No abstract

The aim of this study is to review the different histological and clinical characteristics of glioblastoma multiforme (GBM) with and without cysts (cystic and noncystic GBM, respectively). Thirty-seven GBM were collected; these were tumors for which more than 80% of the volume was surgically resected, including a portion of the peripheral parenchyma of the brain. Based on preoperative magnetic resonance (MR) imaging studies, tumors were tentatively classified as cystic GBM if more than 50% of their volume appeared to be liquid; otherwise, they were considered to be noncystic GBM. Tumor volumes were estimated from contrast-enhanced T1-weighted MR images. Edema was deduced from the maximum width of contrast-enhanced edges. Peritumoral pathological analysis showed distinct margins, indicating little or no infiltration of tumor cells into white matter. Five cases were classified as cystic and 32 were noncystic GBMs. There was a statistically significant difference in age (Mann-Whitney U test; P < 0.05) between the patients with cystic tumors (median, 44 years; range, 26-59 years) and those with noncystic tumors (median, 54 years; range, 26-81 years). Four of the cystic tumors and eight of the noncystic tumors were more than 5 cm in maximum diameter. Cystic GBMs had a well-defined tumor interface and less than 2-cm-thick peritumoral edema compared to the noncystic GBMs (Fisher's exact test; P < 0.05). For patients with cystic GBMs, median survival time after surgery was 19.8 months and the 2-year survival rate was 50%. Patients with noncystic GBMs had a median survival time of 12.8 months and a 2-year survival rate of only 17%. Median time to tumor recurrence was 13.3 months for patients harboring cystic GBMs and 8.5 months for those with noncystic GBMs (log-rank test; P < 0.05). Thus, the prognosis for cystic GBM was significantly better than that for noncystic GBM, possibly because cystic GBMs showed comparatively little infiltration of the peritumoral brain parenchyma.

Although p53 overexpression detected by immunohistochemistry has been reported in pituitary adenomas and carcinomas, genetic mutations in the p53 gene have not been previously detected in these tumors. We analyzed a series of eight pituitary adenomas and six pituitary carcinomas by immunohistochemistry, polymerase chain reaction amplification, and sequencing of p53 exon 5 through exon 8 for genetic mutations. Three carcinomas showed more than 20% expression of p53 protein in the tumor cells. One of these tumors with 60% overexpression of p53 protein had a mutation in codon 248, a common "hot spot" for p53 mutation, while the other carcinoma with 90% overexpression of p53 protein had a mutation in codon 135. All adenomas were negative for p53 mutations and had 15% of the cells expressing the p53 protein. Analysis of control tumors including four lung carcinomas with proven p53 mutations also had greater than 85% of the tumor cells overexpressing p53 protein. Two breast carcinoma cell lines with known p53 mutations, MBA-MD 231 and MBA-MD-486, also showed greater than 85% of the tumor cells overexpressing p53. These results show that p53 mutations are present in a subset of pituitary carcinomas and are usually associated with a high percentage of tumor cells overexpressing the p53 protein.

Galectin-3 is expressed in a cell-type specific manner in human pituitary tumors and may have a role in pituitary tumor development. In this study, we hypothesized that Galectin-3 is regulated by RUNX proteins in pituitary tumors. Transcription factor prediction programs revealed several putative binding sites in the LGALS3 (Galectin-3 gene) promoter region. A human pituitary cell line HP75 was used as a model to study LGALS3 and RUNX interactions using Chromatin immunoprecipitation assay and electrophoresis mobility shift assay. Two binding sites for RUNX1 and one binding site for RUNX2 were identified in the LGALS3 promoter region.LGALS3 promoter was further cloned into a luciferase reporter, and the experiments showed that both RUNX1 and RUNX2 upregulated LGALS3. Knock-down of either RUNX1 or RUNX2 by siRNA resulted in a significant downregulation of Galectin-3 expression and decreased cell proliferation Correspondence to: Ricardo V. Lloyd, lloyd.ricardo@mayo.edu. NIH Public Access Author ManuscriptEndocrine. Author manuscript; available in PMC 2010 August 25.

Intracranial germinomas are accompanied occasionally by a significant granulomatous change and abundant fibrous tissue, and this has made their pathological diagnosis difficult. However, the incidence of the granulomatous reaction and the presence of fibrous tissue together with their clinical characteristics in intracranial germinomas have not been fully investigated. Twenty-four germinomas, none of which had received preoperative treatment, were clinicopathologically examined. The location of the tumor was the pineal region (5 cases), the suprasellar region (13 cases), multiple lesions (2 cases), the basal ganglia region (1 case), and other regions (3 cases). Histologically, the germinomas could be divided into two types: (1) type A (18 cases) consisted mainly of large neoplastic cells and small lymphocytes, showing a two-cell pattern; (2) type B (6 cases) consisted predominantly of fibroinflammatory tissues containing occasional neoplastic cells (5/6 cases) and, rarely, neoplastic cells (1/6 cases). Perioperatively, two-cell-pattern germinomas (type A) were characterized as soft tumors and fibroinflammatory germinomas (type B) as hard tumors. Thus, the fibroinflammatory type B accounted for 25% of the intracranial germinomas. Although there were no topographical and clinical differences between the two types, we conclude that immunohistological studies to detect neoplastic germ cells are warranted in cases of small stereotactic biopsies of hard type B tumors.