Fig. 6. GLI gene expression in human brain tumors, cell lines and effects of cyclopamine. (A,B) RT-PCR analyses of independent brain tumor samples. Note the varying levels of GLI1 and GLI2 expression and its general correlation with the levels of PTCH1. HFB, human fetal brain RNA, used here as control; PNETs, primitive neuroectodermal tumors. Varying levels of two GLI2 bands in B represent previously described differentially spliced forms (arrows) (Tanimura et al., 1998). Levels of expression are interpreted in relation to that of discoidin domain receptor 1 (DDR1) mRNA, used to measure the relative amount of tumor in a given sample (Weiner et al., 2000), assuming homogenous expression per tumor cell. Expression of the housekeeping gene GAPDH served as positive control to quantify the total amount of mRNA. A total of 22 DDR1-positive primary tumors were tested by RT-PCR. Of these, nine glioblastoma multiformes from the frontal, parietal or temporal lobes expressed GLI1 (9/9), GLI2 (7/9), GLI3 (9/9), PTCH1 (7/9) and SHH (7/9). Co-expression of GLI1 and SHH could suggest that the origin of these gliomas is the GLI1- and SHH-positive SVZ of the lateral ventricle (N. D., D. Lim, A. Álvarez-Buylla, A. R. A, unpublished); one gliosarcoma from the temporal lobe and one anaplastic oligodendroglioma from the parietal lobe expressed all these genes; one low grade glioma from the insula expressed GLI1, GLI3, PTCH1 but not GLI2 or SHH; four PNETs from the posterior fossa expressed GLI1, GLI3 and PTCH1; one PNET from the thalamus expressed all of these genes; and six PNETs from the cerebellum expressed all genes except SHH. Ages of the individuals with gliomas ranged from 20 to 74 years and those with PNETs from 2 to 38 years. There was no correlation of gene expression with gender. (C-H) In situ hybridization of cortical glioma (C-F) and cerebellar PNET (G,H) tumor sections showing the expression of GLI1 (C,E,G) coincident with that of PTCH1 (F,H). Sense GLI1 RNA probes did not show specific hybridization (D; six tumors tested). The localization of regions with tumor was determined by the high levels of DDR1 expression and the histopathological examination of Hematoxylin- and Eosin-stained sections (not shown). Matched slides of the same tumor are C,D, E,F and G,H. GBM, glioblastoma multiforme; LGA, low grade astrocytoma. A total of 22 DDR1-positive tumors were tested by in situ hybridization. Of these, seven GBMs from the temporal and parietal lobes expressed GLI1 (7/7) and PTCH1 (6/7); one oligodendroastrocytoma from the temporal lobe and one oligodendroglioma from the frontal lobe expressed both genes; two low grade gliomas from the cerebellum and centrum ovale expressed GLI1 (2/2) and PTCH1 (1/2); five juvenile pilocytic astrocytomas from the cerebellum, thalamus and hypothalamus expressed GLI1 (4/5) and PTCH1 (2/5); one anaplastic oligodendroglioma from the frontal lobe expressed both genes; two anaplastic astrocytomas from the frontal lobe expressed GLI1 (2/2) and PTCH1 (1/2); and three PNETs from the cerebellum, posterior fossa and occipital lobe expressed both genes. The age of individuals with glial tumors ranged from 6 to 60 years and those with PNETs from 3 to 17 years. There was no correlation of gene expression with gender. As controls, one ependymoma from the fourth ventricle and one hemangioma from the cerebellum did not express GLI1 or PTCH1. (I) RT-PCR analyses of human brain tumor cell lines testing for the expression of GLI1, PTCH1 and GAPDH. All cells tested express both GLI1 and PTCH1. (J) Percentage inhibition of the proliferation of four human glioma cell lines by cyclopamine at the concentrations indicated as measured by BrdU incorporation.