Cell lines were obtained from PAs of different grades, developed in 3–16 years old patients. Cell lines were established in vitro as monolayer cultures, with doubling times between 24–48°h, and showed a mixture of stellate and bipolar morphologies, with cells presenting polygonal, cuboidal or flattened appearance. Low-grade glioma cell lines were isolated from fresh tissues collected from the surgical theater. Briefly, fresh tissues were collected in complete DMEM medium supplemented with 20% FBS and antibiotics (Sigma-Aldrich, St. Louis, MO, United States), then mechanically minced straight away and left in culture. Once cells started growing, the FBS concentration in culture medium was reduced to 10%. Cells were followed in culture for weeks, and the proliferative capacity was assessed calculating the number of population doublings (PDs), using the formula log₁₀ [total no./start no.]/ml until cells proceeded to senescence [34, 35]. Of all attempts, two cell cultures (named LGG1 and LGG2) reached 11 passages in culture over a period of 5 months. Genotyping studies of these cells showed a good match compared to their corresponding original tissues, indicating that no contamination had taken place in culture. The high-grade glioma (HGG) cell lines, A172 and MG-U87, used for comparative studies on NGF effects, were purchased from ATCC® (Manassas, VA, United States).

Human recombinant NGF was kindly provided by Dr L. Manni (Institute of Translational Pharmacology, CNR, Rome, Italy). Antibodies (Abs) against β-actin, TrkA, p75, GFAP, p‐16, used for Western Blot were from Santa Cruz Biotechnology (Dallas, TX, United States). Abs anti‐extracellular signal regulated kinase (ERK)1/2, anti‐p‐ERK1/2 (Thr202/Tyr204), anti-AKT, anti‐p‐AKT (Ser473), anti-p38 MAPK and anti-pp-38 MAPK (Thr180/Tyr182) were from Cell Signaling Technology (Danvers, MA, United States); anti-MASH-1 Ab was from Thermo-Fisher (Waltham, MA, United States). Abs for immunofluorescence studies were against: GFAP (Thermo-Fisher), vimentin (Santa Cruz Biotechnology), synaptophysin, S100 β, nestin, all from Sigma-Aldrich. Alexa Fluor 488 F (Ab)2 fragment of goat anti-rabbit IgG and Alexa Fluor 488 F (ab)2 fragment of goat anti-mouse IgG were used as secondary Abs for IF (Jackson Immunoresearch, Ltd. Europe). Eukitt from Sigma-Aldrich was employed as mounting solution.

The human epithelial cell line 293T was maintained in DMEM supplemented with 10% FCS. For transfection experiments, 15 μg of the packaging constructs pCMVDR8, 2,5 μg of the VSV-G expressing plasmid pMD.G and 10 μg of the transfer plasmid pHR0-CMV-hTERT-IRES2-GFP or the pHR0-CMV-IRES2-GFP control (a kind gift from Dr A. Cara, ISS, Rome) were introduced into 293T cells using the Calcium Phosphate Method (Promega, Madison, WI, United States). Transduction was performed by overnight incubation of PA cells with viral supernatants. Between 12–16 h after infection, the virus-containing media was removed, and cells were incubated with fresh media.

The expression of the reporter gene GFP was analyzed by evaluation of the integration of GFP into the target cell’s genome, as GFP expression is reduced when it is positioned as the second gene downstream of the IRES sequence, limiting its flow cytometric detection [36]. Integration of GFP into the cell genome was analyzed by polymerase Chain Reaction (PCR) amplification from purified genomic DNA. PCR analysis demonstrating GFP integration (250°bp product) was performed on control and hTERT transduced cells as described [34]. hTERT expression in transduced cells was also assessed by intracellular flow cytometry detection, following cell permeabilization with Cytofix/Cytoperm (BD, Franklin Lakes, NJ, United States) and staining with anti-hTERT Ab (Santa Cruz) and FITC goat anti-mouse IgG as secondary Ab (Thermo-Fisher).

For the wound healing assay, cells were grown to confluence and serum-starved for 24 h in DMEM without FBS. Monolayers were gently scraped to produce a wound area, then washed and cultured in the presence or absence of NGF (100 ng/ml). Progression of wound closing was monitored by taking phase-contrast pictures at 48 h by Canon Power shot G5 (Tokyo, Japan).

β-galactosidase assay, allowing detection of β-galactosidase activity at pH 6 in senescent cells, was performed following instructions provided by the manufacturer.

For immunofluorescence studies, briefly, cells were seeded and, when indicated, treated with NGF (100 ng/ml) for 72°h, fixed in 4%, paraformaldehyde, permeabilized in 0.3% triton X-100, incubated with primary Ab overnight at 4°C or 1 h at 37°C, then with secondary Ab and stained with DAPI before mounting. Slides were stored at −30°C in the dark until acquisition.

Cells were washed in cold PBS, scraped on ice and lyzed in ice-cold lysis buffer (0.5% sodium deoxycholate, 20 mM Tris–HCl, pH 7.4, 0.1 M NaCl, 1% Nonidet P40, 5 mM MgCl₂, 1 mM DTT, with protease inhibitors) for 20 min on ice. Samples were separated by SDS-PAGE, transferred to Hybond-P membranes (GE Healthcare, Chicago, IL, Unites States), probed with specific primary and horseradish peroxidase-conjugated secondary Abs, then binding visualized by chemiluminescence (GE Healthcare) with Chemidoc XRS System (Bio-RaD, Hercules, CA, United States). Western Blot analysis of MASH-1 expression was also performed in the presence of K252a, a pan Trk inhibitor (Sigma-Aldrich), [37]. Densitometry was performed using Imager ChemiDoc XRS System.

Genomic DNA was isolated from LGG1, LGG2 and LGG2-hTert cell lines using DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) following manufacturer’s protocol. Quality and concentration of the DNA samples were examined by NanoDrop ND-1000 (Thermo Fisher Scientific, Waltham, MA) and 25 ng of DNA was used to perform PCR amplification of BRAF exon15 with the following primers: 5’-CCT​AAA​CTC​TTC​ATA​ATG​CTT​GCT-3’ (forward) and 5’- GGC​CAA​AAA​TTT​AAT​CAG​TGG​A -3’ (reverse). PCR conditions were set up according to manufacturer’s instructions of Invitrogen™ Platinum™ II Taq Hot-Start DNA Polymerase (Thermo Fisher Scientific). The amplified products were purified using DNA clean and concentrator™-5 kit (Zymo Research, Irvine, CA) and sequenced in forward and reverse directions on ABI 3730 automated sequencer (Applied Biosystems Inc. Carlsbad, CA, United States).

Statistical significance of differences was determined by using the Student’s t test and two-way ANOVA. Results were considered statistically significant when p < 0.05.

LGG cell lines were established in vitro as monolayer cultures, with doubling times higher than 72 h and a mixture of stellate and bipolar morphologies, showing cuboidal/polygonal or flattened appearance. Immunofluorescence analysis confirmed the astrocytic nature of the cells through a range of astrocyte markers such as GFAP, vimentin, S100 β, synaptophisin and nestin [38, 39] (Figure 1A).

Cyclin D3 activity is required for cell cycle and G1/S transition, and is modestly expressed in LGG cells [40]. LGG1 cells expressed cyclin D3 at the start of the culture, showing a progressive decline along the passages in culture until disappearance (Figure 1B). LGG2 cells showed similar cyclin D3 baseline expression (data not shown).

Recently, the BRAF (v-raf murine sarcoma viral oncogene homolog B) pathway, involved in the highly oncogenic RAS/RAF/MEK/ERK signaling pathway [41], has become a promising molecular target for personalized LGG therapy [42]. The most frequent mutation is a single nucleotide substitution of thymine to adenine at nucleotide 1799 that converts valine (V) to glutamic acid (E) at amino acid 600 (V600E mutation) [43].

Moreover, AKT activation has been shown to be associated with a clinically aggressive pylocitic astrocytoma phenotype [44].

Then, in order to integrate the phenotypical characterization illustrated in Figure 1, we analyzed the levels of AKT and ERK activation by Western Blot, and scrutinized the V600E BRAF mutation state of the two patient-derived LGG1 and LGG2 lines. LGG1 cells showed higher levels of AKT activation (Figure 1C), while absence of BRAF V600E mutation was observed in both LGG lines as shown by the sequencing electropherograms of Figure 1D.

When we analyzed the proliferative potential of the two patient-derived lines, we observed that LGG2 progressed in culture through a lower number of PDs compared to LGG1 cells (Figure 2A). PD refers to the total number of times the cells into a population have doubled since their primary isolation in vitro, providing an evaluation as to when the cells will reach senescence [34]. Considering the lower proliferative capacity of LGG2 line compared to LGG1, we analyzed whether this distinctive behavior was reflected in a different response to NGF treatment in terms of delay of wound healing, as a measure of cell migration and proliferative potential. The wounds induced with pipette tips in the presence of NGF were sealed at a higher extent in LGG1 as compared with LGG2 cells at the time shown in the figure (Figure 2B). These data indicate a delay in the capacity of wound healing in LGG2, providing a qualitative prediction of cell wound healing efficacy over time and suggesting a different behavior of individual LGG samples under NGF challenge.

We investigated whether hTERT transduction could extend the proliferative potential of LGG2 cells without substantially modifying their morphological features or increasing their aggressiveness. LGG cells were grown for 6–7 PDs in vitro before transfection with either pHR0-CMV-hTERT-IRES2-GFP or pHR0-CMV-IRES2-GFP control lentiviral vectors. Transduction successfully produced a cell line derived from LGG2 cells, which we named hTERT-LGG2, endowed with enhanced proliferative capacity as compared with the LGG2 control cells (up to 50 PDs, Figure 3A), which reached replicative senescence rapidly, after limited rounds of duplication (Figure 3A).

Transduction with hTERT slightly increased cyclin D3 expression (Figure 3B). In particular, hTERT-LGG2 cells analyzed after 20 PDs showed higher cyclin D3 levels compared to those expressed by control cells evaluated after 10 PDs, when their proliferation dramatically decreased (Figure 3A,B). However, levels of cyclin D3 observed in hTERT-LGG2 cells were still much lower compared to those observed in the HGG cell line A172, used as positive control (Figure 3B). Remarkably, hTERT-LGG2 cells did not show the characteristic enlarged and flattened morphology typical of senescent cells (Figure 3C). Accordingly, while LGG2 control cells expressed high levels of β-galactosidase after 10 PDs, in correspondence of the cessation of their persistence in culture, hTERT-LGG2 expressed significantly lower levels of the senescence-related marker for at least 20 PDs (Figure 4A).

The analysis of the integration of the GFP sequence of hTERT expressing vector into the LGG2 cell genome confirmed that transduction had occurred efficiently (Figure 4B). Overexpression of hTERT in LGG2 cells was confirmed by intracellular FACS analysis, as shown by the levels of integrated Mean Fluorescence Intensity (iMFI), a measure that combines the relative frequency of molecule expression with the Mean Fluorescence Value (MFV), and by a representative FACS histogram plot showing high levels of intracellular hTERT expression after 35 PDs (Figure 4C). Absence of BRAF V600E mutation in LGG2 cells was confirmed after transduction with hTERT as shown by the sequencing electropherogram illustrated in Figure 4D.

Treatment with increasing concentrations of NGF produced distinctive responses in the different glioma cells analyzed, in terms of cell survival. HGG cells, used as control, showed some increase in cell growth, either with non-dose-dependent (U87) or dose-dependent behavior (A172) (Figure 5A).

Differently, LGG cells showed either little proliferation increase (LGG1) or dose-dependent growth inhibition (LGG2 and hTERT-LGG2) (Figure 5A), confirming the results obtained with the “wound healing retard” assay. No statistically significant difference was found between LGG2 and hTERT-LGG2 response to NGF in terms of cell growth, while significant divergence was observed between LGG1 and LGG2/hTERT-LGG2. Noteworthy, both hTERT-LGG2 and parental LGG2 cells exhibited growth inhibition following NGF treatment, providing a first evidence that hTERT transduction did not affect the biological response of LGG cells to NGF treatment.

We analyzed whether the potential NGF-associated differentiation resulted in the induction of neurite outgrowth in LGG cells. Immunofluorescence staining confirmed the induction of cell neurite outgrowth in the presence of GFAP astrocyte marker and differentiation indicator S100β, both in LGG2 control and h-TERT-LGG2 cells under NGF treatment (Figure 5B), demonstrating that hTERT-transduced glioma cells retained the susceptibility to NGF-mediated differentiation shown by the parental counterpart.

Then, to further analyze whether hTERT-transduced cells could provide a valid tool to study the behavior of LGG cells in vitro, we compared the effect of NGF in LGG2 and its hTERT-transduced counterpart on critical biochemical features involved in cell differentiation and survival.

TrkA and p75 levels were not substantially affected by hTERT transduction (Figure 6A). Moreover, despite differences in the basal levels of ERK, both LGG2 and hTERT-LGG2 showed an increase in ERK phosphorylation. p16, evaluated as a marker of cell senescence, was decreased in hTERT-LGG2 cells, according to the strong inhibition of senescence observed in these cells, and not modified by NGF treatment both in hTERT- and control LGG2 cells (Figure 6A). Remarkably hTERT transduction strongly reduced the expression of p38 MAPK (Figure 6A).

By increasing the limited lifespan culture, transduction with hTERT provided supplementary LGG2 cells to study additional potential targets of NGF treatment. In particular, our attention focused on Achaete-scute complex homolog 1 (ASCL1 or MASH-1), a basic helix-loop-helix transcription factor essential for neuronal differentiation, which has been associated with astrocytoma progression [45, 46]. Treatment of hTERT-LGG2 cells with NGF was associated with a reduction of MASH-1 (Figure 6B). This effect was counteracted by cell-pretreatment with the pan Trk inhibitor K252a, demonstrating that TrkA receptor activation by NGF was required for MASH-1 modulation.

It should be emphasized that the goal of the present study was the increase of the proliferative capacity of low growing LGG cells in order to have a greater amount of biological material to be employed in preclinical studies of molecular characterization. It is well known the limitation of growing LGG cells, likely due to oncogene-induced senescence [20, 27]. So, the reason why we could conduct a limited amount of experiments especially in terms of molecular characterization, is that LGG2 control cells provided an insufficient amount of material to reproduce a greater number of experiments.