Treating Cancer Through Vascular Normalization

The importance of vasculature for the growth and development of solid tumors has been well established for close to a century. Nonetheless, it was only following Judah Folkman’s 1971 discovery of tumor angiogenic factor that researchers began to explore the therapeutic potential of antiangiogenic strategies (1). Folkman and his colleagues coined the term “antiangiogenesis” and believed that, when this strategy could be clinically achieved, it would be “a powerful adjunct to present methods of cancer therapy.” They theorized that metastases might not arise from a non-vascularized tumor and that such tumors were more vulnerable to chemotherapy and cell-mediated immunologic attack.

Early promise?
The field of angiogenesis progressed (see “Advances in Angiogenesis”) until, in 2004, its theoretical promise became testable with the development of bevacizumab (Avastin), an antibody targeting vascular endothelial growth factor (VEGF) (2). Initially, antiangiogenic agents appeared to deliver on their potential; in the absence of neovascularization, tumors neither grew beyond 2 mm nor underwent metastasis (3). The treatments demonstrated efficacy in many cancer types, including breast, colon, renal, and ovarian – and, in early studies, researchers found no major toxicities (3). In contrast, traditional chemotherapy was characterized by lack of specificity, potential for severe side effects, variable dosing regimens, and development of treatment resistance. As a result, antiangiogenic agents gained appeal and a large number were designed, tested, and made available for clinical use (see Table 1).

As the use of antiangiogenic compounds increased, though, the aura of invincibility began to give way. Cardiovascular side effects – hypertension, conduction abnormalities, QT prolongation, left ventricular systolic dysfunction (LVSD), and even heart failure – were noted, along with risks of bleeding, thrombotic events, proteinuria, leukopenia, lymphopenia, and hyperthyroidism (3). However, the most discouraging clinical observation was the recurrence of more aggressive tumors following a course of antiangiogenic treatment. Research has demonstrated that antiangiogenic drugs can accelerate metastatic tumor growth and decrease overall survival in mice receiving short-term therapy in various metastasis assays (4). Acceleration of metastasis was also seen in mice receiving sunitinib prior to intravenous implantation of tumor cells, suggesting possible “metastatic conditioning” in multiple organs. Similar findings with additional VEGF receptor tyrosine kinase inhibitors suggest a class-specific effect for such agents. In fact, Marta Pàez-Ribes and colleagues found that angiogenesis inhibitors targeting the VEGF pathway demonstrate antitumor effects in mouse models of pancreatic neuroendocrine carcinoma and glioblastoma, but concomitantly elicit tumor adaptation and progression to stages of greater malignancy, with heightened invasiveness and, in some cases, increased lymphatic and distant metastasis (5).

A decade after the advent of these agents, the concept of an ideal antiangiogenic that destroys tumor vessels without harming normal vessels remains elusive. Although it is conceivable that higher doses of currently available antiangiogenic agents could yield complete tumor regression, such doses are likely to adversely affect the vasculature of normal tissues. Excessive vascular regression is counterproductive because it compromises the tissue delivery of drugs and oxygen (6) – and antiangiogenic therapy is associated with an increased risk of arterial thromboembolic events that could be more pronounced with increased doses (7). Lastly, researchers now believe that supratherapeutic doses or scheduling of antiangiogenic agents might lower tumor oxygenation and drug delivery – antagonizing, rather than augmenting, the response to radiotherapy or chemotherapy. This delicate balance between normalization and excessive vascular regression emphasizes the need for careful prescribing of antiangiogenic agents.

A new approach: vascular normalization
To avoid the pro-metastatic and other deleterious impacts of antiangiogenic therapy while enhancing the efficacy of combination therapy, antiangiogenic compounds can be used at low dosage to induce vascular normalization (8). Two murine breast cancer models have empirically demonstrated that lower doses of anti-VEGF receptor 2 (VEGFR2) antibody induce breast tumor vascular normalization (9), including a more homogeneous distribution of perfused tumor vessels, an increase in pericyte vascular coverage (PVC), and a decrease in hypoxia.

The potential benefits of vascular normalization as a treatment strategy have been experimentally demonstrated. Tumor vasculature is more complex, dilated, tortuous, hyperpermeable, and disorganized than normal blood vessels. These complex, leaky vessels represent a limiting factor for the efficacy of combination therapy – but normalizing doses of an anti-VEGFR2 antibody helps reverse this phenotype in favor of a more homogeneous distribution of functional tumor vessels.

Not only that, but lower doses are superior to high doses in polarizing tumor-associated macrophages from an immune-inhibitory M2-like phenotype toward an immune-stimulatory M1-like phenotype, as well as in facilitating CD4⁺ and CD8⁺ T-cell tumor infiltration (10). Based on this mechanism, lower-dose anti-VEGFR2 therapy, with T cell activation induced by a whole cancer cell vaccine, may increase anti-cancer efficacy in a CD8⁺ T cell-dependent manner in both immune-tolerant and immunogenic murine breast cancer models (10). Recently published animal studies using elegant genetic models have also shown that normalizing tumor vasculature can improve anti-tumor immunity. In a spontaneous pancreatic insulinoma mouse model, RGS5 deficiency normalized tumor vessels, increasing PVC and perfusion, which in turn led to increased delivery of adoptively transferred T cells and improved survival (11). In a more recent study, overexpression of the histidine-rich glycoprotein (HRG) induced a normalized vessel phenotype in solid tumors, evidenced by increased PVC, greater perfusion, and reduced hypoxia (12). Together, these findings show a mechanistic link between vessel normalization and enhanced immune cell infiltration and function.

The clinical challenge
Despite enormous theoretical and experimental support for vascular normalization, effectively achieving it in the clinic remains complex and daunting. Perhaps the clearest hurdle is the need to achieve a delicate balance between normalization and excessive vascular regression. Such a precise approach requires careful selection of the dose and administration schedule for antiangiogenic agents (9). Tumors are also highly heterogeneous, so each one requires that we target different proangiogenic factors at different times to induce vascular normalization – and the optimum dose varies by patient and by disease status, so generalizing is impossible. A reliable biomarker would serve a central role in helping to choose a “vascular-normalizing” or “pruning” dose – but identifying such predictive biomarkers remains difficult; after all, we must first elucidate the mechanism of action, which is poorly understood for currently approved antiangiogenic agents.

Another challenge in developing effective biomarkers is establishing adequate criteria for response – especially problematic for antiangiogenic agents, which target the stroma. The standard lesion size evaluation may not optimally assess treatment response, particularly in monotherapy with agents such as sunitinib or sorafenib. Anti-VEGF therapy has primarily cytostatic effects, might prune and normalize the tumor vasculature, and can have substantial systemic effects (such as modulation of circulating proangiogenic and proinflammatory cells and cytokines). These effects might stabilize, rather than shrink, the tumor.

The schedule of drug administration matters, too. Different types of immunotherapies – for example, whole tumor cell vaccine, dendritic cell vaccine, and adoptive T cell transfer – take different amounts of time to boost anticancer immunity, so we must optimize the schedules of antiangiogenic treatments to achieve the best possible anticancer efficacy.

A further complexity is the existence of a “normalization window” – that is, a period during which the addition of radiation therapy yields the best therapeutic outcome. This window appears short-lived (about six days) and is characterized by an increase in tumor oxygenation, which in turn increases the concentration of reactive oxygen species created by the radiation to enhance its efficacy. During the normalization window, VEGFR2 blockade was found to increase PVC in a human brain tumor grown in mice (8). 

In short, we face three major challenges in translating therapies based on the vascular normalization model to the clinic.

1. Determining which antiangiogenic therapies lead to vascular normalization.
2. Identifying suitable surrogate markers of changes in the structure and function of the tumor vasculature – and developing imaging technology to determine the timing of the normalization window during antiangiogenesis therapy.
3. Filling the gaps in our understanding of the molecular and cellular mechanisms of the vascular normalization process.

These challenges will take time to overcome – but, in doing so, I hope we will develop novel approaches to enhance cancer treatment efficacy.

Blood as a biomarker?
As noted, to evaluate vascular normalization as a therapeutic tool, we need a biomarker. An ideal biomarker should be reliable, easy to measure, cost-effective, and allow the clinician to both select and follow the patient’s response to therapeutic intervention. The ability to identify tumor-specific changes rapidly after treatment may also allow tailoring of therapy to the patients most likely to benefit – and early discontinuation of ineffective treatment in others. Currently, no such marker exists. But without it, could we evaluate vascular normalization by measuring the actual flow of blood to the tumor and surrounding organ? The answer is yes; we can use a PET-CT scanning device for molecular imaging – something we’ve done in animal models. With a flow tracer like ¹⁵O-labeled water, it is possible to measure blood flow in an organ like the liver or breast in humans as well as in model organisms (13).

I theorized the following experimental model to evaluate vascular normalization. Using a murine breast cancer model, it should be possible to measure the degree of flow in the breast with tumor and compare it with the blood flow in the contralateral breast. Due to the hyperemia induced by the tumor, I expect this ratio to be greater than 1. After proof-of-concept in a small population of experimental models, the experiment will move to its next stage. One cohort of animals with breast cancer, treated with standard chemotherapy and fixed dosages of an antiangiogenesis agent, will serve as our control. In the experimental arm, we will use standard chemotherapy, but will adjust the dosage of antiangiogenic treatment based on blood flow imaging data acquired by serial PET scanning. The goal is to normalize the blood flow between the affected and non-affected breasts using blood flow imaging as our surrogate marker for achieving vascular normalization. With serial CT scanning, it should be possible to evaluate the treatment’s impact on tumor size, metastasis, and longevity. This will also be a valuable opportunity to evaluate promising biomarkers by assaying them against tumor vascularity. If our experiments succeed, we will have a way of testing the vascular normalization hypothesis – and evaluating how well antiangiogenic treatments work for patients.

Vascular normalization represents a novel way to use antiangiogenic therapies at lower dosages, simultaneously reducing side effects and improving the efficacy of adjuvant therapies. Because of the relatively small opportunity window and the need for careful dosage regulation to balance normalization and inhibition, this approach is difficult to translate to the clinic – but its potential in cancer and a variety of other medical conditions warrants further study. With the development of a more effective biomarker and a standard experimental procedure, we may be able to move vascular normalization from the bench to our patients’ bedsides.