Oncotarget, 2015, Vol. 6, No. 28, p. 26192-26215

Combined expressional analysis, bioinformatics and targeted proteomics identify new potential therapeutic targets in glioblastoma stem cells

 

Biljana Stangeland1,2, Awais A. Mughal1, Zanina Grieg1,4, Cecilie Jonsgar Sandberg1, Mrinal Joel1,4,5, Ståle Nygård3, Torstein Meling1, Wayne Murrell1, Einar O. Vik Mo1, Iver A. Langmoen1,2,4

1 Vilhelm Magnus Laboratory for Neurosurgical Research, Institute for Surgical Research and Department of Neurosurgery, Oslo University Hospital, Oslo, Norway

2 SFI-CAST Biomedical Innovation Center, Oslo University Hospital, Oslo, Norway

3 Bioinformatics Core Facility, Institute for Medical Informatics, Oslo University Hospital and University of Oslo, Oslo, Norway

4 Norwegian Center for Stem Cell Research, Department of Immunology and Transfusion Medicine, Oslo University Hospital, Oslo, Norway

5 Laboratory of Neural Development and Optical Recording (NDEVOR), Department of Physiology, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway

 

Correspondence to: Biljana Stangeland, e-mail: Biljana.Stangeland@labmed.uio.no

Keywords: glioblastoma, GBM, glioblastoma stem cells, GSCs, therapeutic targeting

 

ABSTRACT

Glioblastoma (GBM) is both the most common and the most lethal primary brain tumor. It is thought that GBM stem cells (GSCs) are critically important in resistance to therapy. Therefore, there is a strong rationale to target these cells in order to develop new molecular therapies.

To identify molecular targets in GSCs, we compared gene expression in GSCs to that in neural stem cells (NSCs) from the adult human brain, using microarrays. Bioinformatic filtering identified 20 genes (PBK/TOPK, CENPA, KIF15, DEPDC1, CDC6, DLG7/DLGAP5/HURP, KIF18A, EZH2, HMMR/RHAMM/CD168, NOL4, MPP6, MDM1, RAPGEF4, RHBDD1, FNDC3B, FILIP1L, MCC, ATXN7L4/ATXN7L1, P2RY5/LPAR6 and FAM118A) that were consistently expressed in GSC cultures and consistently not expressed in NSC cultures. The expression of these genes was confirmed in clinical samples (TCGA and REMBRANDT). The first nine genes were highly co-expressed in all GBM subtypes and were part of the same protein-protein interaction network. Furthermore, their combined up-regulation correlated negatively with patient survival in the mesenchymal GBM subtype. Using targeted proteomics and the COGNOSCENTE database we linked these genes to GBM signalling pathways.

Nine genes: PBK, CENPA, KIF15, DEPDC1, CDC6, DLG7, KIF18A, EZH2 and HMMR should be further explored as targets for treatment of GBM.

KEYWORDS: GBM; GSCs; glioblastoma; glioblastoma stem cells; therapeutic targeting

PMID: 26295306

 

Additional information:

Immune therapy and related medicines present a promising new opportunity for the treatment of variety of malignancies. Identification of novel potential molecular targets is therefore an imperative. Glioblastoma (GBM) is one of the deadliest human cancers that cannot be successfully treated with traditional therapies that include combined surgery, chemotherapy and radiation. The average patient survival after diagnosis is 14 months. We have previously compared gene expression in glioblastoma stem/initiating cells (GSCs) with normal adult brain-derived cells with stem-like properties (NSCs) using microarrays [1]. This study identified genes and pathways that were differentially regulated in GSCs compared with NSCs. In our recent study, that is highlighted here, we used combined expressional analysis, bioinformatics and targeted proteomics to identify genes suitable for immune therapy of GBM [2]. The most important finding in this study is a discovery of 9 gene co-expression module highly up-regulated in GSCs that impacts patient survival. A flow-chart of our study is shown in Figure 1.

 

Figure 1 Additional information small

Figure 1. A flow-chart of the study.

 

Firstly, potential candidate genes were identified using bioinformatic analysis of the expression of ~33000 genes represented on an Illumina microarray chip. We used rather simple mathematical filtering to select genes expressed in GSCs (log2 of the expression value>0 on a logarithmic scale) and those whose expression was very low in normal stem cells (log2 of the expression value<0 on the logarithmic scale). This analysis yielded only 20 genes. To confirm the expression of the selected candidate genes at mRNA and protein levels we used qPCR, western blot and immunolabeling. For qPCR analysis we used an extensive number of controls. These controls included both primary NSCs and a neural fetal cell line (NFCs). NSCs were in addition cultured as (a) spheres, as (b) an adherent monolayer on retronectin [3] and as (c) an adherent monolayer in 1% serum (Failsafe medium[4]). For all cultured cells we determined the cell population doubling time. GSCs had cell PDT of 2-7 days. The cell PDT for normal controls was in the same range. By qPCR we could confirm the differential expression of   15 genes (Figure 2).

Western analysis was performed using ~90 different antibodies. As many of the antibodies displayed non-specific binding this study ended up being very expensive and time consuming. In 2010 we could study the expression of very few of the selected proteins, as specific antibodies were not available yet. Fortunately, as the study progressed new antibodies were continuously becoming available. It took almost four years to complete the protein analysis collecting antibodies of sufficient specificity for the selected 20 candidates. By western blot we could finally confirm the differential regulation of 15 proteins (Figure 2).

 

 Figure 2 Additional information

Figure 2. Summary of expression analysis. (a) Gene expression of the selected 20 genes has been analyzed using several experimental methods and bioinformatics analysis. (b) Venn diagram showing the number of genes that were confirmed with the experimental methods: qPCR, Western and correlation between RNA and protein levels (left). On the right Venn diagram showing a summary of the bioinformatics analysis: correlation between gene expression and survival (here shown as “survival”), expression in GBM tissues and differential expression between GBM and LGG. The middle intersection of the two circles is shown indicating the number of genes (9) confirmed by both methods.

 

Further, we performed data mining of public databases. At the time we accessed The Cancer Genome Atlas (TCGA) this database contained expression data of 200 clinical GBM tissue samples [5]. Our bioinformatics analysis of 200 GBM tissues from TCGA then identified a co-expression module consisting of nine (PBK, CENPA, KIF15, DEPDC1, CDC6, DLG7, KIF18A, EZH2 and HMMR) genes. The 9 gene co-expression module was highly up-regulated in proneural GBM samples. Interestingly, increased expression of the 9 gene co-expression module correlated with lesser survival-time especially in mesenchymal patients. How could this be explained? One possible explanation could be that the proneural features in combination with mesenchymal characteristics could lead to a worse type of tumor. Alternatively, it could be that the subtypes defined by Verhaak et al., 2010 [5] (mesenchymal, proneural, classical and neural) might contain cell fractions of other subtypes. This would be in agreement with a recent study that used RNA-seq analysis to study gene expression in individual GSCs [6].

Interestingly, all identified genes were functionally related and their products were part of the same protein-protein interaction network. Furthermore several of the identified nine genes were shown to be interesting candidates for therapeutic targeting in GSCs by using gene knock-down studies. Previously it has been shown that gene knockdowns of EZH2 [7] and HMMR [8] both lead to reduction in tumorigenic features of GSCs. Our own study revealed that PBK was essential for propagation of GSCs [9]. Recently we have also performed a gene knockdown of an additional gene from the 9 gene list that also resulted in a significant reduction of GSC growth (Stangeland, unpublished).

 

Acknowledgements

Biljana Stangeland would like to thank the “UNIFOR” and the “Legatet til Henrik Homans Minde” at the University of Oslo for financial support.

Biljana Stangeland, PhD

The Norwegian Center for Stem Cell Research, Department of Immunology, Oslo University Hospital and Department of Molecular Medicine, Institute of Basic Medical Sciences, The Medical Faculty, University of Oslo, Oslo, Norway

Wayne Murrell, PhD

Vilhelm Magnus Lab, Institute of Surgical Research, Oslo University Hospital, Rikshospitalet, Oslo, Norway

 

References

  1. Sandberg CJ, Altschuler G, Jeong J, Stromme KK, Stangeland B, Murrell W, Grasmo-Wendler UH, Myklebost O, Helseth E, Vik-Mo EO, Hide W and Langmoen IA. Comparison of glioma stem cells to neural stem cells from the adult human brain identifies dysregulated Wnt- signaling and a fingerprint associated with clinical outcome. Experimental cell research. 2013; 319(14):2230-2243.
  2. Stangeland B, Mughal A, Grieg Z, Sandberg C, Joel M, Nygård S, Meling TR, Murrell W, Vik-Mo E and Langmoen IA. Combined expressional analysis, bioinformatics and targeted proteomics identify new potential therapeutic targets in glioblastoma stem cells. Oncotarget. 2015; Epub ahead of print July the 20th.
  3. Rappa G, Kunke D, Holter J, Diep DB, Meyer J, Baum C, Fodstad O, Krauss S and Lorico A. Efficient expansion and gene transduction of mouse neural stem/progenitor cells on recombinant fibronectin. Neuroscience. 2004; 124(4):823-830.
  4. Murrell W, Palmero E, Bianco J, Stangeland B, Joel M, Paulson L, Thiede B, Grieg Z, Ramsnes I, Skjellegrind HK, Nygard S, Brandal P, Sandberg C, Vik-Mo E, Palmero S and Langmoen IA. Expansion of multipotent stem cells from the adult human brain. PloS one. 2013; 8(8):e71334.
  5. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G, Lawrence M, O’Kelly M, Tamayo P, Weir BA, Gabriel S, et al. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 2010; 17(1):98-110.
  6. Patel AP, Tirosh I, Trombetta JJ, Shalek AK, Gillespie SM, Wakimoto H, Cahill DP, Nahed BV, Curry WT, Martuza RL, Louis DN, Rozenblatt-Rosen O, Suva ML, Regev A and Bernstein BE. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science. 2014; 344(6190):1396-1401.
  7. Suva ML, Riggi N, Janiszewska M, Radovanovic I, Provero P, Stehle JC, Baumer K, Le Bitoux MA, Marino D, Cironi L, Marquez VE, Clement V and Stamenkovic I. EZH2 is essential for glioblastoma cancer stem cell maintenance. Cancer research. 2009; 69(24):9211-9218.
  8. Tilghman J, Wu H, Sang Y, Shi X, Guerrero-Cazares H, Quinones-Hinojosa A, Eberhart CG, Laterra J and Ying M. HMMR maintains the stemness and tumorigenicity of glioblastoma stem-like cells. Cancer research. 2014.
  9. Joel M, Mughal AA, Grieg. Z, Murrell W, Palmero S, Mikkelsen B, Fjerdingstad HB, Sandberg CJ, Behnan J, Glover JC, Langmoen IA and Stangeland B. Targeting PBK/TOPK decreases growth and survival of glioma initiating cells in vitro and attenuates tumor growth in vivo. Molecular Cancer. 2015; 14:121.

 

Multiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier SchönmannMultiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier Schönmann