Sensitive (Gli36ΔEGFR-1 and L0627) or resistant (Gli36ΔEGFR-2) to TMZ GBM cells representative of classical subtype were used in this study. Human GBM Gli36ΔEGFR cells (kind gift of Dr. David Louis, Molecular Neurooncology Laboratory, MGH, Boston, MA) (26, 27) carry a mutant epidermal growth factor receptor (Δ2-7, EGFR). Gli36ΔEGFR cells were called Gli36ΔEGFR-1 to underline the sensitivity to Temozolomide (TMZ) treatment compared to the cell line obtained after treatment with sub-lethal doses of TMZ (50 µM of TMZ for 1 month) defined as Gli36ΔEGFR-2 (28). Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) with high glucose supplemented with 10% heat-inactivated Foetal Bovine Serum (FBS), and 50 IU/ml Penicillin/Streptomycin (P/S), 2 mM glutamine (all Euroclone, UK) at 37°C in a 5% CO₂/95% air atmosphere. L0627 GBM CSCs, established at the Neural Stem Cell Biology Unit, San Raffaele Scientific Institute, Milan, Italy and validated in Narayanan et al. (29) and Mazzoleni et al. (30) were cultured under the conditions of the NeuroSphere Assay (NSA) (31). GBM cells were carefully cultured and monitored and in vitro displayed a typical growth pattern and phenotype.

10,000 cells/cm² Gli36ΔEGFR-1 and Gli36ΔEGFR-2 cells were exposed to different concentrations of TMZ (0.1, 1, 5, 10, 25, 50, 100, 200 µM) (Sigma Aldrich, St. Louis, MO, USA) to determine the optimal dose able to distinguish TMZ sensitivity. Then, 10 mM of MET (Sigma Aldrich, St. Louis, MO, USA) alone or in combination with 25 µM of TMZ was added to the medium once at the beginning of the experiment and cell growth was monitored after 24, 48 and 72 hours (h). Cell viability was evaluated by Trypan blue exclusion test. The effect of MET, TMZ or MET plus TMZ was determined as growth inhibition rate and measured as: [1-(C_f/C₀)_A/(C_f/C₀)_V]*100, where C_f is the cell number at the point analyzed, C₀ is the cell number at the beginning of treatment, A is the corresponding drug and V is the vehicle as previously described (12). For L0627 cells, short-term proliferation/survival studies were performed as previously described (32). Apoptosis or necrosis were assessed by Real time-Glo Annexin Apoptosis and Necrosis Assay (Promega Corporation, Madison, Italy).

RNA was extracted using the commercially available illustra RNAspin Mini Isolation Kit (GE Healthcare, Italy), according to manufacturer’s instructions. Total RNA was reverse-transcribed to cDNA using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientifics, USA). Real-time PCR was performed in duplicate for each data point by using the Sybr Green technique and the oligonucleotides used were: β-actin (FRW: TCAAGATCATTGCTCCTCCTG, REV: CCAGAGGCGTACAGGGATAG); bax (FRW: ATG GAC GGG TCC GGG GAG; REV: ATCCAGCCCAACAGCCGC); bad (FRW: CCCAGAGTTTGAGCCGAGTG; REV: CCCATCCCTTCGTCGTCCT); bcl-2 (FRW: GATTGTGGCCTTCTTTGAG, REV: CAAACTGAGCAGAGTCTTC); sox2 (FRW: GCACATGAACGGCTGGAGCAACG; REV: TGCTGCGAGTAGGACATGCTGTAGG). Changes in the target mRNA content relative to housekeeping (β-actin) were determined with the ΔΔct method. Basal level expression of sox2 gene was expressed as difference between the target mRNA content and the housekeeping (β-actin) (Δct).

Animal experiments were carried out in compliance with institutional guidelines for the care and the use of experimental animals, which have been authorized by the Italian Ministry of Health (n°220/2016-PR and n°378/2019-PR). Seven to eight weeks old female nude mice (Envigo RMS srl, San Pietro al Natisone, Italy) were housed at constant temperature (23°C) and relative humidity (40%) under a regular light/dark cycle, with food and water freely available. The orthotopic tumor models were obtained by the stereotactic injection of 3-5*10⁵ cells (Gli36ΔEGFR-1, Gli36ΔEGFR-1 or L0627) in 2 µl of plain DMEM with a 10 µl Hamilton syringe as previously described (10). After cells injection, mice were monitored every day for body weight and clinical signs of disease (fur, eye, motor impairment) and sacrificed at the appearance of evident signs of illness or at the loss of more than 25% of the initial body weight. Firstly, we performed a pilot study on mice inoculated with Gli36ΔEGFR-1 and with Gli36ΔEGFR-2 (n = 6 per each cell line) to monitor tumor growth with MRI at day 5 after surgery and the sensitivity to TMZ. Based on data obtained on pilot study, we decided to start drug administration 7 days after cells inoculation and perform overall survival studies. Gli36ΔEGFR-1 (R-1), Gli36ΔEGFR-2 (R-2), and L0627 tumor bearing mice were randomly assigned to 4 groups of treatment, according to the following scheme: Group A (n = 5 R-1; n = 13 R-2; n=10 L0627) received daily oral administration of Temozolomide (TMZ, 70 mg/kg) in 10% DMSO, 5 days for a 28 days cycle and repeated with this scheme (5/28) until sacrifice of animal; group B (n = 6 R-1; n = 5 R-2; n=5 L0627) received intra peritoneal (i.p.) administration of Metformin (MET, 250 mg/kg) in saline for 5d/wk for the entire treatment period; group C (n = 5 R-1; n = 6 R-2; n=8 L0627) received the combination of daily oral administration of TMZ (70 mg/kg) 5 days for a 28 days cycle and i.p. daily administration of MET (250 mg/kg); group D (n = 9 R-1; n = 9 R-2; n=10 L0627), as vehicle group, received vehicle administration (10% DMSO in saline by oral gavage and 100% saline i.p.). The treatment schedule was decided on the basis of previous studies (12) and of the dose regimen used in clinical practice adapted to mice body surface ( Figure S1 ) (33). The tumor mass presence was confirmed in a subset of animals (n = 3 per treatment group) using MRI as described below. At the onset of signs of severe illness (weight loss > 25% or hemiplegia), mice were sacrificed under anesthesia and brains collected for histological analyses. Treatment efficacy was evaluated as time to sacrifice indicated as “overall survival” using the Kaplan-Meier estimator.

Mice brains were collected at sacrifice and fixed in 10% buffered formalin (Sigma-Aldrich) as described (34). After standard histological samples processing, serial 3 um-thick brain sections were cut and stained with hematoxylin and eosin (H&E) for morphological evaluation or probed with the following primary antibodies: Ki67 (Agilent Technologies, Santa Clara, US.), Sox2 (Cell signaling technologies Leiden, The Netherlands), cleaved-Caspase 3 (Cell Signaling Technologies, Palo Alto, CA, USA), and Iba1 (Wako pure Chemical Ind. Ltd.). Afterword, slides were incubated with Rodent Block R immunohistochemical reagent (Biocare Medical) before secondary antibody addition. Finally, slides were revealed using DAB as chromogen and counterstained with hematoxylin. Stained slides were scanned using Aperio digital scanner instrument (Leica Microsystems) and Iba1 expression was analyzed using the cytoplasmic algorithm available within ImageScope software (Leica Microsystems) after optimization of cell recognition parameters (12). Briefly, Iba1 quantification was performed by drawing regions of interest of about 0.21 mm² inside the tumor, in the peripheral inner and outer part of the mass to evaluate intratumor, peritumor or distant active microglia, respectively. Ki67 marker was quantified by drawing the same regions of interest only inside the tumor. The mitotic index in each sample was scored as the average of mitotic cells per 10 High power fields (HPFs; 40X objective) on H&E slides as previously described (35). If the tumor area was smaller than 10 HPF, then the whole tumor tissue was examined for presence of mitotic cells. In supplementary table 1 and 2 mice used in the study are summarized indicating for each animal, treatment condition and day of sacrifice.

MRI was performed with a 7T small animal magnetic resonance scanner (Bruker, BioSpec 70/30 USR, Paravision 5.1, Germany), equipped with 450mT/m gradients (slew-rate: 3400–4500T/m/s; rise-time: 140ms). A phased-array mouse-head coil with four phased-array channels was used as receiver, coupled with a 72 mm linear-volume coil as transmitter. Mice were anesthetized with isoflurane (2% in oxygen) and positioned prone on a dedicated heated apparatus, to prevent hypothermia. A coronal 2D High Resolution (HR) Rapid Acquisition with Relaxation Enhancement (RARE) T2 and a RARE T1 were acquired. After the injection of 0.2 µl/g of gadobutrol (Gadovist, Bayer Schering Pharma, Berlin-Wedding, Germany), acquisition of the RARE T1 was repeated. Tumor volume was calculated by manual contour of the post-contrast RARE T1 sequence made by an expert neuroradiologist. Gd-T1-weighted MRI was conducted to verify the presence of lesion before the beginning of treatment and for manual co-registration with PET images performed using PMOD 3.2 software.

PET imaging was performed with 3′-deoxy-3′-[¹⁸F]fluorothymidine ([¹⁸F]FLT) and [¹⁸F]VC701 to assess proliferation related to TK1 and TSPO receptor expression, respectively. [¹⁸F]FLT and [¹⁸F]VC701 uptake was evaluated by PET in distinct groups of Gli36ΔEGFR-1, Gli36ΔEGFR-2 and L0627 mice during control condition ([¹⁸F]FLT: n = 11 for R-1, n = 11 for R-2 and n = 9 for L0627; [¹⁸F]VC701: n = 9 for R-1, 11 for R-2 and n = 11 for L0627) and 1 week after the beginning of treatment (vehicle, TMZ, and TMZ plus MET). MET treated animals did not perform imaging. The sample size of each group is indicated in the figures.

Mice were injected via the tail vein with 4.18 ± 0.28 MBq of [¹⁸F]FLT and 4.79 ± 0.91 MBq of [¹⁸F]VC701. PET acquisitions were performed at 60 min ([¹⁸F]FLT) or 120 min ([¹⁸F]VC701) after tracer injection using the YAP-(S)-PET II small animal tomograph (ISE s.r.l., Pisa, Italy) or X-ß-CUBE (Molecubes, Gent, Belgium) as already described (36, 37). All the radiopharmaceuticals injected had a radiochemical purity greater than 99%. PET images were acquired in 3D mode. All the images were co-registered to MRI and quantified with PMOD 3.2 software (Zurich, Switzerland). The quality of co-registration was judged and confirmed by two independent experts in the field (S.V. and I.R.). Two different volumes of interest (VOI) were defined: (i) a control VOI covering the left striatum (volume 7 mm³) was drawn on the axial MR images, adjusted on the other imaging planes and then copied on the PET images of each mouse; (ii) a second glioma-covering VOI was drawn in the tumor-affected brain hemisphere and centered on mice lesions. For quantification, Standardized Uptake Value (SUV) was calculated according to the formula: SUV = tumor concentration activity [MBq/g]/(injected activity [MBq]/animal weight [g]). Maximum tracer uptake in tumor was normalized to the corresponding mean values of uptake of the contralateral control VOI (background) and indicated as tumor to background ratios (T/B ratio). A similar analysis for normal brain parenchyma was performed on normal mice.

In vitro experiments were repeated three times giving reproducible results. Data are presented as mean values ± standard deviation (SD) or standard error of the mean (SEM) of three independent experiments. For statistical analysis, non-parametric t-test, one-way analysis of variance (ANOVA), followed by Tukey’s multiple comparison test, or two-way ANOVA, followed by Bonferroni’s multiple comparison test, were performed using Prism 5 (Graph Pad Software Inc., CA, USA). Log-rank Mantel-Cox test was performed for survival comparison followed by Holm-Sidak method for multiple comparisons correction. Differences were considered statistically significant when p < 0.05.

We identified the minimal in vitro dose of TMZ able to reduce Gli36ΔEGFR-1 (TMZ-sensitive cells) but not Gli36ΔEGFR-2 (TMZ resistant cells) viability. This dose was defined as 25 μM for 48 h ( Figure S2 ). Then, 10 mM of MET alone or in combination with 25 µM of TMZ was added to the medium once at the beginning of the experiment and growth of Gli36ΔEGFR-1 and Gli36ΔEGFR-2 cells was monitored after 24, 48 and 72 hours. The treatment with MET alone induced a significant reduction in cell growth rate only in Gli36ΔEGFR-1 compared to vehicle (p < 0.01 at 48 h and p < 0.001 at 72 h). Furthermore, the growth rate of Gli36ΔEGFR-1 was significantly reduced after TMZ treatment at any time (p < 0.001 at 24, 48 and 72 h). The combination of TMZ and MET (TMZ+MET) displayed an additive effect at 72 h for Gli36ΔEGFR-1 (p < 0.01 vs TMZ) and reverted TMZ resistance of Gli36ΔEGFR-2 cells already after 24 h of combined treatment (p < 0.01 at 24 and 48 h and p < 0.05 at 72h) ( Figure 1A ), confirming the potential efficacy of MET in reducing cell resistance to TMZ ( Figure 1B ). The combination of TMZ+MET significantly reduced Gli36ΔEGFR-1 cell growth also compared to MET alone (p < 0.01 at 24 h, p < 0.001 at 48 and 72 h). These results were confirmed in a patient-derived Cancer Stem Cell (CSC) line (L0627) that shows features typical of the classic molecular subgroup, such as the overexpression of EGFR gene (30). Also in this case, the combination of TMZ+MET significantly decreased cells survival compared to single therapy at 72 h (p = 0.004 vs TMZ and p = 0.012 vs MET) ( Figure S3 ).

In line with the observed effect on cell viability, the association of TMZ and MET treatments deregulated the balance between pro-survival (bcl2) and pro-apoptotic (bax/bad) Bcl-family members in both Gli36ΔEGFR-1 and Gli36ΔEGFR-2 cells ( Figures 1C, D ) as indicated by the increase of bad/bcl-2 and bad/bcl-2 ratios. On the contrary, TMZ alone increased bad/bcl-2 mRNA expression only in Gli36ΔEGFR-1 cells and to a lower extent compared with TMZ+MET treatment. To evaluate the apoptosis induced by treatments, cells were assessed for exposure of Phosphatidylserine (PS) on the outer leaflet of the cell membrane. Supplementary figure 4 showed that only TMZ in Gli36EGFR-1 displayed secondary necrosis with PS translocation to the outer leaflet and loss of membrane integrity. This result suggested that MET and TMZ+MET showed early apoptosis characterized by PS translocation to outer leaflet but no cell membrane disruption ( Figure S4 ).

SOX2 has been reported to play a pivotal role in developing drug resistance of glioma (11, 38) and its expression was correlated with the grade of malignancy and favored the maintenance of an undifferentiated state of cancer stem cells (39, 40). Furthermore, recently its role as glioma stem cell biomarker has been described in association with NANOG and OCT (41, 42).

We performed qRT-PCR to investigate if the different treatments could modulate sox2 expression in TMZ-sensitive and -resistant cells. The basal sox2 mRNA level was similar in Gli36ΔEGFR-1 and Gli36ΔEGFR-2 cells (0.082 and 0.079 respectively). Single-drug treatments did not alter sox2 expression in Gli36dEGFR-1 cells. In contrast, the combination of the two drugs significantly reduced sox2 (p = 0.0005) levels ( Figure 1E ). On the contrary, in Gli36ΔEGFR-2 all treatments led to a dramatic increase of sox2 expression. However, the increment of sox2 transcript was significantly less pronounced with the combinatorial treatment ( Figure 1E ). These findings suggest that in TMZ-sensitive cells, the chemotherapeutic perturbed cell viability and partially reduced the determinants of therapy resistance. In this system, the addition of MET was synergistic. On the contrary, TMZ-resistant cells showed an enhanced multi-drug resistant phenotype and MET administration was only able to partly reduce TMZ-induced sox2 overexpression. sox2 is also expressed on L0627 cell (personal communication from R. Galli) but the effects of drugs were not evaluated.

In a pilot study, after 5 days from cell injection, all animals were positive for intracerebral lesions at MRI (tumor volume: 4.38 ± 1.67 mm³ and 6.12 ± 2.11 mm³ for Gli36ΔEGFR-1 and Gli36ΔEGFR-2 respectively). In the same study we confirmed in vivo the different sensitivity of Gli36ΔEGFR-1 and Gli36ΔEGFR-2 to TMZ (% tumor increase after treatment: 22.7 vs 54.9).

Then, we analyzed the efficacy of MET and TMZ treatments alone or in combination in orthotopic mouse models of glioma obtained by Gli36ΔEGFR-1 or Gli36ΔEGFR-2 cells inoculation. The MET protocol alone did not show any effect on survival of both GBM models. As expected, in the TMZ-sensitive Gli36ΔEGFR-1 model, TMZ treatment significantly increased mice overall survival compared to animals treated with vehicle (median survival 63 days vs 17 days, p = 0.0066). Among TMZ-treated mice only one mouse was histologically confirmed as disease-free, while others mice showed tumor relapse detected by MRI or IHC ( Figures S5A, B ). More interestingly, the combination of TMZ+MET extended the overall survival of Gli36ΔEGFR-1 mice up to 90 days, when we decided to sacrifice all mice which were histologically confirmed as disease-free ( Figure 2A ). In mice bearing TMZ-resistant glioma, TMZ alone did not prolong mice survival compared with that of vehicle-treated mice (18 days vs 17 days). In this case, TMZ+MET protocol slightly increased median survival time up to 23 days ( Figure 2B ). EGFR-2 bearing mice regardless of treatment displayed tumor increase at MRI after one week ( Figures S5C, D ).

When CSC line L0627 was tested in the same settings, orthotopically transplanted mice displayed tumor masses at later times (tumor onset at approximately 40 days) in comparison with Gli36ΔEGFR mice. Confirming our previous data, also in the L0627 model, treatment with MET alone did not show any therapeutic efficacy (median survival 54.5 days and 54 days for control and MET, respectively) but, most remarkably, a significant increase of survival rate could be detected in animals treated with the MET+TMZ combination in comparison with TMZ alone group (p = 0.0058) ( Figure 2C ). In addition, after an initial response, all mice treated with TMZ showed tumor relapse at the MRI, whereas only 3 out of 8 TMZ+MET treated animals developed tumor recurrence ( Figures S6A , B ).

To gain insights into in vivo effects of the different therapeutic protocols, we evaluated glioma proliferation index and inflammation (i.e. microglia/macrophage activation) in mouse brains collected after cancer-related death or sacrifice. Accordingly, Ki67 and Iba1 staining were performed and quantified in different brain areas ( Figure 3A ).

Gli36ΔEGFR-1-tumors treated with vehicle showed a mean proliferation index of 58.4%. In the TMZ arm one mouse did not display tumor mass whereas the two tumors processed displayed a proliferation index of 44.4% and 45.3%. No tumor mass was detected in all samples from the TMZ+MET arm (n = 5). For this reason, we also performed STEM121 staining, a marker able to identify human cells in a mouse tissue, which confirmed the absence of human cells. Gli36ΔEGFR-2 tumors from vehicle-treated animals had a mean proliferation index of 40.1%, whereas those from the TMZ or the TMZ plus MET arms showed a mean Ki67 staining of 48.6% and 40.9%, respectively ( Figures 3B–D, F ).

When we looked at microglia/macrophage activation, we did not observe any difference between vehicle and TMZ-treated Gli36ΔEGFR-1 tumors ( Figure 3E ); after treatment with TMZ plus MET no signs of tumor or Iba1 markers were visible. In Gli36ΔEGFR-2 generated glioma, TMZ+MET significantly reduced infiltration of Iba1 positive myeloid cells only at the tumor-brain border compared to TMZ (p = 0.005 versus TMZ, Figure 3G ). On the other areas data are very heterogeneous.

For the L0627-generated tumors we scored the mitotic count in three high power fields per mouse and tumor cell viability using an antibody toward cleaved-caspase 3, because Ki67 staining was surprisingly absent.

This analysis showed that tumor proliferation was decreased by TMZ or MET individual treatments (20.88 for vehicle, 14.38 for TMZ, 13.90 for MET), but only TMZ marginally increased glioma cells apoptosis (0.68% for vehicle, 4.35% for TMZ and 2.56% for MET) ( Figures S7A, C ). Again, the combinatorial treatment significantly impaired the ability of the glioma cells to in vivo grow, one mouse negative-defined at MRI showed a small tumor mass at sacrifice. In regards of the glioma stem cell markers, we observed a high Nestin staining in all L0627 tumor samples and positive staining for Iba1 within and in the peripheral areas of the tumor and in cerebral hemisphere contralateral to tumor implantation ( Figure S7B ). Interestingly, both MET+TMZ-treated analyzed mice (both that with a small tumor mass and that disease-free) displayed Iba1-positive staining only in the cerebral hemisphere contralateral to tumor implantation indicating an activation of macrophages/microglia.

We used PET with [¹⁸F]FLT to determine if this tracer could identify responder from non-responder mice after treatment. Gli36ΔEGFR-1 and Gli36ΔEGFR-2 generated tumors displayed similar [¹⁸F]FLT pre-therapy uptake values, i.e. a SUVmax value of 0.28 ± 0.06 and of 0.32 ± 0.07, respectively ( Figures 4A–D ).

Early after therapy start (day 7), [¹⁸F]FLT uptake increased in all vehicle-treated mice (either carrying a Gli36ΔEGFR-1 or Gli36ΔEGFR-2 tumors) compared to basal levels (p < 0.001; Figure 4C ). All treatment protocols significantly reduced the increase of [¹⁸F]FLT uptake both in Gli36ΔEGFR-1 bearing mice (p < 0.001 versus vehicle; Figures 4A, C ) and in Gli36ΔEGFR-2 glioma, although to a lower extent ( Figures 4B, D ).

We successfully identified a cut-off value of [¹⁸F]FLT PET-SUVmax able to discriminate TMZ-responder (considering both TMZ- and TMZ+MET-treated animals) from therapy-resistant tumors. Appling Receiver Operating Characteristic (ROC) curve to post-treatment [¹⁸F]FLT PET data, we identified a threshold value of SUVmax (corresponding to 0.3295) with a sensitivity and specificity of 77.8% and 100%, respectively, in distinguishing animals response to treatments ( Figures 4E, F ). Therefore our data suggest that [¹⁸F]FLT-PET may be used to monitor treatment responses in glioma and to identify patients as responders or not after treatment. Unfortunately, [¹⁸F]FLT PET was not able to predict response duration.

This study was repeated also in the CSC-transplanted mice. In agreement with the results on Ki67, we didn’t observe any significant uptake of [¹⁸F]FLT in L0627 tumors independently from the lesions’ dimension or the treatments’ protocol, possibly due to the lowly proliferative, highly infiltrative nature of CSCs and to the maintenance of Blood Brain Barrier (BBB) integrity in this specific GBM model (43).

After 7 days from the beginning of treatment, vehicle- or TMZ-treated Gli36ΔEGFR-1 glioma mice displayed the same uptake of [¹⁸F]VC701 of the pre-treatment mice. On the contrary, the TMZ plus MET protocol in Gli36ΔEGFR-1 glioma led to a significant decrease of [¹⁸F]VC701 uptake (0.15 ± 0.07 SUVmax) not only in comparison with mice treated with vehicle (0.35 ± 0.07, p < 0.01) or TMZ alone (0.31 ± 0.08, p < 0.01) but also in comparison with pre-treatment uptake (0.29 ± 0.07, p < 0.05) ( Figures 4B , 5A ). Moreover, in this model we observed an increased uptake of [¹⁸F]VC701 in brain regions contralateral to tumor mass regardless of treatment conditions compared to that of healthy brains (p < 0.05), except animals treated with TMZ plus MET (SUVmax values: pre-treatment = 0.27 ± 0.05; vehicle = 0.28 ± 0.06; TMZ = 0.29 ± 0.08; MET plus TMZ = 0.14 ± 0.07 and healthy brain = 0.16 ± 0.05) ( Figures S8A, B ). In Gli36ΔEGFR-1 samples, the expression of Iba1 staining was very heterogeneous in the cerebral hemisphere contralateral to tumor implantation ( Figure 3B ) and it was correlated to expression of Iba1 in outer border of tumor (R2 = 0.68, p = 0.02). No expression of Ki67 proliferation marker was detected in these areas. The increased uptake observed in the normal brain parenchyma of tumor bearing animals suggests that Gli36ΔEGFR-1 tumors enhance the inflammation status of the whole brain, an effect that is blocked by the combination of MET to TMZ. Interestingly, within the TMZ-responder Gli36ΔEGFR-1 group (TMZ- and TMZ+MET-treated animals) identified with [¹⁸F]FLT PET we observed a significant inverse correlation between [¹⁸F]VC701 uptake and survival (r = -0.8674, p = 0.0252) with long-term survived mice (disease-free mice at 90 days independently by type of treatment) displaying lower [¹⁸F]VC701 uptake values ( Figures 5C, D ).

On the contrary, in Gli36ΔEGFR-2 tumor bearing mice, [¹⁸F]VC701 uptake was lower in comparison with Gli36ΔEGFR-1 bearing mice and not affected by treatment ( Figure 5B and Figures S8A, B ).

Also in this case, L0627 tumors displayed a different behavior. We observed a slight but significant decrease of [¹⁸F]VC701 uptake in the tumor after 7 days of treatment with TMZ plus MET in comparison with vehicle animals and no significant modifications in the contralateral part of the brain both in untreated and treated mice ( Figure S9 ).