Neurochemistry International 129 (2019) 104485

Contents lists available at ScienceDirect

Neurochemistry International
journal homepage: www.elsevier.com/locate/neuint

Phospho-mTOR expression in human glioblastoma microglia-macrophage T

cells
Lucia Lisia,∗, Gabriella Maria Pia Ciottia, Marta Chiavaria, Michela Pizzoferratoa,
Annunziato Mangiolab, Sergey Kalininc, Douglas L. Feinsteinc,d, Pierluigi Navarraa,e
a
Institute of Farmacologia, Università Cattolica del Sacro Cuore, L.go F. Vito 1, Rome, Italy
b
Department of Neuroscience, Imaging and Clinical Sciences, Università degli Studi G. D'Annunzio Chieti-Pescara, via Colle dell’Ara 100, Chieti, Italy
c
Department of Anesthesiology, University of Illinois at Chicago, Chicago, IL, USA
d
Department of Veterans Affairs, Jesse Brown VA Medical Center, Chicago, IL, USA
e
Fondazione Policlinico Universitario Agostino Gemelli, L.go F. Vito 1, Rome, Italy

A R T I C LE I N FO A B S T R A C T

Keywords: The glioblastoma (GBM) immune microenvironment is highly heterogeneous, and microglia may represent
Glioblastoma 30–70% of the entire tumor. However, the role of microglia and other specific immune populations is poorly
Microglia characterized. Activation of mTOR signaling occurs in numerous human cancers and has roles in microglia-
mTOR glioma cell interactions. We now show in human tumor specimens (42 patients), that 39% of tumor-associated
TSC2
microglial (TAM) cells express mTOR phosphorylated at Ser-2448; and similar mTOR activation is observed
Tumor microenvironment
Molecularly targeted therapies
using a human microglia-glioma interaction paradigm. In addition, we confirm previous studies that microglia
express urea and ARG1 (taken as M2 marker) in the presence of glioma cells, and this phenotype is down-
regulated in the presence of a mTOR inhibitor. These results suggest that mTOR suppression in GBM patients
might induce a reduction of the M2 phenotype expression in up to 40% of all TAMs. Since the M2 profile of
microglial activation is believed to be associated with tumor progression, reductions in that phenotype may
represent an additional anti-tumor mechanism of action of mTOR inhibitors, along with direct anti-proliferative
activities.

1. Introduction macrophage-monocyte lineage belongs to the same conceptual frame-

work (Liang et al., 2018; Amin et al., 2018; Tsutsumi-Kuroda et al.,
Glioblastoma (GBM - grade IV glioma) is the most aggressive pri- 2018). These immunotherapeutic drugs include antibody-drug con-
mary brain tumor, characterized by a high rate of therapeutic failures jugates, peptide vaccines, autologous infusions of modified chimeric
despite aggressive treatments, and a poor prognosis. Tumor classifica- antigen receptor-expressing T cells, autologous dendritic cell vaccines,
tion is based on malignancy, nuclear atypical and infiltration of the oncolytic viruses, immune-stimulatory viruses, checkpoint blockade
surrounding brain parenchyma. Unfortunately, the tumor is often di- inhibitors, and drugs acting on innate immune cells (Miyauchi and
agnosed once the patients become symptomatic, at which time the le- Tsirka, 2018).
sion is already widely extended. The gold standard of therapy, estab- The GBM immune microenvironment is heterogeneous and can
lished over a decade ago, consists of fractionated radiotherapy in differ in GBM subtypes (Chen and Hambardzumyan, 2018). Although
combination with classical chemotherapy with temozolomide; how- manipulation of the tumor microenvironment was shown to be able to
ever, in almost all cases, chemotherapeutic resistance arises and re- repress GBM tumor progression (Arbab et al., 2017), the role of specific
currence is common after initial therapy (Weller, 2018; Lu et al., 2018). immune populations remains poorly characterized. Microglia are the
The clinical success of cancer immunotherapies targeting T-cell principal resident immune cells in the central nervous system (CNS) and
immune checkpoint receptors PD-1/PD-L1 has demonstrated the im- are thought to be versatile players in both inflammatory and physio-
portance of immune-evasion as a hallmark of cancer. On the side of logical contexts. Although the past decade has seen important progress
innate immune responses, the recent preclinical and clinical develop- in the understanding of multi-tasking microglia in GBM pathology, the
ment of novel CCR2 antagonists to control the trafficking of relative importance of microglia at different disease stages and how

∗
Corresponding author. Institute of Pharmacology, Università Cattolica Del Sacro Cuore, L.go F. Vito 1, 00168, Rome, Italy.
E-mail address: [email protected] (L. Lisi).

https://doi.org/10.1016/j.neuint.2019.104485
Received 23 January 2019; Received in revised form 1 May 2019; Accepted 3 June 2019
Available online 10 June 2019
0197-0186/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

microglia could be targeted for optimal therapeutic efficacy remain (Berlin, Germany). Bacterial endotoxin LPS (Salmonella typhimurium)
largely enigmatic. Glioma associated microglia (GAMs) constitute the was from Sigma–Aldrich (St. Louis, MO, USA). The rat recombinant
largest portion of tumor infiltrating cells, contributing between 30 and interferon-γ (IFNγ) was purchased from Endogen (Pierce
70% to the glioma mass (Wood and Morantz, 1979; Roggendorf et al., Biotechnology, Rockford, IL, USA). The human recombinant interleukin
1996; da Fonseca and Badie, 2013). Different profiles of activation co- 1β (IL1β), human IFNγ and recombinant human Tumor necrosis factorα
exist in the same tumor, depending on GAM location or stage of disease (TNFα) were purchased from R&D System. β-actin (clone AC-74) mouse
(Dello Russo et al., 2017). In particular, GAMs are exposed to factors monoclonal antibody was from Sigma Aldrich; rabbit polyclonal anti-
inducing them to produce cytokines and chemokines, which contribute phospho [ser-2448] mTOR and rabbit polyclonal anti-total mTOR were
to tumor growth and to maintain a pro-tumorigenic, im- purchased from Novus Biological (Littleton, CO, USA); mouse mono-
munosuppressed microenvironment. Therefore, a bi-directional inter- clonal p70 S6 was purchased from Santa Cruz Biotechnology; rabbit
action exists between microglia and GBM. In this context, we previously polyclonal 4EBP1 was purchased from Bethyl Laboratories; monoclonal
investigated the in vitro interactions between rat microglia and C6 rabbit Tuberin/TSc2 was purchased from Cell Signaling Technology,
glioma cells (Lisi et al., 2014a). Exposure to conditioned media ob- goat monoclonal AIF-1/Iba1 was purchased from Novus Biological.
tained from C6 cells taken under baseline conditions induced a pre- Arginase I (E−2) mouse monoclonal antibody was purchased from
dominant M2-like phenotype in the microglia. Conversely, if C6 cells Santa Cruz Biotechnology.
were exposed to a medium containing pro-inflammatory stimuli, the
subsequent exposure of microglia to such medium was followed by a 2.2. Cell cultures
shift towards an M1-like phenotype (Lisi et al., 2014a). In addition, we
investigated the status of microglia/macrophage activation in surgical The human microglia cell line (CHME-5; RRID:CVCL_5J53) was
specimens from 41 patients diagnosed with grade IV GBM. For each kindly provided by professor Pierre Talbot (Janabi et al., 1995). Human
patient we analyzed both the center of the tumor and the surrounding U87MG and T98G glioma cell lines were kindly provided by professor
parenchyma. Four different markers, namely IBA-1, CD163, iNOS and Grazia Graziani (Tor Vergata University- Rome). CHME-5, U87MG and
ARG-I were investigated and we showed that M2 markers (in particular T98G cells were grown in DMEM containing 10% FCS and antibiotics
CD163) rather than M1 markers could be envisioned as a prognostic and studies carried out at a low concentration of FCS (1%). Cells were
marker (Lisi et al., 2017a). passed twice a week and plated for experiments at 2 × 104 cells/well.
The phospho-inositide 3 kinase (PI3K)/protein kinase B (AKT)/ All the experiments received institutional approval.
mammalian target of rapamycin (mTOR) pathway is a well-investigated Conditioned media from activated glioma cells was generated fol-
signaling pathway that regulates diverse cellular functions including lowing a protocol which removed the proinflammatory stimulus from
proliferation, metabolism and transcription (Tee, 2018). Activation of the medium. Briefly, in preliminary experiments glioma cells were in-
mTORC1 is observed in numerous human cancers due to gain-of-func- cubated in the presence of mixture of cytokines (10 ng/ml TNFα, 10 ng/
tion mutations in oncogenes (PI3K, AKT or Ras) and/or loss-of-function ml IL1β, 10 UI/ml hIFNγ (TII) for 4, 8 or 24 h (Fig. 3). Proinflammatory
mutations in tumor suppressors (PTEN, LKB1, or TSC1/2) (Foster and mRNA levels were evaluated by real time quantitative PCR (qPCR).
Fingar, 2010). In addition, a direct role of mTOR in the modulation of Based on the results, the glioma-CM were prepared as follows:
glial functions has been described. Data from ourselves and other
groups support the notion that mTOR is involved in glial pro-in- A) B-CM: 4 h incubation in plain medium, followed by 3 washes with
flammatory activation (Dello Russo et al., 2009; Lisi et al., 2011) and in phosphate buffered saline (PBS) and addition of fresh plain medium
microglia-glioma interactions (Lisi et al., 2014b). In an in vitro rat for 24 h. After this second period of incubation, the CM was col-
model, we showed that the inhibition of mTOR polarizes glioma-acti- lected, centrifuged to remove cellular debris and stored as B-CM.
vated microglial cells towards the M1 phenotype, and in parallel pre- B) PS-CM: 4 h incubation with TII, followed by 3 washes with PBS and
vents the induction of the M2 status that promotes tumor growth (Lisi addition of fresh plain medium for 24 h. After this second period of
et al., 2014b). Although the relevance of mTOR in glioma biology is incubation, the CM was collected, centrifuged and stored as PS-CM.
now well established, initial studies with mTORC1 inhibitors, such as
rapamycin (RAPA) and its analogs, showed limited efficacy in clinical Both CM were stored at −80 °C until the experiments on CHME-
trials (Martelli et al., 2018). Despite such initial failures, newer mTOR 5 cells were performed.
inhibitors are being investigated; ClinicalTrials.gov currently lists 478 A flow chart of the studies shows the experimental design (Scheme
ongoing studies on mTOR and cancer, 28 out of them being focused on 1).
glioma. Most of these trials are testing ATP-competitive mTOR in-
hibitors, alone or in combination with monoclonal antibodies; this 2.3. Viability assay
strategy targets both GBM cells and GAM functions, thereby raising the
need for further research into this area. The cell viability was measured using a specific luminescence kit:
In the present work, we adopted two different approaches, i.e. CellTiter-Glo® Luminescent Cell Viability Assay (Promega). Cell mor-
human tumor specimens and human cell models, to characterize the tality was detected using a specific fluorescence kit: RealTime-Glo™ MT
activation of the mTOR pathway in GAMs. In particular, using im- Cell Viability Assay (Promega). The assays were carried out according
munostaining analysis of surgical GBM specimens obtained from 42 to the manufacturer's instructions.
patients, we evaluated phosphorylated-mTOR expression and localiza-
tion to microglial cells and using three human cell lines, we in- 2.4. Nitrite assay
vestigated the involvement of mTOR in an in vitro microglia-glioma
interaction paradigm. iNOS activity was assessed indirectly by measuring nitrite accu-
mulation in the incubation media. Briefly, an aliquot of the cell culture
2. Methods media (80 μL) was mixed with 40 μL Griess Reagent (Sigma-Aldrich)
and the absorbance measured at 550 nm in a spectrophotometric mi-
2.1. Materials croplate reader (PerkinElmer Inc., MA, USA). A standard curve was
generated during each assay in the range of concentrations 0–100 μM
Cell culture reagents [Dulbecco's modified Eagle's medium (DMEM), using NaNO2 (Sigma-Aldrich). In this range, the assay was linear and
DMEM-F12 and Fetal calf serum (FCS)] were from Invitrogen the minimum detectable concentration of NaNO2 was ≥3.12 μM. The
Corporation (Paisley, Scotland). Antibiotics were from Biochrom AG protein content in each sample was determined by Bradford's method

2
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Scheme 1. Metodological flow chart of the study.

(Biorad, Hercules, CA, USA) using bovine serum albumin as standard. machine (Stratagene). Primers used for the evaluation of gene expres-
sion are reported in Table 1. Relative mRNA concentrations were cal-
2.5. Urea assay culated from the take-off point of reactions (threshold cycle, Ct) using
the comparative quantitation method provided by Stratagene software
Urea levels in CHME-5 cells were detected by the QuantiChrom Urea and based upon the -ΔΔCt method. This analysis approximates a given
Assay kit (BIOassay System, Hayward, CA, USA), used according to the sample's target mRNA (e.g. IL6) level relative to the mean of the target
manufacturer's instructions. Briefly after 48 h of incubation with the mRNA levels in untreated controls (‘‘calibrator’’ value), thus permitting
test substances, an aliquot of cell culture media (50 μl) was mixed with statistical analysis of deviation from the mean even among the controls.
200 μL Urea Reagent (Bioassay system) and the absorbance measured at Ct values for α-tubulin expression served as a normalizing signal. In
430 nm in a spectrophotometric microplate reader (PerkinElmer Inc., each assay, the PCR efficiency was also calculated using serial dilution
MA, USA). A standard curve was generated during each assay in the of one experimental sample; efficiency values between 94 and 98%
range of concentrations 0–100 μg/ml using Urea as standard. In this were found for each primer set and taken into account for the com-
range, standard detection was linear and the minimum detectable parative quantitation analysis (Dello Russo et al., 2009).
concentration of Urea was 3.12 μg/ml. The protein content in each
sample was determined by Bradford's method (Biorad, Hercules, CA, 2.7. IL1β, IL6 and PGE2 quantification
USA) using bovine serum albumin as standard.
IL1β, IL6 and PGE2 levels in the incubation medium were detected
2.6. mRNA analysis in real time PCR using a specific Enzyme Immunoassay kit (EIA) for PGE2
(ELABscience), and a specific enzyme-linked immunosorbent assay
Total cytoplasmic RNA was extracted using the RNeasy Micro kit (ELISA) for IL1β and IL6 (R&D System). The assays were carried out
(Qiagen, Hilden, Germany), which included 15 min DNase treatment. according to the manufacturer's instructions.
RNA concentration was measured using the Qubit™ RNA HS Assay Kit
(Thermo Fisher Scientific). Aliquots (0.5 μg) of RNA were converted to 2.8. Western immunoblot
cDNA using random hexamer primers. Quantitative changes in mRNA
levels were estimated by real time PCR using the following cycling The cells were lysed in RIPA buffer (1 mM EDTA, 150 mM NaCl, 1%
conditions: 35 cycles of denaturation at 95 °C for 20 s; annealing and igepal, 0.1% sodium dodecyl sulfate, SDS, 0.5% sodium deoxycholate,
extension at 60 °C for 20 s; using the Brilliant III Ultra-Fast SYBR® Green 50 mM Tris–HCl, pH 8.0) (Sigma-Aldrich) containing protease inhibitor
QPCR Master Mix (Stratagene, La Jolla, CA, USA). PCR reactions were cocktail diluted 1:250 (Sigma–Aldrich). The protein content in each
carried out in a 20-μL reaction volume in Ariamx real time PCR sample was determined by Bradford's method (Biorad, Hercules, CA,

Table 1
Primers used for the evaluation of gene expression.
Genes Forward Reverse Product length

ARG1 TTCTCAAAGGGACAGCCACG AGCACCAGGCTGATTCTTCC 272

ARG2 ACCTCAGAGGAAGAGGCGAA AAATGTCCCCATTAAGGCAGT 276
COX2 TG CTG GCA GGG TTG CTGGTGGTA CAT CTG CCT GCT CTG GTC AAT CGA A 86
IL10 GGCACCCAGTCTGAGAACAG CGCCTTGATGTCTGGGTCTT 255
IL1β AGC CAT GGC AGA AGT ACC GT TCC ATG GCC ACA ACA ACT GA 219
IL6 GGC TCA TTC TGC CCT CGA GCC GGA CCG AAG GCG CTT GTG GAG 100
iNOS CTGCATGGAACAGTATAAGGCAAA C CAGACAGTTTCTGGTCGATGTCATGA 230
Raptor GGT GCT GTT AAG CCA AGT GC TGT TCA GCT GGC ATG TAG GG 279
Rictor TGG GCG AGG TTT CCG GT GTC ATT CCG CCC TCG TAC TC 99
TGFβ CATG GAG CTG GTG AAA CGG A CGG GTG ACT TCT TTG GCG TA 219
Tubuline CCC TCG CCA TGG TAA ATA CAT ACT GGA TGG TAC GCT TGG TCT 110

3
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

USA) using bovine serum albumin as standard. A 100 μg aliquot of 3–4 μm thick, were cut and collected on the superfrost plus slides
protein was mixed 1:2 with 2X Laemmli Buffer (Biorad), boiled for (Fisher). The PT Link (Dako) was used to deparaffinize and rehydrate
5 min, and separated through 10% polyacrylamide SDS gels. Apparent the sections and unmask antigen sites. Slides were immersed in 10 mM
molecular weights were estimated by comparison to colored molecular citrate buffer, pH 6.0, for 10 min at 97 °C and then cooled and washed
weight markers (Sigma Aldrich). After electrophoresis, proteins were in PBS or TBS. Endogenous peroxidase activity was inhibited by in-
transferred to polyvinylidene difluoride membranes by semi-dry elec- cubating the slides with Peroxide Block (ScyTek Laboratories, Utah,
trophoretic transfer (Biorad). The membranes were blocked with 10% USA) for 7 min. At this time, point slides washed with PBs underwent
(w/v) low-fat milk in TBST (10 mMTris, 150 mMNaCl, 0.1% Tween-20, single staining procedures, whereas slides washed with TBS underwent
pH 7.6) (Biorad) for 1 h at room temperature, and incubated in the double staining procedures.
presence of the primary antibody overnight with gentle shaking at 4 °C.
Primary antibodies for rabbit Polyclonal anti-phospho mTOR (NB600- 2.12. Single-staining
607-novus), rabbit polyclonal anti-mTOR (NB-novus), mouse mono-
clonal anti β-actin clone AC-74 (A531-Sigma), mouse monoclonal anti After washing in distilled water and then in PBS, slides were in-
p-p70 S6 kinase α (A-6) (sc-8416-Santa Cruz), rabbit Polyclonal anti- cubated with Avidin/Biotin Blocking System (Spring) and washed 3
4EBp1 (A300-501A-Bethyl), and rabbit monoclonal anti Tuberin/TSC times in PBS. Nonspecific binding was blocked by 5 min incubation
(D93F12) (4308 – Cell Signaling) were used at the final concentration with the Super Block Solution (ScyTek Laboratories, Utah, USA).
of 1:1000. Primary antibodies were removed, membranes washed 3 Sections were incubated for 30 min at room temperature with rabbit
times in TBST, and further incubated for 1 h at room temperature in the anti-human mTOR polyclonal antibody (Novus Biologicals) 1:100, or
presence of specific secondary antibody diluted 1:15,000. Following over-night at 4 °C with goat anti-human Iba1 polyclonal antibody
three washes in TBST, bands were visualized by incubation in ECL re- (Novus Biologicals) 1:250. Sections were washed extensively with PBS
agents (GE Healthcare) and exposure to Hyperfilm ECL (GE Healthcare and subsequently treated with the Ultra Tek HRP Anti-Polyvalent kit
NY, USA). The same membranes were washed 3 times in TBST, blocked (ScyTek Laboratories). Finally, after 3 washes in PBS, sections were
with 10% (w/v) low-fat milk in TBST for 1 h at room temperature and treated with 3,3′-diaminobenzidine (ScyTek Laboratories) as chro-
used for β-actin immunoblot. mogen, contrasted with Hematoxylin and mounted.

2.9. Immunostaining 2.13. Double-staining

Cover glasses of 13 mm of diameter were coated with Collagen I The PT Link (Dako) was used to deparaffinize and rehydrate the
0,1 mg/ml or Poli-L-lysine 50ug/ml for 1 h at 37 °C, after that three sections and unmask antigen sites. Slides were immersed in 10 mM ci-
washes in PBS with calcium and magnesium were needed and used to trate buffer, pH 6.0, for 10 min at 97 °C and then cooled and washed in
seed the CHME-5 at 20000 cells per well concentration. After 24 h from TBS. Endogenous peroxidase activity was inhibited by incubating the
the treatment, cells were blocked with PAF at 4% concentration in PBS slides with Peroxide Block (ScyTek Laboratories, Utah, USA) for 7 min.
with calcium and magnesium for 20 min at room temperature. After After washing in distilled water and then in TBS, nonspecific binding
three washes in PBS with calcium and magnesium, cells were blocked was blocked by 10 min incubation with Background Punisher (Biocare-
with BSA and incubated in presence of primary antibody. The incuba- Medical). Sections were incubated for 30 min at room temperature with
tion time was overnight for ARG1 1:50; after three washes in PBS with Rabbit Anti-Human mTOR polyclonal antibody (Novus Biologicals)
calcium and magnesium in gentle shaking, cells were incubated with 1:100 and over-night at 4 °C with Goat Anti-Human Iba1 polyclonal
secondary antibody (anti-mouse) for 1 h and mounted with Vectashield antibody (Novus Biologicals) 1:250. Sections were washed extensively
with DAPI (Vector Laboratories). in TBS and subsequently incubated with the MACH 2 Rabbit HRP-
Polymer (Biocare-Medical) for mTOR and with Ultratek HRP kit
2.10. Patients and specimens (ScyTek Laboratories) for Iba1. Finally, after 3 washes in TBS, sections
were treated with 3,3′-diaminobenzidine (Biocare-Medical) as chro-
We enrolled 42 adults [mean age 60.51 (34–79), 27 males/15 fe- mogen for IBA1 and with Vina Green (Biocare-Medical) as chromogen
males], who underwent surgery for primary GBM at the Neurosurgery for p-mTOR and then contrasted with Hematoxylin and mounted.
Department, Fondazione Policlinico Gemelli” (Rome, Italy), from
March 2005 to September 2011. Diagnosis of GBM was established on 2.14. Immunostaining analysis
histological examination according to the WHO classification (grade IV)
of tumors of the CNS. In all cases a total tumor removal was achieved, Two examiners who were blinded as to the antibody used evaluated
allowing us to obtain tissues samples from both the tumor and the staining of human specimens. Both qualitative and quantitative ana-
surrounding macroscopic normal brain tissue (between 1 cm and 2 cm lyses were carried out. For qualitative analysis the intensity of staining
from the tumor border; larger resections were performed in tumors that was evaluated. In particular, staining was scored on a scale from 0 to 5
grew far from eloquent areas). The demographic characteristics for where score 0 indicates no significant staining, score 1 very low
single patient are reported in Table 2. All patients provided written staining, score 2 low staining, score 3 significant staining, score 4 strong
consent to use their specimens for research and the research proposal staining and score 5 very strong staining. For quantitative analysis, the
was approved by the local ethics Ethical Committee (Lisi et al., 2017a). number of phospho-mTOR+, IBA1+, or both phospho-mTOR+ and
IBA1+ cells were counted in at least 50 cells total. In particular, two
2.11. Tissue preparation and immunohistochemistry blinded examiners have examined three different areas of the same
slides and have counted 50 cells that included the number of positive
Human tumor tissue obtained from surgical resection of patients cells for each antibody, the number of positive cells for both antibodies
with grade IV GBM were fixed in 4% paraformaldehyde in 0.1 M and the number of negative cells (Atzori et al., 2017). In total, the
phosphate buffer pH 7.6 overnight at 4 °C. Dehydration of tissue was average of six counts was reported as percentage.
through a series of 80%, 95% ethanol one hour each followed by 100%
ethanol overnight. Two 100% xylene washes were done for 1 h each 2.15. Statistical analyses
and then 1 h in 60 °C Paraplast Plus (Tyco/Healthcare, Mansfield, MA).
After a change of Paraplast Plus, tissue was incubated in a 60 °C vacuum Statistical comparison of the differences between pairs of groups
oven for 2 h prior to placing in molds to cool and solidify. Sections, was performed by Student's t-test. For multiple comparisons ANOVA

4
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Table 2
Demographic characteristic of GBM patients.
GBM patients Tumor location Primary (P) Overall survival (months) WHO classification Tumor Peryphery
Recurrent (R)

1 NAa P 13 IV -b −/+
2 Temporal P 19 IV +++ ++
3 Temporal P 60 IV +++ +++
4 Frontal R NA IV + +
5 Frontal P 7 IV −/+ −/+
6 NA P NA IV - +
7 Frontal P 15 IV +++ +
8 Frontal P 2 IV +++ −/+
9 Temporal P 4 IV +++ +
10 Occipital P 33 IV ++ +++
11 Temporal P 14 IV - -
12 NA P NA IV + +
13 NA R NA IV ++ ++
14 Temporal P 53 IV + -
15 Parietal P 53 IV - −/+
16 NA P 9 IV - −/+
17 Frontal P NA IV - −/+
18 Frontal P NA IV + +
19 Tempo-Parietal R 53 IV +++ +
20 Temporal P C IV - +
21 NA P NA IV + −/+
22 Temporal P NA IV −/+ +++
23 Parietal P 6 IV + -
24 NA P NA IV - +
25 Temporal P NA IV ++ +
26 NA R NA IV +++ +
27 Frontal R NA IV ++ +
28 Occipital P NA IV ++ ++
29 NA P NA IV ++ +
30 Temporal P NA IV + −/+
31 NA P NA IV +++ +
32 Temporal P NA IV + +
33 Temporal P 6 IV - +
34 NA P NA IV −/+ -
35 Frontal P 12 IV −/+ +++
36 Fronto-temporal P NA IV - +
37 Occipital P 8 IV +++ +
38 Frontal P NA IV ++ ++
39 Temporal P NA IV +++ +
40 NA P 38 IV ++ +++
41 NA R 24 IV +++ ++
42 Frontal P 12 IV ++ ++

a
NA: not available.
b
The number of positive mTOR cells was counted in a total of 50 cells. mTOR staining was scored as percentage of positively stained cells: -, ≤10%; −/+,
11–25%; +, 26–50%; ++, 51–75%; +++, > 75%.

analysis, followed by Bonferroni's post-test or Sidak's test, was used. 22% of cells in the glioma specimens were microglia-macrophages; 39%
Statistical significance was determined at α = 0.05 level. Differences out of these cells are IBA1+ cells expressing phosphorylated-mTOR.
were considered statistically significant when p < 0.05. Conversely, in periphery of the tumor, 15% of cells are microglia-
macrophages, 21% out of these cells express phospho-mTOR (Fig. 2).
3. Results
3.2. Characterization of U87MG and T98G glioma cell lines
3.1. mTOR in human glioma specimens
Two different human glioma cell lines, U87MG and T98G, were
We enrolled 42 patients diagnosed with GBM IV (Table 2). For each used to examine their responses to an inflammatory stimulus. Cells
patient, we were able to examine glioma specimens as well as the tissue were exposed to a mixture of pro-inflammatory cytokines: 10 ng/ml
surrounding the tumor by 1–2 cm (periphery). More than 35% of the TNFα, 10 ng/ml IL1β, 10 UI/ml IFNγ (TII), after which gene expression
cells within the glioma tissue and more than 25% of the peripheral cells was evaluated at different time points (4, 8 and 24 h). IL1β, IL6 and
showed mTOR activation, indicated by phosphorylation at Ser2448. COX2 was significantly increased by TII, with maximal levels at 8 h
Although the percent of phospho-mTOR-activated positive cells was after which expression decreased (Fig. 3). In contrast, TGFβ and ARG2
higher within the tumor, we found no difference in staining intensity expression was not affected by pro-inflammatory stimuli. Neither iNOS,
between tumor and peripheral tissues (Fig. 1). In addition, in a sub- IL10 nor ARG1 mRNAs were detectable either before or after pro-in-
group of 15 patients showing a higher percentage of phospho- flammatory stimulation (data not shown).
mTOR + cells (39% on average), we carried out a double staining for Based on these results as well as previous data (Lisi et al., 2014a),
phospho-mTOR and IBA1 (Fig. 2), a marker of microglia-macrophage we prepared glioma conditioned medium (CM) by pre-incubating
cell type. Within the tumor, about 9% of all cells were phospho- glioma cells for 4 h with TII or plain medium as control. After pre-in-
mTOR + microglia-macrophages compared to 4.4% within the tissue cubation, cells were carefully washed and incubated for a further 24 h
surrounding the tumor. Looking at IBA1+ cells, we found that about in plain medium only to generate basal-conditioned medium (B-CM) or

5
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

pre-stimulated conditioned medium (PS-CM).

To characterize the CMs, we measured levels of end-products of the
above genes, namely IL1β, IL-6, PGE2, urea and NO, in the incubation
medium. We found that NO (expressed as nitrite levels) was un-
detectable in either B-CMs or PS-CMs obtained from either U87MG or
T98G cells (not shown). In contrast, significant urea levels could be
detected at baseline both in T98G CMs (1 μg/ml on average) and in
U87MG CMs (3 μg/ml on average), but those levels were not different
when the U87MG or T98G were pre-stimulated with the mixture of
cytokines (data not shown). We observed significant 1.5-fold, 4-fold
and 5-fold increases in PGE2, IL6 and IL1β levels, respectively, in
U87MG PS-CM compared to B-CMs (Fig. 4). In T98G PS-CM we also
found a 3-fold and 6-fold increase in IL1 β and IL6 levels, respectively,
as compared to B-CMs, whereas no change was observed in PGE2 levels
(Fig. 4).

3.3. Effects of B-CM and PS-CM obtained from glioma cell lines on human
microglia CHME-5 cells

Fig. 1. Evaluation of mTOR pathway in GBM specimens. Panel A shows a re- We next conducted a series of experiments in which the human
presentative image of phospho-mTOR staining in GBM sample (magnitude 40x), microglia CHME-5 cells were exposed to CMs. Cells were incubated
Panel B shows a representative image of phospho-mTOR staining in periphery with graded concentrations of CMs (i.e. 25% CM + 75% plain medium,
around the tumor (magnitude 40x). Red arrow shows positive cells; yellow 50% CM + 50% plain medium, 100% CM) over a period of 0–72 h
arrow shows negative cells. Panel C shows the percentage of cells expressing (Fig. 5). CHME-5 cell growth in plain medium reached a peak within
phospho-mTOR, whereas panel D shows the average of intensity score of
16–24 h, and remained stable up to up to 48 h; thereafter, an increase in
phospho-mTOR positive cells.
the number of dead cells was observed (Fig. 5A). This growth profile
remained unchanged when the percent of CMs in the media was in-
creased, with 2 exceptions: at all times tested, 100% B-CM from
U87MG, but not T98G cells increased the total number of live cells,
whereas an increase in lethality was observed after exposure to 100%
PS-CM from T98G cells (Fig. 5B–C). We found no dose-dependent effect
of CM on cell growth; however, a dose-dependent effect was observed
on lethality induced by PS-CM from T98G cells (not shown). Because of
the increase in lethality observed after 48 h of incubation, all sub-
sequent experiments were carried out between 0 and 48 h.
To investigate the phenotypic profile induced in CHME-5 cells by
the different CMs, we measured levels of urea and NO released in the
medium, as markers of L-arginine metabolism through the ARG/NOS
pathways, as well as expression of TNFα, iNOS, TGFβ and IL10 mRNAs.
Fig. 2. Evaluation of microglia-macrophages expressing phospho-mTOR. Panel We found that both B-CM and PS-CM elicited dose-dependent increases
A shows a representative image of phospho-mTOR and IBA1 double staining in in urea levels, which reached statistical significance compared to plain
glioma specimens. Brown arrow shows IBA1 positive cells, green arrow shows medium from 50% of CMs from T98G cells, and at 100% of CMs from
phospho-mTOR positive cells, and black arrow shows phospho-mTOR and IBA1 U87MG cells (Fig. 6). On the contrary, we found no difference in NO
positive cells. Panel B shows the percent of microglia-macrophage phospho- levels compared to controls, nor any change in iNOS mRNA levels, or in
mTOR + expressing in GBM specimens (black box) and in periphery (grey box).
any of the other genes tested (not shown).

Fig. 3. Evaluation of pro-inflammatory gene ex-

pression elicited by cytokines in glioma cell lines.
Panel A, B and C show time-dependent up-regulation
of IL-1β, IL-6 and COX2 gene expression respectively
in U87MG (blue line) and in T98G (red line).
*,**,***,****p < 0,05, p < 0,01, p < 0,001 or
p < 0,0001 respectively T98G vs U87MG in each
time point. Two-way ANOVA analysis was carried
out.

6
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Fig. 4. Evaluation of pro-inflammatory products in

CM obtained after pre-stimulation with cytokines
challenge. Panel A, B and C show the level of IL-1β,
IL-6 and PGE2. *, ***p < 0,05 or p < 0,001 re-
spectively vs U87 MG control or °, °°°p < 0,05 or
p < 0,001 respectively vs T98 G control. Student's t-
test was carried out (basal vs pre-stimulated).

Moreover, we sought to investigate the effects of CMs on CHME- both resting microglia (Fig. 6) and in pre-stimulated M1 microglia
5 cells that were previously stimulated with TII for 4 h; indeed, such (Figs. 7 and 8), although in the latter with a transient response.
pre-stimulation drives CHME-5 cells into an activated M1 profile (Lisi
et al., 2017b). Under these conditions, we found significant change in
3.4. Involvement of mTOR pathway in glioma-microglia interaction
urea levels elicited by T98G CMs (Fig. 7), however by comparing the
urea releases promoted by all CMs into not pre-stimulated cells (or pre-
We next tested if the mTOR pathway is activated in human micro-
treated with medium) versus cells pre-treated with TII, all CMs sig-
glia after exposure to human glioma-derived CMs. We found that both
nificantly increase urea releases (Fig. 7). In contrast to non-prestimu-
T98G B-CM and PS-CMs, but not U87MG CMs, significantly increased
lated cells, all CMs significantly increased iNOS mRNA levels in CHME-
phosphorylation of mTOR at Ser2448 with respect to controls after 2-h
5 cells that were prestimulated with TII for 4 h (Fig. 8). The iNOS mRNA
incubation periods (Fig. 9A). In parallel, mTOR downstream factor
levels were increased after 4 h stimulation with CMs, but after longer
4EBP1 was significantly modified by B-CMs (Fig. 9B). Similarly, an-
exposure times (i.e. 24 or 48 h), increases were no longer observed.
other mTOR downstream factor, p-P70S6k was significantly increased
After 48 h incubation, the NO levels were increased by pre-treatment
in the presence of all CMs (Fig. 9C). We also looked at the mechanisms
with TII (Fig. 8C), but were consistently reduced by the presence of any
of mTOR activation; interestingly, the upstream factor TSC2 was sig-
CMs. Taken together the data suggest that CMs induce a M2 profile in
nificantly decreased in the presence of CMs (Fig. 10).

Fig. 5. Evaluation of cell viability and cell mortality

of CHME-5 Cells. Panel A shows the growth of
CHME-5 cell line (20.000 cells/well) under basal
conditions in plain growth media. The fluorescence
data indicates the not viability cells whereas the lu-
minescence data indicates the viability cells. Panel B
and C show the cell viability and cell mortality re-
spectively in the presence of the indicated CMs. For
each time point the data are expressed relative to the
control, which is taken as 100%. ****p < 0,0001
CM vs basal condition. Two-way ANOVA analysis
was carried out.

7
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

origin or pre-stimulation state-effected that gene expression (data not

shown), nor did incubation with the mTOR inhibitor RAPA modify ei-
ther Raptor and Rictor gene expression in the presence of any of the
CMs (data not shown).
Despite absence of effects on mTOR components, RAPA
(10–100 nM) decreased urea production (Fig. 11) and ARG1 expression
(Fig. 12) from CHME-5 cells exposed to B-CMs or PS-CMs. In addition,
incubation with RAPA had no effects on either IL6, ARG1, or IL10 gene
expression. RAPA increased gene expression of iNOS 4-fold versus iNOS
levels due to the presence of CMs only (data not shown); however that
increase was not associated with any increase in NO levels at 48 h.
Fig. 6. Effects of CMs on urea release. Left panel shows the urea content in These data are consistent with our previous observations in the rat
CHME-5 cells, under basal condition (white), after U87MG B-CM challenge model (Lisi et al., 2014a), and reinforce the idea that the blockade of
(yellow) and after U87MG PS-CM stimulation (blue). Right panel shows the
mTOR is associated with a modification of M2/M1 ratio.
urea content in CHME-5 cells, under basal condition (white), after T98G B-CM
challenge (yellow) and after T98G PS-CM stimulation (blue). The data are ob-
tained after 48 h of incubation. *, ***p < 0,05 or p < 0,001 respectively vs
4. Discussion
control. One-way ANOVA analysis was carried out.
In the present study, we used two different approaches (i.e. human
tumor specimens and human cell models) to study the activation of the
mTOR pathway in glioma-associated microglia. In particular, by im-
munostaining analysis of surgical GBM specimens from 42 patients, we
showed phosphorylated-mTOR expression and localization in micro-
glia-macrophage cells. In addition, these findings were confirmed in
functional experiments. Using U87MG or T98G cells as a model for
human glioma and CHME-5 cells to model human microglia, we in-
vestigated the involvement of mTOR in in vitro paradigms of microglia-
glioma interaction under two different conditions (Figure Suppl 1).
In the last decade, the role of microglia in GBM pathology is be-
coming clearer. Under the influence of GBM, microglia tend to assume a
Fig. 7. Effects of CMs on M1 CHME-5. CHME-5 were pre-stimulated with TII for M2 phenotype, which has predominantly a pro-tumor role. However, in
4 h. After 4 h cytokines were removed, CMs were administrated for the sub-
addition to this role, evidence that is more recent indicates that some
sequent 48 h and urea content were measured. As indicated in figure boxes 2-3-
cells with microglial characteristics seen in GBMs are actually part of
4-5-6 were pre-treated with TII, boxes 7-8-9-10 were pre-treated with medium.
*p < 0,05 vs TII CMs. One-way ANOVA analysis followed by Bonferroni's post-
the tumor cell population. According to these findings, the expression of
test was carried out. CD163, an M2 microglia marker, is also a marker for fusion hybrids
between immune cells (macrophages or microglia) and neoplastic
tumor cells (Huysentruyt et al., 2011; Lindström et al., 2017; Gast et al.,
2018). In this context, we recently found that in surgical GBM speci-
mens CD163 expression is higher within GBM specimens than in the
surrounding periphery in both male and female patients. (Lisi et al.,
2017a). Taken together these findings lead to the hypothesis that under
specific conditions microglial cells may lose characteristics of the im-
mune system and acquire tumor-cell characteristics. However, addi-
tional studies are required to confirm or negate these possibilities.
In our studies, we used production of urea to serve as an index of
arginine metabolism (an M2 type phenotype). The main source of urea
in eukaryotes is from conversion of arginine to ornithine by the enzyme
arginase. However, urea can also be generated from glutamine via ac-
tion of glutaminase which produces ammonia which can then be con-
densed with CO2 to generate urea (Bach and Smith, 1956); a process
which mainly occurs in liver. However, most glial derived tumor cells
depend upon glutamine for growth (Szeliga et al., 2009; Szeliga and
Fig. 8. Evaluation of iNOS pathway after CMs challenge. CHME-5 were pre-
Albrecht, 2016), associated with expression of a phosphate activated
stimulated with TII for 4 h. After 4 h cytokines were removed, CMs were ad-
glutaminase (Majewska et al., 2017) to generate glutamate as an energy
ministrated for the subsequent 4 or 24 h and iNOS gene expression were
source. It is therefore possibly that treatments of T98G (and U87MG)
measured. Panel A show the effects of U87MG CMs on iNOS gene expression in
CHME-5. Panel B show the effects of T98G CMs on iNOS gene expression in cells that modify urea production may be due in part to changes in
CHME-5. Panel C shows nitrite content in 48-h experiments in CHME-5. White phosphate activated glutaminase. Moreover, the suppression of urea
bar indicates the control group whereas the red bar the TII group. release due to mTOR inhibitors could be due to reductions in gluta-
***p < 0,001 vs control. One-way ANOVA analysis followed by Bonferroni's minase expression, which could contribute to the ability of these in-
post-test was carried out. hibitors to reduce glioma growth. This would be consistent with find-
ings that up-regulation of glutaminase in certain glioblastoma cell lines
3.5. Effects of mTOR inhibitors on the glioma-microglia paradigm renders them resistant to mTOR inhibition (Tanaka et al., 2015).
Using the human microglia-glioma interaction paradigm, we
Under basal conditions, CHME-5 cells express high levels of Raptor showed that the mTOR pathway is fully activated in microglia cells
A (major component of mTORC1) and Rictor (a major component of under conditions mimicking the human GBM pathology. Such mTOR
mTORC2) gene expression. None of the CMs used -regardless of cell activation, observed as increases in phosphorylated mTOR fraction, is
in full agreement with data from human tumor specimens, showing that

8
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Fig. 9. Evaluation of mTOR pathway in in vitro mi-

croglia-glioma interaction. Western blot analysis of
CHME-5 lysates prepared 2 h after the following
treatments: Lane 1, control, 2, U87MG B-CM; 3,
U87MG PS-CM; 4, T98G B-CM; 5, T98G PS-CM; 6,
TII. Panel A shows western blot analysis for phospho-
mTOR Serine 2448 (top image) and total mTOR
(bottom image), panel B shows 4EBP1 quantification
and panel C shows p-P70S6K quantification. *, **,
***p < 0,05 p < 0,01 or p < 0,001 respectively
vs control. One-way ANOVA analysis followed by
Bonferroni's post-test was carried out.

important to emphasize that microglia may represent 30–70% of the

entire tumor (Wood GW and Morantz, 1979; Roggendorf et al., 1996; da
Fonseca and Badie, 2013).
In the present study, we reported data using two different glioma
cell lines. Preliminary data (Figs. 3 and 4) showed that conditioned
media obtained by U87MG or T98G are similar but not identical in term
of IL6, IL1β and PGE2 release. In fact, CMs obtained from T98G contain
more IL6 and IL1β than that from U87MG but contain less PGE2. These
differences could explain why in some conditions (i.e phospho-mTOR
activation) CMs give different results.
Here we showed that mTOR activation could be elicited through
different pathways, depending on different experimental conditions. In
fact, when CHME-5 cells were treated with B-CM [a condition mi-
micking an early stage of pathology (Lisi et al., 2014b)], the activation
of mTOR pathway was associated with a significant blockade of TSC2.
On the contrary, when CHME-5 cells were exposed to PS-CM [a con-
Fig. 10. Evaluation of TSC2 in in vitro microglia-glioma interaction. Western dition that mimics a late stage of pathology (Lisi et al., 2014b)], mTOR
blot analysis of CHME-5 lysates for TSC2 (top image) and β-actin (bottom activation was in part independent from TSC2 inactivation. While TSC2
image) prepared 2 h after the following treatments: Lane 1, control; 2, U87MG inhibition is considered as the ‘classical’ pathway of TOR activation,
B-CM; 3, U87MG PS-CM; 4, T98G B-CM; 5, T98G PS-CM. **, ***,****p < 0,01,
TSC2-independent mechanisms have also been described. Upstream
p < 0,001 or p < 0,0001 respectively vs control. One-way ANOVA analysis
oncogenic mutations activate mTORC1 kinase via at least two alter-
followed by Bonferroni's post-test was carried out.
native mechanisms: down-regulation/inactivation of TSC2 and phos-
phorylation of the mTORC1 inhibitor PRAS40 by AKT (Huang and
39% of microglia-macrophage in glioma specimens express mTOR Manning, 2009; Lv et al., 2017). The TSC2 protein integrates signals
phosphorylated at Ser-2448, taken as marker of mTORC1 activation from three pathways (AKT, ERK/RSK, and LKB1/AMPK); each of these
(Rosner et al., 2010). In addition, and consistent with previous data pathways can lead to phosphorylation of several serines and threonines
from our group and others, here we show that microglia express an M2 on TSC2. The latter in turn acts as a GTPase-activating protein for the
pro-tumor phenotype in the presence of glioma cells, and that this M2 small Ras-like protein, Rheb. Whereas Rheb bound to GDP is thought to
phenotype is down-regulated in the presence of an mTOR inhibitor. be inactive, Rheb-GTP is an essential activator for mTORC1 kinase
Based on the above data, we hypothesize that the blockade of mTOR in (Dibble and Cantley, 2015). Phosphorylation of PRAS40 by AKT in-
GBM patients might induce a reduction in M2 phenotype expression in hibits the binding of PRAS40 to mTORC1 and the subsequent mTORC1
about 40% of all tumor-associated microglia. In this regard, it is kinase activation (Lv et al., 2017).

9
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Fig. 11. Effects of rapamycin on urea levels in CHME-5 after stimulation with CMs (48 h experiments). **p < 0,01 vs control; ° p < 0,05 vs CM. One-way ANOVA
analysis followed by Bonferroni's post-test was carried out.

Fig. 12. Effects of rapamycin on ARG1 expression in CHME-5 after stimulation with CMs (24 h experiments). Panel A, control; B, B-CM; C, B-CM + Rapa 100 nM; D,
PS-U87; E, PS-U87 + Rapa 10 nM; F, B-T98; G, B-T98 + Rapa 10 nM; H, PS-T98; I, Ps-T98 + Rapa 10 nM.

Because of their availability as immunosuppressive agents, mTOR had limited success in clinical trials for various other tumor types (in
inhibitors have been extensively investigated in clinical trials in on- glioma but also in major solid tumors). The reasons for such limited
cology, based on their ability to control cell proliferation. So far, the success are yet to be clarified, but may be related to the inhibition of a
efficacy of everolimus (RAD0001) has been demonstrated as a second large number of mTORC1-regulated signaling systems normally in-
line treatment in renal, breast and neuroendocrine cancers of lung, volved in tumor suppression, such as the activation of receptor tyrosine
pancreatic or gastrointestinal origin. Overall, mTOR inhibitors have kinases (RTKs), PI3K-Akt signaling, and Ras-ERK pathway (Nussinov

10
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

et al., 2018). In order to overcome these limitations, alternative stra- 1186/s13046-017-0577-2.

tegies have been explored in the past few years and a number of ATP- Bach, S.J., Smith, M., 1956. Glutamine; a nitrogen source in urea synthesis. Biochem. J.
64 (3), 417–425.
competitive mTOR inhibitors have been developed, which block both Chen, Z., Hambardzumyan, D., 2018. Immune microenvironment in glioblastoma sub-
mTORC1 and mTORC2 activity. Unlike RAPA, which is a specific al- types. Front. Immunol. 8 (9), 1004. https://doi.org/10.3389/fimmu.2018.01004.
losteric inhibitor of mTORC1, these ATP-competitive inhibitors target da Fonseca, A.C., Badie, B., 2013. Microglia and macrophages in malignant gliomas: re-
cent discoveries and implications for promising therapies. Clin. Dev. Immunol. 2013,
the catalytic site of the enzyme, thus promoting a broader, more potent 264124. https://doi.org/10.1155/2013/264124.
and sustained inhibition of mTOR, thereby preventing the activation of Dello Russo, C., Lisi, L., Tentori, L., Navarra, P., Graziani, G., Combs, C.K., 2017.
PI3K/Akt caused by the de-repression of negative feedbacks (Martelli Exploiting microglial functions for the treatment of glioblastoma. Curr. Cancer Drug
Targets 17 (3), 267–281. https://doi.org/10.2174/1568009616666160813191240.
et al., 2018). In addition, it is important to consider that glucose and Dello Russo, C., Lisi, L., Tringali, G., Navarra, P., 2009. Involvement of mTOR kinase in
glutamine are drivers of mTOR activation (Tanaka et al., 2015) and in cytokine-dependent microglial activation and cell proliferation. Biochem. Pharmacol.
particular, glucose drives GBM through the PI3K-AKT pathway (Shukla 78 (9), 1242–1251. https://doi.org/10.1016/j.bcp.2009.06.097.
Dibble, C.C., Cantley, L.C., 2015. Regulation of mTORC1 by PI3K signaling. Trends Cell
et al., 2018). Therefore another possible way to escape the limited
Biol. 25 (9), 545–555. https://doi.org/10.1016/j.tcb.2015.06.002.
success of mTOR inhibitor in clinical trials could be use therapies that Foster, K.G., Fingar, D.C., 2010. Mammalian target of rapamycin (mTOR): conducting the
target glucose and glutamine to affect the mTOR pathway. However, cellular signaling symphony. J. Biol. Chem. 285 (19), 14071–14077. https://doi.org/
while the inhibition of mTOR pathway remains a meaningful target for 10.1074/jbc.R109.094003.
Gast, C.E., Silk, A.D., Zarour, L., Riegler, L., Burkhart, J.G., Gustafson, K.T., Parappilly,
the treatment of GBM, novel and more specific molecules are needed. M.S., Roh-Johnson, M., Goodman, J.R., Olson, B., Schmidt, M., Swain, J.R., Davies,
In conclusion, a main finding of this study the observation that the P.S., Shasthri, V., Iizuka, S., Flynn, P., Watson, S., Korkola, J., Courtneidge, S.A.,
mTOR pathway is activated in 39% of microglia-macrophage within Fischer, J.M., Jaboin, J., Billingsley, K.G., Lopez, C.D., Burchard, J., Gray, J.,
Coussens, L.M., Sheppard, B.C., Wong, M.H., 2018. Cell fusion potentiates tumor
GBM tumors, compared to 21% in peripheral tissues. We also showed heterogeneity and reveals circulating hybrid cells that correlate with stage and sur-
that pharmacological inhibition of mTOR reduce urea levels and ARG1 vival. Sci. Adv. 4 (9) eaat7828. https://doi.org/10.1126/sciadv.aat7828.
expression -taken as a M2 marker-in in vitro models of both early- and Huang, J., Manning, B.D., 2009. A complex interplay between Akt, TSC2 and the two
mTOR complexes. Biochem. Soc. Trans. 37 (Pt 1), 217–222. https://doi.org/10.
late stage glioma. Since the M2 profile of microglial activation is widely 1042/BST0370217.
thought to be associated to tumor progression (Annovazzi et al., 2018), Huysentruyt, L.C., Akgoc, Z., Seyfried, T.N., 2011. Hypothesis: are neoplastic macro-
the increase in the ratio of M1 microglia bearing cytotoxic and anti- phages/microglia present in glioblastoma multiforme? ASN Neuro. 22 (4), 3. https://
doi.org/10.1042/AN20110011.
tumor activity brought about by mTOR inhibition might be envisioned Janabi, N., Peudenier, S., Héron, B., Ng, K.H., Tardieu, M., 1995. Establishment of human
as an additional antitumor mechanism of mTOR inhibitors, along with microglial cell lines after transfection of primary cultures of embryonic microglial
direct anti-proliferative activity. cells with the SV40 large T antigen. Neurosci. Lett. 195 (2), 105–108. https://doi.
org/10.1016/0304-3940(94)11792-H.
Liang, F., Giordano, C., Shang, D., Li, Q., Petrof, B.J., 2018. The dual CCR2/CCR5 che-
Author contribution mokine receptor antagonist Cenicriviroc reduces macrophage infiltration and disease
severity in Duchenne muscular dystrophy (Dmdmdx-4Cv) mice. PLoS One 13 (3),
Study concepts: LL; PN. e0194421. https://doi.org/10.1371/journal.pone.0194421.
Lindström, A., Midtbö, K., Arnesson, L.G., Garvin, S., Shabo, I., 2017. Fusion between M2-
Study design: LL; AM. macrophages and cancer cells results in a subpopulation of radioresistant cells with
Data acquisition: GMPC; MC; MP; SK. enhanced DNA-repair capacity. Oncotarget 8 (31), 51370–51386. https://doi.org/10.
Data analysis and interpretation: LL; SK; DLF. 18632/oncotarget.17986.
Lisi, L., Ciotti, G.M., Braun, D., Kalinin, S., Currò, D., Dello Russo, C., Coli, A., Mangiola,
Statistical analysis: LL; DLF; PN. A., Anile, C., Feinstein, D.L., Navarra, P., 2017. Expression of iNOS, CD163 and ARG-
Manuscript preparation: LL. 1 taken as M1 and M2 markers of microglial polarization in human glioblastoma and
Manuscript editing: PN. the surrounding normal parenchyma. Neurosci. Lett. 645, 106–112. https://doi.org/
10.1016/j.neulet.2017.02.076.
Manuscript review: AM, DLF and PN. Lisi, L., Laudati, E., Navarra, P., Dello Russo, C., 2014b. The mTOR kinase inhibitors
polarize glioma-activated microglia to express a M1 phenotype. J.
Funding Neuroinflammation 11, 125. https://doi.org/10.1186/1742-2094-11-125.
Lisi, L., Navarra, P., Feinstein, D.L., Dello Russo, C., 2011. The mTOR kinase inhibitor
rapamycin decreases iNOS mRNA stability in astrocytes. J. Neuroinflammation 8 (1),
Fondi Ateneo 2017 (PN) financed the research. 1. https://doi.org/10.1186/1742-2094-8-1.
Lisi, L., Pizzoferrato, M., Miscioscia, F.T., Topai, A., Navarra, P., 2017b. Interactions
between integrase inhibitors and human arginase 1. J. Neurochem. 142 (1), 153–159.
Conflict of interest
https://doi.org/10.1111/jnc.14039.
Lisi, L., Stigliano, E., Lauriola, L., Navarra, P., Dello Russo, C., 2014. Proinflammatory-
Any conflicts of interest to disclose. activated glioma cells induce a switch in microglial polarization and activation status,
from a predominant M2b phenotype to a mixture of M1 and M2a/B polarized cells.
ASN Neuro. 6 (3), 171–183. https://doi.org/10.1042/AN20130045.
Appendix A. Supplementary data Lu, V.M., Jue, T.R., McDonald, K.L., Rovin, R.A., 2018. The survival effect of repeat
surgery at glioblastoma recurrence and its trend: a systematic review and meta-
Supplementary data to this article can be found online at https:// analysis. World Neurosurg. 115, 453–459. e3. https://doi.org/10.1016/j.wneu.
2018.04.016.
doi.org/10.1016/j.neuint.2019.104485. Lv, D., Guo, L., Zhang, T., Huang, L., 2017. PRAS40 signaling in tumor. Oncotarget 8 (40),
69076–69085. https://doi.org/10.18632/oncotarget.17299.
References Majewska, E., Rola, R., Barczewska, M., Marquez, J., Albrecht, J., Szeliga, M., 2017.
Transcription factor GATA3 expression is induced by GLS2 overexpression in a
glioblastoma cell line but is GLS2-independent in patient-derived glioblastoma. J.
Amin, S.A., Adhikari, N., Baidya, S.K., Gayen, S., Jha, T., 2018. Structural refinement and Physiol. Pharmacol. 68 (2), 209–214.
prediction of potential CCR2 antagonists through validated multi-QSAR modeling Martelli, A.M., Buontempo, F., McCubrey, J.A., 2018. Drug discovery targeting the mTOR
studies. J. Biomol. Struct. Dyn. 3, 1–20. https://doi.org/10.1080/07391102.2017. pathway. Clin. Sci. 132 (5), 543–568. https://doi.org/10.1042/CS20171158.
1418679. Miyauchi, J.T., Tsirka, S.E., 2018. Advances in immunotherapeutic research for glioma
Annovazzi, L., Mellai, M., Bovio, E., Mazzetti, S., Pollo, B., Schiffer, D., 2018. Microglia therapy. J. Neurol. 265 (4), 741–756. https://doi.org/10.1007/s00415-017-8695-5.
immunophenotyping in gliomas. Oncol. Lett. 15 (1), 998–1006. https://doi.org/10. Nussinov, R., Tsai, C.J., Jang, H., 2018. Oncogenic Ras isoforms signaling specificity at
3892/ol.2017.7386. the membrane. Cancer Res. 78 (3), 593–602. https://doi.org/10.1158/0008-5472.
Arbab, A.S., Rashid, M.H., Angara, K., Borin, T.F., Lin, P.C., Jain, M., Achyut, B.R., 2017. CAN-17-2727.
Major challenges and potential microenvironment-targeted therapies in glioblastoma. Roggendorf, W., Strupp, S., Paulus, W., 1996. Distribution and characterization of mi-
Int. J. Mol. Sci. 18 (12) pii, E2732. https://doi.org/10.3390/ijms18122732. croglia/macrophages in human brain tumors. Acta Neuropathol. 92 (3), 288–293.
Atzori, M.G., Tentori, L., Ruffini, F., Ceci, C., Lisi, L., Bonanno, E., Scimeca, M., Eskilsson, https://doi.org/10.1007/s004010050520.
E., Daubon, T., Miletic, H., Ricci Vitiani, L., Pallini, R., Navarra, P., Bjerkvig, R., Rosner, M., Siegel, N., Valli, A., Fuchs, C., Hengstschlager, M., 2010. mTOR phosphory-
D'Atri, S., Lacal, P.M., Graziani, G., 2017. The anti-vascular endothelial growth factor lated at S2448 binds to raptor and rictor. Amino Acids 38 (1), 223–228. https://doi.
receptor-1 monoclonal antibody D16F7 inhibits invasiveness of human glioblastoma org/10.1007/s00726-008-0230-7.
and glioblastoma stem cells. J. Exp. Clin. Cancer Res. 36 (1), 106. https://doi.org/10. Shukla, A., Gupta, P., Singh, R., Mishra, D.P., 2018. Glycolytic inhibitor 2-Deoxy-d-

11
L. Lisi, et al. Neurochemistry International 129 (2019) 104485

Glucose activates migration and invasion in glioblastoma cells through modulation of 1591–1602. https://doi.org/10.1172/JCI78239.
the miR-7-5p/TFF3 signaling pathway. Biochem. Biophys. Res. Commun. 499 (4), Tee, A.R., 2018. The target of rapamycin and mechanisms of cell growth. Int. J. Mol. Sci.
829–835. https://doi.org/10.1016/j.bbrc.2018.04.001. 19 (3) pii: E880. https://doi.org/10.3390/ijms19030880.
Szeliga, M., Albrecht, J., 2016. Glutamine metabolism in gliomas. Adv. Neurobiol. 13, Tsutsumi-Kuroda, U., Inoue, T., Futakuchi, A., Shobayashi, K., Takahashi, E., Kojima, S.,
259–273. https://doi.org/10.1007/978-3-319-45096-4_9. Inoue-Mochita, M., Fujimoto, T., Tanihara, H., 2018. Decreased MCP-1/CCR2 axis-
Szeliga, M., Obara-Michlewska, M., Matyja, E., Łazarczyk, M., Lobo, C., Hilgier, W., mediated chemotactic effect of conjunctival fibroblasts after transdifferentiation into
Alonso, F.J., Márquez, J., Albrecht, J., 2009. Transfection with liver-type glutaminase myofibroblasts. Exp. Eye Res. 170, 76–80. https://doi.org/10.1016/j.exer.2018.02.
cDNA alters gene expression and reduces survival, migration and proliferation of 008.
T98G glioma cells. Glia 57 (9), 1014–1023. https://doi.org/10.1002/glia.20825. Weller, M., 2018. Next generation neuro-oncology. Eur. J. Cancer 96, 1–5. https://doi.
Tanaka, K., Sasayama, T., Irino, Y., Takata, K., Nagashima, H., Satoh, N., Kyotani, K., org/10.1016/j.ejca.2018.03.016.
Mizowaki, T., Imahori, T., Ejima, Y., Masui, K., Gini, B., Yang, H., Hosoda, K., Sasaki, Wood, G.W., Morantz, R.A., 1979. Immunohistologic evaluation of the lymphoreticular
R., Mischel, P.S., Kohmura, E., 2015. Compensatory glutamine metabolism promotes infiltrate of human central nervous system tumors. J. Natl. Cancer Inst. 62 (3),
glioblastoma resistance to mTOR inhibitor treatment. J. Clin. Investig. 125 (4), 485–491.

12