87 and the lowest median value was 0. 62. For class II, the highest median was 61. 51 and the lowest was 0. 56. HDAC11, the only member of class IV, had the lowest median value for normal brain tis sue and the highest for glioblastoma. Comparison of the HDAC mRNA levels in low and high grade gliomas and normal brain tissue new We compared mRNA levels of HDAC genes in low and high grade gliomas and also in normal brain. Among class I HDACs, significant differences in gene expression between tumor groups were not observed. Seven of 8 class II HDAC genes were expressed at lower levels in high grade astrocytomas compared to low grade astrocytomas. The same significant difference for low grade and high grade gliomas was observed for HDAC11.

We compared the most malignant form of astrocytoma with the other 3 tumors and normal brain in order to establish a correlation between HDAC expression and tumor grade. As mentioned above, no significant difference in HDAC expression was observed for class I. however, for classes II and IV, there was a decrease in expression of these genes in glioblastoma compared to that in other groups. Moreover, this downregulation appeared to fol low a pattern in which lower grade tumors had a larger number of HDAC genes at lower expression levels. Com parison between glioblastoma and grade III astrocytoma showed that 4 of 8 genes were expressed at lower levels in glioblastoma samples. Compar ison with low grade astrocytoma showed that the expression of 6 of 8 genes was lower in glioblast omas.

Finally, when we com pared glioblastoma with normal brain, 7 of 8 genes studied, with the exception of HDAC7, were expressed at lower levels in glioblastoma. Protein analysis Acetyl H3 but not Acetyl H4 correlates with mRNA levels In order to validate the data obtained from qRT PCR, western blot analysis was performed for HDAC9b protein. This protein was chosen because we found the highest lev els of mRNA expression for HDAC9b. The results obtained for HDAC9b western blot analysis confirmed the data obtained in quantitative mRNA analysis. The pro tein levels of HDAC9a were higher in normal brain tissue and low grade astrocytoma than in the grade III astrocy toma and glioblastoma. Anti acetyl histone H3 and anti acetyl histone H4 anti bodies were also used to verify the level of acetylated his tones H3 and H4 and to correlate the findings with histone deacetylase activity in the groups studied.

Considering the large number of HDAC genes with low levels of expression in glioblastomas, we expected that the levels of acetylated histones were higher in those Entinostat tumors. Interestingly, when we analyzed the acetylation levels of the H3 and H4 histones, H3 histone acetylation, but not H4 histone acetylation, correlated with the data obtained by qRT PCR. Glioblastoma samples showed higher levels of acetylated histone H3 than normal brain and low grade gliomas.