Intratumor heterogeneity is a primary feature of high-grade
gliomas, complicating their
therapy. As accumulating evidence suggests that intratumor heterogeneity is a consequence of cellular subsets with different cycling frequencies, we developed a method for transcriptional profiling of
gliomas, using a novel technique to dissect the
tumors into two fundamental cellular subsets, namely, the proliferating and non-proliferating cell fractions. The
tumor fractions were sorted whilst maintaining their molecular integrity, by incorporating the
thymidine analog
5-ethynyl-2'-deoxyuridine into actively dividing cells. We sorted the actively dividing versus non-dividing cells from cultured
glioma cells, and parental and clonally derived orthotopic
tumors, and analyzed them for a number of transcripts. While there was no significant difference in the transcriptional profiles between the two cellular subsets in cultured
glioma cells, we demonstrate ∼2-6 fold increase in transcripts of
cancer and neuronal stem cell and
tumor cell migration/invasion markers, and ∼2-fold decrease in transcripts of markers of
hypoxia and their target genes, in the dividing
tumor cells of the orthotopic
glioma when compared to their non-proliferative counterparts. This suggests the influence of the brain microenvironment in transcriptional regulation and, thereby, the physiology of
glioma cells in vivo. When clonal
glioma cells were derived from a parental
glioma and the resultant orthotopic
tumors were compared, their transcriptional profiles were closely correlated to
tumor aggression and consequently, survival of the experimental animals. This study demonstrates the resolution of intratumor heterogeneity for profiling studies based on cell proliferation, a defining feature of
cancers, with implications for treatment design.