Nanotechnological Strategies for Therapeutic Targeting of Tumor Vasculature

Nanomedicines Targeting Tumor Vasculature

Parameters for Design of Tumor Vascular-targeted Nanoparticles

It becomes evident that tumor vasculature is a highly accessible and convenient target for drug delivery. However, most therapeutic agents, especially chemotherapy, have no intrinsic affinity for tumor vessels, which can cause severe drug resistance and systemic side effects. An increasing number of studies are, therefore, focusing on the design of anti-tumor nanoparticles to target tumor endothelial cells and noncellular blood components.

Generally, nanotechnology-based tumor vascular targeting includes 'passive' and 'active' targeting. The ability of nanoparticles to target tumor vasculatures depends on both biological parameters and particle properties.^[13] Dynamic parameters of tumor vasculatures, including abnormalities of blood flow, hematocrit alterations, permeability enhancement and angiogenic marker expression, can affect the targeting efficiency of circulating nanoparticles.^[8,14,15] In addition to composition, nanoparticles need to be tailored in terms of suitable size, shape and surface modification against the tumor vascular characteristics. Currently, the integration of computational modeling, with in vitro microfluidic testing and in vivo validation, has been employed for the rational design of nanoparticles.^[16]

Vascular targeting requires effective margination followed by strong targeting avidity. Compared with normal vasculatures, tumor blood vessels are characterized by slow blood flow, dilation, high red blood cell (RBC) flux and overexpression of angiogenic markers.^[14,15] Similar to leukocyte and platelet margination, which requires an interaction with RBCs, vessel dilation and blood flow reduction, proper theoretical design can achieve nanoparticle accumulation in close proximity to vessel walls. RBCs are located in the blood vessel's center, forming a 'cell-free layer' near to the vessel walls, which vary in thickness depending on channel size and blood velocity. Nanoparticles can be designed to preferentially repel RBCs and accumulate in the cell-free layer.^[17] Shear stress, representative of blood flow, is commonly evaluated to determine the flow level that nanocarriers must withstand to bind with vessel walls.^[18] The margination dynamics under flow can be controlled by changing particle size, shape and aspect ratio. Surface modifications of nanoparticles mediate their subsequent adhesion to endothelial cells, which is mainly affected by ligand–receptor association.^[14]

Particle size remains one of the most remarkable features for tumor targeting since size dramatically influences the biodistribution, pharmacokinetics and safety of nanomedicine. Generally, circulating nanoparticles smaller than 20–30 nm are inclined to be eliminated by rapid renal clearance, while particles larger than 200 nm are more efficiently sequestered by the reticuloendothelial system (RES) in the liver, spleen, lung, lymph nodes and bone marrow.^[19] Compared with normal vasculatures, tumor blood vessels are highly permeable, with fenestrations ranging between 200 and 1200 nm. Considerable studies, especially research on subcutaneous tumor models, have demonstrated that leaky capillaries and defective lymphatic drainage in tumors promote tumor accumulation of nanoparticles (maximum size: 100–300 nm depending on nanomaterial type) through the enhanced permeability and retention (EPR) effect.^[14] Hence, size should be taken into account for the design of tumor vascular-targeted nanoparticles. However, extensive investigations have shown that certain nanoparticles in tumors do not exhibit the EPR effect and the high permeability of blood vessels varies within a tumor.^[20,21] This mechanism has been called into doubt in most nonsubcutaneous tumor models, larger mammalian preclinical models and humans, hindering its clinical application.^[22]

The vast majority of nanoparticles are spherical. Recent developments in fabricating particulates with diverse shapes have provided new design solutions for tumor vascular targeting. Mathematical models have revealed that the shape of circulating nanocarriers is essential for their margination.^[23–25] Unlike spherical particles, the margination efficiency of nonspherical particulates is significantly affected by aspect ratio. The tumbling effect of nonspherical particles under flow can also facilitate their adhesion to the vascular interface.^[26] Meanwhile, elongated nanoparticles are capable of escaping nonspecific uptake by different types of cells.^[27] Consistent with the simulation and in vitro studies, nonspherical nanomaterials, such as carbon nanotubes, worm-shaped micelles and iron oxide nanoparticles, have shown increased accumulation and retention in tumor interstitium, resulting in enhanced therapeutic or imaging efficiency.^[28–31]

Surface chemistry is another key determinant for nanoparticle performance.^[32] Nanoparticles can be surface modified with tunable charge, hydrophobicity and functionality, which is essential for their stability and targeting effectiveness. Surface charge mainly influences tissue penetration, drug release and cellular internalization of nanoparticles. Particles with a hydrophobic layer tend to be absorbed by RES. Certain polymers, such as PEG, can provide a hydrophilic surface for these nanoparticles to evade RES phagocytosis. Moreover, active targeting can be achieved by surface conjugation with ligands that recognize specific receptors on tumor vessels without the requirement of vessel permeability, particle size and shape. In practice, passive and active targeting are synergistically utilized to improve vascular targeting efficiency.

Nanoparticles Targeting Proangiogenic Factors & Tumor Endothelial Cells

Since tumor angiogenesis is primarily controlled by serial endothelial cell transitional behaviors that are stimulated by proangiogenic factors, the endothelium and multiple proangiogenic factors are considered to be the chief targets for antiangiogenic and anti-tumor therapies. Nanomedicines that activate endogenous angiogenesis inhibitors, repress proangiogenic factors, or directly destruct endothelial cells and tumor cells have been designed for the prevention of angiogenesis, tumor growth and metastasis (Figure 1). Various nanomaterials possessing specific characteristics or carrying diversified antiangiogenic/anti-tumor agents have been investigated (Figure 2). Small chemicals, peptides, proteins, antibodies, genes, siRNAs and miRNAs that suppress different aspects of endothelial cell behaviors or simultaneously eradicate tumor cells can be assembled to in the nanocarriers (Figure 1).

Passive Targeting. Some nanoparticles possess passive target ability and preferentially accumulate in tumor tissues, potentiating the activity, selectivity and stability of current antiangiogenic and anti-tumor strategies. For instance, liposome-mediated systemic administration of the angiostatin gene significantly suppressed and delayed melanoma growth by reducing neovessel formation, which represents an alternative approach to antiangiogenesis gene therapy.^[33] PEG-overlaid liposomes encapsulating glucocorticoids were shown to inhibit tumor angiogenesis, inflammation and tumor cell proliferation through substantial accumulation in tumors via the EPR effect.^[34] Similarly, colloidal gold nanoparticles surface coated with PEG and recombinant human TNF-α were constructed to selectively deliver TNF-α to solid tumors by passive extravasation. The resulting nanomedicine (CYT-6091 [Aurimune™; CytImmune Sciences, Inc., MD, USA]) disrupted tumor endothelial cells, tumor cells and other cells harboring the TNF-α receptor in the tumor interstitium.^[35] A smart dendrigraft poly-L-lysine nanoparticle dual loaded with VEGF shRNA and doxorubicin caused effective blood vessel shutdown and cell apoptosis within tumors, which was activated by low extracellular pH and high MMP-2 expression after its passive targeting (Figure 3).^[36] Interestingly, certain nanomaterials themselves can serve as angiogenic inhibitors. Fullerenic nanoparticles with multiple hydroxyl groups were found to downregulate more than ten proangiogenic factors, which significantly decreased tumor microvessel density and tumor progression.^[37] This 'particulate' form of medicine may be superior to traditional 'molecular' medicines that usually target single or a few angiogenic factors. Perfluorocarbon-based paramagnetic nanoparticles enhanced tumor vascular targeting specificity due to a size-mediated intravascular constraint, which is a prospective entity to monitor drug efficacy on the basis of simultaneous angiogenetic imaging and therapy.^[38] Notably, various viral nanoparticles, including vaccinia virus vectors, adenoviral vectors, adeno-associated viral vectors, retroviral vectors, lentiviral vectors, measles virus and herpes simplex viral vectors, have also been applied for delivery of antiangiogenesis genes as well as cytotoxic agents.^[39] As natural nanocarriers they are easily manufactured, have good biocompatibility and biodegradability, and can be engineered for active targeting, representing an effective strategy for anti-tumor treatment, although some safety concerns remain. These studies illustrate that nanotechnology has the capacity to reduce drug off-target effects, and achieve combination therapy and theranostic medicine through passive targeting.

Active Targeting. Emerging studies have revealed that passive targeting efficiency is limited. Thus, functionalized nanoparticles with active targeting residues have subsequently been developed in order to achieve minimal side effects, optimal pharmacokinetics and maximal efficacy. Induced by tumor-secreted factors, endothelial cells express a series of angiogenic markers that provide opportunities for the design of actively targeted nanoparticles (Table 2). Ligands, such as affinity peptides, proteins, aptamers and antibodies, which specifically recognize and bind to angiogenic markers are chemically arrayed on the surface of nanoparticles. Among them, peptides are widely utilized owing to their easy availability and simple structures, which have been largely screened and identified by in vivo phage display technology from complex combinatorial peptide libraries.^[40]

Integrins, especially α_vβ₃, are key players in promoting endothelial cell migration, survival and growth in a variety of cancer settings and, therefore, they have been widely used as targeting moieties.^[41] The first integrin-targeted nanoparticle, coupled to a small organic integrin α_vβ₃ antagonist, was reported by Hood et al. a decade ago, which induced tumor endothelium apoptosis, and the regression of primary and metastatic tumors by selectively delivering a mutant Raf gene to angiogenic blood vessels.^[42] Inspired by this study, Xie et al. synthesized a neopentyl derivative of an antagonist that had superior binding affinity to integrin α_vβ₃. Lipid nanoparticles conjugated to this derivative showed stronger and more selective accumulation in angiogenic vessels, which provided an optimized platform for targeted delivery of anti-tumor drugs.^[43] Cyclic or linear derivatives of the RGD peptide for integrin targeting were subsequently developed in conjugation with various nanocarriers, including polymers, dendrimer, carbon nanotubes and metal nanoshells, in which chemotherapeutic agents, siRNAs or miRNAs were loaded to be selectively targeted to tumor vasculatures. These nanotherapeutic agents significantly decreased vascular density and tumor growth without eliciting toxicity in a broad range of tumor models (Table 2).^[44–50] Remarkably, an α_vβ₃-targeted nanoparticle encapsulating doxorubicin exhibited antimetastatic activity, again revealing the vital role of angiogenesis in tumor metastasis.^[51] Tumstatin has been widely accepted as a ligand of integrin α_vβ₃. Strikingly, a tumstatin-conjugated iron oxide nanoparticle showed enhanced penetration and selective targeting to endothelial cells, and resulted in significant tumor neovascularization inhibition.^[52]

Nucleolin, another angiogenic marker, is highly expressed on tumor blood vessels relative to normal vasculatures, mediating endothelial cell proliferation and migration.^[53] Both a 31-amino-acid (F3 peptide) and a DNA aptamer (AS1411) that recognize cell surface nucleolin have been developed and widely utilized for the formulation of nucleolin-targeting nanoparticles.^[54–57] Winer et al. showed that cisplatin-loaded polyacrylamide nanoparticles modified by the F3 peptide were primarily bound to tumor endothelial cells in vivo. Tumor regression and vascular necrosis were observed in ovarian cancer models (Figure 4).^[54] Prickett et al. functionalized the single-walled carbon nanotubes with the F3 peptide, which was selectively internalized by actively dividing endothelial cells and caused significant cell death.^[55] In a separate study, PEG–poly(lactide) nanoparticles, loaded with paclitaxel and surface conjugated by F3 peptide, were coadministered with a tumor-homing peptide that recognizes neuropilin to improve drug penetration across angiogenic vasculature into glioma parachyma.^[56] Consistently, PEG–poly(lactide) nanoparticles coupled to AS1411 exhibited prolonged circulation and enhanced tumor accumulation of paclitaxel via tumor vascular targeting, leading to a higher rate of tumor inhibition and animal survival.^[57] Together, these studies demonstrate that nucleolin-targeted nanoparticulate drug delivery systems hold great promise for improving the specificity, efficacy and half-life of current cancer treatments.

The signal pathway of VEGF–VEGFRs is one of the most crucial mediators in tumor angiogenesis. Among all of the VEGF isoforms, VEGF-A exhibits the most powerful proangiogenic effects by interacting with VEGFR-2, which promotes endothelial cell proliferation, migration, lumen formation and vascular permeability. Peptides targeting signaling molecules, especially VEGFR-2, have been identified and applied in nanoparticle conjugation. One peptide (HTMYYHHYQHHL) that binds VEGFR-2 with high affinity and specificity was conjugated to paclitaxel-loaded nanoparticles, contributing to rapid, long-term and accurate tumor vascular inhibition.^[58] Polymeric nanoparticles modified by a cyclopeptide encompassing residues 79–93 of VEGF have shown high binding affinity with VEGFR-2 and low toxicity, and may represent useful vectors for targeting tumor angiogenesis.^[59]

Aminopeptidase-N (CD13), a membrane-bound enzyme, is associated with angiogenesis and has been used for tumor vascular targeting via the NGR/CD13 ligand–receptor system.^[60] One successful example was a liposomal formulation of doxorubicin bearing the NGR peptide at the outer surface, which led to dramatic vascular damage and neuroblastoma eradication.^[61] Dual-targeting liposomes modified with APRPG and GNGRG peptides were further developed by Murase et al. to enhance the potential of endothelial active targeting. These peptides were observed to cooperatively facilitate the association of liposomes to proliferating endothelial cells. Moreover, doxorubicin encapsulated in the dual-targeting liposomes was demonstrated to significantly suppress angiogenesis and tumor growth.^[62]

CD105 is an ideal tumor angiogenic marker since it is almost exclusively expressed on proliferating endothelial cells and not readily detectable in resting endothelial cells or normal organs, making it an attractive tumor vascular target. Liposomes and nanographene oxide conjugated with anti-CD105 antibodies have been utilized for noninvasive tumor imaging.^[63–65] The nanoparticles exhibited excellent stability and improved tumor uptake that was specific for the neovasculature. These studies warrant further investigation, and development of CD105-targeted nanomedicine for the efficient delivery of antiangiogenic and anti-tumor drugs is required.

Monomeric vascular endothelial cadherin (VE-cadherin) is highly expressed on the surface of angiogenic endothelial cells, and its dimerization with another VE-cadherin on an adjacent endothelial cell leads to the formation of intercellular tight junctions. The antibody designated E4G10 was exploited to bind monomeric VE-cadherin rather than the homodimeric form, thus, conferring specificity for targeting angiogenic vessels with poor junctions rather than normal connective endothelium. This antibody was utilized to covalently functionalize single-wall carbon nanotubes, which enabled improved vascular inhibition, tumor volume reduction and blood clearance kinetics.^[66]

Inflammation is an essential hallmark of cancer, which fosters the functions of other hallmarks, including angiogenesis. Many proinflammatory factors in the tumor microenvironment, such as lipopolysaccharide, TNF-α and interleukins, can markedly upregulate the transmembrane glycoprotein E-selectin in endothelial cells. E-selectin on the endothelial cell surface specifically recognizes the tetrasaccharide sialyl-Lewis X, which mediates leukocyte or tumor cell adhesion to blood vessels. A recent study has revealed that tetrasaccharide sialyl-Lewis X functionalized nanoparticles can be selectively internalized by TNF-α-stimulated endothelial cells, thus, offering great potential for tumor vascular targeting, treatment of inflammation and inhibition of tumor metastasis.^[67] Further evidence has been provided by liposomes decorated with thioated oligonucleotide aptamers against E-selectin, which showed efficient uptake by endothelial cells and accumulation at the vasculatures of breast tumor xenografts.^[68] Therefore, insights into the association between inflammatory cells and endothelial cells in the microenvironment will provide more insight into the formulation of tumor vascular-targeted nanoparticles.

Considering the interaction between endothelial cells and the extracellular matrix, MT1-MMP has emerged as a targeting moiety of endothelial cells. MT1-MMP is a key enzyme for the activation of MMP-2, which is closely related to tumor angiogenesis as well as metastasis. Liposomes modified with a peptide (GPLPLR) that targets MT1-MMP showed high binding affinity to endothelial cells in vitro and elevated localization in the tumor sites compared with unmodified liposomes, resulting in improved therapeutic effect.^[69] Peptides derived from other extracellular matrix proteins that recognize certain angiogenic receptors on endothelial cells can also be utilized to generate nanoparticles against tumor vasculatures.

Apart from the aforementioned well-known vascular-targeted moieties, many other angiogenic-homing peptides have also been discovered by phage display technology. For instance, by treating oral cancer bearing mice with a phage-displayed peptide library, two peptides (PIVO-8 and PIVO-24) were identified to selectively recognize certain unidentified receptors on tumor vessels. After conjugation with them, liposomal doxorubicin showed tumor-targeted delivery, enhanced therapeutic efficacy and improved pharmacokinetics.^[70]

Taken together, these studies demonstrate that nanoparticles that actively target tumor vasculatures, especially tumor endothelial cells, offer great potential for improving the selectivity and efficacy of current antiangiogenic and anti-tumor strategies. As tumor blood vessels are characterized by multiple angiogenic markers, further investigation is required to enhance the targeting specificity achieved by combining several ligands on one nanocarrier.

Well-placed Shutdown of Tumor Blood Flow by VDAs

Apart from the antivascular strategies mentioned above, another method of attacking a tumor's blood supply is to preferentially destroy the established tumor vessel network with the use of VDAs.^[71] These VDAs differ from antiangiogenic agents not only in the mechanism of action but also in the therapeutic application. Antiangiogenic therapy is administered continually over months or years, while VDAs are only used in an intermittent manner. Most solid tumors can be affected and occasionally eradicated when their blood supply is shutdown. VDAs have shown anti-tumor effects by selectively clamping off the tumor-feeding blood supply, and subsequent tumor necrosis. Most developed VDAs are small chemicals comprising of tubulin-depolymerizing agents and flavonoids. The leading candidate compounds of tubulin-depolymerizing agents include CA4DP, ZD6126 and AVE8062A, which can rapidly disrupt endothelial tubulin cytoskeletons, induce a coagulation cascade and cause intratumoral vessel occlusion.^[71] By contrast, flavonoids such as 5,6-dimethylxanthenone-4-acetic acid exert their effects on tumor blood vessels via the localized release of TNF-α from activated macrophages within a tumor, leading to vascular occlusion and subsequent necrosis.^[72] More importantly, increasing evidence indicates that VDAs can be combined with conventional chemotherapeutic agents and irradiation therapy to improve tumor response.

Nanotechnology can potentiate the anti-tumor properties of VDAs. Polymer nanocarriers have been utilized to sequentially release CA4DP and paclitaxel, showing apparent effects on both tumor vasculature disruption and tumor cell apoptosis. Therefore, the nanodrug-based strategy is a promising example for combination therapy in a single nanocarrier system.^[73] Furthermore, gold nanoparticles were applied as adjuvants to brachytherapy to allow lower energy sources via utilizing radiation-induced photo/Auger electrons from the gold nanoparticles, serving as a novel tumor VDA.^[74]

Well-placed Thrombotic Occlusion of Tumor Vessels Induced by Coagulation Factor

Although VDAs can be used to effectively damage tumor vessels, this particular treatment was too toxic for systemic application. By contrast, creating procoagulants that specifially and effectively induce coagulation in tumor vasculatures exhibits great advantages. Through such a strategy, coagulation can be induced directly, rather than subsequent to endothelial cell damage. For example, tissue factor (TF), a member of the cytokine receptor family group 2, is the initiator of the extrinsic blood coagulation cascade. Normally the truncated form of TF (tTF), which is missing a transmembrane domain, is powerless to trigger thrombus formation.^[75] However, its coagulation induction activity can be recovered by localizing it to phospholipid membranes, such as the membrane of the tumor vessel wall. The prothrombotic state of tumor tissues also supports the possibility that tumor blood coagulation can be rapidly induced by low levels of procoagulants, which cannot cause coagulation in normal vessels.^[76]

As mentioned above, studies based on phage-displayed peptide libraries in vivo and ex vivo have screened a series of tumor-homing peptides that can specifically recognize the markers on tumor vessels. According to these advances, novel approaches for shutting down blood supply by targeting tTF have been proposed. When tTF was conjugated with a bispecific antibody targeting an artificially induced endothelial cell antigen, the MHC class II antigen I-A^d, it initiated tumor blood coagulation and showed dramatic anti-tumor effects.^[77] Subsequent studies have demonstrated that conjugates of tTF with VCAM-1 antibody, NGR motif or ligand-targeting prostate-specific membrane antigen can well home to tumor vessels and cause rapid intraluminal blood coagulation, resulting in tumor necrosis in a variety of tumor models.^[78–80] In particular, clinical first-in-man application of low dosages of NGR–tTF fusion complexes revealed good tolerability and tumor inhibition.^[80]

Nanoparticles offer an opportunity to alter the pharmacokinetic profile of drugs, reduce off-target adverse effects and improve the therapeutic index. Therefore, if tTF could be brought into contact with tumor endothelium by engineered nanocarriers, the resulting formulation would have the potential to add a powerful punch in the fight against cancer. On the other hand, using nanoscale carriers to encapsulate and site-specifically deliver tTF raises the possibility of overcoming various barriers faced by antibody or peptide–tTF complexes, such as short circulation half-life in cancer patients, poor tumor tissue targeting or molecular defects involved in thrombogenic cascades initiated by the TF (e.g., deficiencies in factors V, VII, VIII or X).

To date, however, no literature for the introduction of nanocarriers to target procoagulants to tumor-associated vessels is available. Only one study has demonstrated that selectively blocking the blood flow of tumor vessels could be achieved to some extent via physical accumulation effects of nanoparticles in tumor tissues.^[81] How we can construct 'smart' nanocarriers to specifically deliver TFs to tumor sites should be considered as a major challenge in the application of TF for cancer therapy. Furthermore, using nanocarriers to package high-procoagulant factors, such as thrombin, collagen and ADP, may establish a realistic new approach for cutting off the blood supply of a broad spectrum of tumors.

Nanomedicine. 2013;8(7):1209-1222. © 2013 Future Medicine Ltd.