Tumor angiogenesis is crucial for the progression and metastasis of cancer. The vasculature of tumor tissue is different from normal vasculature. Therefore, tumor vascular targeting therapy could represent an effective therapeutic strategy with which to suppress both primary tumor growth and tumor metastasis. The use of viral vectors for tumor vascular targeting therapy is a promising strategy based on the unique properties of viral vectors. In order to circumvent the potential problems of antiviral neutralizing antibodies, poor access to extravascular tumor tissue, and toxicities to normal tissue, viral vectors need to be modified to target the tumor endothelial cells. Viral vectors that could be used for tumor vascular targeting therapy include adenoviral vectors, adeno-associated viral vectors, retroviral vectors, lentiviral vectors, measles virus, and herpes simplex viral vectors. In this review, we will summarize the strategies available for targeting viral vectors for tumor vascular targeting therapy.
Normal angiogenesis and tumor angiogenesis
Angiogenesis is a multistep process of new blood vessel formation from pre-existing vasculature. This process usually takes 2 forms: sprouting and unsprouting. Sprouting angiogenesis is the development of new blood vessels through proteolytic degradation of the extracellular matrix, migration/proliferation of endothelial cells (ECs), new organization of the luminal membrane, and maturation of endothelial cells to functional capillaries.1 New vessels can arise from postcapillary venules (angiogenesis) or from precapillary arterioles (arteriogenesis). Nonsprouting angiogenesis occurs by intussuception, which means the splitting of primary vessels by transcapillary pillars.2
Most normal angiogenesis occurs in the embryo. In adults, normal angiogenesis occurs during the ovarian cycle, pregnancy, and during physiologic repair processes such as wound healing or endometrial regrowth.3 In normal circumstances, angiogenesis is a highly ordered process that is under tight regulation by both angiogenesis-inducing factors and angiogenesis-inhibiting factors. These factors include soluble growth factors secreted from cells, such as vascular endothelial growth factor (VEGF), angiopoietins (Ang's), fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), transforming growth factor-β (TGF-β), and membrane-bound molecules, such as integrins, cadherins, and ephrins. Other accessory cells, such as pericytes, play a supportive role in the development of the vasculature. In addition to these factors, mechanical forces mediated by blood flow on vascular endothelium also contribute to the angiogenesis process. Vessels that are not perfused with blood eventually regress.4
In 1971, Dr Judah Folkman first proposed the hypothesis that tumor growth is angiogenesis dependent.5 Now it is widely recognized that tumor angiogenesis is crucial for the progression and metastasis of tumor. Tumor cells may begin growing around existing blood vessels.6 Tumor cells produce Ang2, which induces the regression of blood vessels in the absence of VEGF or in the presence of low levels of VEGF.7 This results in central tumor hypoxia and necrosis. Tumor nodules cannot grow beyond 2 to 3 mm in diameter without the supply of the blood.1 Central tumor hypoxia up-regulates the expression of the transcriptional factor hypoxia-inducible factor (HIF), which upregulates the secretion of VEGF by tumor cells.8 This high level of VEGF in tumor tissue confers upon tumor vasculature its unique properties. The interaction of VEGF with its receptors presented on the endothelial-cell surface leads to endothelial-cell migration, proliferation, and vascular permeability.9 Moreover, the expression of Ang2 in the presence of high levels of VEGF in tumor tissues facilitates the process of tumor angiogenesis by rendering endothelium unstable and plastic.10
A number of studies have led to the hypothesis that tumor cells can form blood vessels and contribute to the functional vasculature themselves. This is called vascular mimicry.2 Masood et al11 reported the expression of VEGF, Flk-1, and Flt-1 in a variety of human tumor cell lines of nonendothelial origin, including melanoma and carcinomas of the ovary, pancreas, and prostate. Antisense agents or antibodies directed against these factors led to the inhibition of tumor growth in vitro.11 These data led to the hypothesis that certain types of tumor cells may have the ability to form primary blood vessels by themselves.
Tumor cells induce angiogenesis in a process that mimics in some respects normal angiogenesis. Tumor angiogenesis is also mediated by molecules generated by both tumor cells and surrounding cells such as pericytes and monocytes. Some of these factors, such as VEGF, Ang2, and FGF are involved in normal angiogenesis but are present in the tumor tissue at concentrations that are higher than in the environment in which normal tissue angiogenesis takes place. Others are not established as playing a role in normal angiogenesis, but are important in tumor angiogenesis, such as interleukin-8 (IL-8).12 The up-regulated expression level of IL-8 in tumor tissues stimulates the production of matrix metalloproteinase 2 (MMP-2), which degrades the basement membranes, remodels the extracellular matrix, and facilitates tumor angiogenesis.13 The presence of circulating CD34+ endothelial-cell precursors and monocytes in areas of angiogenesis in wounds and tumors suggest the hypothesis that these circulating hematopoietic precursors may partly contribute to the regulation of angiogenesis.14,15
Tumor angiogenesis is a result of a shift in the balance between angiogenesis stimulators and angiogenesis inhibitors.16 Several endogenous inhibitors of angiogenesis have been well characterized, such as interferons, IL-4, tissue inhibitors of metaloproteinases, and the proteolytic fragments angiostatin and endostatin.17 The identification of the endogenous angiogenesis inhibitors and tumor angiogenesis–specific molecules may contribute greatly to the development of new approaches for cancer treatment.
Specific markers of tumor angiogenesis
Although tumor cells induce the formation of blood vessels in a process that resembles normal angiogenesis, the ultrastructure of tumor vessels is abnormal. The neovasculature is dilated and convoluted, and much of this vascular structure lacks functional pericytes.18 Moreover, the walls of tumor vessels are thin and they may be made up of both endothelial cells and tumor cells.19 These vessels are also extremely permeable in part due to the presence of fenestrae, transcellular holes, and the lack of a complete basement membrane.20 The tumor is nutrient starved, acidic, and under oxidative stress. The endothelial cells of tumor vessels have been shown to respond transcriptionally to all of these stimuli, which induce an increase in the levels of specific markers of angiogenesis. These markers can be used to distinguish the endothelial cells of tumor vessels from the endothelial cells of normal vessels.21
Many technologies have been applied to identify specific markers in tumor vascular endothelial cells or to identify ligands that specifically bind to tumor endothelial cells. Oligonucleotide and cDNA microarray technologies are currently being used to characterize genes that are up-regulated or down-regulated in the normal endothelium versus the tumor endothelium.22
However, these technologies cannot provide functional information about the candidate genes. Therefore, functional genomics or proteomic studies are used to describe how proteins encoded by unknown genes are associated with angiogenesis. Another recent method used to identify endothelial-specific genes is gene-trap expression screening. The p53-binding protein ASPP1 was found to be endothelial specific by this method.23 In order to identify peptides or ligands that might specifically bind to the endothelial cells, the in vivo phage display system has been used.24 Koivunen et al25 discovered a cyclic peptide His-Try-Gly-Phe (HWGF), which specifically targets angiogenic blood vessels in vivo by phage display. This peptide was shown to selectively inhibit MMP-2 and MMP-9 metalloproteinases.
There are many phenotypic differences between mature, quiescent vessels and the newly formed, immature tumor vessels. The up-regulated molecules on tumor endothelial cells include growth factor receptors, such as the tyrosinase receptors Flt-1 and Flk-1 for VEGF, the Eph receptors, and the Tie2 receptor;26 extracellular matrix (ECM)–binding integrin receptors, such as αvβ3 and αvβ5;27 adhesion molecules such as E-selectin and endoglin; and other molecules such as annexin A1.28 The expression level of tissue factor (TF) is also up-regulated in the tumor vascular endothelial cells.29 Moreover, the specific markers of tumor angiogenesis are not limited to the endothelial cells of tumor tissue. The supporting pericytes and smooth muscle cells in the vessel wall as well as the ECM also carry distinct markers, such as the glycoprotein “prostate-specific” membrane antigen, the NG2 proteoglycan,30 the ED-B domain of fibronectin, and various proteases.31 In some cases, molecules that are not expressed in endothelium, such as Her-2-Neu, might also be useful for vascular targeting.32
Targeting tumor vasculature is a potential approach for cancer treatment
Tumor growth is highly dependent on angiogenesis. Weidner et al found that the greater the degree of angiogenesis detected in a primary tumor, the worse the prognosis, suggesting that there is a direct relationship between metastasis and angiogenesis.33 Indeed, development of veins from tumor angiogenesis as well as tumor-derived lymphatics provide channels through which cancer cells can metastasize. Therefore, antiangiogenesis treatment could represent an effective therapeutic strategy with which to suppress both primary tumor growth and tumor metastasis. Tumor growth can be inhibited by blocking tumor-derived angiogenic signals, or by directly targeting the tumor vascular endothelial cells.
Since the vasculature of tumor tissue is different from normal vasculature, it is possible to develop therapeutic agents that specifically target tumor vasculature. There are also other features that make tumor blood vessels an attractive target compared with tumor cells. (1) A single vessel supports the survival of many tumor cells by providing of oxygen and nutrients, as well as providing a main route for metastatic spread. Thus, destroying one vascular endothelial cell in vessels of tumor tissue may result in the death of many more tumor cells. (2) Endothelial cells in the tumor vasculature have a lower mutation rate compared with tumor cells, which means that endothelial cells will be unlikely to acquire resistance to the therapeutic drugs. (3) Angiogenesis is infrequent in the adult, which means that therapies which target angiogenic endothelial cells in tumor vasculature may not damage normal endothelial cells and therefore may have minimal toxicities. (4) Tumor vascular targeting therapy as well as antiangiogenesis therapy could control tumor growth independently of tumor-cell type. In other words, one vascular targeting agent may be effective in many tumor types. (5) The endothelial cells are easily accessible by intravenously administrated therapeutic agents, thus circumventing the formidable problem of transcytosis across the vascular wall and diffusion throughout tumor tissue.
Many antiangiogenesis agents that target one or more events of the angiogenesis cascade have been developed for preclinical or clinical testing. These include endogenous inhibitors for angiogenesis such as angiostatin, endostatin, and platelet factor-434 ; agents neutralizing angiogenic peptides or their receptors, such as antibodies to FGF, VEGF, and soluble receptor for VEGF; chimeric recombinant biologics35 ; and extracellular matrix MMP inhibitors. Tumor vascular targeting agents, including antibodies directed to intracellular proteins as well as to intercellular adhesion molecules such as anti–integrin αvβ3 and αvβ5 antibodies and specific inhibitors of endothelial-cell growth such as IL-12 and TNP-470 have been studied clinically.36
Using viral vectors for tumor vascular targeting therapy of cancer
The use of viral vectors for delivery of antiangiogenesis genes as well as cytotoxic agents represents an effective strategy for anticancer treatment because: (1) most gene therapy vectors can be easily manufactured and purified to high titers; (2) they are stable and can be stored long periods of time, thus permitting the production of thousands of single-use therapeutic vials; (3) they can be engineered to bind to markers on tumor vascular cells; (4) vectors can be engineered to selectively express genes in tumor vascular endothelial cells for long periods of time; and (5) therapeutic viral vectors can be engineered to directly destroy tumor vascular endothelial cells once they bind.
Lin et al37 used an adenoviral vector to deliver a recombinant soluble Tie2 receptor. A single intravenous injection of this construct resulted in a high circulating level of the soluble receptor protein, and the growth of 2 different primary tumors was inhibited significantly.37 A variety of viral vectors are being studied for possible use in tumor vascular targeting therapy, including adenoviral vectors, adeno-associated viral vectors, retroviral vectors, lentiviral vectors, measles virus, and herpes simplex viral vectors.
To minimize side effects, viral vectors need to be modified so that they can target markers that are uniquely presented on the luminal membrane of tumor vascular endothelial cells. Engineering vectors to target the tumor endothelial cells has several advantages: (1) circumvention of neutralizing antibodies present against the vectors in some human subjects; (2) reduction of toxicity to normal tissue; and (3) targeting of the neighboring tumor cells, which are often difficult to contact through the intravascular route. Several strategies have been used to construct targeted viral vectors: (1) targeting the vector at the transcriptional level such that vector replication and transgene expression occurs only in the target luminal endothelial cells of tumor vasculature or in tumor cells; and (2) targeting the vector at the level of cellular binding and infection (transduction), which allows the vector to bind selectively to target cells by modifying the viral attachment protein. The transductional-targeted viral vector can be obtained by pseudotyping, in which one species of virus is made to incorporate the envelope protein of another virus38 by bifunctional antibodies that block the vector's native tropism and redirect the vector to other cellular receptors, or by genetic modification of the vector coat protein to eliminate native receptor interactions and redirect the vector to a novel receptor.
The adenovirus is a double-stranded DNA virus. There are 42 serotypes of adenovirus known to infect humans and most of the adenoviral vectors used for gene therapy are based on serotypes 2 and 5. The first step for infection of cells by Ad5 is the binding of the head domain (knob) of the viral fiber protein to a specific cell-surface receptor called the coxsackievirus-adenovirus receptor (CAR).39 Next, the penton base of the virus interacts with the cellular αvβ3 and αvβ5 integrin receptors, which leads to virus internalization and transport of the viral DNA into the cell nucleus.40
The adenoviral vector is a popular vector for gene therapy because of its high gene transfer efficiency and high levels of transgene expression. However, the transient expression of the transgene, the presence of neutralizing antibodies in human subjects, the inflammatory and immune response elicited by the vector, and its receptor-independent uptake by the reticuloendothelial system (RES) and the liver are its main drawbacks. Song et al41 constructed the transcriptionally targeted adenoviral vector Ad-hVEGFR2-iCaspase-9, in which inducible caspase (iCaspase-9) was under the regulation of the VEGFR2 promoter. The human VEGFR2 promoter induced reporter gene expression primarily in proliferating human dermal microvascular endothelial cells (HDMECs). The activation of the iCaspase-9 mediates apoptosis of neovascular endothelial cells, and overcomes the prosurvival effect of vascular endothelial growth factor or basic fibroblast growth factor. They observed that local delivery of the Ad-hVEGFR2-iCaspase-9 vector followed by intraperitoneal injection of a cell-permeable dimeric drug AP20187 resulted in endothelial-cell apoptosis and local ablation of microvessels in the severe combined immunodeficient (SCID) mouse model of human angiogenesis.41 Other endothelial specific promoters have been used in adenoviral vector, including that of the VEGF receptor Flt-142 and the murine preproendothelin-1 (PPE-1) promoter.43,44
Replacing the Ad5 fiber with that of Ad16, a subgroup B adenovirus whose receptor is CD46, increases transduction efficiency of vascular cells in culture and in intact human saphenous vein segments.45 Bispecific antibodies have also been used to target adenoviral vector to alternative receptors on endothelium such as E-selectin or angiotensin-converting enzyme (ACE).46-48 Reynolds et al49 established an adenoviral system for achieving cell-specific transgene expression in pulmonary endothelium. They used an endothelial-specific promoter Flt-1 to achieve transcriptional targeting and bispecific antibody targeting to ACE to achieve transductional targeting. This dual-targeting vector improved the specificity of transgene expression in lung endothelium.49
Recently, the Curiel group50 has re-engineered the adenoviral fiber protein so that it can accommodate single-chain antibodies (scFv) without destabilizing the packaging of the vector. They have engineered a framework for ScFv that is stable in the cytosol. Moreover, they have developed a variation of the yeast 2 hybrid screening systemic which allows the selection of complementarity determining regions (CDRs) of the heavy and light chains of the scFv, which are thermodynamically stable and fold correctly in the cytosol. These technological breakthroughs have enabled the Curiel group to target adenoviral vectors to the vascular endothelial cells with scFv on the fiber.
One of the challenges to using the adenoviral vector is to reduce the level of uptake by the RES so that the engineered vector can reach its new target. Strategies for temporarily blocking the function of reticuloendothelial cells and hepatic uptake during intravascular administration of vectors have been reported already by the laboratories of Seymour51,52 and Gerritsen,53 respectively. These include: bisphosphonates and Gadolinium to inhibit the function of the reticulendothelial cells, reactive polymers to cover the charges of the adenoviral vector,51,52 and mutational change of the adenoviral vector fiber protein and the penton base proteins, to decrease uptake into the hepatocytes.53
Adeno-associated viral vectors
The adeno-associated virus (AAV) is a single-strand DNA virus and is replication defective. In the absence of a helper virus like the adenovirus, the wild-type AAV genome can integrate into the host-cell chromosomal DNA to maintain a latent state. The preferred integration site is located on the short arm of chromosome 19. Integration of AAV only occurs in the presence of the viral replication proteins (Rep).54 In the presence of helper virus, the AAV viral genome replicates episomally, followed by viral protein synthesis. The intact AAV vectors are able to integrate into the host-cell genomic DNA. In the case of the recombinant vectors used in gene therapy studies, the AAV vector is predominantly episomal. Although the vector in not integrated, it can be expressed from the episomal unit, and in nondividing cells, the expression of this episomal DNA can extend over months if not years. AAV generates either a low or undetectable level of immune response from the human host. Features that limit the use of AAV are the small capacity (5 kb) and the difficulties encountered in production of the vector for clinical trials.
There are at least 8 serotypes of AAV. Each AAV serotype interacts with a specific cellular receptor and has a distinct tissue tropism. Among them, AAV serotype 2 (AAV-2) is the most extensively studied serotype. Cell-surface heparan sulfate proteoglycan (HSPG) serves as the primary cellular attachment receptor for AAV2. The αvβ5 integrin receptor, the hepatocyte growth factor receptor c-Met, and the fibroblast growth factor receptor 1 (FGFR1) have been proposed as coreceptors.55 Platelet-derived growth factor receptor (PDGFR)–α and PDGFR-β were receptors for AAV serotype 5.56
The AAV-mediated delivery of antiangiogenesis factors, such as a truncated form of the VEGF receptor Flk-1, or antisense mRNA against VEGF, can be used as an antitumor strategy.57,58 However, AAV-2 has a relatively poor tropism for endothelial cells, which makes directly targeting the vector to endothelial cells difficult. This poor transductional efficiency is partly due to the sequestration of AAV-2 with the extracellular matrix around endothelial cells, which prevents cell binding and entry, and the degradation of the internalized AAV-2 particles in the proteasome.59 Recent studies have shown that AAV1 or AAV5 serotypes can transduct endothelial cells efficiently, and that sialic acid residues are required for rAAV1 transduction in endothelial cells.60
Emphasis has been placed on improving the transductional efficiency of the AAV vector to endothelial cells. AAV tropism can be genetically engineered by the incorporation of endothelial-cell-targeted peptides discovered by phage display into the capsids of the vector. This modification increases the efficiency and specificity of the vector to infect endothelial cells both in vitro and in vivo.61 Nicklin et al59 engineered AAV2 vector by incorporating a heptamer peptide SIGYPLP into position I-587 of the AAV2 capsid. The transductional efficiency for endothelial cells was increased by such a modification. Muller et al62 developed an AAV-2 peptide library displaying random peptides within the AAV-2 capside and isolated such modified AAV vectors selective for endothelial cells in vitro.
The great virtue of the AAV is the absence of a vigorous host immune response against the vector-infected cells. This results in the potential for prolonged periods of transgene expression in the target cells following infection. Another potential advantage of the AAV vectors for therapy is that transgenes can be expressed for years in nondividing infected cells. Given the enormous surface area of the tumor vasculature, the AAV could be used to promote the in vivo production of small-molecular-weight biologic therapeutics for chronic treatment of systemic diseases.
The retrovirus buds from the plasma membrane of infected cells and therefore contains a lipid covering within which is found homodimers of linear single-stranded RNA. The infection of cells by retrovirus is mediated by attachment of the viral envelope glycoprotein to the target cell-surface receptors. Once entry into the cytoplasm is completed, viral RNA is reverse-transcribed into DNA and then randomly integrated into the host-cell genome.63
Most of the retroviral vectors are based on the murine leukemia virus. Their cDNA can be incorporated into the DNA of dividing cells with high efficiency and therefore they can provide long-term expression of transgenes because of the integration of the expression cassette of the transgene into the host-cell DNA. However, this process of random integration may also cause leukemia.64 Leukemic transformation has been observed even when the retroviral vectors are replication incompetent both in mouse models64 and in human subjects with severe combined immunodeficiency disease who received transplants of bone marrow transduced with a therapeutic retroviral vector.65 These reports appear to be due in part to integration mutagenesis.65 The long time interval between transplantation with the retrovirally modified autologous bone marrow cells and the emergence of leukemia has suggested to some workers that integration mutagenesis is interacting with other factors. One of these may be that some preparations of retroviral vectors may carry “passenger” noncoding regulatory RNAs which can up-regulate genes involved in cellular proliferation.66
Retroviral vectors have been engineered to bind to specific cell types.67 Retroviral vectors are suitable for targeting endothelial cells in the tumor neovasculature because of their intrinsic selectivity for proliferating cells. Several specific promoters, such as Flt-1, ICAM-2, and KDR have been used in the development of transcriptionally targeted retroviral vectors to tumor endothelial cells.68 A retroviral vector in which the herpes simplex virus thymidine kinase (HSV-TK) gene was driven by a hybrid endothelial-specific PPE-1 long terminal repeat (LTR) has been constructed. Treatment of xenograft tumor models by this vector combined with chemotherapeutic agents resulted in widespread vascular disruption and tumor-cell apoptosis. In this report, they also demonstrated that vascular targeting combined with appropriate chemotherapy is more effective than either therapy alone.69
Liu et al has engineered a retroviral vector by displaying tumor vasculature targeting motifs containing Asn-Gly-Arg (NGR) sequence within the context of Moloney murine leukemia viral (MuLV) envelope (ENV) “escort” proteins. Such modification improved the binding efficiency and transduction of the vector to both human umbilical vein epithelial cells (HUVECs) and KSY1 endothelial cells.70 Immunoglobulin G (IgG)–binding peptides have been inserted into the MuLV ENV protein as well. Then an antihuman VEGF receptor Flk-1/KDR antibody served as a molecular bridge to direct the MuLV vector to the endothelial cells.71
The lentivirus is a subclass of retroviruses. Lentiviral vectors are derived from the human immunodeficiency virus (HIV) by removal of the nonessential regulatory genes and sequences through which homologous recombination could lead to the recombination of therapeutic vector with HIV. Unlike retroviral vectors, lentiviral vectors can integrate their cDNA into both dividing and nondividing cells. Thus they can also transduce terminally differentiated cells such as neurons, macrophages, and hematopoietic stem cells.72 Lentiviral vectors expressing MMP-2, angiostatin, and endostatin have been developed for antiangiogenesis therapy.73,74
De Palma et al75 generated a panel of vesicular stomatitis virus–pseudotyped lentiviral vectors (LVs) engineered for endothelial-cell (EC)–specific expression. They cloned a panel of EC-specific transcription regulatory sequences (Tie1, Tie2, Flk-1, VE-Cad, and ICAM-2) into self-inactivating LVs to drive expression of the marker gene encoding green fluorescent protein (GFP), and compared them with the ubiquitously expressing phosphoglycerate kinase (PGK) and cytomegalovirus (CMV) promoters. They found that vectors containing promoter and enhancer sequences from the Tie2 gene achieved remarkable specificity of expression in ECs in vitro and in vivo.75 Cefai et al76 showed that multiply attenuated, self-inactivating lentiviral vectors transduce both proliferating and growth-arrested HUVECs with high efficacy. Shichinohe et al73 have also developed lentiviral vectors for antiangiogenesis gene therapy.
Measles viral vectors
The measles virus is an enveloped negative strand RNA virus. The Edmonston vaccine strain of measles viruses (MV-Edm) has oncolytic effects against Hodgkin and non-Hodgkin lymphomas.77 There are 2 proteins which project from the measles virus envelope: the fusion protein (F) and the attachment protein (H). Both the H and F proteins are required for infection and subsequent cell-to-cell fusion. The H protein determines the viral binding specificity.78 The MV-Edm infects cells through the CD46 or CDw150 receptor.79,80
The great advantage of the measles virus is that the tumor cells have no defense against the mechanism through which it destroys cells, and that the measles virus can be easily engineered to target tumor vasculature by inserting single-chain antibodies (scFv) into the coat proteins. The major obstacles for clinical use of the measles viral vectors include neutralizing antibodies to the vector in human subjects and difficulties in production for gene therapy trials. The Russell laboratory at the Mayo Clinic has overcome these production problems and may have found a way to circumvent the neutralizing antibodies through engineering new targeting ligands into the virus. They developed an Echistatin-targeted measles virus vector (MV-ERV), which binds the αvβ3 integrin receptor with a high affinity, was displayed on the COOH terminal of the viral attachment H protein.81 They have been experimenting with cellular delivery systems as well. The MV-ERV vector can target to activated endothelial cells in tumor tissue by Echistatin binding to the αVβ3 integrin, and it can also target to tumor cells by the H protein binding to the native CD46 receptor. The MV-ERV vector has potential use in gene therapy for targeting tumor-associated vasculature for the treatment of solid tumors.81
Herpes simplex viral vectors
The herpes simplex virus (HSV) is an enveloped, double-strand DNA virus. The envelope glycoproteins gB and gC of HSV bind to the HSV receptor heparan sulfate molecules or nectin-1.82 Internalization of the virus involves the viral envelope gD protein and the cellular FGF receptor.83 HSV can infect lytically or can establish latency. HSV vector has the capability to establish latency in neurons, a wide host range, and a large capacity for insertion of transgenes. However, the toxicity of the vector limits its use.
Several reports indicated that endothelial cells are susceptible to HSV-1 infection.84 An oncolytic HSV vector has been shown to infect and kill tumor endothelium and thus exert antiangiogenic effects in vitro and in vivo.85 Oncolytic HSV vectors which expressed IL-12 have also been shown to enhance the therapeutic effect of the vector on squamous-cell carcinoma (SCC) through antiangiogenic mechanisms.86 One of the major obstacles to the clinical use of HSV therapeutically is the inflammatory response it induces at the site of initial injection, and at sites distant from the initial infection of neurons at which it is released through axonal transfer.87 It is important to note the deaths that occurred among test mice following treatment with amplicon vectors derived from HSV.88
Conclusion and discussion
Advangtages of vector targeting of tumor vasculature
The development of vasculature is a rate-limiting factor for tumor growth. Endothelial cells of tumor vessels display specific features that can be exploited for the selective delivery of cancer therapeutics. The tumor vasculature has already been targeted by antibodies and small chemicals. The limitation of these approaches is that their effect is stoichiometric. The introduction of a vector into endothelial cells is by contrast exponential or geometric. The effect of the vector can be multiplied many millions-fold due to the release of infective viral particles, in the case of a replication competent vector, and through the release of proteins produced from the transgenes that the vector encode. Compared with other gene transfer methods, viral vectors provide a relatively efficient means of transporting DNA and expression of transgenes in tumor vascular endothelial cells.
Moreover, the application of viral vectors allows the combination of other therapeutic strategies with antiangiogenesis and tumor vascular targeting therapy. Zhang et al89 used a tumor-specific replication-competent adenoviral vector in which the E1A gene was driven by human telomerase reverse transcriptase (hTERT) promoter as an antiangiogenic gene delivery vector. The endogenous angiogenesis inhibitor endostatin delivered by this vector takes the advantage of viral therapy and antiangiogenic gene therapy and offers apotent therapeutic effect.89
Another major advantage of engineering vectors so that they can target tumor vascular endothelial cells is that the vectors can easily access the target cells without having to cross through the vessel wall to reach the target. Finally, the combination of vector-targeted antiangiogenesis therapy with cytotoxic chemotherapy, radiotherapy, or immunotherapy may ultimately offer a selective and potent method on controlling the growth of neoplastic cells over a prolonged time period.
Disadvangtages of vector targeting of tumor vasculature and obstacles to overcome for clinical implementation
Although vectors engineered to bind to and infect tumor vascular endothelial cells and tumor cells specifically work in vitro, only very limited targeting is seen in vivo. One of the major reasons for this is the uptake of viral vectors, especially the adenoviral vector, by the RES. A million-fold decrement in the circulating level of vector particles can be mediated by RES uptake within 30 minutes following intravascular injection.
The RES uptake may be one of the major reasons why the transduction of tumor vascular endothelial cells has been so low to date. Thus, RES uptake is one of the major problems to overcome if viral vectors are to be applied systemically. As outlined by Fisher et al,51 the use of dose fractional and polymer coating for charge neutralization appears to have reduced uptake into the RE and liver cells.51-54 It will be important to study strategies that block receptor independent uptake of these vectors by the RES and by hepatocytes. Moreover, to further evaluate the therapeutic effect of tumor vascular targeting vectors, spontaneous tumors in addition to transplantable tumor models should be used.
Another problem confronting the field of tumor vascular targeting with vectors is the need to demonstrate that a candidate marker of tumor vascular endothelial cells is truly specific for the tumor vascular endothelial cells. The fact that the concentration of proangiogenic proteins in tumor tissues may be much higher than those present in expanding vasculature in healing normal tissue may create the opportunity for truly tumor specific vascular markers. Moreover, the mouse models used for the initial testing of tumor vascular targeting strategies may not be totally predictive of the outcome in clinical trials in human subjects. One of the most important reasons for this is that the vasculature in a 2-month-old test mouse is very different from the state of the vascular endothelial cells in a 50-year-old human subject with cancer. In addition, human subjects may have a vasculature that has been extensively modified throughout life by hypertension, atherosclerosis, or immune complex disease.
It is also worth noting that tumor vascular targeting and antiangiogenesis therapy induce tumor regression. Theoretically, hypoxia caused by such therapy could increase the efficiency of tumor angiogenesis through induction of HIF and VEGF, although there is no indication that antiangiogenic therapy promotes metastasis and progression.90,91
Finally, studies are needed to establish the porosity of the tumor vasculature for vectors of different sizes. The tumor vasculature is a very open porous structure due to the disconnection of the neighboring endothelial cells, which arises from the very high percentage of endothelial cells that are undergoing cellular division in tumor vasculature. Because the distances between neighboring endothelial cells may therefore vastly exceed the dimensions of the vectors, differences in the size of the various candidate vectors may not be important.
In addition, one of the attributes of targeting the tumor vascular endothelial cells with vectors is that extravasation into the extravascular space for transduction of tumor cells no longer is a prerequisite for a decisive therapeutic effect since the destruction of the tumor vascular endothelial cells becomes the major focus of the therapy. The destruction of the tumor cells becomes a secondary consequence of the destruction of the tumor vasculature. This change, from targeting of the extravascular tumor tissue to targeting of the luminal membrane of the tumor vasculature, may be the change that will open the door to success with cancer gene therapy strategies based on intravascular administration of therapeutic vectors.
Prepublished online as Blood First Edition Paper, December 22, 2005; DOI 10.1182/blood-2005-10-4114.