Angiogenesis as a marker of preneoplastic lesions of the human breast

Induction of angiogenesis during the transition from hyperplasia to neoplasia

It is now well established that unrestricted growth of tumours is dependent upon angiogenesis. Previous studies on tumour growth, however, have not revealed when or how the transition to an angiogenic state occurs during early tumour development. The advent of transgenic mice carrying oncogenes that reproducibly elicit tumours of specific cell types is providing a new format for studying multi-step tumorigenesis. In one of these models, transgenic mice expressing an oncogene in the beta-cells of the pancreatic islets heritably recapitulate a progression from normality to hyperplasia to neoplasia. We report here that angiogenic activity first appears in a subset of hyperplastic islets before the onset of tumour formation. A novel in vitro assay confirms that hyperplasia per se does not obligate angiogenesis. Rather, a few hyperplastic islets become angiogenic in vitro at a time when such islets are neovascularized in vivo and at a frequency that correlates closely with subsequent tumour incidence. These findings suggest that induction of angiogenesis is an important step in carcinogenesis.

Control of angiogenic activity in carcinogen-initiated and neoplastic hamster pouch keratinocytes and their hybrid cells

This study was undertaken to test the hypothesis that deregulated expression of the angiogenic phenotype by tumor cells is due to loss or inactivation of an angiogenesis suppressor gene(s). We used the technique of somatic cell hybridization to test the ability of untreated or chemical carcinogen-initiated hamster pouch keratinocytes, when fused to squamous epithelial neoplasms, to suppress tumor angiogenic activity by assaying hybrid-conditioned media (CM) in the avascular cornea of rat eyes. A non-angiogenic keratinocyte line, CL-2, derived from cultures of untreated epithelium and 3 lines of carcinogen-initiated keratinocytes, PN3, 5, and 7, of varying angiogenic potential were fused, using polyethylene glycol, to 3 tumorigenic, potently angiogenic, drug-resistant, hamster pouch carcinomas cell lines: HCPC-1, AW16E1-1, and AW16 E1-2. Serum-free 48-h CM from hybrid clones was prepared and assayed for angiogenic activity in rat corneas. CM from 5 hybrid clones derived from normal x neoplastic keratinocytes failed to induce an angiogenic response in 28 of 29 (97%) corneas tested. In contrast, CM from 4 hybrid clones derived from fusions between carcinogen-initiated and tumor cells were potently angiogenic in 24 of 25 (96%) corneas tested. Two angiogenesis suppressed hybrids clones were propagated in culture for an extended period of time, to permit chromosome segregation, and were found to re-express the angiogenic phenotype. These result indicate that angiogenesis is a recessive trait in normal hamster keratinocytes which is regulated in trans in these hybrid cells. It would also appear that loss or inactivation of angiogenesis suppressor function occurs early in the neoplastic process.

Morphometric evaluation of microvessels surrounding hyperplastic and neoplastic mammary lesions

Size and density of microvessels in close proximity to normal, hyperplastic, and in situ neoplastic mammary epithelium were evaluated with a morphometric method. Human tissue samples were obtained from breast mammoplasty or biopsy specimens with diagnosis of either epitheliosis or ductal carcinoma in situ. The mean vessel size was significantly enhanced in hyperplastic and neoplastic lesions as compared to mammoplasty specimens considered as 'normal'. On the contrary, the density of the microvessels around the epithelium was comparable in the three groups of tissues. It appears that as long as the basement membrane remains intact, sprouting of new vessels around neoplastic lesions is not manifested.

A study on the vascular proliferation in tissues around the tumor in breast cancer

In order to study the vascular proliferation in human breast cancer, blood vessels were counted, per square millimeter, in the tissue immediately around tumors. Mastectomized specimens of 84 patients with breast cancer and specimens from 10 patients with benign mammary diseases were stained by hematoxylin eosin and, where required, by the avidin biotin peroxidase complex method for laminin staining. The vascular density around the breast cancer tissue was 20.35 +/- 8.40/mm2, which was significantly higher than the value of 13.44 +/- 5.85/mm2 for noncancerous mammary tissues (p less than 0.001) or the value of 12.65 +/- 4.12/mm2 for benign mammary disease tissues (p less than 0.01). Among the breast cancers, noninvasive carcinoma had a higher vascular density (28.44 +/- 6.15/mm2) than invasive carcinoma (19.73 +/- 8.22/mm2, p less than 0.02). According to the Japan Mammary Cancer Society Classification of invasive ductal carcinoma, vascularity was higher in the papillotubular type of cancer than in the solid-tubular or scirrhous types of cancer (p less than 0.02), although the papillotubular type had the lowest rate of nodal metastasis and vascular invasion as compared with the scirrhous and solid-tubular types. The vascular density around the tumors did not change in association with an increase in tumor size and it was suggested that blood vessels around a tumor would increase almost in proportion to the square of the tumor diameter.

Chemical identification of a tumor-derived angiogenic factor

Neoplasms produce substances that induce blood vessel formation (angiogenesis). Fractions from ethanol extracts of the Walker 256 carcinoma were isolated by silica column chromatography and C18 reversed-phase high-performance liquid chromatography. Two of the isolated fractions induced neovascularization when tested in the rabbit corneal micropocket assay. One of the fractions was identified as nicotinamide by desorption-electron impact mass spectrometry, nuclear magnetic resonance spectroscopy, and gas chromatography-mass spectrometry. The second active fraction contained nicotinamide as part of a more complex, as yet unidentified, molecular arrangement. Microgram quantities of commercial nicotinamide induced neovascularization in the corneal micropocket assay and in the chick chorioallantoic membrane assay.

Progression from hormone dependence to autonomy and angiogenesis in mouse mammary tumours

The transplantable pregnancy-dependent mammary tumour (TPDMT-4), the related hormone-dependent (TPDMT-4EP) and autonomous (T4-0I320 and T4-0I96) subline tumours, and the mammary glands from DDD mice were compared for angiogenic activity on the rabbit cornea by tissue implantation. The TPDMT-4EP tumour was established by serially transplanting TPDMT-4 tumour fragments in oestradiol plus progesterone treated mice. The T4-0I320 and T4-0I96 tumours directly derived from the TPDMT-4 and TPDMT-4EP tumours, respectively. Angiogenic activity was graded by macroscopic and microscopic examinations into 3 classes; negative, partial and complete angiogenesis. These tumours were comparable to mammary glands in activity and induced complete angiogenesis in only 15-23% of the implants. However, when partial and complete responses were combined as positive angiogenesis, TPDMT-4, T4-0I320, TPDMT-4EP and T4-0I96 tumour implants were angiogenic in 25, 29, 42 and 54%, respectively. The T4-0I96 tumour was significantly more angiogenic than the parent tumour but this was not so for the TPDMT-4EP tumour. Spontaneous C3H mouse mammary tumours, human gliomas from nude mice, rat Walker 256 carcinomas and rabbit VX-2 tumours induced complete angiogenesis in 54, 63, 59 and 92% of the implants, respectively. The results suggest that the TPDMT-4 tumour is unique in being weakly angiogenic and able to progress toward greater autonomy with or without augmented angiogenic activity in different conditions.

On the origin and progression of human breast cancer

The mature human female breast contains thousands of hormone-sensitive, potentially milk-producing microorgans, called lobules. In the nonpregnant state they vary in size from 1 to 8 mm but most are 1 to 2 mm in diameter. Each lobule is drained by a terminal duct attached to the main duct system. It is called the terminal ductal-lobular unit, which normally regresses at menopause. Most breast diseases except papillomas in major ducts arise in terminal ductal-lobular units. Disease processes such as hyperplasias and neoplasia alter the terminal ductal-lobular units to such a degree that conventional light microscopic appearances make the structures appear as though they were ducts. However, the basic microarchitecture in three dimensions is preserved, proving the origin of cancer is the terminal ductal-lobular unit.

Carcinoma of the breast with stromal multinucleated giant cells

Three cases of invasive mammary carcinoma with stromal osteoclast-like giant cells are reported. All the specimens were grossly of reddish-brown colour. On microscopical examination two were invasive ductal carcinomas and the third was of mixed ductal and mucoid type. In the stroma all specimens showed numerous osteoclast-like giant cells, vascular proliferation, and accumulation of haemosiderin pigment. Supplementary immunohistochemical and electron microscopical investigations indicated that the giant cells originated from mononuclear stromal cells with only slight histiocytic properties. The giant cell formation was probably induced by the infiltrating carcinomatous tissue which is also believed to induce new blood vessel formation. However, there is no evidence of a direct relationship between these features.

Tumor angiogenesis

The hypothesis that tumors are angiogenesis dependent has, in the past decade, generated new investigations designed to elucidate the mechanism of angiogenesis itself. Many laboratories are now engaged in this pursuit. Some are studying angiogenesis that occurs in physiological situations, whereas others are interested in angiogenesis that dominates pathological conditions. These efforts have led to (1) the development of bioassays for angiogenesis; (2) the partial purification and, in one case, the complete purification of angiogenic factors from neoplastic and non-neoplastic cells; (3) the development of new polymer technology for the sustained release of these factors and other macromolecules in vivo; (4) the cloning and long-term culture of capillary endothelial cells; (5) the demonstration of the role of nonendothelial cells, such as mast cells in modulating angiogenesis; (6) the discovery of angiogenesis inhibitors; and (7) the demonstration that certain animal tumors will regress when angiogenesis is inhibited. The effects of angiogenesis inhibitors provide perhaps the most compelling evidence for the role of angiogenesis in tumor growth. It is conceivable that the original effort to understand the role of angiogenesis in tumor growth will also lead to the use of angiogenesis inhibitors as a new class of pharmacologic agents in a variety of non-neoplastic diseases such as arthritis, psoriasis, and ocular neovascularization. However, much work remains to be done before it will be possible to understand (1) the regulatory systems that govern capillary density in normal tissues; (2) the factors that maintain the viability of microvascular endothelium; (3) the development of the vascular system itself; and (4) the mechanism by which vascular regression occurs, both in the embryo and in the postnatal organism. A knowledge of the mechanisms which underlie these normal processes may help to enlarge our comprehension of tumor angiogenesis.

Towards a psychobiological model of cancer: biological considerations

The control of tissue homeostasis is extremely complex and many factors contribute to the growth and development of tumours. Although the immune system has been regarded as an essential intermediary between putative psychological factors and the development or restraint of malignant tumours, this review indicates that many other possible mechanisms also exist. Current aspects of tumour biology, immunology and hormonal control systems are reviewed, and detailed psychobiological mediating mechanisms are considered at each stage of tumour development. An approach to the future investigation of this difficult field is proposed.

Mammary hyperplastic and preneoplastic changes: taxonomy and grading

Mammary lobular and ductal carcinoma in situ are thought to originate from preneoplastic precursors. Most of our knowledge on mammary tumorigenesis is based on experimental data obtained in rodents, which are focused on the biology of hyperplastic alveolar nodules (HAN). The morphology of HAN is analyzed and compared with possible analogous lesions in humans. The microarchitecture of the terminal duct lobular unit (TDLU) is described with reference to descriptions and nomenclature appearing in the literature. A comparison is intended between existing classifications of human hyperplastic and preneoplastic changes.

Detection of angiogenesis activity in malignant bladder tissue and cells

Neoplastic and non-neoplastic mammalian urothelium was tested for angiogenesis activity in the rabbit iris assay. Tiny fragments of tissue were placed on a rabbit's iris and observed through a slit lamp stereoscope for as long as 4 days after implantation. Angiogenesis activity was demonstrated either by the appearance of new capillary growth or, following an injection of fluorescein, by a green fluorescence around an implant. Capillary proliferation was stimulated by 99 per cent of the specimens containing malignant urothelium, while only 9 per cent of normal urothelial specimens had this effect (p less than .001). Cells grown in vitro were similarly tested for angiogenesis by implanting 1 mm. fragments of collagen coated either with malignant or normal cells. A total of 92 per cent of the specimens covered with neoplastic cells induced angiogenesis, while only 15 per cent of the implants covered by normal cells caused this response (p less than .001). Benign reversible hyperplasia was induced in a rabbit bladder by systemic administration of cyclophosphamide, and an angiogenic response could not be detected in a small number of specimens. The results suggest that angiogenesis activity may be a reliable method for distinguishing between benign and malignant bladder tissue.

Invasiveness of transformed bladder epithelial cells

The invasiveness of bladder tumors has been studied in man, experimental animals, and in tissue culture by numerous authors. The prognostic importance of cellular markers for invasiveness is stressed, and the usefulness of histopathological and cytologic grading, cytogenetic studies, antigenic investigations, and enzymatic characterization is discussed. The invasiveness of bladder cells has frequently been examined in transplantation and explantation experiments. In human urothelial cell cultures three grades of transformation are defined, and a correlation has been established between the invasiveness of these cell lines in a three-dimensional in vitro model and their tumorigenicity in nude mice. The mechanism of tumor invasion is discussed, and it is recommended in future research to make a distinction between invasion en bloc and cellular infiltration.

Tumour angiogenesis

A variety of different factors has been implicated in inducing angiogenesis. Their identity and precise mechanisms of action remain elusive. The present treatise summarises the state of our knowledge not only in relation to tumour-induced capillary growth but also for non-neoplastic situations and mechanisms.

Tumor angiogenic activity (TAA) production in vitro and growth in the nude mouse by human malignant melanoma

The production of angiogenic activity by eleven human melanoma lines in vitro was compared with the extent of tumor growth in the nude mouse. Angiogenic activity was assayed by measuring the vascular response of the chick chorioallantoic membrane to serum-free supernatants. Growth in the nude mouse was determined after subcutaneous injection of cells 60 days later. Angiogenic activities ranged from negative to highly positive. In five lines angiogenic activity in vitro correlates with the extent of tumor growth in the nude mouse. In contrast, two lines did not show such a correlation, e.g. they produced large tumours without any detectable angiogenic activity. Histological examination of these two tumors revealed moderate degrees of vascularization and only low degrees of necrosis. It is concluded that the extent of tumor growth in the nude mouse is partly independent of the production of angiogenic activity by the tumor cells themselves.

Angiokines, angiogenesis and angiolymphoproliferative syndromes (ALPS)

In earlier experiments we had noted that transformed and leukemic leukocytes produced an RNA-rich angiogenic lymphokine. The formation of capillaries is a stepwise process in which reticulum cells first become detached and attracted to a site (mobilization and migration along a reticulin network). This is followed by local proliferation and finally by elongation and alignment against a basal membrane in tubular geometry. Coincidental with the last step is a biochemical and immunochemical differentiation of the endothelial cells manifested by the appearance of alkaline phosphatase, angiotensin-converting enzyme, factor VIII and the generation of receptors for thrombin as well as the capacities to produce prostacyclin and fibronectin on demand. It is postulated that there may be not one but several angiogenic lymphokines (angiokines) for each step of capillary development. Angiokine 1 (AK1) for the mobilization-chemotactic-migration, AK2 for the local proliferative, and AK3 for differentiating-morphogenic events. The above postulate aids in the classification and understanding of a number of angiolymphoproliferative syndromes since these reflect different disorders of the stepwise vessel formation. The association and the simultaneous proliferation of vascular and lymphoid elements is a feature that a number of lymphoproliferative disorders, of otherwise differing nature, have in common. To this effect they have been grouped in this study as angiolymphoproliferative syndromes (ALPS). These are a group of prelymphomatous or prelymphomogenic clinicopathologic entities in which proliferation of a lymphoid element (cell) is coupled with the accelerated development of blood capillaries and postcapillary venules.

Basement membrane changes in human breast tissue in explant culture: a morphometric and ultrastructural study

Basement membrane has been studied in vitro using human mammary lobules maintained in explant culture for periods up to 4 weeks. Morphologic similarities with basal lamina in malignant tumors in vivo have been observed, such as decrease in numbers of hemidesmosomes as determined by morphometry, reduplication and multilayering of the basal lamina, and failure to produce basal lamina after migration to the surface of the explant. Light microscopy using special stains for basement membrane has shown increasing fragmentation and loss with time in culture. These findings suggest that basement membrane changes seen in malignant tumors in vivo may not be limited to malignancy, per se, but may also be the result of other conditions.

Angiogenesis: initiation and control

From in vivo experiments using new methods such as the rabbit cornea, it is now becoming clear that the growth of a capillary involves an ordered sequence of events that includes lysis of the basement membrane of a parent venule, directional migration of capillary endothelial cells toward the angiogenic stimulus, lumen formation, development of branches, and anastomosis of the tip of one tube with another to form a loop. It is also clear that diffusible angiogenic stimuli can be released not only from most solid tumors, but also from at least three non-neoplastic cells. These include activated macrophages, sensitized lymphocytes, and adipocytes. Other normal tissues can also stimulate angiogenesis, but the type of cell giving rise to the angiogenic stimulus is unknown, and the period of angiogenic stimulation is brief. With the recent ability to clone capillary endothelial cells and to carry them in long-term culture, it has been possible to further delineate the mechanism of capillary growth. In vitro studies have shown that the mast cell seems to behave as a helper cell for capillary endothelial cells, in some way speeding up their rate of directional migration. At this writing, heparin appears to be the principal mast cell factor responsible for this effect on capillary endothelial cells. One theoretical possibility is that mast cells may prepare the matrix, perhaps by slow release of heparin, so that capillary sprouts can more easily move through it toward their angiogenic target. While the study of angiogenesis as a phenomenon is still in an early phase, it has become possible, by using a combination of in vitro and in vivo techniques, to more thoroughly understand the initiation and control of capillary growth.

Angiogenesis induced by "normal" human breast tissue: a probable marker for precancer

Normal human breast lobules, freshly isolated by precision microdissection of tissue stained with methylene blue chloride, were assayed for their ability to induce neovascularization (angiogenesis) in rabbit irises. Histologically, normal lobules from cancerous breast induced angiogenesis twice as often as lobules from noncancerous breasts, suggesting that preneoplastic transformation is diffuse.

Microvasculature of human colorectal epithelial tumors. An electron microscopic study

The ultrastructure of the microvasculature in human colorectal adenomas and carcinomas was studied, and compared with that of normal tissue. Blood vessels in adenomas were generally of normal structure, whereas carcinoma vessels displayed a variety of structural alterations. Endothelial proliferation was frequently observed in all the eleven carcinomas examined, which confirms the reports of new vessel formation in experimental malignant tumors. The presence of fenestrations in obviously abnormal endothelium was tentatively attributed to hypoxia, vascular regression or immaturity of the cells, in spite of the fact that normal capillaries in colonic mucosa are fenestrated. Some vessels showed multilayered basement membranes, various types of activated cells containing numerous lysosomal granules, and thickened perivascular tissue. These features may be a vascular response to repeated damage or a manifestation of vascular remodelling. Dilated thin-walled vessels, which were usually found at the periphery of carcinomas, were identified as venules. Small vessels located between the closely apposed carcinoma tubules were not obviously abnormal. They were nevertheless considered to be tumor-induced vessels which had undergone differentiation. The differences in vascular morphology between adenomas and carcinomas were considered to be due to their different growth patterns, growth rates and degrees of maturity.

Stimulation of S-9 fraction metabolism in rat liver and breast by vital dyes

The vital dyes neutral red (NR), methylene blue (MB) and trypan blue (TB) induced microsomal enzymes that metabolize the chemical carcinogens 2-aminoanthracene (2AA), 6-aminochrysene (6AC) and diaminofluorene (DAF). The Ames test was used as a method of assessing the amount of microsomal enzyme activity that occurred in rat breast and liver when treated with NR, MB, and TB. Livers were more readily induced than mammary glands of midpregnant rats. Liver S-9 fraction from rats treated with vital dyes showed an increase in revertant colonies for 2AA, DAF, and 6AC over that from untreated rats whereas breast S-9 fraction from rats treated with vital dyes showed an increase in revertant colonies for 2AA over untreated.

Angiogenic activity as a marker of neoplastic and preneoplastic lesions of the human bladder

Angiogenic activity has been assessed in biopsy specimens from 49 patients with transitional cell carcinoma, benign prostatic hypertrophy and vesicoureteral reflux. Over 95% of the specimens containing malignant and premalignant transitional epithelium stimulated capillary proliferation on rabbit iris. In contrast, less than 10% of normal tissues had this effect. Sixty-one per cent of specimens with inflammatory round cells were angiogenic but this was reduced to 10% by preincubation with rabbit antihuman lymphocyte serum. Cystitis cystica also induced capillary proliferation even if incubated with the antilymphocyte serum. Angiogenic capacity may be an early marker of preneoplastic and neoplastic lesions of the human bladder.

The developmental pathology of breast cancer

The early stages in the development of breast cancer, from epithelial inception to initial invasion, comprise a predictable, fairly uniform series of events. Once invasion has occurred, behavior is much more random. For this reason, local treatment methods are most effective in these early stages and efforts directed at early detection are thoroughly justified.

The influence of pathologic factors on breast cancer management

The pathologist routinely provides to the therapist data which are used in the management of breast cancer patients. Clinical as well as gross and microscopic examination provides information used for staging and treatment selection. Biologic neoplasia precedes the usual morphologic and cytologic changes that characterize precancer and in situ carcinoma to the pathologist. Minimal breast cancer, including in situ carcinoma and small (0.5 cm) infiltrating cancers, is now a recognized entity separable from clinical cancer, although therapy is not yet uniform. The pathologist can routinely report the gross size and contour of the cancer; microscopic evaluation of the primary cancer adds information on the histologic type, differentiation (histologic or cytologic grade), and such other data as blood vessel invasion and cellular infiltration. One of the most useful bits of information is the status of axillary lymph nodes: whether or not there are metastases, the number of nodes with metastases, and whether they are micrometastases or macrometastases that extend through the capsule and involve pericapsular vessels. All of these data can be recorded routinely and are useful in developing management criteria. The pathologist, as any other consultant, reports to the attending physician who then uses these data with all other pertinent facts to formulate an individualized therapeutic program.