Optimizing the Delivery of Cancer Drugs That Block Angiogenesis

See allHide authors and affiliations

Science Translational Medicine  20 Jan 2010:
Vol. 2, Issue 15, pp. 15ps3
DOI: 10.1126/scitranslmed.3000399


Drugs that block angiogenesis are important components of first-line therapies for a number of human cancers. However, some of these agents have undesirable effects on the patient. Optimal delivery systems must be developed to maximize clinical benefits and minimize adverse effects in cancer patients. In this Perspective, we discuss these drug-related issues and propose ways to optimize antiangiogenic therapy by the development of new drug delivery systems.

When tumors reach a critical mass, they require a blood supply to deliver nutrients needed for their further growth. The process of generating new blood vessels from existing ones—angiogenesis—is a crucial step in the conversion of a dormant tumor into a fast-growing or metastatic one (1). Drugs that block angiogenesis have become important components of the first-line therapies for a number of human cancers, such as colorectal, renal, and metastatic breast cancers. Recent preclinical and clinical experiences with some of these drugs have raised several issues concerning the potential undesirable host, or off-therapy, effects related to short-term antiangiogenic therapy. These off-therapy–induced host responses include (i) rapid-rebound revascularization of tumors after withdrawal of the antiangiogenic drug, and (ii) compensatory enhancement in the expression of alternative angiogenic factors that are not targets of the antiangiogenic drug at the core of a patient’s therapy regimen. These host changes can render tumors refractory to the anticancer agent in use or can induce evasive escape of the tumor from the original drugs (2, 3). These findings suggest that optimal delivery systems must be developed to maximize clinical benefits to and minimize adverse effects on the host. In this Perspective, we discuss these drug-related issues and propose ways to optimize antiangiogenic therapy by the development of new drug delivery systems.


In 1971, surgeon and scientist Judah Folkman predicted that the inhibition of tumor blood vessel development would be a new therapeutic approach for treating a wide variety of cancers (4). The first angiogenesis inhibitor was isolated just a few years later with the use of a novel slow-release drug delivery system developed with ethylene–vinyl acetate copolymers (5). However, it was not until nearly three decades later that the U.S. Food and Drug Administration approved the first specific antiangiogenic drug, bevacizumab, on the basis of clinical benefit measured in metastatic colorectal cancer patients (6). Bevacizumab is a humanized antibody to vascular endothelial growth factor (VEGF) that neutralizes VEGF’s ability to stimulate the growth of new blood vessels. Direct clinical evidence has shown that a single intravenous infusion of bevacizumab results in decreases in tumor perfusion, microvessel density, and interstitial fluid pressure (7). Today, bevacizumab and several other antiangiogenic drugs, including sunitinib and sorafenib, are key components of cancer treatment regimens (8).


Despite these successes, in a number of cases the clinical benefits of antiangiogenic therapy have been modest, either because of troublesome side effects or the development of drug resistance through a variety of mechanisms (2). For example, several recent preclinical and clinical studies have shown that circulating amounts of VEGF and placental growth factor (PlGF) are significantly increased in animals or patients receiving anti-VEGF drugs (including bevacizumab and sunitinib), relative to the amounts seen in controls (9, 10). In contrast, endogenous VEGF inhibitors, such as soluble VEGF receptor–2, are down-regulated in patients on antiangiogenesis drugs (11), relative to healthy people. In addition, these findings suggest that the withdrawal of such drugs from cancer patients may cause a rebound effect, in which elevated amounts of growth factors that persist in circulation after drug withdrawal stimulate angiogenesis and thus facilitate tumor growth and invasion (3). In support of this notion, discontinuation of anti-VEGF therapy has been shown to result in rapid re-neovascularization in tumors (3) (Fig. 1).

Fig. 1. Antiangiogenic cancer therapy.

Anti-VEGF drugs, including bevacizumab, sunitinib, and sorafenib, induce vascular regression and normalization in tumors, and the withdrawal of these drugs could lead to rapid re-neovascularization. Short-term delivery of antiangiogenic drugs also leads to increased tumor invasiveness and metastasis as a result of tissue hypoxia.


Another side effect of large amounts of circulating VEGF may be the induction of cancer-associated systemic syndromes that manifest as cachexia (fatigue, loss of appetite and weight, and muscle weakness and atrophy) or some other paraneoplastic disorder (a condition caused by the systemic rather than the local presence of cancer cells) that impairs the functions of multiple organs and jeopardizes both the quality of life and the survival of the patient (12). One possible way to prevent rebound angiogenesis and its accompanying effects after cessation of anti-VEGF therapy is to develop a system that allows persistent delivery of anti-VEGF drugs, such as controlled-release microspheres (13).

Although the molecular mechanisms that underlie rebound angiogenesis remain elusive, anti-VEGF–induced tumor hypoxia is known to play a crucial role in the up-regulation of VEGF and several other angiogenic factors in the host. Two recent studies showed that short-term treatment of mice with VEGF inhibitors augmented tumor cell invasion and metastasis after drug withdrawal, possibly via treatment-induced hypoxia (14, 15). Antiangiogenesis-triggered hypoxia in normal tissues also is responsible in part for the compensatory increase in the expression of multiple angiogenic growth factors (Fig. 2) (16, 17), and these hypoxia-induced changes contribute to the development of antiangiogenic drug resistance in tumors (2, 16, 18). The intimate interplay among these angiogenic factors can lead to synergistic angiogenic responses, which means that the expression levels of tumor-derived angiogenic factors may not be reliable surrogate markers for drug resistance. For example, small amounts of fibroblast growth factor–2 (FGF-2) and platelet-derived growth factor–B (PDGF-B) can initiate a robust angiogenic response that facilitates tumor invasion and metastasis (19, 20).

Antiangiogenic drug–induced tumor hypoxia is not the only factor that causes compensatory escape from or evasive resistance to (drug-induced resistance caused by switching to alternative angiogenic factors that are not targets of the original antiangiogenic drug) antiangiogenic therapy. Indeed, treatment of tumor-free healthy animals with sunitinib results in elevated amounts of multiple circulating angiogenic factors, including VEGF, PlGF, granulocyte colony-stimulating factor, and stroma-derived factor–1 (21).

Fig. 2. Antiangiogenic drug–induced host responses.

Antiangiogenic drugs induce hypoxia in both tumors and healthy tissues, and hypoxia switches on the expression of VEGF and other non-VEGF angiogenic factors, including FGFs, PDGFs, angiopoietin (Ang), and PlGF, all of which induce blood vessel growth and tumor metastasis. HIF, hypoxia-inducible factor.



Given the fact that most antiangiogenic drugs are systemically delivered to cancer patients, it is plausible that the current delivery approaches for antiangiogenic drugs result in global increases in angiogenic factors by targeting healthy tissues. The development of microchip delivery systems (22, 23) that can simultaneously or sequentially deliver cocktails of angiogenesis inhibitors that block various signaling pathways could improve therapeutic efficacy and reduce resistance to antiangiogenic drugs (Fig. 3). Such a system, sometimes called a “pharmacy on a chip” (22, 23), has been shown to house, and deliver on demand, many different drugs in an experimental setting. These systems can potentially be preprogrammed or externally regulated to release drugs at a given time and rate and in a particular pattern. Such a system might also one day enable novel combination therapies—for example, first attacking tumor cells with angiogenesis inhibitors, then destroying the remaining tumor cells with chemotherapeutic drugs, and finally maintaining the patient on long-term antiangiogenic therapy (24). Such systems have already been used in canine models to deliver tumor agents—such as luteinizing hormone–releasing hormone analogs, which are used in the treatment of prostate cancer—on demand for more than 6 months by means of radiofrequency control (25). It is also possible to obtain pulsatile delivery of proteins (such as angiogenesis inhibitors) from polymer systems, including microspheres, using an appropriate polymer matrix design (26).

These types of controlled-release systems may be particularly useful to resolve the dose-related issues of antiangiogenic agents. It is known that several endogenous angiogenesis inhibitors, such as interferon-a, endostatin, and a peroxisome proliferator–activated receptor-γ ligand, produce biphasic dose efficacies: a U-shaped dose efficacy curve in which the drugs' effects are optimal between very low and very high doses (8). Thus, biphasic delivery of these antiangiogenic agents may potentially be achieved using the novel microspheres (26) or smart microchip systems (22, 23) discussed above to release optimal doses of individual drugs based on their maximal antiangiogenic activity. Such biphasic delivery systems have already been developed (26, 27).


Several angiogenic factors such as VEGF and FGF-2 activate survival signals in endothelial cells, which form the inner layer of blood vessels and thus play a central role in angiogenesis (19, 28). Inhibition of these angiogenic factors would in theory result in the regression, or chewing back, of angiogenic factor–dependent tumor blood vessels. Indeed, agents that specifically block VEGF as well as those that target multiple angiogenic factors cause marked enhancement of vascular regression in both preclinical and clinical settings (3, 29, 30). Furthermore, preclinical and clinical studies show that some antiangiogenic drugs exhibit vascular normalization activity in tumors, and this normalizing of the abnormal tumor vasculature allows efficient oxygen and drug delivery to the tumor. Therefore, when a well-devised multidrug delivery regimen is employed, the alleviation of drug-induced hypoxia can boost the effectiveness of anticancer therapies (31, 32).

Fig. 3. Antiangiogenic drug delivery regimens for cancer therapy.

(A) Advantages of antiangiogenic drug delivery include the achievement of long-term therapy, overcoming of drug resistance, optimization of drug dosages, controlled drug release, simultaneous release of drug cocktails, reduction of adverse effects, sequential drug release, and targeted drug release. (B) Development of microchip controlled–release systems. A microchip that houses a variety of drugs can potentially be introduced into patients. This microchip would simultaneously or sequentially release the same or different drugs based on a specific computer program devised from patient-specific clinical information. The clinical protocol would theoretically be dictated by molecular, cellular, and physiological biomarkers. Anti-Ang1, 2, antibody to angiopoietin 1, angiopoietin 2, respectively; Anti-PlGF, antibody to PIGF; anti-PDGF, antibody to PDGF; anti-VEGF, antibody to VEGF; Anti-FGF, antibody to FGF. Chemo 2 and Chemo 3, two different cancer chemotherapeutic drugs.


Therapeutic efficacy can be improved by targeting the multiple pathways that induce angiogenesis in tumors and normal tissues through the simultaneous delivery of a cocktail that contains several angiogenesis inhibitors. This organizing principle has successfully been used in a variety of preclinical models (8). For example, simultaneous delivery of three angiogenesis inhibitors that target the integrin, E-cadherin, and VEGF angiogenesis pathways almost completely inhibited pathological angiogenesis in malignant and nonmalignant tissues (33). The use of multiple microspheres, controlled-release microchips, or layer-by-layer polymer systems (34) may be useful in achieving intricately choreographed multidrug delivery.

Taken together, these and other preclinical and clinical findings suggest that long-term, targeted, and optimized antiangiogenic drug delivery is desirable for cancer therapy. Other critical issues for the improvement of antiangiogenic drug delivery are discussed elsewhere (2, 10, 35, 36).


Current antiangiogenic drugs are administrated to cancer patients systemically, and thus they nonspecifically target blood vessels in all tissues and organs. Considering the tiny tumor burden relative to body weight, most drug molecules are distributed to nonmalignant tissues upon systemic delivery. There are clinical indications that antiangiogenic drugs (for example, bevacizumab) can cause adverse effects such as hypertension and proteinuria (37). Pharmacological data show that antiangiogenic drugs such as sunitinib cause dose-dependent adverse effects in animals, including bone marrow depletion, pancreatic toxicity, and adrenal hemorrhage (38). These findings suggest that targeted delivery of antiangiogenic drugs to the tumor vasculature may help to avoid these harmful side effects. In addition, clinical data show that high doses of antiangiogenic drugs may not necessarily augment the clinical benefits (38). Furthermore, different types of tumors should probably be treated with different doses of the same drug. Thus, optimization of dosages and delivery systems is vital if we are to achieve improvements in current clinical practice with antiangiogenic drugs.

An approach to making the effects of such drugs more specific is to target the drugs to the vasculature (29). One method might be to couple low–molecular-weight antiangiogenic drugs to polymers, which alters the biodistribution of the drug after systemic intravenous administration. This approach is based on the concept that uncoupled low–molecular-weight drugs will penetrate most tissues, because they pass rapidly through cell membranes. Therefore, the drug is distributed quickly throughout the body with no selectivity toward the tumor. In the case of polymer-bound drugs, the polymer-to-drug linkages can be designed so that they are stable in the bloodstream, permitting the polymer-associated drug to persist in the circulation for a longer period of time than the unbound drug. This extended circulation time occurs because the higher–molecular-weight polymer/drug system can gain entry into the cells only through endocytosis. Because most normal tissues have nonleaky microvasculatures, the polymer/drug complex accumulates to a greater extent in tumor tissue, which has a leaky vasculature.

An example of this approach involves TNP-470, a low–molecular-weight synthetic analog of fumagillin, a compound secreted by the fungus Aspergillus fumigatus Fresenius (39). TNP-470 is a potent inhibitor of endothelial cell functions in vitro, and, in animal models, TNP-470 has the broadest anticancer spectrum of any known agent. However, when it was used clinically, many patients experienced neurotoxicity at the same drug doses that produced antitumor activity. Thus, any modifications of TNP-470 that retain or increase its antiangiogenic activity while reducing its toxicity are desirable. Satchi-Fainaro and Folkman first suggested that angiogenesis inhibitors be coupled with water-soluble synthetic polymers [such as hydroxyl propyl methyl acrylamide (HPMA) copolymers] that can be internalized in tumor vessels (40). HPMA copolymers are biocompatible, nonimmunogenic, and nontoxic, and their body distribution is well characterized. These researchers hypothesized that conjugation with polymers would increase the accumulation of TNP-470 in angiogenic tissues, such as the tumor vascular bed, and restrict its passage through the normal blood/brain barrier. They also suggested that conjugation of angiogenesis inhibitors to such carriers would prolong their circulating half-life. Finally, Satchi-Fainaro and Folkman showed that the TNP-470–HPMA copolymer conjugate is effective in controlling tumor growth in several animal tumor models but without neurotoxicity (40).

Another approach to the targeted delivery of drugs involves using liposomes as drug carriers. Liposome-based drug delivery has been substantially improved by manipulating the liposomal physicochemical properties in ways that are responsive to specific physiological features of a tumor, with the goal of optimizing drug accumulation in tumor vessels (41). The increased permeability and complex three-dimensional architecture of the tumor vasculature have also been exploited to deliver tumor-suppressing drugs with increased efficiency (42).


In clinical practice, the therapeutic benefits of antiangiogenic drugs in cancer patients have been realized when these drugs are used in combination with conventional chemotherapeutic agents, and antiangiogenic monotherapy rarely produces favorable effects (6, 8). The molecular mechanisms underlying the clinical benefits seen with combination therapy are unknown. However, it has been suggested that antiangiogenic drugs function at least in part by normalizing disorganized tumor blood vessels and thus improving blood perfusion (31, 43). Normalization of the tumor vasculature by antiangiogenic agents leads to the reversal of tissue hypoxia and the repair of vascular leakage, which reduces interstitial fluid pressure. When chemotherapeutic drugs are simultaneously administered, antiangiogenic drug–induced vascular normalization increases cytostatic drug delivery. Vascular normalization is an attractive mechanism to explain why combination therapy, but not antiangiogenic monotherapy, is clinically beneficial. In certain types of experimental tumor models, such as glioblastoma, anti-VEGF drug-induced vascular normalization persists within a limited period of time. Reversal of this normalization probably occurs as a result of the compensatory switching-on of other proangiogenic factors, leading to drug resistance or escape (32). These findings suggest that the optimal timing of antiangiogenic treatment is within the vascular normalization window and before the administration of cytostatic drugs, so that cytotoxic drug release occurs within the normalization window. This type of sequential drug delivery may be achieved by using a controlled-release microchip (22, 23) or a layer-by-layer polymer/drug system (34).

Folkman’s laboratory also reported that tumors that are resistant to the DNA-alkylating drug cyclophosphamide regain sensitivity to this same drug if it is delivered at frequent intervals and at a low dose (44). This finding helped to define a novel chemotherapeutic mechanism that involves the targeting of the endothelial cell compartment by alterations in dosage and drug delivery scheduling. Dose and delivery variation with antiangiogenic drugs was validated later by others, who christened this approach “metronomic” chemotherapy (45, 46). Early clinical studies demonstrated that chemotherapy-refractory cancer patients show improved clinical benefits when they are shifted to metronomic chemotherapy (47). It is intriguing that combinations of vasculature-targeted antiangiogenic drugs and metronomic cytotoxic chemotherapy yield particularly promising clinical benefits with significantly reduced toxicity (35). Achieving the improved therapeutic benefits of metronomic chemotherapy in combination with targeted antiangiogenic drugs requires further development of long-acting controlled-release systems that allow simultaneous or sequential delivery of two or more anticancer drugs (22, 23, 34).


  • Citation: Y. Cao and R. Langer, Optimizing the delivery of cancer drugs that block angiogenesis. Sci. Transl. Med. 2, 15ps3 (2010).

References and Notes

  1. We thank M. Moses, R. Jain, and M. Toi for critical reading of the manuscript and S. Lim for help with the figures. The laboratory of Y.C. is supported through research grants from the Swedish Research Council, the Swedish Cancer Foundation, the Karolinska Institute Foundation, the Torsten and Ragnar Söderberg’s Foundation, the European Union Integrated Projects of Metoxia (project no. 222741), and a European Research Council advanced grant (250021). R.L. is funded by NIH. Y.C. is a Chiang Jiang scholar at Shandong University, Jinan, China. Y.C. declares that he is a stockholder for Clanotech. R.L. declares that he is a stockholder for Microchips and Alkermes.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article