Angiogenic and angiostatic microenvironment in tumors--role of gangliosides.

Gangliosides are important components of the cell membrane that are usually shed in the surrounding microenvironment by neoplastic cells. Gangliosides can also modulate the angiogenic response of microvessels stimulated by angiogenic factors. The experiments reported here make a contribution to the assessment of the nature of this angiogenic modulation, by demonstrating that a) GM3 gangliosides can block the proliferation of endothelium induced by neoplastic cells from human tumors of five different origins; b) this block also occurs when the endothelial cells are preincubated with GM3 and disappears when the cells are returned to a medium poor in GM3; c) in the presence of GM3 the capacity of the endothelial cells to bind to fibronectin and to collagen types I and IV was sharply reduced; d) concentrations of GM3 able to block endothelial cell growth are counteracted by addition to the medium of GT1b ganglioside. The data suggest that the prevalence of a microenvironment rich in GM3 prevents proliferation of vascular endothelium, but the appropriate presence of another ganglioside, such as GT1b, nullifies the effect. Modulation of the angiogenic response of vascular endothelium to angiogenic factors released by tumors is probably dependent on the distribution and activity of growth factor receptors on the endothelial cell surface. The nature and concentration of the gangliosides in the endothelial microenvironment have a decisive influence on this event and possibly on the progression of tumor-induced angiogenesis.

Prostaglandins and gangliosides of tumor microenvironment: their role in angiogenesis.

In the solid tumor the microenvironment is the space limited by the basement membrane of the microvessels and the neoplastic cells membrane. It includes the stroma and a liquid phase, the tumor interstitial fluid (TIF). We developed a method to sample TIF in vivo and found it rich in prostaglandins. In the rabbit cornea PGE1 induces neovascularization and acts as an angiogenesis factor. Before angiogenesis appears the ganglioside content of the cornea doubles with sharp reduction of the GM3/GD3 ratio. These gangliosides are not angiogenic but they influence the endothelial cell behavior. In particular when the PGE1 dose is insufficient to induce angiogenesis, enrichment of corneal tissue with GD3 or GM1 stimulates angiogenesis. However, when the corneal tissue is enriched with GM3, doses of PGE1, normally angiogenic fail to do so. The same was observed when bFGF substituted PGE1 as an angiogenesis trigger. The gangliosides tested acted as modulators of the angiogenic response by promoting the angiogenic capacity of molecules, such as PGE1 or bFGF, normally present in the tissue microenvironment. Several neoplastic cells, especially melanomas, shed gangliosides in the microenvironment. Their modulatory effect on angiogenesis may influence metastatic and/or primary tumor growth.

Angiogenesis can be stimulated or repressed in vivo by a change in GM3:GD3 ganglioside ratio.

BACKGROUND: We had previously observed that rabbit cornea stimulated by an angiogenic factor 1) became richer in total gangliosides and 2) reduced the GM3:GD3 ganglioside ratio. Moreover, experimentally induced global enrichment of corneal gangliosides favors angiogenesis. EXPERIMENTAL DESIGN: The objective of this work was to explain the possible relationship between angiogenic response and changes in the GM3:GD3 ratios observed in vivo. Cornea was utilized because it is avascular and transparent; i.e., the onset of opacity permitted exclusion of angiogenesis produced by a generic inflammatory response. Prostaglandin E1 or basic fibroblast growth factor were applied as angiogenesis triggers. Angiogenesis in vivo and mobilization and growth of microvascular endothelium in vitro were taken as parameters to indicate whether differences in GM3:GD3 ratios could modify the extent of the angiogenic response. RESULTS: In vivo angiogenesis, whether prostaglandin E1 or basic fibroblast growth factor induced, was repressed by GM3 and enhanced by GD3 or GM1 enrichment of the cornea. In vitro growth and motility of microvascular endothelium were reduced by GM3 addition to the medium and returned to normal levels by addition of GD3. CONCLUSIONS: Formation of new vessels induced by two different angiogenic factors could be stimulated or repressed in the cornea by reduction or enhancement of the GM3:GD3 ratio of tissue gangliosides. Changes in the relative proportion of molecules normally present in adult tissues, like prostaglandin E1, basic fibroblast growth factor, GM3, GD3, were sufficient to modulate or even block angiogenesis.

Gangliosides, copper ions and angiogenic capacity of adult tissues.

The report summarizes the work of our laboratory aimed at improving the understanding of the angiogenic response of adult tissues, an event that transforms a micro-embolus of neoplastic cells into a growing metastasis. Attention has been focused on tumor-induced angiogenesis. The following aspects of the subject are discussed: (a) relationship between size of vascular network and tumor growth rate or tumor cell population; (b) angiogenic capacity of tumors and role that prostaglandin E1 may have as an angiogenesis factor; (c) relationship between acquisition of angiogenic capacity and neoplastic transformation of a cell population; (d) modification of tissue composition at the onset of angiogenesis; (e) behaviour of copper ions and copper carriers in the course of the angiogenic response; (f) the influence of gangliosides on endothelial cell motility, survival and growth in vitro; (g) modulation of the angiogenic response by gangliosides (GM1, GT1b) in vivo.

Retinyl acetate-induced arthritis in C3H-A(vy) mice.

Severely impaired musculoskeletal mobility in C3H-A(vy) mice was noted during a pharmacologic trial evaluating the antitumorigenic properties of retinyl acetate (RAc). To determine the etiology of this impairment, we studied 103 female C3H-A(vy) mice that were fed RAc in daily doses of 75-300 micrograms or placebo and were killed after 3-16 months. Whole-body radiographs and histologic sections of the hindlimbs were scored for presence and severity of arthritis. C3H-A(vy) mice treated with RAc in any dose had a significantly higher incidence of arthritis than placebo-treated mice. Histologic evidence of enthesopathic disease closely paralleled the radiographic changes and ranged from small enthesophytes at tendinous and capsular insertions to complete periarticular bony bridging. Articular cartilage was not grossly affected. The incidence and severity of arthritis were significantly correlated with the total dose of RAc administered. The bony metaplasia induced by RAc was similar to the pathologic changes caused by other retinoids. This model may be useful for studying the pathogenesis of periarticular bone formation in diffuse idiopathic skeletal hyperostosis and related syndromes.

Influence of gangliosides on primary and metastatic neoplastic growth in human and murine cells.

The influence of gangliosides on tumor growth and frequency of metastasis in vivo as well as on growth and motility of neoplastic cells in vitro was tested utilizing human and rodent cell populations. In mice receiving injections of a ganglioside mixture twice daily the tumor volume, the number of spontaneous metastases per animal, and the number of mice with metastasis was approximately double that of controls. Preincubation of neoplastic cells with the ganglioside mixture doubled the number of metastatic foci in the lungs of mice receiving the cells by i.v. injection. Addition of a ganglioside mixture to the culture medium enhanced motility of neoplastic cells about 3-fold. This finding was similar to that observed for capillary endothelium. The presence of gangliosides in the culture media for a 48-h incubation period about doubled the number of neoplastic cells as compared to controls; the same was observed for capillary endothelium. The data are interpreted to indicate that gangliosides improve growth and mobilization of capillary endothelium and neoplastic cells. Both events may concur in enhancing tumor growth in vivo, the first by improving angiogenesis, the second by direct action on the neoplastic cell population.

Considerations on the mechanism of the angiogenic response.

The principal objective of our work is to sufficiently understand the mechanism of angiogenesis in the adult organism to allow interference with the process on a rational basis. It is apparent that several "factors" can trigger angiogenesis. To test these, we used the rabbit cornea mostly because it is avascular (i.e., the background is zero) and transparent (i.e., the newly formed capillaries that invade the cornea are easily visible in the unanaesthetized animal). Under these conditions, it was found that the cornea ready to be colonized by capillaries under the action of an angiogenesis effector becomes rich in copper ions and sialic acid. Motility of bovine capillary endothelium was utilized to analyze the angiogenesis process on the ground that mobilization of capillary endothelium is the first morphological event observed during angiogenesis in vivo and the methods to measure cell motility are reasonably accurate. With this approach it was found that heparin, fibronectin, and gangliosides are involved in the mobilization of capillary endothelium. The precise interaction among these three components is not yet clear.

Interaction of gangliosides with fibronectin in the mobilization of capillary endothelium. Possible influence on the growth of metastasis.

Mobilization of the capillary endothelium is one of the first events observed during angiogenesis, and the study of conditions that control or influence the mobilization of the endothelium in vitro has been assumed to offer information relevant to the understanding of angiogenesis in vivo. In vitro mobilization of the bovine capillary endothelium was substantially enhanced by addition of gangliosides to the culture medium. Optimal mobilization was obtained when the endothelium incorporated the gangliosides first and was then seeded on fibronectin anchored to collagen type I. Preincubation of the capillary endothelium with gangliosides, trisialoganglioside in particular, doubled the amount of fibronectin bound to the cells and enhanced the migration about 5-fold. 'Blockage' of ganglioside binding with cholera toxin or gamma-interferon substantially reduced migration. Rabbit corneas, treated in vivo with a variety of angiogenesis effectors to induce neovascularization, consistently showed an increase in sialic acid content just prior to the time the tissue would be penetrated by the capillaries. This finding was interpreted to indicate that an increment of the ganglioside content of the capillary endothelial cell membranes may play a determinant role in the mobilization of the capillary endothelium in vivo as shown here to take place in vitro. Since the formation of a tumor from a micrometastasis requires formation of new capillaries and highly metastasizing tumors very frequently have high levels of sialic acid on the cell surface, it is hypothesized that production and shedding of gangliosides from the surface of neoplastic cells may be a factor in promoting angiogenesis and metastatic growth.

Effect of pregnancy and nursing on the growth of metastases from N-nitroso-N-methylurea-induced mammary carcinomas.

Pregnancy increased the survival times of inbred BUF/N rats bearing occult metastasis of a mammary carcinoma at the time of conception. Nursing following delivery nullified the effect of pregnancy. The similarity of the controlled experimental data with the limited clinical observations is noted. An experimental model of rat mammary carcinoma has been described that possesses a highly metastasizing capacity and can be utilized to study the behavior of clinically silent metastasis.

Cooperation between fibronectin and heparin in the mobilization of capillary endothelium.

Angiogenesis is indispensable to sustain promotion and growth of metastases. As a contribution to the understanding of the angiogenesis process, the experiments reported showed that: (a) fibronectin is involved in the mobilization of capillary endothelium which is the first event in angiogenesis; (b) antifibronectin serum can block the mobilization, and neutralization of the antiserum can restore it; (c) the combination of fibronectin + heparin is a powerful mobilizer of capillary endothelium, and (d) fragments of the fibronectin and heparin molecules in combination can mobilize capillary endothelium as effectively as the intact molecules. The results are interpreted to indicate that molecules normally present in the extracellular matrix like heparin and fibronectin, may act as angiogenesis effectors when the physiological structure of the tissue is altered, for instance by lytic enzymes released by metastatic neoplastic cells.

Expression of kappa-casein in normal and neoplastic rat mammary gland is under the control of prolactin.

An 820-nucleotide-long cDNA clone for the kappa-casein (the casein micelle-stabilizing protein) from rat mammary gland was isolated, and its nucleotide sequence was determined. The deduced amino acid sequence from the nucleotide sequence revealed a signal peptide, 21 amino acids long, and a mature protein of 157 amino acids. The signal peptide of rat kappa-casein was highly homologous to that of the precursor to ovine kappa-casein. However, little homology was apparent when the mature kappa-casein protein sequences from ovine or bovine sources were compared with rat kappa-casein. The kappa-casein mRNA content of the mammary tissue was found to increase during its functional differentiation. Prolactin appears to modulate the production of kappa-casein mRNA. Mammary glands of virgin females had no detectable kappa-casein mRNA; however, a marked induction of kappa-casein mRNA was obtained by intravenous infusion of prolactin. Mammary carcinomas did not follow the same pattern. 7,12-Dimethylbenz[a]anthracene-induced mammary carcinomas had normally low levels of kappa-casein mRNA, but intravenous prolactin infusion increased the levels by 2-fold. The MTW9 mammary carcinoma that grows only in the presence of high levels of mammotropic hormones had kappa-casein mRNA content equivalent to that in 10-day lactating rat mammary gland. Continuous venous infusion of prolactin to MTW9 mammary carcinoma did not modify the kappa-casein mRNA levels. Nitrosomethylurea-induced mammary carcinomas had no detectable kappa-casein mRNA, and intravenous prolactin infusion was unable to induce it.

Angiogenesis in vivo and selective mobilization of capillary endothelium in vitro by heparin-copper complex.

The objective of this work was to contribute to the interpretation of the mechanisms of capillary formation in the adult tissues. We have observed that the heparin-copper complex is angiogenic in vivo and stimulates migration of capillary endothelium in vitro. This effect is specific for capillary endothelium since aortic endothelium or fibroblasts from the rabbit cornea or human skin were unresponsive.

Characterization of a chemoattractant for endothelium induced by angiogenesis effectors.

The mechanism of neovascularization was further explored by the use of chemically defined angiogenesis effectors. The vascularization of the rabbit cornea was selected as an experimental approach that permits comparison of one cornea treated by the angiogenesis effector with the contralateral cornea of the same subject treated by the same molecule deprived of angiogenic capacity. Under these conditions, we observed that neovascularization was initiated by the appearance of a chemoattractant for the bovine capillary endothelium only in the cornea treated by the angiogenesis effector. The chemoattractant was purified about 150-fold by a single-step procedure, using gelatin:Sepharose affinity chromatography. Chemoattraction resulted from the combined effect of a chemotactic factor(s) and an activating factor(s). The association of the two enhanced 5- to 8-fold the motility of the capillary endothelium in a concentration-dependent manner with optimum at 0.2 mg/ml. The activating factor(s) does not have chemotactic capacity, but without it, chemotaxis is reduced to about one half. The chemotactic complex was present in the cornea regardless of the nature of the angiogenesis effector used as the triggering device. Heat and proteases eliminated chemotaxis and destroyed the chemotactic complex. Thus, neovascularization may be triggered by effectors able to induce in the cornea proteins, normally not present, that influence angiogenesis via mobilization of capillary endothelium.

Mobilization of capillary endothelium in vitro induced by effectors of angiogenesis in vivo.

An assay to measure endothelial cell mobilization on a gelatin substratum has been developed. Utilization of the gelatin-agarose and Boyden chamber assays established that: (a) fragments or extracts of corneas treated with several effectors of angiogenesis in vivo acquired the capacity to mobilize the capillary endothelium in vitro; (b) this mobilization was selective for the capillary endothelium; endothelium from aorta and fibroblasts from human skin or rabbit cornea were unresponsive; and (c) among the effectors of angiogenesis utilized alone; i.e., without the intermediary action of the cornea, none were able to mobilize the capillary endothelium in vitro, except for the heparin-copper complex. The data are interpreted to indicate that new formation of capillaries in vivo is the end result of a cascade of events of which heparin and copper are important components.

Ceruloplasmin, copper ions, and angiogenesis.

The ability to induce new formation of capillaries in the cornea was tested for ceruloplasmin, the copper carrier of serum, for fragments of the ceruloplasmin molecule with and without copper, for heparin, and for glycyl-L-histidyl-L-lysine, bound or not bound to copper ions. Male or female 2- to 3-kg New Zealand White rabbits were used. These experiments were prompted by the previous observation of copper accumulation in the cornea during angiogenesis and by the inability of copper-deficient rabbits to mount an angiogenic response. The results showed that the three different molecules were all able to induce angiogenesis provided that they were bound to copper. Fragments of the ceruloplasmin molecule also induced angiogenesis but only when copper was bound to the peptides. The data are interpreted to indicate that copper ions are involved in the sequence of events leading to angiogenesis and that the carrier molecules may be of quite a different nature.

Role of prostaglandin E1 and copper in angiogenesis.

The interstitial fluid of MTW9A and Walker carcinomas and their ethanol extract induced strong angiogenic response in the rabbit (New Zealand White) corneal test. The fluid collected in vivo was rich in E-type prostaglandins, and prostaglandin E1 (PGE1) in particular was strongly angiogenic at the lowest dose as compared with the angiogenic responses of prostaglandins E2, I2, and F2 alpha. Neoplastic fibroblasts also induced a strong angiogenic response, but in indomethacin-treated rabbits neovascularization failed to occur. Copper was concentrated in the cornea during PGE1-induced neovascularization, and copper-deficient rabbits were unable to mount an angiogenic response in the corneal test. Ceruloplasmin, the copper carrier of plasma, was found to be angiogenic at high doses. In indomethacin-treated rabbits, however, ceruloplasmin at the same high doses failed to induce angiogenesis. The experiments are interpreted to indicate that angiogenesis is the end result of a sequence of events, two of which are PGE1 production and copper mobilization in the tissue where neovascularization occurs.

Angiogenesis and neoplastic progression in vitro.

In vivo cell populations at high risk of neoplastic transformation have been shown to acquire the ability to induce new formation of vessels. The present experiments tested whether the same change occurred during neoplastic transformation in vitro. In four cell populations (human HBL 100 mammary epithelium, BALB/c fibroblasts, C57BL-MG epithelium, and Syrian golden hamster embryo cells), angiogenic capacity appeared during their cultivation in vitro and was evident long before a neoplastic transformation could be recognized. The data were interpreted to support the hypothesis that acquisition of angiogenic capacity by a cell population normally devoid of this capacity indicates an increased risk of neoplastic transformation.

Temperature gradients and local perfusion in a mammary carcinoma.

Temperature gradients of mammary tumors in randombred Sprague-Dawley rats under normothermia, hypothermia, and hyperthermia were determined, and their experimental modifications were utilized to assess differences in perfusion rates within the neoplastic tissue. Normothermic tumors showed a circadian rhythm with zenith at midnight and nadir at midday. Differences between highest and lowest temperatures recorded during the 24-hour period reached up to 3 degrees C. Similar oscillations were observed in subcutaneous tissue without tumor. An average temperature increment of 0.5-1.0 degrees C was observed when a tumor was transferred from the subcutaneous to the abdominal location. Gradients larger than 1 degrees C were observed within the same tumor in locations only a few millimeters distance from each other. The nonuniformity in temperature within normothermic tumors was exaggerated during hyperthermia. No appreciable change in temperature gradients was seen within a normothermic tumor when tumor blood flow was doubled or reduced to one-third of the basal level. Hyperthermia increased both volume and temperature of tumor efferent blood. As expected, decrease or increase in blood flow during hyperthermia increased or decreased tumor temperature, respectively, but substantial temperature gradients up to 2 degrees C still persisted within adjacent regions. The extent of temperature changes in the tumor could not be correlated with a known change in blood supply. A pulse of cold serum into the tumor afferent artery produced a substantial reduction of tumor blood flow, but only a small depression in tumor temperatures, and a very small change in tumor temperature gradients. No appreciable modification could be brought about in tumor temperature levels and temperature gradients within the tumor by pulses of cold serum in the afferent artery during hyperthermia. After external cooling of the tumor, the time necessary to compensate for temperature depression did not correlate with either the reduction of temperature or with the thickness of the tumor tissue separating the thermistor from the cold source. The results indicate extensive anisotropy of temperature and blood distribution within growing neoplastic tissue and suggest that heat transfer by convection within the tumor is much less effective than it is commonly assumed.