Search

Additional file 1: Figure S1. (A) Differentiation of mitochondria-related genes in A172 TMZ resistance cells (A172-R) with shSOD2 or shControl is shown in the heatmap graphs. A total of 84 genes was assessible from the assay. Among them, however, only 71 of the items were found in the database of mitochondria-related genes according to MitoCarta 2.0. (B) The correlation between SH3GLB1 and SOD2 was also shown in TCGA-GBM dataset and the 14 paired patients. (C) Nine cases showing expression of SH3GLB1 in the paired primary- and recurrent-tissues using IHC staining (100x and 400x magnification). (D) Kaplan-Meier curves of TCGA-GBMLGG (GBM and low-grade glioma) database showed higher SH3GLB1 levels caused poor survival. Figure S2. Ingenuity Pathway Analysis was applied for analysis of the major mitochondria-related molecular and cellular functional alterations in the transcriptome data from 14 RNA-seq data from paired recurrent and treatment-naïve high-grade glioma samples. As shown in the figure, sample numbers of down- (left column) or up-regulated (right column) genes. In each grid, the color is determined by the rank percentile, and the number represents sample that fulfills the criteria. The levels and significance of the genes ratio from CGGA glioma database (recurrent versus primary) are also shown aside. Note that none of the genes in complex II had CGGA database significance better than **, and the genes in complex V and CoQ synthesis was not higher in the resistant groups in the 14 paired samples. ***p < 0.001, **p < 0.01, *p < 0.05, NS: not significant. NA: data not available. Figure S3. (A) Five naïve glioblastoma tumor samples were used for single-cell transcriptome and were sorted into nine clusters according to their gene expression (Fig. 2A). The heatmap graph shows the common related genes of GBM and tumor-initiating cells among the clusters. Note that clusters 3, 7, 8, and 9 are identified as immunocytes because of the presence of the markers such as CD3E/G, GZMA, ITGAM, ITGAX. (B) The clusters reclassified by the studies of Neftel et al. (Cell, 2019) were divided into four cellular states. Figure S4. (A) Cell metabolic phenotype profile were measured using Seahorse Analyzer (OCR and ECAR). (B) The individual OCR parameters were measured using the indicated reagents and plotted in Fig. 5B, with the statistic graphs showing the parameters of maximal respiration, spare respiratory capacity, and nonmitochondrial respiration. (C) The parental cells transfected with SH3GLB1-overexpressing vector were stained with a JC-1 dye, showing the enhanced fluorescent ratio. Figure S5. (A) IACS-010759 effect to cell density for 72 hours of the parental cells with or without SH3GLB1 overexpression are shown in the bar graph. (B) IACS-010759 effect to cell density for 72 hours of the resistant cells with or without SH3GLB1 knockdown are shown in the bar graph. (*P < 0.05) Figure S6. (A) The U87MG-R luciferase-carrying cells (U87MG-R-luc-EGFP) were transfected with shSH3GLB1 used in animal experiment. SOD2 levels did not be affected by SH3GLB1 deficiency. (B) The tumor volume was measured at Day 20 and Day 23 according to the National Cancer Institute formula. (C) The body weight was recorded at the indicated day as the statistic graph. (*P < 0.05) (D) The image shows the tumor extracted from euthanized mice (scale bar = 1 cm) (E) The SH3GLB1 expression in the tissue using an IHC staining is shown (scale bar = 200 μm). (F) The body weight was recorded and shown in the statistical graph. Figure S7. Kaplan-Meier curves of GBM database from CGGA showed the survival of the indicated subunits genes of OXPHOS complex. Figure S8. A172 or A172-R cells are used in the results including roles of SH3GLB1 on Sp1 promoter, Sp1 expression, 36 autophagy levels, OXPHOS levels, cell density and OCR analysis. Figure S9. Pt#5 or Pt#5-R cells are used in the results including roles of SH3GLB1 on Sp1 promoter, Sp1 expression, autophagy levels, OXPHOS levels, cell density and OCR analysis.

Additional file 1. Additional figures and tables.

Additional file 1: Figure S1. MPT0B291 commonly alters two lncRNAs (LINC00461 and LINC01559) in stem-like and TMZ-resistant U87MG cells. Volcano plots show overlapping lncRNAs between two microarray data (MS1040791 and MS1060987). Relative expression level (log2) to control (DMSO) of each lncRNA is shown in the x-axis. y-axis indicates SNR (signal-to-noise ratio) value (log2) of each lncRNA. Genes with significant upregulation (red) and downregulation (green) in both microarray datasets are highlighted. Figure S2. MPT0B291 decreases Ki-67 expression in vivo. (A) IHC detected the protein expressions of proliferation (Ki-67) and apoptosis (cleaved caspase-3) marker. Serial sections of control and MPT0B291-treated xenografts were hybridized with either an antibody against Ki-67 (upper two panels) or cleaved caspases-3 (lower two panels), and cell nuclei were stained with hematoxylin. Scale bars indicate 20 ��m and 200 ��m. Mean intensity of Ki-67 and cleaved caspases-3 staining in the cells of control and MPT0B291-treated xenografts were semi-quantitative determined. Unpaired Student���s t-test. Figure S3. Interaction between HDAC6 and RNA-binding proteins (CNOT6 and FUS). (A) Flag-tagged CNOT6-expressed U87MG cells were treated with 10 ��M MPT0B291, 10 ��M Trichostatin A (TSA), or DMSO for 2 h. Flag-CNOT6 protein was immunoprecipitated and the precipitated samples were then analyzed by immunoblotting analysis with antibodies of HDAC6 or Flag-M2. (B) Enrichment of HuR from the lysate by desthiobiotin-labeled androgen receptor (AR) 3���-UTR validates the efficiency of RNA-protein pull-down assay (left panel). Protein lystaes from TMZ-resistant Pt#3 cells were also used for the RNA-protein pull-down assay with desthiobiotin-labeled LINC00461, and both poly(A)-binding protein (PABP) and HuR were slightly detected in the pull-down sample, but desthiobiotin-labeled LINC00461 did not bind HDAC6 directly. (L = lysate; FT1 = flow-through in first wash; FT2 = flow-through in second wash; E = eluate). (C) Quantification of CNOT6 and FUS protein expression from the parental and TMZ-resistant Pt#3 cells either treated with MPT0B291 or si-HDAC6. Unpaired Student���s t-test. The results are shown as mean �� SEM for triplicate samples in each group. Figure S4. Cell division-related genes show poor prognostic implications in high-grade glioma patients. (A) Kaplan���Meier curves compare the survival outcomes in high-grade glioma patients with high (red) and low (green) expressions of cell division-related genes. All survival curves were obtained from PROGgeneV2. Log-rank test. Figure S5. Functional analysis of the conserved markers in cluster 5 from scRNA-seq data reveals gene classes associated with cell division and survival. Core analysis using IPA revealed the top ten molecular and cellular functions of highly conserved marker genes in cluster 5 of patient-derived GBM cells. Log(p-value) indicates the significance of enrichment for highly expressed marker genes from our dataset. The threshold for significance was set at a p-value