Extensive intra-tumor heterogeneity in primary human glial tumors as a result of locus non-specific genomic alterations.

Genomic changes are a hallmark of the neoplastic process. These range from alterations at specific loci and defined karyotypic changes which influence tumor behavior to generalized alterations exemplified by microsatellite instability. Generalized genomic changes within a tumor would be evidence in favor of the mutator hypothesis which postulates a role for such extensive changes during tumorigenesis. In this report, we have used the DNA fingerprinting technique of randomly amplified polymorphic DNA (RAPD) analysis to study genomic alterations within primary human astrocytic tumors (gliomas) in a locus non-specific manner. The RAPD fingerprinting profile of consecutive segments of tumors 2 mm across was studied; 17 astrocytic (high- and low-grade) tumors were sectioned end to end. Tissue from 50 consecutive sections, 40 microm thick (total 2 mm across), was pooled and taken to be a tumor compartment. DNA was subjected to RAPD amplification by 15 random 10-mer primers. A tumor segment was taken to have a DNA fingerprinting pattern different from others in the same specimen when its RAPD profile differed from others by at least one band of one RAPD reaction. All but one of the tumors showed compartments with a unique genetic profile, indicating genomic instability leading to widespread intra-tumor genetic heterogeneity. Eight tumors were also studied for loss of heterozygosity (LOH) of the p53 and D17S379 loci in the different segments as examples of alteration of specific tumor influencing loci. Three showed LOH of p53, which was limited to only one compartment of each tumor. The extensive intra-tumor genetic instability detected in this study is suggestive of the overall high rate of change in the genomes of tumors including those of a lower grade. It is hypothesized that some of these altered clones, which manifest as zones of heterogeneity in a solid tumor, may accumulate changes at loci known to influence tumor behavior, and thus clinical outcome.