Is There a Role for Nitric Oxide in Tumor Angiogenesis?

In this issue of the Journal, Gallo et al. (1) report an interesting analysis of nitric oxide production and angiogenesis in head and neck cancers. They observed the following: 1) Tumor levels of total nitric oxide synthase activity (including inducible nitric oxide synthase activity) and of 38,58 cyclic guanosine monophosphate (cGMP) (a known mediator of cellular response to nitric oxide) were statistically significantly higher than those found in normal mucosa; 2) tumor specimens from patients with cervical lymph node metastases were more angiogenic (i.e., they had greater microvessel density) than specimens of nonmetastatic tumors; 3) when tumor specimens with high levels of nitric oxide synthase were implanted in the rabbit cornea, angiogenesis was induced, but this angiogenesis was suppressible by inhibitors of nitric oxide synthesis; and 4) angiogenesis at the tumor edge was an independent predictor of metastasis, a finding that is consistent with previous reports [see(2)] of patients with other types of cancer. These results reveal a strong positive correlation between the expression of nitric oxide synthase and tumor angiogenesis and tumor progression. Beyond this correlation lies the question of whether nitric oxide has a central role in tumor angiogenesis, and, if so, by what mechanism [see also (3)]? We can ask, “What is known about a putative mechanism for nitric oxide in angiogenesis, and what experimental questions remain unanswered?” It is now well established that tumor growth is angiogenesis dependent (4,5). It is also clear that the angiogenic phenotype is the result of a net balance between positive and negative regulators of angiogenesis (6,7). Positive angiogenesis regulators (e.g., vascular endothelial growth factor [VEGF] and basic fibroblast growth factor [bFGF]) and negative regulators (e.g., thrombospondin-1 and angiostatin) themselves may be modulated by other mediators or act through them. Nitric oxide is one example. Addition of exogenous VEGF to endothelial cells in vitro causes them to increase synthesis of nitric oxide (8,9). Furthermore, nitric oxide appears to be critical for the mitogenic effect of VEGF on endothelial cells, since inhibition of nitric oxide synthesis by L-NAME (Nv-nitro-L-arginine-methyl ester) blocks the mitogenic effect of VEGF (8,9). Sprout formation at the beginning of the angiogenic process is commonly preceded by vasodilation. This vasodilation is most easily observed during the induction of corneal angiogenesis after implantation of a tumor or of an angiogenic protein. We speculate that vasodilation of microvessels, by allowing endothelial cell elongation and spreading, may permit mitosis and migration of endothelial cells. When sparse cultures of capillary endothelial cells are maintained so that the cells do not make contact and cell spreading is controlled by substrata of varying adhesivity, DNA synthesis increases with increasing elongation of the cells (10). This situation would be analogous to endothelial cells in a maximally dilated microvessel (Fig. 1). In contrast, as cells become foreshortened and more rounded, DNA synthesis ceases, even in the presence of saturating levels of bFGF, a potent endothelial cell mitogen (11) (Fig. 2). This latter configuration would be analogous to a resting microvessel. Furthermore, VEGF increases directional migration in endothelial cells, and this migration is blocked by inhibitors of nitric oxide synthesis (12). L-NAME blocks VEGF-induced angiogenesis in the rabbit cornea (8). Expression of endothelial nitric oxide synthase is elevated in other human tumors (1) in addition to the head and neck tumors discussed by Gallo et al. (1). Inhibition of nitric oxide synthesis in tumors decreases blood flow and possibly vasodilation of the tumor vasculature (1). Production of angiogenic activity by macrophages is nitric oxide dependent (15). Of particular interest is the demonstration by Gallo et al. that angiogenesis induced in the rabbit cornea by an implant of human squamous cell carcinoma can be blocked by inhibition of nitric oxide synthesis. While it is not clear that these human tumors are producing VEGF, the potent suppression of tumorinduced angiogenesis by nitric oxide inhibition suggests that nitric oxide production is necessary but not sufficient for tumor angiogenesis. Certain experiments, which remain to be done, could be of great explanatory value. For example, is there compelling evidence that endothelial cells produce VEGF, and, if so, could there be an autocrine loop between VEGF and nitric oxide synthesis within the endothelium? Because vasculogenesis is not impaired in mice whose inducible nitric oxide synthase genes have been knocked out, but wound healing is delayed and can be restored by introducing inducible nitric oxide synthase via gene therapy (16), it would be of interest to know if tumor angiogenesis is suppressed in these animals. Are the activities of angiogenic factors other than VEGF blocked by the inhibition of nitric oxide synthesis? Also, do the activities of other angiogenic factors depend on VEGF? What is the effect on nitric oxide synthesis after VEGF is administered locally or systemically, as in the case of treatment of patients with ischemic limbs by therapeutic administration of VEGF (17)? Is there a role for nitric oxide in the development of collateral vessels in these patients? Are both of the major receptors for VEGF (flk and flt) dependent on nitric oxide for the angiogenic and vasodilator effects of VEGF? With respect to the latter question, the following information is known. An antibody that activates only the flk-1 receptor induces angiogenesis but not hypotension in the rabbit. In contrast, VEGF administration in the rabbit produces both angiogenesis and hypotension (18). Proposed model in which an angiogenic stimulus, e.g., vascular endothelial growth factor (VEGF), leads to early nitric oxide (NO)-mediated vasodilation of microvessels (a→b). A dilated vessel permits elongation (or greater spreading) of endothelial cells, such as the one labeled E in a, thereby increasing their sensitivity to an endothelial mitogen and resulting in migration and proliferation (b→c). Endothelial cells in vitro increase DNA synthesis (measured in disintegrations per minute [dpm] per cell of [3H]thymidine incorporation) in response to an endothelial mitogen (i.e., basic fibroblast growth factor [bFGF]) as they increase spreading (measured as surface area in mm2). Adapted from Ingber (11) with permission from the author. Because of the increasing effort to translate antiangiogenic therapy from the laboratory to the clinic, it is important to understand the mechanisms of action of angiogenesis inhibitors. Furthermore, the antitumor effect of certain conventional anticancer chemotherapeutic agents may have an antiangiogenic component. For example, it has recently been demonstrated that the 90-kd heat shock protein (Hsp90), which is present in endothelial cells, is an activator of endothelial nitric oxide synthase (19). Therefore, it is possible that chemotherapeutic agents that block Hsp90 signaling could have an antiangiogenic component by inhibiting nitric oxide synthesis. It remains to be seen whether any of the known selective or specific angiogenesis inhibitors affect nitric oxide production by the vasculature of tumors. The article by Gallo et al. (1) is provocative, not only because it demonstrates a correlation of nitric oxide production with angiogenesis (as quantified by microvessel density) in human head and neck cancers, but also because it calls attention to the need for further molecular, genetic, and functional experiments to elucidate a precise role for the nitric oxide pathway in tumor angiogenesis.