Oligodendroglial neoplasms are a subgroup of gliomas with distinctive morphological characteristics. In the present study we have evaluated a series of these tumors to define their molecular profiles and to determine whether there is a relationship between molecular genetic parameters and histological pattern in this tumor type. Loss of heterozygosity (LOH) for 1p and 19q was seen in 17/23 (74%) well-differentiated oligodendrogliomas, in 18/23 (83%) anaplastic oligodendrogliomas, and in 3/8 (38%) oligoastrocytomas grades II and III. LOH for 17p and/or mutations of the TP53 gene occurred in 14 of these 55 tumors. Only one of the 14 cases with 17p LOH/TP53 gene mutation also had LOH for 1p and 19q, and significant astrocytic elements were seen histologically in the majority of these 14 tumors. LOH for 9p and/or deletion of the CDKN2A gene occurred in 15 of these 55 tumors, and 11 of these cases were among the 24 (42%) anaplastic oligodendrogliomas. Comparative genomic hybridization (CGH) identified the majority of cases with 1p and 19q loss and, in addition, showed frequent loss of chromosomes 4, 14, 15, and 18. These findings demonstrate that oligodendroglial neoplasms usually have loss of 1p and 19q whereas astrocytomas of the progressive type frequently contain mutations of the TP53 gene, and that 9p loss and CDKN2A deletions are associated with progression from well-differentiated to anaplastic oligodendrogliomas.

De novo glioblastomas develop in older patients without prior clinical history of less malignant tumors. Progressive glioblastomas are common among younger patients and arise through progression from lower-grade astrocytomas. CDKN2A deletions, PTEN alterations, and EGFR amplification are more prevalent among de novo glioblastomas, whereas p53 mutations are more common among progressive glioblastomas. Loss of heterozygosity (LOH) for chromosome 10 is seen uniformly among both de novo and progressive high-grade astrocytomas. The inactivation of the PTEN gene is found in approximately 30% to 40% of astrocytomas with chromosome 10 loss, and LOH pattern in the remaining astrocytomas strongly supports the presence of another yet unidentified tumor suppressor gene telomeric to PTEN. More than 80% of oligodendrogliomas exhibit LOH for 1 p and 19q alleles. Oligoastrocytomas with 1p/19q LOH are related to oligodendrogliomas, and those with p53 mutations are related to astrocytomas.

Gastrin and its carboxyl-terminal homolog cholecystokinin (CCK) exert a variety of biological actions in the brain and gastrointestinal tract that are mediated in part through one or more G protein-coupled receptors which exhibit similar affinity for both peptides. Genomic clones encoding a human gastrin/CCKB receptor were isolated by screening a human EMBL phage library with a partial-length DNA fragment which was based on the nucleotide sequence of the canine gastrin receptor. The gene contained a 1356-bp open reading frame consisting of five exons interrupted by 4 introns and was assigned to human chromosome llpl5.4. A region of exon 4, which encodes a portion of the putative third intracellular loop, appears to be alternatively spliced to yield two different mRNAs, one containing (452 amino acids; long isoform) and the other lacking (447 amino acids; short isoform) the pentapeptide sequence Gly-Gly-Ala-Gly-Pro. The two receptor isoforms may contribute to functional differences in gastrin-and CCK-mediated signal transduction. (12,13). They encode homologous proteins of similar length (453 and 450 aa, respectively) with seven deduced transmembrane regions, comparable to other G protein-coupled receptors. The recombinant receptors, when transfected into COS-7 cells, manifest high (nanomolar) affinities for both gastrin and CCK and are coupled to an increase in intracellular calcium (12,13). Structurally, the canine gastrin receptor displays nearly 90% amino acid identity with the putative CCKB receptors from human and rat (9-11).The sequence and organization of the full-length genes encoding members of the gastrin/CCK receptor family have not been identified. Such information is essential for analysis of their transcriptional regulation and structural relationships to other receptor genes. We have cloned and sequenced a gastrin/CCKB receptor gene from a human genomic library**. The gene appears to be alternatively spliced to yield two different receptor isoforms which differ by 5 aa in the putative third intracellular loop. The gastrin/CCKB receptor gene was assigned to human chromosome 11p15.4 by fluorescence in situ hybridization. METHODSGenomic Library Screening and DNA Sequencing. A human genomic library in the bacteriophage AEMBL-3 (Clontech) was screened with a 32P-labeled cDNA fragment generated by the polymerase chain reaction (PCR) with primers based on the cDNA sequence of the canine gastrin receptor (see below). Two positive clones were digested with Sst I and the resulting restriction fragments of length 0.7-3.2 kb (Fig. 1) were subcloned into phage M13mpl8 or -mpl9 and sequenced in both directions by the dideoxy method (14). In Abbreviation: CCK, cholecystokinin. tPresent address:

Cytogenetic and molecular genetic studies of glioblastoma multiforme (GBM) have shown that the most frequent alterations are gains of chromosome 7, losses of 9p loci and chromosome 10, and gene amplification, primarily of the epidermal growth factor receptor (EGFR) gene. Although this profile is potentially useful in distinguishing GBM from other tumor types, the techniques used tend to be labor intensive, and some can detect only gains or losses of genetic loci. Comparative genomic hybridization (CGH) is a powerful technique capable of identifying both gains and losses of DNA sequences. The present study compares the CGH evaluation of 22 GBM with classic cytogenetics, loss of heterozygosity by allelotyping, and gene amplification by Southern blot analysis to determine the reliability of CGH in the genetic characterization of GBM. The CGH and karyotypic data were consistent in showing gain of chromosome 7 accompanied by a loss of chromosome 10 as the most frequent abnormality, followed by a loss of 9p in 17 of 22 GBM cases. Loss of heterozygosity of chromosomes 10 (19/22) and 9p (9/22) loci confirmed the underrepresentation by CGH. Genomic amplifications were observed by CGH in 5 of the 10 cases where gene amplification was detected by Southern blot analysis. The data show that CGH is equally reliable, compared with the more established genetic methods, for recognizing the prominent genetic alterations associated with GBM and support its use as a plausible adjunct to glioma classification.