Background
Glioblastoma multiforme (GBM) is the most common and the most aggressive primary brain cancer. The average annual age-adjusted incidence rate is ~3 per 100.000 person-years [
1]. Despite multimodal therapy and inclusion of temozolomide, the overall survival remains dismal [
2,
3]. This grim scenario mandates a pivotal shift in our approaches to develop new therapy as well as understanding the underlying disease mechanisms. During the last decade the GBM field has shown interest in a subgroup of cells harbored within these tumors with stem cell characteristics [
4‐
6] that forms invasive tumors, similar to the tumor of origin, upon orthotopic xenotransplantation [
7]. These cells are commonly referred to as glioblastoma-initiating cells (GICs) and deemed important for the characteristics of GBM. They possess enhanced invasive properties [
8,
9], promote tumor angiogenesis [
10‐
12] and are resistant to irradiation [
13] and chemotherapy [
14,
15]. Thus, there is a strong rationale for validating potential molecular targets in GICs.
Dysregulation of co- or post-translational protein modifications is a trait of many human cancers. Nα- or N-terminal (Nt-) acetylation is one of the most common covalent modifications of all soluble human proteins [
16] and occurs predominantly co-translationally [
17]. N-terminal acetyltransferases (NATs) or Nα-acetyltransferases (NAAs) are arranged in complexes and believed to target ~80-90 % of all soluble human proteins [
17,
18]. Six NAT complexes have been identified in humans, NAT-A – NAT-F and each complex consists of one catalytic subunit and auxiliary subunit(s) [
16]. With an increasing body of knowledge about their diverse functions, substrates and downstream targets, NATs are emerging as potential targets in several cancers [
16]. Knockdown studies targeting NATs in various malignancies including colon [
19], thyroid [
20] and hepatocellular cancers [
21] have shown that depletion of the catalytic subunit causes a less aggressive phenotype with reduced cell proliferation and/or increased apoptosis.
NAT12/NAA30 is the catalytic subunit of the NAT-C complex [
22]. Knockdown of each of the NAT-C subunits, led to p53-dependent apoptosis in HeLa and colon carcinoma cell lines with the strongest phenotype observed with depletion of
NAT12/NAA30 [
23]. The latter study also pointed out mammalian target of rapamycin (MTOR) as a substrate for NAT-C
. Another report also suggested TOR as a target of NAT-C activity [
24].
In the present study we investigated the expression of NAT12/NAA30 in GBM tissue samples, GICs, normal brain tissue, and neural stem cells (NSCs) from the adult human brain as well as in a neural fetal cell line (NFCs). Using immunolabeling, we revealed a hitherto undescribed nuclear localization of NAT12/NAA30 protein. To study the function of NAT12/NAA30, we performed gene knockdown using RNA interference (RNAi) technology. Knockdown of NAT12/NAA30 resulted in markedly reduced cell viability and sphere-forming ability of GICs. To study genes and pathways downstream of NAT12/NAA30, we used microarray analysis and western blot. This enabled us to identify several pathways such as p53, ribosomal assembly, hypoxia response and cell proliferation regulated by NAT12/NAA30. Furthermore, we documented a reduction of phospho-MTOR (Ser2448) and increased levels of p53 and glial fibrillary acidic protein (GFAP) in the knockdown cultures. We show that intracranial transplantations into severe combined immunodeficient (SCID) mice of GICs featuring NAT12/NAA30 knockdown, resulted in a significant prolongation of animal survival compared to controls.
Discussion
GBM remains one of the most aggressive human cancers. Accumulating evidence supports the role of GICs in disease progression, tumor cell invasion and the tumor’s notorious resistance to chemo- and radiotherapy. Consequently, new molecular therapies against GBM would also need to target this cell population.
Recently, protein Nt-acetylation has been implicated in cancer development [
16]. In the present study we analyzed the expression of several NATs in GIC cultures and GBM tissues showing which of these are co-expressed and thus might be functionally redundant (Fig.
1b-c and Additional file
4: Figure S2). We also show that the increased expression of several NATs in GBM tissues is associated with poor patient survival and that these enzymes may be relevant as new potential therapeutic targets. To further reinforce the validity of our study we used NSC cultures isolated from several different regions of the brain as controls. We identified NATs whose expression was significantly down- or up-regulated in GBM and GICs (Additional file
3: Figure S1).
By detailed expression analysis of
NAT12/NAA30 in tissues and cell cultures (GIC and NSC cultures, and NFCs), we detected the presence of alternative transcript(s) that had 3’UTR regions of variable length (Fig.
1d). We have demonstrated that these were differentially regulated in GICs compared to NSCs. Especially the distal 3’UTR of
NAT12/NAA30 was strongly down-regulated in GBM and GIC cultures as shown by two independent methods (microarray and qPCR) (Fig.
1a and d). Moreover, public database mining revealed that the decreased expression of the distal 3’UTR region correlated with a shorter survival of glioma patients (Additional file
5: Figure S3A). Several studies have shown that mRNA isoforms with differences in their 3’UTRs have different stability and translational activity [
34‐
36]. In yeast it was found that shorter mRNAs more frequently formed a closed-loop structure that enhanced protein translation [
37]. Mayr et al. studied 27 cancer cell lines from different tissues and showed that shortening of the 3’UTR in the tested mRNAs seemed to be important for activation of oncogenesis [
38]. Shortening of the 3’UTRs resulted in enhanced translation in some of these cancer cell lines. More importantly, shorter mRNAs had greater stability and produced more protein [
38]. Another study reported that shortening of 3’UTRs correlated with poor patient survival in breast and lung cancer [
39]. To our knowledge we are the first to report the existence of an analogous mechanism associated with GBM and GICs. In the present work we show that the up-regulation of NAT12/NAA30 at the protein level in GICs is not caused by the increased levels of the steady state mRNA but rather by other mechanisms such as shortening of the 3’UTR and possibly greater stability of mRNA. Other processes involved in the complex regulation of
NAT12/NAA30 expression might be alternative splicing, alternative poly-adenylation sites (Stangeland, unpublished), and the presence of regulatory and/or coding sequences within the 3’UTR.
Human NAT12/NAA30 is known to be present in the cytoplasm [
23‐
26]. Our data show that it is also located in the nuclei of GICs and some NSCs (Fig.
2i and
k). The transport of proteins from the cytosol to the nucleus is a tightly regulated process facilitated by importins [
40]. NAT12/NAA30 contains a nuclear localization signal (RRGYIAMLAVDSKYRRN at 243 according to pSORT II,
http://psort.hgc.jp/) that confers a high probability of being imported to the cell nuclei. It has also been reported that 3’UTRs of other genes contain regulatory regions that influence nuclear export and subcellular localization [
41]. The observed differential subcellular localization of NAT12/NAA30 in the nucleus is novel and might imply an additional role for NAT12/NAA30. Another member of the NAT family, NAT-D, is known to acetylate histones H2A and H4 [
42]. Whether NAT12/NAA30 protein can have a similar role remains to be shown.
Furthermore, evaluating the staining pattern of NAT12/NAA30 in GBM biopsies led to an interesting observation (Fig.
2b-f). NAT12/NAA30 was found to be expressed predominantly in cells surrounding blood vessels. In this perivascular region the GBM cells have a much stronger expression of nestin than in other parts of the tumor specimen. Co-staining with these two markers showed a large degree of overlap in the GBM cells around blood vessels. Recently, it was shown that nestin regulates stemness, cell growth and invasion in GBMs [
43]. The authors reported that the overexpression of
NESTIN increased cell growth, sphere formation and cell invasion while depletion of
NESTIN resulted in decreased expression of stem cell markers. The expression of nestin has been found to be an independent prognostic factor in glioma patients [
44]. Since GICs are believed to reside in the perivascular niche of the tumor [
10‐
12], our immunolabeling might indicate that NAT12/NAA30 is predominantly expressed in an immature cell types in GBM.
RNAi-mediated gene silencing of
NAT12/NAA30 enabled us to obtain a stable knockdown at both transcript and protein level as determined by qPCR, western blot and immunolabeling (Fig.
3). NATs are evolutionarily highly conserved [
17,
22]. The partial knockdown we obtained might be explained by selection mechanisms having enabled GICs to avoid the lethality that would result from total
NAT12/NAA30 knockdown. Nevertheless, functional analysis of the knockdown cultures revealed a clear phenotype with restricted cell growth in vitro and in vivo (Fig.
4a-d). Reduced sphere-forming ability in the knockdown cultures was stable across passages. This finding is especially important because sphere formation has been reported to be an independent predictor of clinical outcome [
45].
The number of spheres is assumed to be both an indicator of the aggressiveness of the tumor and the number of tumor initiating cells. The ability to form spheres strongly correlates with tumor growth and survival also in animal models [
46]. Our results from intracranial transplantation into SCID mice were in agreement with this as mice transplanted with GICs featuring
NAT12/NAA30 gene knockdown survived significantly longer than controls. Altogether 3/11 mice did not form tumors until the end of the observation period in the
NAT12/NAA30 knockdown group while all 6/6 mice died from tumors in the control group.
There are few known substrates of NAT-C, but based on the Nt-amino acid composition of target proteins NAT-C can potentially Nt-acetylate up to 14.5 % of all cytoplasmic human proteins [
47]. The N-terminus of MTOR protein begins with Met-Leu and is a strong candidate as a direct substrate [
23,
24]. In a study in zebrafish TOR was suggested as a downstream target of NAT-C [
24]. The authors also demonstrated that overexpression of TOR rescued the effect of depletion of a NAT-C subunit [
24]. Knockdown of MTOR is a critical effector of the PIK3-AKT pathway which is dysregulated in many cancers [
48]. In GBM AKT-signaling activity is significantly correlated with phosphorylation of MTOR [
49].
In our study, knockdown of
NAT12/NAA30 led to decreased levels of phospho-MTOR (Ser2448), while the total MTOR levels remained unchanged (Fig.
5a). Phospho-MTOR (Ser2448) binds to the effector complexes mTORC1 and mTORC2 and is important for mTORC1 activity [
50]. Implication of MTOR as a (direct or indirect) target of
NAT12/NAA30 was further supported by the expression analysis that showed increased levels of
IGFBP3, an inhibitor of IGF1/MTOR pathway, in the cell cultures featuring
NAT12/NAA30 knockdown (Additional file
8: Figure S6).
Our results also indicate that
NAT12/NAA30 acts via the p53 pathway in GICs. In addition to an increment in p53 protein levels in the knockdown cultures we also provide evidence for dysregulation of the p53 pathway genes. Genes involved in cell-cycle regulation, apoptosis and DNA repair were those that suffered the most substantial expressional alteration (Additional file
8: Figure S6). We also show that
NAT12/NAA30 affects a considerable number of genes that regulate cell proliferation and protein kinases.
We investigated the expression of stem cell related genes in NAT12/NAA30 knockdown cultures. Of the tested stem cell markers, we could detect reduced levels of nestin and SOX2 at the protein level. The knockdown of NAT12/NAA30 also reduced the percentage of CD133+ cells. Our data thus indicate the effect of NAT12/NAA30 knockdown on stemness in GICs.
A previous study has shown that mitochondrial proteins are substrates of NAT12/NAA30 [
22]. Interestingly, we found that knockdown of
NAT12/NAA30 in GICs causes a more abrupt and severe mitochondrial membrane depolarization compared to the NS control cultures when these cells are exposed to hypoxia. This indicates a reduced mitochondrial tolerance to acute hypoxia upon
NAT12/NAA30 knockdown. This notion was further supported by microarray analysis showing that the role of
NAT12/NAA30 in hypoxia response involves
HIF1α and several other genes (Additional file
10: Figure S9). Importantly, western blot analysis confirmed the reduced expression of HIF1α protein (Fig.
5d). Hypoxia induced factors are key mediators in the response of cancer stem cells to hypoxia [
32,
51]. Several studies have pointed out that GICs reside in the hypoxic regions of GBMs [
33] and that hypoxia-induced molecular changes regulate and maintain the phenotype of GICs [
52]. This might imply that targeting
NAT12/NAA30 increases the vulnerability of GICs to hypoxic conditions.
Materials and methods
Tissue specimens and cell culture
Tissue specimens were harvested from consenting patients after approval by the Norwegian National Committee for Medical Research Ethics. Tumor biopsies were obtained as a part of surgical procedures for treating GBM. Normal brain tissue (SVZ, HPC, GM and WM) was obtained from fresh human temporal lobes surgically resected to treat medically refractory epilepsy. All biopsy specimens were evaluated by neuropathologists.
Tumor biopsies underwent mechanical dissociation and Trypsin-EDTA (Gibco, Life Technologies, NYC, NY, USA) was added for enzymatic dissociation. Subsequently, 2 mg/ml human albumin (Octapharma pharmazeutika produktionges, Vienna, Austria) was used to block the Trypsin effect and the cells were washed in L-15 (Lonza, Basel, Switzerland) before being plated in serum-free neurosphere medium containing 10 ng/ml basic fibroblast growth factor (bFGF) and 20 ng/ml epidermal growth factor (EGF) (both R&D Inc., Minneapolis, MN, USA), B27-supplement (1:50, Invitrogen, Carlsbad, CA, USA), 100 U/ml Penicillin/streptomycin (Lonza), 1 ng/ml Heparin (Leo Pharma, Ballerup, Denmark) and 8 mM Hepes (Lonza) in Dulbecco’s modified essential medium with nutrient mix F-12 and Glutamax (DMEM/F12, Invitrogen) [
24,
25]. Dissociated GBM biopsies grown as free-floating spheres in serum free-medium containing mitogens (EGF and bFGF) are highly enriched for GICs [
7,
54]. The GIC cultures used in this current work (T65, T08, T59) were analyzed for the expression of stem cell markers by flow cytometry and showed high expression of CD44, CXCR4, CD166 and CD9 while the expression of CD133 and CD15 was variable [
55]. Immunolabeling showed that most cells were nestin and Sox2 positive (Additional file
6: Figure S4H-I, Stangeland, B. et al., in press). The cells were cultured in 75 cm
2 non-treated flasks (Nunc, Roskilde, Denmark) at a density of 10
5 cells/ml and supplemented with EGF and bFGF twice a week. When the spheres reached approximately 100 μm in diameter they were dissociated into single cells as previously described [
28]. We have previously shown that the GIC cultures T65, T08, TC3 and TC4 formed invasive tumors upon orthotopic transplantation to SCID mice [
55,
56].
NFCs (ReNcell VM Human Neural Progenitor Cell Line, SCC008, Merck Millipore, Darmstadt, Germany) were cultured as spheres in serum-free Neurobasal A medium (Gibco) containing B27 (Gibco), 2 mM L-glutamine, 10 ng/ml bFGF, and 20 ng/ml EGF (both from R&D Systems).
Tissue specimens from the adult human brain were dissociated into single cells and cultured according to our
FAILSAFE protocol (1 % FBS, 10 ng/ml bFGF and 20 ng/ml TGF α) that ensures robust long-term propagation of multipotent stem cells from the adult human brain [
57].
Microarray analysis and public database mining
Microarray analysis was performed using Illumina gene chip. The data were quantile normalized and analyzed using J-Express (Molmine, Bergen, Norway) analysis software. We used published microarrays [
57] with submission numbers GSE41470 (encompassing GSE41390, and GSE41394), GSE53800 (GSM1301030, GSM1301033, GSM1301042) as well as the microarrays submitted in connection with this work GSE60818 (encompassing GSE60705 and GSE60706). For more information on microarray analysis see Additional file
1: Figure S1 and Additional file
2: Figure S2 and Additional file
3 and Additional file
7.
REMBRANDT: Microarray data from the Repository for Molecular Brain Neoplasia Data (REMBRANDT, National Cancer Institute, 2005,
http://rembrandt.nci.nih.gov) were accessed on January the 15th 2014. Hierarchical clustering with distance matrix and the K-means clustering of the NAT genes’ expression were performed using J-Express software (Molmine). For calculation of statistical parameters, we used Graphpad Prism (
www.graphpad.com).
Western blot and quantification of protein expression
The cells were homogenized by triturating in Cell Extraction Buffer (Mammalian cell extraction kit, Biovision, Milpitas, CA, USA) and centrifuged through a QIAshredder (Qiagen, Venlo, Netherlands). For HIF1α and the western blots with phospho-proteins, the cells were homogenized in 10 mmol/L Tris–HCl (pH 7.4), 1 % SDS, 10 mmol/L NaF, and 2 mmol/L Na3VO4 as previously described [
58]. 20–40 μg of whole protein extract were mixed with loading buffer (NuPAGE, Invitrogen) and loaded onto a 4–12 % gradient Nu-PAGE gel (Invitrogen). Protein gels were blotted onto 0.45 μm PVDF membranes (Invitrogen). Membranes were blocked with 5 % skimmed milk in TBS/0.1 % Tween 20 (TBST) and probed with primary antibodies diluted in the same solution. Primary antibodies from Cell Signaling Technologies (CST, Danvers, MA, USA) were incubated in bovine serum albumin (BSA) according to recommended procedures. For a complete list of the antibodies used in this study see Additional file
12. For a complete list of the antibodies used in this study see Additional file
12. The blots were developed using the Lumiglo Reserve CL Substrate kit, and detected using the Epi Chemi II Darkroom (UVP-Laboratory Products, Upland, CA, USA). The intensities of the protein bands were quantified using Photoshop (Adobe, San Jose, CA, USA), background corrected and normalized to the intensities of the corresponding ACTB bands, so that the relative protein expression (RPE) could be calculated (RPE
protx = PE
protx/PE
ACT).
RNA isolation and real-time quantitative reverse-transcription PCR (qPCR)
Total RNA was isolated using an RNeasy Mini Kit and QIAshredder (Qiagen). For cDNA synthesis, experimental set up and oligonucleotide design we used the procedure previously described [
59]. For expression analysis of
NAT12/NAA30 we used seven sets of oligonucleotides covering translated and non-translated regions of the reference sequence NM_001011713.2. For oligonucleotide sequence information see Additional file
3. qPCR was performed on an ABI PRISM 7900HT PCR machine (Applied Biosystems, Life Technologies, Foster City, Ca, USA) using SYBR Premix Ex Taq™ (Takara, Otsu, Japan) or Taqman probes (Applied Biosystems) according to the manufacturer's protocol. For analysis of the
NAT12/NAA30 coding regions we used the following Taqman probes: ex2-3 (Hs02340852), ex3-4(Hs02340853) and ex4-5(Hs02340854). Crossing point (CP) values were generated using second-derivative calculation software (SDS2.2). Gene expression levels in GIC cultures were calculated using two house-keeping genes and multiple controls (values obtained for all tested NSCs and NFCs). For analysis of qPCR results we used 2-∆∆ CT-method and REST software [
60].
GIC culture T65 was used to establish stable NAT12/NAA30 knockdown cultures using shRNA1 (RHS4430-99149700; Clone ID V2LHS_180063), shRNA2 (RHS4430-99137454; Clone ID V2LHS_180058) and a non-silencing (NS) shRNA as a control (RHS4346) (all from Open Biosystems, Thermo Scientific, Huntsville, AL, USA). Cell cultures harboring shRNA constructs 1 and 2 are referred to as KD1 and KD2 respectively.
Production of virus in the 293 FT cell line using plasmid DNA and concentration of virus was done as following: Nine μg of plasmid DNA was transfected into the 293FT cell line using Arrest-In transfection reagent (according to the manufacturer’s protocol). Viral supernatants were collected after 48 and 72 h, centrifuged at 3000 rpm for 20 min at 4 °C and filtered through a sterile 0.45 μm low protein binding filter (Sarstedt, Nümbrecht, Germany). The virus was then concentrated in sterile SW28 ultracentrifuge tubes by ultracentrifugation (Beckman Optima™ LE-80 K ultracentrifuge, Fullerton, CA, USA) equipped with a SW-28 rotor at 23,000 rpm for 1.5 h at 4 °C. The pellet was resuspended in 200 μl of DMEM and aliquots of the concentrated virus were stored at −80 °C. GSC cultures (10×104 cells/ well) were then transduced in 24-well plates by adding 10 μl concentrated virus/well and the cells were incubated for 48 h at 37 °C in 5 % CO2. Three to five days after transduction, GFP positive cells were selected by FACS sorting using a FACS Diva cell sorter (Becton Dickinson, Franklin Lakes, NJ, USA) equipped with an argon ion laser, ‘TurboSort Plus’ option, and Diva software (Becton Dickinson). Alternatively, the cells were selected in 2 μg/ml Puromycin (Sigma-Aldrich, St. Louis, MO, USA) for 3–4 weeks before use for functional assays.
Cell viability, apoptosis and sphere-forming assay
For analysis of cell viability we used a colorimetric test based on tetrazolium salt – XTT (Roche Diagnostics, Indianapolis, IN, USA). Briefly, after sphere dissociation single cells were plated in neurosphere medium at a density of 1×104 cells per well in a flat-bottom 96-well plate (Sarstedt). Cells were incubated overnight and XTT-reagents added as recommended by the manufacturer. The colorimetric changes (absorbance) were measured after 24 h at 490 nm using a plate reader (Victor™ X5 Multilabel Plate Reader, Perkin Elmer 2030, Waltham, MA, USA). Five wells were evaluated for each cell culture. All results were presented as a mean of three independent experiments ± standard deviation. P-values were calculated using unpaired t-test with Welch’s correction.
Apoptosis was assessed using a fluorimetric assay which measures the activity of activated caspases (Roche Diagnostics). 1×104 cells were plated per well in a flat-bottom 96-well plate (Sarstedt) and incubated overnight before adding the reagents as recommended by the manufacturer. The fluorimetric changes were measured after 24 h at 485/535 nm using a plate reader (Victor™ X5 Multilabel Plate Reader). Five wells were evaluated for each cell culture.
The sphere-forming assay was done by seeding single cell suspensions containing 500 cells per well in ultra-low attachment flat bottom 96-well plates for 10 days (Sarstedt). Subsequently, the plates were imaged using GelCount™ (Oxford Optronics, Abington, UK). Only spheres >50 μm were taken into consideration and 10 wells were evaluated for each cell culture. The number and the size (average area) of spheres were measured using software supplied by the manufacturers. All results were presented as a mean of five independent experiments ± standard deviation. P-values were calculated using unpaired t-test with Welch’s correction.
Intracranial transplantation to SCID mice
All animal procedures were approved by the National Animal Research Authority. C.B.-17 SCID male mice (8–9 weeks old) were obtained from Taconic (Ejby, Denmark). They were acclimatized for > 1 week and divided into 3 groups: (i) T65 NS cells (n = 6), (ii) T65 KD1 cells (n = 6) and (iii) T65 KD2 cells (n = 6).
T65 GIC cultures were sorted by FACS after viral transduction and cultured for a few passages before transplantation. After tumor sphere dissociation, cells were counted and plated at a low density for overnight incubation. Prior to the inoculation of cells, mice were anesthetized and placed in a stereotactic frame (Kopf Instruments, Tujunga, CA, USA) and ~3×10
5 incubated cells were inoculated into the right striatum (AP 0 mm, RV 2 mm, ML 2 mm) using a Hamilton syringe and needle (Hamilton, Bonaduz, Switzerland) [
28].
All animals were monitored for adverse effects and examined regularly for general appearance, signs of distress, and weight measurements taken. Mice were sacrificed when they lost >10 % of their bodyweight and developed neurological symptoms. Brains were harvested and further processed for HE-staining as described previously [
28]. Results for all groups are presented as median survival in days. Log-rank test was used to calculate the level of significance.
Mitochondrial depolarization assay
After passage, KD1, KD2 and NS cultures were seeded on fibronection-coated cover slips, and cultured in neurosphere medium in separate culture dishes for 1–6 days before experiments.
For each experiment, a cover slip was removed from its cell culture dish, and mounted in the microscopy chamber. Artificial CSF (aCSF) with the fluorescent cationic dye Rhodamine 123 (30 μM) (Invitrogen) was added to the chamber, and cells were loaded with dye for 15 min at 37 °C on the microscope. The chamber was covered with a lid, and subsequently perfused with air and 5 % CO
2 for maintenance of appropriate pH. After loading of dye, the chamber was perfused with aCSF at a flow rate of approximately 1 ml/min. Imaging was started after at least 10 min of flow. Healthy looking cells were chosen by the criteria clear cell membrane and intact processes. Live imaging was performed using an Olympus IX81 inverted microscope equipped with a MT20 fluorescence light unit and Olympus
xcellance software. Images were captured every 15 s. Rhodamine 123 (Rh123) used at 30 μM concentration accumulates in healthy mitochondria and is released to the cytosol upon depolarization of the mitochondria [
61]. This release is shown by increment in fluorescence from the cell. Since all cells expressed GFP, the red spectrum of Rh123 was recorded by using filters suitable for red fluorescent protein.
The cells were exposed to severe hypoxia for 15 min. Hypoxia was obtained by removing most of the oxygen by nitrogen perfusion of the flow solution. The oxygen scavenger Sodium Dithionite (0.75 mM) (Sigma-Aldrich) induced severe hypoxia. Control experiments were done with Sodium Dithionite perfusing the fluid with oxygen to control for any toxic effects other than caused by hypoxia. These control experiments did not show toxic effects on mitochondrial function. The protonophore Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) (1 μM) (Sigma-Aldrich) was applied before the end of all experiments. FCCP fully depolarizes mitochondria, and thereby causes release of any remaining dye from the mitochondria. In this way any remaining mitochondrial membrane potential in the cell is revealed by an increase in Rh123 fluorescence.
Immunolabeling
GIC cultures were plated in 24-well plates (Sarstedt) or chamber slides (Sigma-Aldrich) pretreated with Retronectin 50 μg/ml (Takara) and incubated overnight to facilitate cell adhesion. Cells were fixed with 4 % PFA and washed with PBS. Tissue slides were fixed in 10 % buffered formalin and washed with PBS. Immunolabeling was performed as previously described [
62]. Briefly, permeabilization in 0.1 % Triton, was followed by blocking with 5 % BSA and 5 % blocking serum, and incubations firstly with primary antibody in 0.1 % Tween-20 (o/n) and then with secondary antibody (2 h). The protocol was slightly modified for the co-staining of NAT12/NAA30 and CD31 and 0.1 % Triton was replaced with 0.1 % (w/v) Saponin (Sigma-Aldrich). After staining of nuclei with Hoechst 33342 (Invitrogen) the slides were mounted with antifade reagent (ProLong Gold, Invitrogen).
For detection of NAT12/NAA30 protein we used the primary antibody #AV48508 [(Sigma-Aldrich), rabbit 1:200]. For confirmation of the obtained results we used #NBP1-70631 [(Novus Biologicals, Littleton, CO, USA), rabbit 1:200]. Immunofluorescence images shown in the article were obtained with #AV48508. We used primary antibodies against human nestin (NES) [#ab6320 (Abcam, Cambridge, UK), mouse 1:400], CD31 [#ab54211 (Abcam), mouse 1:100] and SOX2 [#AF2018 (R&D), goat 1:100].
Secondary antibodies were: Alexa Fluor 488 [#A-11059 (Invitrogen), donkey anti-mouse, 1:500], Alexa Fluor 555 [#A-31572 (Invitrogen), donkey anti-rabbit, 1:500], Alexa Fluor 594 [#R37119 (Invitrogen), donkey anti-rabbit, 1:500], and Alexa Fluor 594 [#A-11058 (Invitrogen), donkey anti-goat, 1:500].
Flow cytometry
Spheres from GIC culture T65 were dissociated into single cells, and incubated overnight to recover. Cells were pelleted, and incubated with primary conjugated antibodies against CD133/2 –PE conjugated (Miltenyi biotec) and SSEA1-PE (Becton-Dickinson). Cells were then analyzed using a LSRII flow cytometer (Becton-Dickinson).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AAM participated in the design of the study, carried out cell culture work, immunolabeling, in vivo work, statistical analysis and drafted the manuscript. ZG participated in cell culture work and performed western blots. HS carried out the mitochondrial depolarization experiments and helped to draft the manuscript. AF participated in the work related to the transplantation experiments. ML participated in cell culture work and functional assays. MJ participated in the preparation of plasmids and viral transduction. MSA participated in planning and performing western blots. WM provided the NSC cultures and their microarray data and proofread the manuscript. EVM designed and performed flow cytometry, participated in discussion of individual experiments and helped to draft the manuscript. IAL initiated the study of stem cells, provided the biological material, had the general supervision of the lab, and participated in discussion of individual experiments and edited the manuscript. BS concieved the study, participated in the design of the study, established the shRNA technology, conducted qPCR, microarray and the bioinformatic analysis, participated in cell culture work and western blots and drafted the manuscript. All authors read and approved the final manuscript.