The unfolded protein response in glioblastoma stem cells: towards new targets for therapy
Peñaranda Fajardo, Natalia
DOI:
10.33612/diss.118411504
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Peñaranda Fajardo, N. (2020). The unfolded protein response in glioblastoma stem cells: towards new targets for therapy. Rijksuniversiteit Groningen. https://doi.org/10.33612/diss.118411504
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
GBM stem cells by comparative transcriptomic analysis
Natalia M. Peñaranda Fajardo1Coby Meijer1 Raul Aguirre-Gamboa2 Frank A. E. Kruyt1
1Department of Medical Oncology, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
2Department of Genetics, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
Glioblastoma multiforme (GBM) is the most aggressive brain tumor with poor prognoses that has been ascribed to presence of GBM stem cells (GSCs). Eradication of GSCs is therefore a major therapeutic aim. The ER stress/ unfolded protein response (UPR), involved in the pathophysiology of GBM as well as therapy resistance, may provide new targets for GSCs. Here, we examined the effect of PERK inhibition on GSC self-renewal in the absence of acute ER stress. Interestingly, both a pharmacological PERK inhibitor (GSK2606414) and genetic PERK ablation reduced self-renewal in GSC models in vitro. To obtain insight in the underlying mechanisms comparative transcriptomics was performed. First, using mRNA NGS, the transcriptomic response to PERK inhibition was determined in two GBM neurosphere models, GG16 and GSC23, revealing only a small number of significant differentially expressed genes (DEGs). Second, DEGs were identified between GBM neurospheres and their serum-differentiated counterparts allowing identification of genes possibly relevant for GSC regulation. A large number of DEGs were identified of which a minority (approximately 25%) was overlapping between the two models, indicating heterogenic responses. This was also illustrated by differences in known stem cell/ differentiation markers and gene ontology (GO) pathway enrichment. Third, DEGs found in PERK inhibitor-treated and differentiation-induced datasets for each individual model were compared to identify common PERK-regulated genes that control stemness. Only for GG16 overlapping DEGs were found, reflecting upregulation of apoptotic, TNFα/NFκB and epithelial-to-mesenchymal transition (EMT) signaling as shown by GSEA. In conclusion, PERK inhibition in the absence of acute stress reduced GSC self-renewal. This was accompanied by minor effects on gene transcription. GG16 and GSC23 models showed heterogeneity in transcriptional responses induced by PERK inhibition and serum-differentiation. In GG16 DEGs were found that may be involved in PERK-dependent control of stemness. Further exploration is required to determine the relevance of the identified DEGs in PERK signaling and stemness.
Glioblastoma multiforme (GBM) is the most common and lethal brain tumor in adults [1]. Current treatment consists of surgery and chemo-radiotherapy and has limited therapeutic benefit resulting in a median survival for primary IDH wt GBM patients of only 12-15 months and a 5-year survival rate of less than 5% [2]. A subpopulation of tumor cells, named GBM stem cells (GSCs), are considered responsible for driving tumor formation, progression and resistance to therapy and consequently poor prognosis of GBM patients [3,4]. GSCs, like other cancer stem cells, are mainly characterized by sustained self-renewal, persistent proliferation and high tumor initiating properties [4]. Failure of current therapy to eradicate GSCs may explain the poor outcome in the clinic. Thus, targeting GSCs is essential for successful therapy in order to deplete the GSC compartment [5,6]. Hence, the identification of molecular mechanisms that control self-renewal and differentiation of GSCs will be key for developing better treatments for this deadly disease.
Endoplasmic reticulum (ER) stress aggravation and targeting of the unfolded protein response (UPR) have been identified as interesting novel therapeutic approaches for treating cancer, including GBM. Drugs that directly or indirectly disrupt protein homeostasis in the ER will activate the UPR, an adaptive mechanism that aims to restore ER homeostasis. When ER stress is overwhelming the UPR in cancer cells will activate cell death programs resulting in therapeutic benefit. Particularly, the UPR sensors inositol-requiring protein α (IRE1α) and protein R-like ER kinase (PERK) have attracted interest for developing targeted strategies to disrupt UPR functioning in GBM [7,8]. The ER transmembrane proteins IRE1α and PERK both contain kinase domains for which small molecule inhibitors have been developed that are able to block down-stream signaling [8-12]. Although mostly known for balancing ER stress, the UPR is increasingly recognized as a regulator of basal cellular homeostasis important for controlling normal fluctuations in cellular processes such as a high secretory demand or differentiation leading to temporally variation in protein production [13]. Moreover, ER stress and the UPR have been linked with regulation of stemness and differentiation in normal and malignant stem cells [14]. Thus, the UPR is activated also during normal physiological conditions involving more subtle mechanisms of sensor activation, although these mechanisms remain largely elusive.
Previously (see Chapter 3 [15]), we found that PERK is an important mediator of ER stress-induced cytotoxicity in GSC-enriched neurospheres. Importantly, we identified a novel noncanonical function for PERK as a regulator of SOX2 expression and differentiation of GBM neurospheres. In the current study we examined if PERK also affects GSC self-renewal in the absence of extrinsic ER stress in GG16 and GSC23 neurospheres. Indeed, PERK inactivation reduced self-renewal potential of GSCs. To identify possible PERK-regulated genes relevant for stemness control in GG16 and GSC23 neurospheres, we used comparative transcriptional profiling to obtain possible mechanistic clues. For this we first identified differentially expressed genes (DEGs) in control and PERK inhibitor-treated GG16 and GSC23 cells. Secondly, we compared transcriptional profiles between GBM neurospheres and differentiated cells. Thirdly, a cross comparison was made between the above identified DEGs in order to identify potential PERK target genes that control stemness.
Patient-derived GG16 and GSC23 neurospheres were cultured in serum-free neural basal medium supplemented with EGF, bFGF, B27 and Glutamine as previously described [16]. GG16 and GSC23 were previously assigned a mesenchymal (MES) and proneural (PN) subtype, respectively, determined by qRT-PCR-based PN/MES metagene analysis [16], which was confirmed by the Verhaak-based transcriptional signature [17] (Supplementary Figure 1). Differentiated counterparts were obtained by culturing neurospheres in medium containing 10% FCS without supplements for 7 days [18]. GSC23 was kindly provided by Krishna Bhat, Ph.D. (Translational Molecular Pathology, Department of Pathology, MD Anderson Cancer Center, Texas University, USA).
The indicated GBM neurosphere cells were pelleted, washed with PBS and
Materials and Methods
Cell Culturesdissociated with accutase (Sigma-Aldrich, Zwijndrecht, the Netherlands) and careful repeated pipetting in medium. When indicated seeded cells were exposed for 24 h to 1µM PERK inhibitor GSK2606414 (GSK414; Tocris Bioscience, Bristol, UK). Single cell suspensions were sorted based on forward and side scatter pattern using a flow cytometer (BD Biosciences, Breda, The Netherlands). Sorted cells were seeded in 96-well plates at a density of 10, 20, 40 or 80 cells/well in a volume of 100 μl NSM; cells were replenished with 50 μl of NSM every 5-7 days. After 3 weeks, the number of neurospheres per well was counted. Experiments were performed in duplicate.
crRNAs were designed and cloned in pSpCas9(BB)-2A-GFP(pX458) (Addgene Teddington, UK) by the iPSC/CRISPR Centre, ERIBA, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands. GG16 cells were transformed with the plasmid using FuGENE® HD Transfection Reagent (Promega Corporation, Leiden, the Netherlands) followed by clone selection. For more details please see Chapter 3 methods section [15].
RNA was isolated from GG16 and GSC23 neurospheres, GSK414 exposed neurospheres and serum-differentiated counterparts using TRIzol® Reagent (Life Technologies, Thermo Fisher Scientific, Bleiswijk, the Netherlands) following the manufacturer’s protocol. RNA quantification and cDNA synthesis was performed as previously described [19]. Quality check, RNA quantification and Illumina Next Generation Sequencing (NGS) was performed by the Genome Analysis Facility (GAF), Genomics Coordination Centre (GCC) at University Medical Center Groningen (the Netherlands) following the specifications as described in Chapter 3 [15].
For RNA sequence analysis, low gene transcript counts (arbitrary set at less than 40 reads) were kept out of the analysis. Transcript counts were normalized using
Cloning of guides into pX458 and genome editing of GG16 cells using the CRISPR-Cas9 system
trimmed mean of the M-values method. Differential expression analysis between conditions was done using the DESeq2 package [20] for R (http://www.r-project.org/). Gene counts were normalized using the variance stabilizing transformation (VST) or Rlog from the DESeq2 package in order to perform principle component analysis (PCA) and clustering by samples, as well as to visualize and compare expression levels across samples. Each differential expression analysis was performed using paired samples and included library size as covariate. Genes that had a p-adjusted (-adj) value ≤ 0.05, and a Log2 Fold Change more or less than 1 were defined as significant differentially expressed genes (DEGs). Gene ontology (GO) pathway enrichment analysis was performed with output of differential expression analysis using generally applicable gene-set enrichment (GAGE) [21]. Besides, common DEGs upon differentiation induction in GG16 and GSC23 and common DEGs in cross comparison per model were evaluated using Gene Set Enrichment Analysis (GSEA). GSEA was performed with GSEA 2.0 (Broad Institute, Cambridge, MA, USA) web tool, as described previously [22], hallmark gene sets were used in the comparisons. The stem cell-related gene set from QIAGEN RT² Profiler™ Human Stem Cells (Qiagen, the Netherlands) was used to evaluate stem cells gene expression in the models.
The pharmacological inhibitor GSK414 is a potent inhibitor of PERK phosphorylation and activation [9]. We employed GSK414 to determine if PERK inhibition in the absence of extrinsic ER stress affects the self-renewal potential of GSCs using previously generated GG16 and GSC23 neurospheres. Cells were exposed for 24h to GSK414 and the effect on neurosphere formation was evaluated. Of note, the concentration of 1 μM was used as previously found to be sufficient for effective inhibition of thapsigargin-induced PERK activation [15]. GSK414 treatment resulted in approximately 50% reduction in the number of neurospheres in both GG16 and GSC23 (Figure 1A). To corroborate this finding we employed the previously generated CRISPR/Cas9 EIF2AK3 (PERK) gene knockout in GG16 cells (see Chapter 3 [15]).
Results
Figure 1B shows reduced neurospheres formation potential of the GG16-PERK-ko GSCs compared to control GG16 cells, although to a lesser extent than observed with GSK414. This may be related to abrupt PERK inhibition being more detrimental for GSCs then genetic ablation of PERK by CRISPR/cas9, which is more time consuming allowing adaptive/ rewiring responses to occur that may dampen the effect. Regardless of this, these findings indicate that PERK plays a role in regulating GSC self-renewal.
Figure 1: PERK inhibition reduces self-renewal potential in GBM neurospheres. (A)
Limiting dilution assays of GG16 and GSC23 neurosphere cells treated for 24h with GSK414 (1 µM). (B) Limiting dilution assays of GG16 PERK-ko cells compared to control. Two independent experiments were performed; # (numbers); error bars indicate standard deviations. *p-value<0.05.
Comparative transcriptomic analyses
In order to identify PERK-dependent mechanisms that may regulate GSCs, first mRNA NGS was performed on untreated and GSK414-treated GG16 and GSC23 neurospheres, followed by comparative transcriptional analysis. Secondly, transcriptional profiles were generated from GG16 and GSC23 neurospheres and their differentiated counterparts to identify genes involved in GSC self-renewal and differentiation.
PCA was performed demonstrating clustering based on cell line, GSK414 exposure and differentiation status as shown for one of the principle components in Figure 2. Distinct clustering was clearly seen for untreated (control) GG16 and GSC23 neurospheres and their differentiated counterparts. Control and GSK414-treated GBM neurospheres displayed less distinct, mostly overlapping PCA clustering indicating small transcriptional differences.
Figure 2: Principal component analysis (PCA) of the obtained transcriptional profiles.
PCA comparing transcriptional profiles (n=3) from GG16 and GSC23 neurospheres: control neurospheres (Nsp; green), GSK414-treated neurospheres (Nsp + GSK414; blue), and differentiated counterparts (Diff; red).
Effects of PERK inhibition on transcription in GG16 and GSC23
Differential transcriptomic analyses between control and GSK414-treated cells revealed 174 and 10 DEGs in GG16 and GSC23 neurospheres, respectively (Figure 3A and B, Table 1, 2 and 3). In GG16 130 DEGs were up- and 44 downregulated upon GSK414 treatment. The small number of DEGs (10) in GSK414-exposed GSC23 neurospheres were all downregulated.
GO analysis (on output of differential expression analysis) was performed and Figure 3C and D show top 15 significantly up- and downregulated pathways in GG16 and GSC23 neurospheres. Mainly, DNA repair and catabolic processes were downregulated in GG16 neurospheres after GSK414 exposure, whereas wound healing and vasculature-related processes were upregulated. In GSC23 neurospheres GOs associated with ER- and UPR-related processes were downregulated, whereas RNA-translation-related processes were upregulated by GSK414 treatment.
Serum-differentiation-induced changes in transcription in GG16 and GSC23
Next, transcriptional profiles obtained from the GBM neurospheres and differentiated counterparts were compared. Comparative analyses revealed 1747 and 2482 DEGs in GG16 and GSC23, respectively (see Figure 4A and B; Tables 4 to 7). In GG16 neurospheres 633 and 1114 DEGs were up- and downregulated upon differentiation, respectively. In GSC23 neurospheres differentiation upregulated 1220 and downregulated 1262 DEGs. GO analysis was performed and top 15 pathways are depicted in Figure 4C and D for GG16 and GSC23, respectively. Differentiation of GG16 neurospheres was predominantly accompanied by downregulation of GTPase and RAS signal transduction, sterol/cholesterol biosynthesis and actin cytoskeleton-related processes. On the other hand, processes cytoskeleton-related to protein translation and ER protein targeting were upregulated. In GSC23 neurospheres differentiation resulted in downregulation of chromatin and histone modification processes, peptidyl-threonine modifications and (neural) development processes, whereas electron transport chain processes, extracellular matrix organization and cytokine and immune-related processes were upregulated.
Figure 3: Comparative transcriptional analyses of control and GSK414-treated GBM neurospheres. Volcano plots of differential expression analysis between control (Nsp) and
GSK414-treated GG16 (A) and GSC23 neurospheres (B). Red dots represent DEGs and names of top 10 DEGs (according to p-adj value significance) are indicated. GO pathway analysis in GG16 (C) and GSC23 (D). The top pathways (according to significance of p-value) are depicted.
Figure 4: Comparative transcriptional analyses of control and differentiated GBM neurospheres. Volcano plots of differential expression analysis between GBM neurospheres
(Nsp) and differentiated (Diff) GG16 (A) and GSC23 (B). Red dots represent DEGs and names of top 10 DEGs (according to p-adj value significance) are indicated. GO pathway analysis in GG16 (C) and GSC23 (D). The top pathways (according to significance of p-value) are depicted.
Then, GG16 and GSC23 neurospheres/ differentiation datasets were compared for overlapping DEGs. In common were 548 downregulated and 268 upregulated DEGs upon differentiation (Figure 5A). GSEA of overlapping downregulated DEGs showed associations with G2/M checkpoint and mitotic spindle regulation, androgen and estrogen responses, TNFα, MTORC, KRAS and IL2/STAT5 signaling (Figure 5B). Upregulated DEGs were associated with pathways related to inflammatory, interferon gamma and alpha responses, and epithelial-mesenchymal transition (EMT) (Figure 5C). It was noted that DEGs linked with IL2/ STAT5 signaling and TNFα signaling via NF-κβ were both up- and downregulated upon differentiation.
Figure 5: Overlapping DEGs in control and differentiated GG16 and GSC23 neurospheres.
(A) Venn diagrams depicting distinct and overlapping differentiation-dependent DEGs in GG16 and GSC23 neurospheres. (B) GSEA identified processes related to upregulated or downregulated DEGs (number of DEGs in white).
Analyses of known stem cell genes
Known stem cell and differentiation markers represented by the Qiagen Human Stem Cells gene set were also examined for altered expression in both the GG16 and GSC23 neurospheres and differentiated counterparts (Figure 6A and B). Overlaps as well as differences in marker expression were seen between the two models in both heatmaps and DEGs. For example, the stem cell markers OLIG2 and MYC/
MYCN were decreased in differentiated GG16 and GSC23 cells, whereas somewhat
surprisingly stem cell marker CD44 increased. In GG16 PROM1 (CD133) decreased and in GSC23 the astrocytic differentiation marker GFAP increased upon differentiation.
Figure 6: Expression of stem cell-related genes. (A) Heat maps depicting differential
expression of stem cell-related genes according to the Qiagen Human cell genes set in GG16 and GSC23 neurospheres versus differentiated data sets (p-adj value ≤ 0.05). Color scale represents z-scores. (B) Genes that have a Log2 Fold Change > 1 or < -1 and p-adj value ≤ 0.05 (DEGs) in GG16 and GSC23 neurospheres.
Identification of overlapping DEGs in PERK inhibition and differentiation datasets
Finally, overlapping DEGs in the GSK414 treatment and differentiation datasets were determined for each GBM neurosphere model. Only GG16 neurospheres showed overlapping DEGs; 7 down- and 23 upregulated (Figure 7A and B, Table 8). GSEA indicated an upregulation of overlapping DEGs linked with apoptosis, EMT and TNF signaling (Figure 7C).
Figure 7: Overlapping DEGs identified in GSK414 treatment and differentiation datasets.
Venn diagrams showing distinct and overlapping downregulated or upregulated DEGs in the GSK414-treatment and differentiation datasets for GG16 (A) and GSC23 (B) GSEA on overlapping upregulated DEGs in GG16. The number of DEGs (in white) belonging to the identified pathway are indicated (see also table 8).
Discussion
In this study we found that PERK inhibition in GG16 and GSC23 neurospheres reduces stem cell potential. Subsequently we aimed to identify PERK-regulated genes and processes possibly involved in the control of stemness.
The observation that PERK inhibition by GSK414 or genetic ablation in the absence of extrinsic ER stress decreased stem cell potential in the GBM neurospheres is a novel finding. Although GSK414 has been recently reported to have off-target effects on both RIPK1 and KIT (CD117, c-KIT) [23,24], we also observed decreased self-renewal in PERK knockout GG16 cells. Commonly acute stress- induced PERK activation involves the sequestering of BiP/GRP78 to misfolded proteins in the ER
lumen leading to dissociation from PERK and subsequent auto-phosphorylation of PERK, phosphorylation of eIF2a, suppression of CAP-dependent mRNA translation and accumulation of downstream effectors such as transcription factor ATF4 [8]. However, in regular cultured GBM neurospheres we were not able to detect activation of the PERK branch such as indicated by accumulation of downstream ATF4 or CHOP (see Chapter 3 [15]). On the other hand, we were able to detect constitutive BiP/ GRP78 expression in the GBM neurospheres, which has been linked with malignant transformation [25]. Indeed a permanent oncogenic and metabolic stressed state in cancer cells is thought to result in a continuous dependency on adaptive UPR signaling [26]. Regardless, our results show that pharmacological inhibition of PERK kinase activity was sufficient to reduce stemness in regular cultured GBM neurospheres, PERK inhibition in the absence of acute stress did not affect SOX2 expression (see Chapter 3 [15]). This indicates involvement of a GSC regulatory mechanism distinct from ER stress- or differentiation-induced PERK-dependent suppression of SOX2. We have not been able to detect PERK phosphorylation under regular culture conditions, although experiments with phospho-specific PERK antibodies remain to be performed, particularly since higher phospho-PERK levels have been reported in breast cancer cells compared to normal tissue [27,28]. As already mentioned, it should be noted that GSK414 has been reported to have additional targets [23] and therefore other available pharmacological PERK inhibitors should also be tested. On the other hand, genetic ablation of PERK also reduced stemness further indicating PERK dependency. Also the impact of PERK inhibition on the activation status of the other two UPR branches, IRE1 and ATF6 needs to be examined for possible compensatory mechanisms that may affect stemness.
In order to obtain insight in the mechanism underlying PERK-dependent regulation of stemness in the GBM neurospheres first comparative transcriptomics was performed between control and GSK414 exposed GG16 and GSC23 neurospheres. Only, a relative small number of genes was significantly affected by PERK inhibition. GO analyses revealed that PERK inhibition downregulated processes related to DNA repair, catabolism and cell cycle regulation in GG16, whereas wound healing, vasculature-related processes and cell proliferation were upregulated. Although in GG16 UPR-associated processes were not among the top 15 GOs, PERK inhibition in GSC23 showed downregulation of ER- and UPR-related processes and upregulation
of RNA-translation-related processes and targeting of proteins to ER (see Figure 3). These findings indicate that PERK inhibition results in differential responses in the two GBM neurosphere models. Moreover, the low number of DEGs identified upon PERK inhibition suggests that in the absence of acute stress direct or indirect transcriptional regulation by PERK is limited.
Transcriptomic comparison between GBM neurospheres and serum-differentiated counterparts was performed to identify possible genes and pathways relevant for regulation of stemness. GO analyses of DEGs indicated that differentiation of GG16 neurospheres cells was accompanied by downregulation of processes associated with GTPase and RAS signal transduction, sterol/cholesterol biosynthesis and actin cytoskeleton-related processes, whereas upregulation of protein translation-related processes and ER protein targeting were seen (Figure 4). Differentiation of GSC23 neurospheres showed downregulation of chromatin and histone modification processes, peptidyl-threonine modifications and (neural) development processes, and upregulation of electron transport chain processes, extracellular matrix organization and cytokine and immune-related processes. Focusing on known stem cell and differentiation markers and regulators (Qiagen Human Stem Cells gene set), significant differences were seen between the models. Overlapping genes were identified, such as OLIG2 and MYC/MYCN that were downregulated upon differentiation in both GG16 and GSC23 cells. Surprisingly, stem cell marker CD44 levels increased upon differentiation. The well-known GSC marker PROM1 (CD133) was downregulated in differentiated GG16 whereas it was not detected as differentially expressed in GSC23. On the other hand, in GSC23 the differentiation marker GFAP was upregulated upon differentiation that was not seen in GG16. Despite these differences in gene expression within the models transcriptional profiling by NGS also identified a considerable number of overlapping genes in GG16 and GSC23, 548 and 268 DEGs being downregulated and upregulated upon differentiation, respectively (Figure 5). GSEA of common downregulated DEGs were involved in G2/M checkpoint and mitotic spindle regulation, androgen and estrogen responses, TNFα, MTORC and KRAS signaling. Pathways related to upregulated DEGs included inflammatory, interferon gamma and alpha responses, and EMT.
Together this illustrates that different stem cell regulatory mechanisms can be involved in GSC maintenance and differentiation, adding up to the known heterogeneity
in GBM and between GSCs [29]. Moreover, various signals derived from niches and microenvironment also greatly affect heterogeneity, which were not included in our experimental set up [30].
In an attempt to identify PERK-dependent genes involved in stem cell-differentiation regulation, the GSK414 treatment and cell-differentiation datasets were compared for overlapping DEGs. Since comparative transcriptomics between GG16 and GSC23 neurospheres and differentiated counterparts revealed considerable cell-dependent differences, only model-specific DEGs were compared. No overlapping DEGs were found in the GSC23 datasets, whereas a total of 30 DEGs were overlapping in the GG16 datasets (Table 8). GSEA of GG16 DEGs were linked with apoptosis, EMT and TNF signaling. All these DEGs are potentially important for PERK-dependent GSC regulation and require further analyses. Particularly interesting may be the cell surface glycoprotein CD44 and the extracellular matrix molecule Tenascin-C (TNC) that were found in the afore mentioned three processes and are known to be involved in GBM malignancy and progression [31,32]. Although CD44 transcripts went up both upon differentiation and GSK414 exposure in our models, it may be interesting to further explore their role in the proliferation, migration and stemness in GBM [33,34]. The same accounts for TNC that has been linked with poor prognosis in GBM patients [35] and to enhance GSC proliferation via induction of NOTCH expression [36], although another study found that TNC knockdown enhanced GBM xenograft proliferation and identified a role for TNC in modulating GBM interactions with stromal, endothelial and microglia cells [37].
In conclusion, we found that PERK inhibition in the absence of acute stress reduces stem cell potential in two GBM neurosphere models. This finding together with our previous finding of PERK being instrumental in ER stress-induced inhibition of self-renewal potential in the same models underscores the importance of PERK in regulating stemness in GSCs. Our findings further suggest therapeutic benefit of PERK targeted strategies in GBM. The underlying mechanisms by which PERK regulates GSCs under regular and acute stress-induced appear distinct. Our attempt to identify possible genes and pathways involved in PERK-dependent regulation of stemness in the absence of extrinsic stress identified possible candidate mechanisms that need to be explored further. It should be noted that the number of DEGs identified after PERK inhibition was low and it is likely that mechanisms at the protein level, independent of
transcription, are also involved. In addition GG16 and GSC23 models showed partly overlapping, but mostly distinct transcriptional responses to PERK inhibition as well as differentiation induction, illustrating heterogeneity in GBM and GSCs. Further studies are required to explore the possible relevance of the identified genes and pathways in PERK-dependent regulation of stemness.
Acknowledgments
CRIPSR/Cas knock outs were generated with help from the iPSC/CRISPR Centre, ERIBA, UMCG, University of Groningen. RNA-Seq was performed with the help of dr. Klaas Kok, Department of Genetics, UMCG, University of Groningen. This research was funded by 617-2013 call Colciencias, Colombia and the Graduate School of Medical Sciences, University of Groningen.
1. Ostrom QT, Bauchet L, Davis FG, Deltour I, Fisher JL, Langer CE, et al. The epidemiology of glioma in adults: a “state of the science” review. Neuro Oncol. 2014; 0: 1-18.
2. Preusser M, de Ribaupierre S, Wöhrer A, Erridge SC, Hegi M, Weller M, et al. Current concepts and management of glioblastoma. Ann. Neurol. 2011; 70(1): 9-21.
3. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, et al. Identification of human brain tumour initiating cells. Nature 2004; 432(7015): 396-401.
4. Lathia JD, Mack SC, Mulkearns-Hubert EE, Valentim CL, Rich JN. Cancer stem cells in glioblastoma. Genes Dev. 2015; 29(12): 1203-1217.
5. Carén H, Beck S, Pollard SM. Differentiation therapy for glioblastoma–too many obstacles? Molecular & cellular oncology 2016; 3(2): e1124174.
6. Campos B, Wan F, Farhadi M, Ernst A, Zeppernick F, Tagscherer KE, et al. Differentiation therapy exerts antitumor effects on stem-like glioma cells. Clin. Cancer Res. 2010; 16(10): 2715-2728.
7. Ojha R, Amaravadi RK. Targeting the unfolded protein response in cancer. Pharmacological research 2017; 120: 258-266.
8. Maurel M, McGrath EP, Mnich K, Healy S, Chevet E, Samali A. Controlling the unfolded protein response-mediated life and death decisions in cancer. 2015; 33: 57-66.
9. Axten JM, Medina JR, Feng Y, Shu A, Romeril SP, Grant SW, et al. Discovery of 7-methyl-5-(1-{[3-(trifluoromethyl) phenyl] acetyl}-2, 3-dihydro-1 H-indol-5-yl)-7 H-pyrrolo [2, 3-d] pyrimidin-4-amine (GSK2606414), a potent and selective first-in-class Inhibitor of protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK). J. Med. Chem. 2012; 55(16): 7193-7207.
10. Maly DJ, Papa FR. Druggable sensors of the unfolded protein response. Nature chemical biology 2014; 10(11): 892.
11. Jiang D, Niwa M, Koong AC. Targeting the IRE1alpha-XBP1 branch of the unfolded protein response in human diseases. Semin. Cancer Biol. 2015; 33: 48-56.
12. Atkins C, Liu Q, Minthorn E, Zhang SY, Figueroa DJ, Moss K, et al. Characterization of a novel PERK kinase inhibitor with antitumor and antiangiogenic activity. Cancer Res. 2013; 73(6): 1993-2002.
13. Rutkowski DT, Hegde RS. Regulation of basal cellular physiology by the homeostatic unfolded protein response. J. Cell Biol. 2010; 189(5): 783-794.
14. Yang Y, Cheung HH, Tu J, Miu KK, Chan WY. New insights into the unfolded protein response in stem cells. Oncotarget 2016; 7(33): 54010-54027.
15. Peñaranda-Fajardo NM, Meijer C, Liang Y, Dijkstra BM, Aguirre-Gamboa R, den Dunnen WF, et al. ER stress and UPR activation in glioblastoma: identification of a noncanonical PERK mechanism regulating GBM stem cells through SOX2 modulation. Cell death & disease 2019; 10(10): 1-16.
16. Joseph J, Conroy S, Tomar T, Eggens-Meijer E, Bhat K, Copray S, et al. TGF-β is an inducer of ZEB1-dependent mesenchymal transdifferentiation in glioblastoma that is associated with tumor invasion. Cell death & disease 2014; 5(10): e1443.
17. Verhaak R, Hoadley K, Purdom E, Wang V, Qi Y, Wilkerson M, 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.
18. Joseph JV, van Roosmalen IA, Busschers E, Tomar T, Conroy S, Eggens-Meijer E, et al. Serum-Induced Differentiation of Glioblastoma Neurospheres Leads to Enhanced Migration/Invasion Capacity That Is Associated with Increased MMP9. PloS one 2015; 10(12): e0145393.
19. Conroy S, Kruyt FA, Joseph JV, Balasubramaniyan V, Bhat KP, Wagemakers M, et al. Subclassification of Newly Diagnosed Glioblastomas through an Immunohistochemical Approach. PloS one 2014; 9(12): e115687.
20. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014; 15(12): 550.
21. Luo W, Friedman MS, Shedden K, Hankenson KD, Woolf PJ. GAGE: generally applicable gene set enrichment for pathway analysis. BMC Bioinformatics 2009; 10(1): 161.
22. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 2005; 102(43): 15545-15550.
23. Rojas-Rivera D, Delvaeye T, Roelandt R, Nerinckx W, Augustyns K, Vandenabeele P, et al. When PERK inhibitors turn out to be new potent RIPK1 inhibitors: critical issues on the specificity and use of GSK2606414 and GSK2656157. Cell Death & Differentiation 2017.
24. Mahameed M, Wilhelm T, Darawshi O, Obiedat A, Tommy W, Chintha C, et al. The unfolded protein response modulators GSK2606414 and KIRA6 are potent KIT inhibitors. Cell Death & Disease 2019; 10(4).
25. Dong D, Ni M, Li J, Xiong S, Ye W, Virrey JJ, et al. Critical role of the stress chaperone GRP78/BiP in tumor proliferation, survival, and tumor angiogenesis in transgene-induced mammary tumor development. Cancer Res. 2008; 68(2): 498-505.
Annelies Verhelst, Xavier Janssens, Sophie Heindryckx,Femke Van Vlierberghe, Hans. The paradox of the unfolded protein response in cancer. Anticancer Res. 2013; 33(11): 4683-4694.
27. Avivar-Valderas A, Salas E, Bobrovnikova-Marjon E, Diehl JA, Nagi C, Debnath J, et al. PERK integrates autophagy and oxidative stress responses to promote survival during extracellular matrix detachment. Mol. Cell. Biol. 2011; 31(17): 3616-3629. 28. McGrath E, Logue S, Mnich K, Deegan S, Jäger R, Gorman A, et al. The unfolded
protein response in breast cancer. Cancers 2018; 10(10): 344.
29. Zhou D, Alver BM, Li S, Hlady RA, Thompson JJ, Schroeder MA, et al. Distinctive epigenomes characterize glioma stem cells and their response to differentiation cues. Genome Biol. 2018; 19(1): 43.
30. Dirkse A, Golebiewska A, Buder T, Nazarov PV, Muller A, Poovathingal S, et al. Stem cell-associated heterogeneity in Glioblastoma results from intrinsic tumor plasticity shaped by the microenvironment. Nature communications 2019; 10(1): 1787. 31. Yamini B. NF-κB, Mesenchymal Differentiation and Glioblastoma. Cells 2018; 7(9):
125.
32. Iwadate Y. Epithelial mesenchymal transition in glioblastoma progression. Oncology Letters 2016; 11(3): 1615-1620.
33. Dong Q, Li Q, Wang M, Hu J, Dai J, Niu L, et al. Elevated CD44 expression predicts poor prognosis in patients with low grade glioma. Oncology Letters 2019; 18(4): 3698-3704.
34. Wang J, Li X, Wu H, Wang H, Yao L, Deng Z, et al. EMP1 regulates cell proliferation, migration, and stemness in gliomas through PI3K‐AKT signaling and CD44. J. Cell. Biochem. 2019.
35. Mock A, Warta R, Geisenberger C, Bischoff R, Schulte A, Lamszus K, et al. Printed peptide arrays identify prognostic TNC serumantibodies in glioblastoma patients. Oncotarget 2015; 6(15): 13579-13590.
36. Sarkar S, Mirzaei R, Zemp FJ, Wei W, Senger DL, Robbins SM, et al. Activation of NOTCH Signaling by Tenascin-C Promotes Growth of Human Brain Tumor-Initiating Cells. Cancer Res. 2017; 77(12): 3231-3243.
37. Xia S, Lal B, Tung B, Wang S, Goodwin CR, Laterra J. Tumor microenvironment tenascin-C promotes glioblastoma invasion and negatively regulates tumor proliferation. Neuro-oncology 2015; 18(4): 507-517.
Supplementary Material
Supplementary Figure 1: Transcriptional classification of GG16 and GSC23 neurospheres.
Transcription profile heatmaps of Verhaak signature genes in GG16 and GSC23 data sets (n=3); the color scale represents z-scores. Subtypes are classical (CL), mesenchymal (MES), neuronal (NL) and proneuroal (PN). Subtypes of GG16 and GSC23 were confirmed as MES and PN, respectively.
