesangioproliferative nephritis results in impaired glomerular
`gene induced by adenovirus.mediated 17. Mack, C.A. et al. Salvage angiogenesis
`
`
`
`
`
`endothelial regeneration and increased endothelial cell death'°.
`
`
`
`
`
`
`
`
`transfer of vascular endothelial growth factor protects against ischemic vascular
`
`
`occlusion./. 1.7, 699-709 (1998).
`
`
`
`Some CD34· hematopoietic progenitor cells mobilized by
`Vose. S11rg.
`18. Pearlman, J.D. et al. Magnetic resonance mapping demonstrates benefits of
`
`
`
`
`
`
`GM-CSF from human peripheral blood, bone marrow, fetal
`
`
`
`
`
`VEGF-induced myocardial angiogenesis. NawreMed. 1, 1085-1089 (1995).
`
`
`liver or umbilical cord blood were shown to express VEGFR-2
`
`
`
`
`
`
`19. Harada, K. et al. Basic fibroblast growth factor improves myocardial function in
`
`
`
`
`
`chronically ischemic porcine hearts./. Clin. Invest. 94, 623-630 (1994).
`
`
`on human on their surface91, and VEGFR-2 is expressed
`
`
`
`
`20. lopez, J.J. er al. VEGF administratiot, i,, Ch(0nic myo(erdia! ischemia in pig$.
`
`
`
`hematopoietic stem cells92• Endothelial progenitor cells expand
`
`Cardiovasc. Res. 40, 272-281 (1998).
`
`
`
`of bFGF and differentiate into endothelial cells after addition
`
`
`
`21. Li, J. et al. VEGF, flk-1, andflt-1 expression in a rat myocardial infarction model of
`
`
`angiogenesi,. Am./. Physiol. 1.70, Hl 803-Hl 811 (1996).
`
`
`to and VEGF to the cultures, and they can thus be considered
`22. Mack, C.A. et o/. Biologic bypass with the use of adenovirus-mediated gene
`
`
`
`
`
`
`provide endothelial progenitor cells0' ... ,. The endothelial prog­
`
`
`
`
`transfer of the complementary deoxyribonucleic acid for vascular endothelial
`
`
`
`
`
`growth factor 121 improves myocardial perfusion and function in the ischemic
`
`
`enitor cells from bone marrow may be mobilized using the stro­
`
`
`
`porcine heart. /. Thor. Cordiovosc. Surg. 115, 168-1 76 (1998).
`
`
`
`or tissue the GM-CSF cytokine ved factor 1 chemokine,
`mal-deri
`F. et o/. lntracoronary
`factor-5
`23. Giordcu,o,
`
`
`gene transfer of fibrobla.st growth in­
`
`
`
`in ac­hypoxia94• As these cells may be capable of participating
`
`
`
`
`
`creases blood flow and contractile function in an ischemic region of the heart.
`
`Not. Med. 2, 534-539 (1996).
`
`
`tive angiogenesis after entry into the circulatory system9', they
`
`after arterial gene transfer of evidence of angiogenesis 24. Isner, J.M. et al. Clinical
`
`
`
`
`
`
`
`
`
`
`provide an interesting possibility for the delivery of cellular or
`
`
`
`
`
`phVEGFI 65 in patient with ischaemic limb. Lancet 348, 370-374 (1996).
`
`
`gene therapy to sites of neovascularization.
`
`
`
`
`65 atter intramuscular expression of phVEGFI 25. Baumgartner, I. et al. Constitutive
`
`
`
`
`
`
`gene transfer promotes collateral