Page 111 - GPD-4-2
P. 111

Gene & Protein in Disease                                             Pediatric glioma circadian clock genes




                         A


















                         B





















            Figure 1. Gene expression in pediatric and adult gliomas. (A) PER2 and PER3 were not significantly different from non-tumor tissue in Griesinger et al.’s
            pediatric dataset.  Pilocytic astrocytoma-ependymoma (p<0.001) and GBM-ependymoma showed (p<0.01) significant differences in PER2 expression,
                       37
            and only GBM-ependymoma showed a significant difference in PER3 expression (p<0.05). (B) Adult GBM PER2 and PER3 expression was significantly
            lower than in non-tumor tissue (asterisks *) in the Rembrandt dataset  (p<0.001). Other tumors showing significant differences for PER2 were GBM-
                                                         40
            astrocytoma and GBM-oligodendroglioma (p<0.001); unknown-non-tumor (p<0.01); and oligodendroglioma-unknown (p<0.05). For PER3, the tumors
            included astrocytoma-unknown, GBM-oligodendroglioma, GBM-astrocytoma, unknown-non-tumor, and GBM-mixed glioma (p<0.05). Derived from
            plots and statistical tests available at the GlioVis website. 27
            significantly repressed in six of the seven GBM groups and   Pediatric GBM and LG gliomas generally showed
            seven of the ten LGG groups examined (Tables 1 and 2).   changes in gene expression in the same direction when
            RORB had a low average EI (−0.870) for datasets in all three   expression differed from normal tissue. However, the
            glioma groups (adult and pediatric GBM and pediatric LG   PER genes were not significantly repressed in pediatric
            tumors) as shown in Figure 2. CRY2 was downregulated in   gliomas, unlike in the adult GBM groups. The one low-
            15 of the 17 groups with a low average EI (−0.794). Other   grade adult glioma dataset examined (PA) did not show
            genes in the set that were highly downregulated in all three   a PER expression pattern consistent with the adult GBM
            glioma  groups  were  DBP,  HLF,  and  TEF,  which  agreed   datasets (Table  2). Furthermore, unlike in the adult
            with their similarly functioning proteins.  TIMELESS was   gliomas,  ARNTL was downregulated in six of the 11
                                            55
            unique because of its overexpression in 11 of the groups   pediatric groups and in all pediatric glioma types but
            with an average EI of 0.610. The other overexpressed genes   ependymal. It is unclear why the results for the clock gene
            include CRY1 and NFIL3, which were mostly in the adult   set differed between the two pHGG datasets. It could have
            GBM groups; PER1 and NR1D1 data were missing in many   resulted from the smaller sample sizes in the Buczkowicz
            groups (Tables 1 and 2). Note that the EI is merely a way to   dataset  (19  non-tumor,  35  tumor)  than  the  Griesinger
            summarize the expression data and shows trends in similar   dataset (13 non-tumor, 117 tumor), which might have
            departures in  expression from normal  tissue  without   reduced chances of detecting a true difference. Another
            carrying a statistical connotation of these changes.  possibility is that the Buczkowicz dataset in GlioVis was


            Volume 4 Issue 2 (2025)                         5                                doi: 10.36922/gpd.4112
   106   107   108   109   110   111   112   113   114   115   116