Kiyoshi Onda

Cerebral glioblastoma with cerebrospinal fluid dissemination: a clinicopathological study of 14 cases examined by complete autopsy

During the last 17 years, complete autopsies were performed on 51 patients who died of cerebral glioblastoma, and 14 were found to have dissemination by cerebrospinal fluid (CSF). In these 14 cases of glioma, the extent of intraparenchymal invasion by the primary tumor and the degree of seeding were studied in connection with histological findings and immunohistochemical staining for glial fibrillary acidic protein (GFAP) as the most reliable marker of astrocytic differentiation. From the findings obtained, the cases were divided into two groups. In one group, consisting of 7 gliomas, autopsy revealed intense seeding, despite only slight invasion by the primary tumor. Among these 7 extensively disseminated gliomas, 4 expressed almost no GFAP, 2 contained only a few GFAP-positive cells, and only 1 displayed an immunohistochemically high degree of astrocytic differentiation. Clinically, 6 of the 7 affected patients developed symptoms attributable to CSF seeding. In the other group consisting of the remaining 7 gliomas, only slight dissemination was seen, despite extensive infiltration of the primary tumor. Each of these 7 gliomas contained many GFAP-positive cells. None of the affected patients developed symptomatic seeding. This study shows the existence of two clinicopathologically distinct groups of disseminated cerebral glioblastomas and suggests that, regardless of morphological features, glioblastomas showing immunohistochemically poor astrocytic differentiation tend to shed tumor cells more vigorously but are less invasive at the primary site than those with many GFAP-positive cells. It is also suggested that, as a consequence, the former glioma type produces symptomatic seeding more frequently than the latter type.

Subcortical hematoma caused by cerebral amyloid angiopathy: does the first evidence of hemorrhage occur in the subarachnoid space?

Six autopsy cases of subcortical hematoma caused by CAA were examined to elucidate the primary site of hemorrhage. Immunohistochemistry for amyloid β‐protein (Aβ) revealed extensive CAA in the intrasulcal meningeal vessels rather than in the cerebral cortical vessels. All of the examined cases had multiple hematomas in the subarachnoid space, mainly in the cerebral sulci, as well as intracerebral hematomas. Each intracerebral hematoma was connected to the subarachnoid hematomas at the depth of cerebral sulci or through the lateral side of the cortex. There was no debris of the cerebral cortical tissue in the subarachnoid hematomas. In case 2, another solitary subarachnoid hematoma, which was not connected to any intracerebral hematoma, was seen. In all of these subarachnoid hematomas, many ruptured Aβ‐immunopositive arteries were observed. These ruptured arteries did not accompany any debris of the brain tissue, some of them were large in diameter (250–300 µm), and several of them were far from the cerebral cortex. Therefore, it was considered that they were not cortical arteries but meningeal arteries. Within the cerebral cortex, there were only a few ruptured arteries associated with small hemorrhages. There were no ruptured vessels within the intracerebral hematomas. There was a strong suggestion that all of the subarachnoid hematomas, including the solitary one in case 2, originated from the rupture of the meningeal arteries. The present study indicates that in some cases of subcortical hematoma caused by CAA, the primary hemorrhage occurs in the subarachnoid space, in particular the cerebral sulci, because of rupture of multiple meningeal arteries. Infarction occurs subsequently in the cortex around the hematoma, the hematoma penetrates into the brain parenchyma, and finally, a subcortical hematoma is formed.

Intracranial malignant lymphomas: clinicopathological study of 26 autopsy cases.

We examined 26 autopsy-proven cases of intracranial malignant lymphoma (IML) in immunocompetent patients to determine the extent of neoplastic involvement of the central nervous system (CNS) and to evaluate the effects of radiation on the tumor and brain tissue. All tumors were identified as diffuse non-Hodgkins lymphomas of B-cell origin. In six patients who had not received radiotherapy, the clinical course of the disease was short and extensive infiltration of the tumor was seen. The remaining 20 patients were treated with radiotherapy and had a longer survival time. Leptomeningeal involvement was common, but extensive subarachnoid proliferation of the tumor was seen in only two cases. The posterior, but not anterior, lobe of the pituitary was involved in 5 of 22 cases, and choroid plexus involvement was seen in 4 of 21. Direct invasion of the tumor into the spinal cord, which tended to occur in patients with posterior fossa masses, was observed in 5 of 21 cases. Following irradiation, coagulation necrosis was frequently found in the invading zone as well as in the tumor mass, and degeneration of the white matter was also seen. We suggest that IML can extensively infiltrate into the CNS, including the posterior lobe of the pituitary and spinal cord, and that radiation injury to the brain appears to occur relatively easily in this disease.

A Comparative Study of Glioma Cell Lines for p16, p15, p53 and p21 Gene Alterations

A total of 10 glioma cell lines were examined for alterations of the p16, p15, p53 and p21 genes, which are tumor suppressor genes or candidates with direct or indirect CDK‐inhibitory functions. Genetic alterations (deletions or mutations) were frequently seen in the p16, p15 and p53 genes in these cell lines, but not in the p21 gene. When the states of the p16, p15 and p53 genes were compared among cell lines, all the cell lines showed abnormalities in at least 1 gene, often in 2 or 3 genes coincidentally, suggesting that dysfunction of these genes is closely related to glioma cell growth. Although alteration of all 3 genes was most frequent, there were cell lines having either p16/p15 or p53 or p16 and p53 gene alterations, suggesting that the time order of these genetic alterations was variable depending on the cell line. Among cell lines examined, one with homozygous p53 gene deletion seemed of particular practical value, since such a cell line might be useful in various studies, including investigation of the functions of various mutant p53 genes in the absence of heteromeric protein formation. On examination of the primary tumor tissues, the same alterations of the p16/p15 and p53 genes as detected in the cell lines were demonstrated in all 6 cases examined: p16/p15 gene deletion in 1, p16 gene mutation in 1 and p53 gene mutations in 5 cases. This suggested that the p16/p15 and the p53 gene alterations and their combinations in at least some glioma cell lines reflected those in the primary glioma tissues.

Symptomatic cerebrospinal fluid dissemination of cerebral glioblastoma

SummaryComputed tomography (CT) findings in eleven patients with symptomatic cerebrospinal fluid (CSF) dissemination from cerebral glioblastoma were analyzed and, in seven cases subsequently autopsied, they were compared with histological observations. Each patient had multiple CT abnormalities including periventricular enhancement (5/11), subarachnoid enhancement (10/11) and progressive hydrocephalus (7/9) by cranial CT, and small filling defects with or without block (5/5) by CT myelography. The areas that showed periventricular or subarachnoid enhancement on CT were confirmed to have macroscopically detectable seeding at autopsy. On the other hand, microscopic deposits were more widely distributed than the enhancement suggested, and were hardly visualized on CT. In association with subarachnoid seeding, we found low-density lesions on CT which had resulted from ischemia or reinvasion of adjacent structures by disseminated glioblastoma and resulting parenchymal edema. By cranial CT, subarachnoid enhancement seems to be a very reliable sign of CSF seeding, whereas periventricular enhancement due to CSF metastases should be carefully distinguished from that due to periventricular tumor infiltration. CT myelography is capable of revealing minute metastatic spinal deposits and may be helpful for ruling out spinal seeding as well as its precise evaluation.

The hemorrhage caused by sporadic-type cerebral amyloid angiopathy occurs primarily in the cerebral sulci.

We examined a solitary hematoma in a patient with sporadic cerebral amyloid angiopathy (CAA). The hematoma affected the middle frontal sulcus, cerebral cortex (CC) and subcortical frontal white matter (sfWM). We embedded the hematoma in four paraffin blocks, each of which was cut serially into 6‐µm‐thick sections. The first section and every 18th section from each block were subjected to Elastica‐Goldner (E‐G) staining, and the distribution and diameter of the ruptured blood vessels (rBVs) were examined. The rBVs were then marked on diagrams representing each E‐G‐stained section. The present study yielded the following important findings: (i), early‐ and recently ruptured Aβ‐positive arteries were present mainly in the intrasulcal hematoma (ISH), rather than in the CC; (ii) many early‐ruptured arteries in the ISH were larger in diameter than those in the CC; and (iii) ruptures of the cortical arteries, even near the cortical surface, did not occur so frequently and the ruptured vessels were small in size. We concluded that in patients with subcortical hematoma caused by sporadic‐type CAA, successive rupturse of the meningeal vessels, mainly arteries, occur in the cerebral sulcus initially, followed by formation of an ISH and development of a fresh hemorrhagic or anemic infarct in the CC surrounding the ISH, the latter in most cases then extending into the brain parenchyma through the necrotic CC at the depth of the sulcus, finally creating a secondary hematoma in the subcortical white matter.

Spinal Metastases of Cerebral Glioblastoma: The Value of Computed Tomographic Metrizamide Myelography in the Diagnosis

Spinal metastases from cerebral glioblastoma via the cerebrospinal pathway are rarely detected when the primary tumors are under apparent control. The authors report two adult patients with cerebral glioblastoma who developed spinal symptoms referable to spinal seeding without neurological and computed tomographic findings of the recurrence of the primary tumors. Computed tomographic metrizamide myelography clearly revealed minute deposits of perispinal metastatic tumors that could not be detected by conventional myelography. Even perispinal mass lesions so minute that they are revealed only by computed tomographic metrizamide myelography can invade the spinal cord and cause clinical symptoms.

Correlation of DNA ploidy and morphological features of human glioma cell cultures with the establishment of cell lines

SummaryEleven gliomas were serially cultivated and examined for DNA distribution by flow cytometry and simultaneously for morphological features by light microscopy at the various passage levels until passage 50 at most. Seven gliomas (four low-grade gliomas, three anaplastic gliomas) showed a similar DNA distribution pattern with a main diploid and small tetraploid peaks at various passages. In this group, only one culture formed a permanent cell line, whereas six cultures showed a limited growth ranging from 6 to 24 passages. In contrast, the other four gliomas (each an anaplastic glioma) showed a marked change of DNA distribution through passages and finally a single DNA aneuploid population prospered. Each of these four gliomas yielded established cell lines. Thus, it is suggested that the change of DNA ploidy and prosperity of DNA aneuploid populations in flow cytometry might be used as early and reliable indices for the later establishment of glioma-derived permanent cell lines. Since the changes of DNA distribution are frequently associated with the morphological changes, as seen in the latter group, careful tracing of morphological features is valuable in determining of the fate of cultures, especially in the absence of a flow cytometer. The correlation between the potential to become established cell lines and histology of the original gliomas is also discussed.

Proliferative potential of recurrent intracranial meningiomas as evaluated by labelling indices of BUdR and Ki-67, and tumour doubling time.

Summary This study was designed to provide the reciprocal relationship among labelling indices of 5-bromodeoxyuridine (BUdR LI), Ki-67 (Ki-LI), and tumour doubling time (Td) of recurrent meningiomas. In our series of 182 primary intracranial meningiomas, 46 cases recurred. The average of BUdR LI and Ki LI for nonrecurrent meningiomas were 0.77±0.13% and 4.71±1.96%, respectively. Recurrent meningiomas had significantly higher LIs at the first operation: BUdR LI was 3.77±1.22% and Ki LI was 14.78±3.17%. The recurrent ratio significantly increased with the degrees of each LI. And the linear regression analysis has demonstrated a significant correlation between BUdR and Ki LI. Td was calculated accurately by NIH, a computer software. Td showed a significant inverse correlation with each of the labelling indices. Consequently, BUdR, Ki LIs and Td of individual tumours correlate mutually well. Of the 46 recurrent cases, 4 received radiation after the operation. Td of the irradiated meningiomas tended to be longer than expected for their higher level of BUdR and Ki LIs before radiation therapy. Thus, it was shown that the radiation therapy delays the regrowth of meningiomas.

Differential expression of glial fibrillary acidic protein in human glioma cell lines.

SummaryWe have obtained a cDNA fragment to human glial fibrillary acidic protein (GFAP) by immunoscreening a λgt11 human brain cDNA library with antibody to bovine GFAP. The highly homologous nucleotide sequence of this clone with that of the mouse GFAP enabled the identification of this cDNA as one encoding GFAP. As this cDNA hybridized with a single major RNA species in Northern blots of RNA from human and mouse brain tissues and gave one or two bands in Southern blots of human genomic DNA, it was considered to be specific for GFAP. Using this cDNA as a probe we investigated the levels of GFAP expression in ten human glioma cell lines. A 3.5-kb GFAP mRNA was detected in five of the ten glioma cell lines, one of which was U-251 MG cell line and the other four were clones derived from the same tumor (CL1, 2, 3, and 4). There was a difference in the amount of GFAP mRNA among U-251 MG and the four clonal cell lines. Quantitative evaluation of this difference by RNA dot blot analysis revealed that the amount of GFAP mRNA expressed in CL3 was about 1/5 and in CL4 about 1/10 the amount expressed in U-251 MG, CL1, and CL2. Semiquantitative Western blot analysis showed that GFAP levels corresponded to the GFAP mRNA levels in these cell lines. By Southern blot analysis of genomic DNA the GFAP gene was similarly detected in all of these cell lines regardless of the level of GFAP expression. Thus, by using a cDNa to human GFAP we have demonstrated the presence of clonal cell lines from human glioma showing different levels of GFAP expression, which may provide a useful basis for further investigations on the regulation of GFAP gene expression in glial cells.