Search

The distribution of proliferating cell nuclear antigen (PCNA)-(clone PC 10)- and Ki-67-(clone MIB-1)-positive nuclei was investigated in 60 medulloblastomas of childhood. Although the labeling index of the two markers did not coincide, both showed a wide range of parallel variations. The percentage of positive nuclei was similar in both classic and desmoplastic tumors. A variable proliferation capacity was found in the different tumor structures. Areas with neuronal and glial differentiation showed very few positive nuclei; these were very abundant in the infiltration areas, and along penetrating vessels from subarachnoidal growths. Pale islands were negative or positive only in their peripheral part. Large-cell areas were richer in positive nuclei than classic ones, accounting for their more malignant character. Hyperchromatic round nuclei, not belonging to necrotic foci and called lymphocyte-like nuclei, differently interpreted in the past, were variably found in every case. They are known, from previous experience, to stain orange with Acridine Orange fluorochroming, like single-stranded DNA. They were not easily distinguishable from mitoses and were stained by in situ end-labeling of DNA strand breaks, as demonstrated by incorporation of labeled nucleotides. They were regarded as possible apoptotic nuclei, representing either a peculiar type of cell death or the preservation of the cell deletion capacity, typical of the embryonal tissue of origin.

Proliferating cell nuclear antigen (PCNA)/cyclin is currently often investigated immunohistochemically in tumors as a marker of cell proliferation, but many problems remain open concerning its reliability as a prognostic factor. PCNA has been studied in a series of 123 brain tumors using the monoclonal antibody PC10. A clear intra- and inter-tumor variability of PCNA-positive nuclei has been found, but taking into account the tumor areas with the highest number of positive nuclei, a positive correlation between this number and the histological malignancy of tumors has been demonstrated. The staining intensity of nuclei was variable; very-intensely positive nuclei, counted separately, are hypothesized to represent nuclei in S-phase of the cell cycle. In ependymomas the investigation included a quantitative statistical analysis. The number of PCNA-positive nuclei correlated with cell density and mitotic index, but only very intensely positive nuclei showed a significant statistical correlation with survival. In spite of the many possibilities of wrong interpretation of PCNA expression, the most important of which is its deregulation, the method is useful in the practice for prognostic purposes. Its important advantages are the possibility of a retrospective application and a visual analysis of the proliferation potential of tumors.

Gliomas induced in the rat by transplacental administration of ethylnitrosourea (ENU) are intensely immunoreactive for vimentin and scarcely for glial fibrillary acidic protein (GFAP). Since tumoral transformation takes place during the late fetal and early postnatal period, the sequential expression of the two glial antigens has been investigated in this age period in ENU-treated and control rats. Immunohistochemical and immunoelectron microscopical methods have been employed. Vimentin was widely expressed starting from embryonal day 14 (E 14) in the processes of radial glia; as long as radial glia was present, vimentin decorated it. GFAP was, at earliest, observed at E 20 and expressed by glial cells with a stellate, i.e., mature shape. No GFAP-positive radial process was observed. No difference was found between ENU-treated and control rats. Since ENU is most effective in producing tumors when administered at the 16-17th day of fetal life, vimentin-positive radial glia is a candidate target of ENU. The similarity of intermediate filament pattern between radial glia in the late fetal life and tumors induced by transplacental ENU suggests that radial glia might be the cell of origin.

Fisher 344 rats born from mothers treated with ethylnitrosourea (ENU) 50 mg/kg intravenously were injured at the 1st and 2nd month of extrauterine life by a transcranial stab. The wound affected cerebral cortex, white matter and basal ganglia. The animals were killed 15 and 45 days and 5 months after injury and cell reaction was studied histologically and immunohistochemically. Bromodeoxyuridine (BrdUrd) was administered 1 h before sacrifice and the labeled cells were evaluated. In ENU-treated rats injured at 1 month of age only minor differences were found in comparison with injured controls. In ENU rats injured at 2 months of age and killed 15 days later, a higher number of BrdUrd-labeled cells was found in comparison with controls; 45 days after injury the cell reaction acquired the aspect of a microtumor, however, no microtumor unrelated with the needle track was present. In ENU rats killed 5 months after the injury, there was no difference between injured and not injured ENU-treated rats, as far as the aspect and the number of tumors were concerned. The tumor phenotype was, thus, anticipated by the cell response to trauma in ENU rats. The interpretation is that the additional cell division, in response to trauma, anticipate not only the phenotypic, but also the cell kinetics changes, as indicated by BrdUrd labeling.

The occurrence and the distribution of GAGs have been studied histochemically in 224 human cerebral tumors by means of Alcian blue techniques. In the normal peritumoral gray matter the alcianophilia is stronger than in the white matter and demonstrated the presence of HA and CS. In the glioma group the alcianophilia, due to HA and CS, is mainly related to the presence of infiltrated cortex. In the other tumors, GAGs are histochemically disclosed in relation to collagen, reticulin, mesodermic areas, etc. The vessels of every tumor show a positive staining for HA, CS and HS. The histochemical findings are consistent with the biochemical ones as reported in Part I, even though the significance of GAGs in cerebral tumors remains unknown.

Neo-vascularization and endothelial hyperplasia have been shown to be very active in malignant gliomas. In this contribution the vascularization of the cortex infiltrated by malignant gliomas is morphometrically studied and the endothelial proliferations are immunohistochemically investigated and reconstructed by a three-dimensional computer-assisted procedure. Vessel density increases after tumor infiltration in some cases only. The diameter of vessels increases and so does the number of nuclei/vessel after the complete invasion of the cortex when vascular glomeruli develop. In completely infiltrated cortex with development of glomeruli and circumscribed necroses, vessel density is very low. No neoformation of vessels takes place before the complete infiltration of the cortex by the tumor. The hyperplastic formations, usually arranged parallel to the deep or outer cortical layers, take origin from the radially penetrating vessels from the meninges and their lateral branching. The hyperplasia deforms the vascular network, making it often inadequate to supply tumor cells. Immunohistochemically, the cells composing the hyperplastic structures are variably positive for factor VIII/RAg and, at a lesser extent, for alpha-smooth muscle actin. The poorness of the vascular network in many instances of completely infiltrated cortex is responsible for the development of circumscribed necroses.

Experimental cerebral tumors have been induced by transplacental ENU. The morphologic study of the brains of treated rats revealed that cellular hyperplasias appear at the 30th day of extrauterine life in the paraventricular white matter, i.e., before the already known "early neoplastic proliferations". Cytofluorimetric investigations failed to demonstrate differences between treated and control rats during the 1st month. On the contrary, adenylate cyclase activity is very high in that period. The duration of the latency period is discussed.

The immunohistochemical distribution of laminin in 70 cases of primary and secondary brain tumors has been studied. The sections were fixed in Carnoy, embedded in paraffin, and treated with collagenase before applying the specific immune serum. In normal neural tissue and in gliomas, the distribution of laminin overlaps with that of basement membrane, as is known from observations on ultrastructure in the literature. The overlapping was observed also in neurinomas, papillomas of the choroid plexus, ependymomas, and meningiomas. Problems arise in the interpretation of the results in the other oncotypes. In hemangiopericytomas and hemangioblastomas, laminin demonstrates a basement membrane around the vessels only; it is absent around stromal cells. In desmoplastic medulloblastomas, laminin forms a network among the tumor cells. In primary lymphomas, concentric rings of laminin are evidenced around some vessels. In discussing the results, the ultrastructural distribution of basement membranes is taken into account.

GAGs accumulate in the oligodendrogliomas and in the oligodendroglial component of mixed gliomas induced in rats by transplacental ENU. We studied the period between birth and the appearance of the earliest tumoral lesions and did not find any accumulation preceding the early neoplastic proliferations. These findings, together with the observation that GAGs are present only in the oligodendroglial component of tumors, are in agreement with our previous interpretation that the tumoral transformation of oligodendroglia takes place after it has acquired the ability to interfere with GAG metabolism during myelinogenesis.

35 brains with malignant gliomas treated by surgery, radiotherapy and chemotherapy and processed by means of the "complete study" technique have been studied. The preliminary results are reported. Many types of changes have been observed. Areas of coagulative necrosis with vessel alterations and teleangectasias are present around the tumor in many cases. Different pathologic aspects are observed distant from the tumor. Edema both in the homo - and contralateral hemisphere is constantly present. Severe edema with demyelination, axon disappearance and loss of cells is frequently found associated with prolonged survival. Often it is accompanied by wide or small areas filled with macrophages. Spongio-necrosis in the white matter of contralateral temporal lobe has been observed in two cases and telangiectasias areas have been found in another one. The relation between changes and treatment is discussed.

GFAP, Factor VIII/RAg, laminin, and fibronectin were immunohistochemically investigated in 15 glioblastomas and 15 gliosarcomas . GFAP was found variably positive in the glial areas. F VIII/RAg characterizes the endothelial cells and in gliosarcomas suggests the origin of the sarcomatous component from the endothelial proliferations. Laminin separates the two components and characterizes the inner and the outer basement membranes in the vessels. It is multiplied and thickened in endothelial proliferations, while it is often fragmented in the larger vessel wall proliferations. Our observations confirm that gliosarcoma represents the last stage of a process which starts with the endothelial hyperplasia of glioblastoma.

The reproliferation of glioblastomas after radiation and chemotherapy has been studied in 25 tumors by means of whole mount histological preparations. The tumors have been subdivided into four groups according to the radiation dose and the distance from the end of radiation. After 4,000 rad vessel proliferations and mitoses stop, while vessel degenerations and astrocytes with deformed nuclei appear. Six months after 6,000 rad have been delivered, repopulation phenomena are clearly evident and are mainly represented by mitoses both in parenchyma and in the vessel walls, circumscribed necroses with pseudopalisadings, proliferations of formerly degenerated vessels, etc. The brain adjacent to tumor (BAT) has a great importance since it is one of the reproliferating sites, even though it may be unrecognizable for the occurrence of radiation damages.

Experimental cerebral tumors have been induced by transplacental ENU. A systematic diachronic study of the brains has been performed starting from the 15th day of extrauterine life. The tumoral lesions firstly appear as "early stage proliferations" or oligodendroglial foci and develop as glial micro- and macrotumors or as isomorphic and polymorphic oligodendrogliomas respectively. Neurinomas appear later than glial tumors. The duration of the different lesions is sketched out. The usefulness of the model for chemotherapy studies is discussed.

The application of BCNU, CCNU and MeCCNU has been shown to greatly reduce the number of tumors induced in rats by transplacentally administered ENU and to inhibit the full development of tumor proliferation. Moreover, no neurinomas appeared after treatment with CCNU. The application of GCNU proved to be much less effective. Our statistics in the form of tables showing these findings are included and discussed.

The distribution of vimentin (VIM) has been histochemically investigated in 53 cerebral tumors and compared in gliomas to that of glial fibrillary acidic protein (GFAP). In gliomas VIM is less positive than GFAP, but shows the same distribution. It cannot be considered as indicating immaturity of glial tumor cells. VIM is also positive in glial processes of cerebellar pilocytic astrocytomas, in Schwann cells of neurinomas and in endothelial cells of all oncotypes. In medulloblastomas, VIM decorates reactive glia cells. A diffuse positive reaction has been observed in meningiomas. In hemangioblastomas, besides intervascular and endothelial cells, groups of polygonal cells are intensely positive for both VIM and GFAP. The interpretation of VIM in cerebral tumors is largely based on the distribution patterns of this intermediate filament in the developing CNS of rodents.