Cross talk between mesenchymal and glioblastoma stem cells: Communication beyond controversies

Abstract Mesenchymal stem cells (MSCs) can be isolated from bone marrow or other adult tissues (adipose tissue, dental pulp, amniotic fluid, and umbilical cord). In vitro, MSCs grow as adherent cells, display fibroblast‐like morphology, and self‐renew, undergoing specific mesodermal differentiation. High heterogeneity of MSCs from different origin, and differences in preparation techniques, make difficult to uniform their functional properties for therapeutic purposes. Immunomodulatory, migratory, and differentiation ability, fueled clinical MSC application in regenerative medicine, whereas beneficial effects are currently mainly ascribed to their secretome and extracellular vesicles. MSC translational potential in cancer therapy exploits putative anti‐tumor activity and inherent tropism toward tumor sites to deliver cytotoxic drugs. However, controversial results emerged evaluating either the therapeutic potential or homing efficiency of MSCs, as both antitumor and protumor effects were reported. Glioblastoma (GBM) is the most malignant brain tumor and its development and aggressive nature is sustained by cancer stem cells (CSCs) and the identification of effective therapeutic is required. MSC dualistic action, tumor‐promoting or tumor‐targeting, is dependent on secreted factors and extracellular vesicles driving a complex cross talk between MSCs and GBM CSCs. Tumor‐tropic ability of MSCs, besides providing an alternative therapeutic approach, could represent a tool to understand the biology of GBM CSCs and related paracrine mechanisms, underpinning MSC‐GBM interactions. In this review, recent findings on the complex nature of MSCs will be highlighted, focusing on their elusive impact on GBM progression and aggressiveness by direct cell‐cell interaction and via secretome, also facing the perspectives and challenges in treatment strategies.

tional properties for therapeutic purposes. Immunomodulatory, migratory, and differentiation ability, fueled clinical MSC application in regenerative medicine, whereas beneficial effects are currently mainly ascribed to their secretome and extracellular vesicles. MSC translational potential in cancer therapy exploits putative anti-tumor activity and inherent tropism toward tumor sites to deliver cytotoxic drugs. However, controversial results emerged evaluating either the therapeutic potential or homing efficiency of MSCs, as both antitumor and protumor effects were reported. Glioblastoma (GBM) is the most malignant brain tumor and its development and aggressive nature is sustained by cancer stem cells (CSCs) and the identification of effective therapeutic is required. MSC dualistic action, tumor-promoting or tumor-targeting, is dependent on secreted factors and extracellular vesicles driving a complex cross talk between MSCs and GBM CSCs. Tumor-tropic ability of MSCs, besides providing an alternative therapeutic approach, could represent a tool to understand the biology of GBM CSCs and related paracrine mechanisms, underpinning MSC-GBM interactions.
In this review, recent findings on the complex nature of MSCs will be highlighted, focusing on their elusive impact on GBM progression and aggressiveness by direct cell-cell interaction and via secretome, also facing the perspectives and challenges in treatment strategies.  Table 1 for details), which represent a promising tool for cell therapy, tissue engineering, and regenerative medicine. More recently, Wharton's jelly (WJ), amnion and corion, 5 and umbilical cord blood (UCB) 6 were proposed as alternative source of MSCs, although in the UCB they were identified at very low frequency as compared to other tissues. 7 MSCs have been harvested also from endometrium, synovium, 8 muscle, 9 skin, placental, 10 and dental pulp. 11 Increasing evidence proposes the use of MSCs as promising therapeutic approach for the treatment of several diseases and applications in the fields of regenerative medicine, neuroscience, oncology, pharmacology, and bioengineering. Currently, more than 500 studies (recruiting, not yet recruiting, active not recruiting on May 2020; search terms: MSC and mesenchymal stromal cell) are registered for MSCs, according to https://ClinicalTrials.gov.
At present, the extension of the concept and term MSC, originally circumscribed to nonhematopoietic BM-derived cells, to cells derived from additional postnatal tissues, rises some concerns. In particular, the MSC concept was questioned as far as two main defining stem cell assumptions, self-renewal and multipotency, and concerning the experimental approaches used for isolation and characterization.
Indeed, MSCs from different tissue sources comprise differences in cell populations, which display distinct characteristics, technical difficulties and advantages, and clinical translation potential. 12 Furthermore, large-scale high quality ex vivo isolation of MSCs is hampered by the low prevalence in human tissues and suboptimal in vitro expansion protocols, which are unable to maintain MSC essential properties required for therapeutic applications.
MSCs in vitro expansion for therapeutic applications might impact proliferative rate, homing molecules, genetic stability, transcriptional processes, multipotency, transformation, and senescence of isolated MSCs, 13,14 rising critical biosafety issues. 15 In this context, the optimization of culture conditions (medium, serum, supplements, substrates) to preserve MSC phenotype, homogeneity, fate, and better mimic should be integrated with processes allowing large-scale production of quality MSC, to be used in both preclinical and clinical studies.
Isolation and ex vivo expansion of MSCs are also crucial steps to ensure adequate material for potential clinical application. Commonly, MSC isolation exploits their plastic-adherence ability, which allows quite easy cell recovery and grow in a defined culture medium; however, this protocol implies low homogeneity of freshly-isolated cells, which increases during long-term expansion in vitro, necessarily combined with maintenance of differentiation ability, essential to make MSC reliable candidate for medical biotechnology.
Alternatively, the separation of MSCs could be performed by mechanical or enzymatic approaches although with higher impact on biological properties of the isolated cells. 16 Therefore, standardized separation techniques, allowing MSC purity or optimal enrichment, could significantly improve the reliability of results from in vitro and in vivo experimental approaches as well as clinical trials.

| Defining criteria: Properties and markers
The key step to verify MSC identity, following plastic adherence, is the assessment of cell immune phenotyping and multipotency.  17 Although MSCs from different sources share the minimum standard criteria set by ISCT, many studies suggest that each tissue of origin has a different MSC content, and influences in vitro MSC features, such as the proliferative rate, the differentiation potential, and the marker expression profile (Table 1). 18,19 On the contrary, the in vivo molecular signature of MSCs is still undefined and scantly investigated, 20 as well as the exact matching between their in vitro and actual native behavior in vivo. 21 In this context, CD271 + /CD140a − MSCs have been proposed to fulfill stringent stem cell criteria of self-renewal and multipotency, also in vivo. 22 STRO-1 marker, not included in ISCT criteria, is used to immuneselect fresh BM-MSCs, although it decreases during in vitro expansion of the culture 23 and it is not univocally present in MSCs, but it is also expressed by endothelial cells. 24  After intravenous inoculation, the vast majority of MSCs is rapidly trapped within the capillary beds in the lungs, due to MSC size and volume, which are bigger than lymphocytes, a phenomenon termed "first-pass" effect. Intra-arterial infusion bypasses the "first-pass" effect, allowing MSCs to spread peripheral tissues before to reach lungs. 43 Of note, endogenous BM-MSCs are smaller in size than ex vivo in vitro-expanded MSCs. 44 Conversely, MSC tropism seems not be limited by the blood-brain-barrier (BBB), formed by cellular interaction between astrocytes, pericytes, neurons, and microvascular endothelial cells. 45 However, the molecular mechanisms responsible for MSC brain  Interestingly, perivascular MSCs also reside in the human brain tissue, within neuro-vascular niches, with endothelial cells, astrocytes, and neurons, 67,68 as well as in mouse brain, where MSCs, morphologically similar to BM-MSCs, have been described. 69 There is increasing evidence of the critical involvement of MSCs in the TME, and therefore in cancer development and progression.
MSCs are recruited within tumors and tightly interact with the other cell types present in the TME, since tumor site is affected by a chronic state of inflammation. Thus MSC homing, fate and reprogramming is driven by cues produced by diverse resident cells composing the tumor stroma (endothelial cells, fibroblasts, pericytes, adipocytes, immune cells), cancer cells, and cancer stem cells (CSCs). 70 CSCs are slowly dividing cells, display a highly invasive phenotype, and are considered the "root" of cancer recurrence.
MSCs exposed to tumor cell-conditioned medium can differentiate into cancer associated fibroblasts (CAF) 71-74 by TGFβ1-mediated mechanisms, promoting tumor invasion, epithelial-mesenchymal transition (EMT), ECM modification, and cancer cell stemness, leading to tumor progression and metastasis. 75 In particular, a back-and-forth crosstalk between MSCs and CSCs has been shown in several tumor types 76,77 ; for example, in breast cancer this interaction can transform MSC into tumor-forming cells. 78 MSCs have been identified in the stroma of many cancers, including human GBM, likely deriving from local sites or being recruited from BM, 79 and represent, altogether with other nonneoplastic stromal components (endothelial cells, pericytes, immune cells, and glial cells), about 50% of GBM mass. 80 GBM, the most common and aggressive brain primary cancer, despite surgery, radiation, and chemotherapy, is invariantly lethal.
GBM represents the paradigm of the role of tumor-initiating CSCs 81,82 in the promotion of tumor aggressiveness, cell heterogeneity, and drug resistance. 83 All these glioblastoma stem cell (GSC) functions are tightly regulated by autocrine/paracrine activation of chemokine receptors (ie, CXCR4/CXCR7), 84 and by TME. 85 In detail, GSC selfrenewal is sustained by reactive tumor-associated microglia and macrophages, astrocytes, endothelial cells and other cell types present in the niches, MSCs included, which thus contribute to tumor recurrence and therapeutic resistance. [86][87][88] Interestingly, in the WHO classification, which divide GBMs in proneural, neural, mesenchymal, and classical subclasses, the majority of GBMs has a mesenchymal phenotype, identified by the incorporation of peculiar molecular signatures (IDH status, ATRX loss, H3K27M mutation, TP53 mutation, 1p/19q codeletion). 89 The clinical behavior of mesenchymal GBMs is extremely aggressive, with resistance to radiotherapy and the poorest prognosis as compared to all the other subtypes. 89,90 In addition, also GSCs from this subgroup of GBM express mesenchymal markers, being highly positive for CD44 and BMI1, and negative for CD133. 91,92 The evolution of GBM toward the mesenchymal phenotype is pushed by several factors, including stromal and immune cells within the TME, and a selective pressure induced by radio-chemotherapy. 93,94 In fact, current GBM therapies Several findings suggest that MSCs are involved in angiogenesis, further contributing to the malignancy of gliomas. 96,101 The ability of GA-MSCs to differentiate into pericytes, driving the maintenance of functional vessels essential for GBM growth, has been also described, 100 and indeed CD105 + MSCs are localized around GBM arterioles. 79 Finally, the ability of GBM-associated endothelial cells to acquire a MSC-like phenotype in GBM TME also contributes to chemoresistance via the activation of Wnt/β-catenin axis and the multidrug resistance-associated protein-1. 102  A comparative analysis of the major studies analyzed in this review is reported in Table 2.

| Direct MSC-GBM cell-to-cell interaction
The crosstalk between MSCs and GBM cells, besides protein and vesicular component exchange, is strictly dependent on and sustained by their close contact that is established in the TME. 93   (It has to be noted, however, that PLoS One Editors 116 retract a previous study from these authors on the same topic 117 138,139 ; and apoptotic bodies (diameter range 50-5000 nm), which represent fragments of apoptotic cells. 140 Besides a certain degree of confusion in particle definition existing in the studies trying to develop a therapeutic approach using MSCderived EVs, a further complication also derives from the potential different activity according with the tissue origin of MSCs (see below).
Thus a rather complex scenario is currently present in literature when approaching the biological role of EVs in intercellular communications and, even more seriously, when considered for potential clinical application. 141 Indeed, while the therapeutic application of MSCderived EVs seems to be simpler than the direct use of whole MSC (ie, as far as production, manipulation and storage, including their engineering to transport pharmacological active molecules, which extend their application fields), several standardization issues are still to be solved to obtain a reliable therapeutic use of these particles. 135 The first evidence of the potential therapeutic role of MSCderived EVs occurred in trials using whole MSCs to induce immunosuppression in Crohn and graft-vs-host diseases. However, also in the presence of beneficial effects for the patients, MSCs were not proved to persist after administration or directly contribute to the observed activity. Thus a paracrine activity, including the production of EVs, was assumed as mechanism of their therapeutic activity. 142 This hypothesis was confirmed in some models of ischemia-reperfusion injury by injection of MSCs conditioned medium, which showed the same efficiency of the administration of the whole cells. 143 Following this evidence, the development of MSC EVs as innovative drugs was extended to many pathological conditions including tumors. To date, the main approach tested is the use of MSC EVs as vehicle for cytotoxic drugs or for molecules with potential antiproliferative activity (eg, miRNAs).
However, in agreement with the modulatory effects on cell proliferation observed by the direct MSC-GSC interaction, 119 a direct modulation of GBM cell functioning by interaction with EVs was also reported. 144 Glioma cells in culture are able to uptake sEVs (particle analysis was reported to contain a predominant population with a diameter <150 nm) into the cytosol (measured by both confocal fluorescence microscopy and FACS). Using U87 glioma cell line, it was shown that tumor cells treated with these particles were highly modified in their proliferation rate. 144 However, in this study, a surprising differential effect was observed using EVs isolated from MSCs of dif- This effect was mediated through the stimulation of the release of several growth factors (VEGF, PDGF, FGF-7, hepatocyte growth factor) by the EV-exposed SHSY5Y cells. 148 Thus, the glioma-dependent MSC conditioning represents another potentially clinically relevant variable to be considered when MSCs are planned to be used as therapeutic agents.

| Cargo analysis of MSC EVs
To TMZ of GBM cells. 159 In another study, MSCs were transduced with a lentivirus vector expressing miR-124a and showed that high levels of miR-124a were specifically accumulated within exosomes. 160 In vitro, GSC treatment with exosomes from transduced cells caused reduction of viability and clonogenicity. Deliver of exosomes containing miR124 to mice bearing intracranial GSC-induced GBM resulted in long-term survival in 50% of the animals, in which no histological evidence of tumor was present. Mechanistic studies showed that miR-124a acts by silencing Forkhead box (FOX)A2, which renders GSCs unable to efficiently metabolize lipids, which intracellularly accumulate to reach toxic levels.
These data suggest that MSCs may act as natural biofactory for exosomes carrying miR-124a with antitumor potential. 160 The EVs secretion of MSCs has been exploited to deliver synthetic miR-124 and miR-145 mimics to impair GSC self-renewal and migration through the inhibition of the expression of SCP-1 and Sox2. Importantly, MSCs were able to deliver miR-124 mimic to GBM xenografts, when intracranially administered. 161

| THE POTENTIAL OF MSC-BASED THERAPY FOR GBM
Tumor-tropic capacity of MSCs gained attention also as a tool to deliver cytotoxic agents into cancer sites. 172 Easy viral vector transduction, extensive protein production and expansion in vitro as well as their immune-inert feature, associated, in the brain tumor context, with their ability to cross the BBB, favor MSC genetic engineering. Cytotoxic effects of MSC loaded with different anticancer drugs, were also prepared using liposomes and nanoparticles.
BM-MSCs were loaded with paclitaxel and tested on T98G glioma cell proliferation in vitro. It was shown that the drug was incorporated in a sufficient amount released at cytotoxic effective concentrations when located in the proximity of the cancer cells, despite a 20% loss of cells due to spontaneous apoptosis. 173 In vivo, the treatment with paclitaxel-loaded MSCs of mice carrying GBM xenografts showed a high tropism of MSCs toward the tumor cells, and the ability to cause drug-dependent cytotoxicity, without interfering with normal astrocyte viability. 174 In another study, BM-MSCs were primed with sorafenib without significant toxicity and were showed to be able to release up to 60% on the loaded drug. These modified MSCs were intranasally administered to mice xenografted with U87 glioma cells, 6 and 10 days after U87 cells. The strong interest in exploiting the tumor-homing ability of MSCs to target tumor cells is increasingly improving the genetic manipulation of these cells, to optimize their migration and effectiveness; however, the underlying mechanisms sustaining tumor tropism and antitumor activity, likely dependent on MSC origin and contextdependent effects, are not completely understood and will require further studies.
As alternative approach, modified EVs from MSCs can be used as selective drug delivery system, 193,194 although in some circumstances a pro-tumorigenic activity was also reported. 195 Several advantages have been, in principle, identified in the use of EVs. 196 In general, they are quite stable and there is the possibility to freeze them for later use. It is also possible to target them to obtain a directional cell-specific uptake, for example by inducing the expression on the cell of origin of ligands, which can be transferred to the surface of the vesicles and bind to specific receptors on target cells. Thus, high performing exosome purification procedures were developed from WJ-MSC to obtain pure populations, a fundamental condition to adopt these vesicles in clinics. 197 Different approaches can be used to modify EV cargo to use them as precision medicine drug carriers, including overexpression of protein of interest in parental cells, antibody or antigen conjugation, chemical modification, or passive and active loading (co-incubation with free drugs, sonication, electroporation, incubation with saponin, antibody binding). 193,198,199 Several approaches were used in this perspective, including EV loading with miRNAs (see previous paragraph) or cytotoxic drugs (doxorubicin, paclitaxel, etc.). Hydrophobic and hydrophilic small therapeutic molecules can be incorporated into EVs, leading, after parental administration, to an improve drug targeting to tumor cells, obtain a slow time-dependent release from the exosomes, and increase in potency and efficacy. 193 While drug delivery, via EVs, and exosomes in particular, was tested in several other tumors 200  (d) define and unify the MSC processing protocols, from the in vitro isolation and expansion steps to the patient treatment.
Due to the complexity of these issues and the absolute requirement for a standardization of the procedures for MSC isolation and characterization to ensure their optimal application not only in oncology, but also in other therapeutic fields, the constitution of expert panels which can carefully dissect the literature, possibly organized in subgroups according the different application of this cell therapy, are highly recommended. We believe that this approach, establishing a F I G U R E 2 Diagrammatic representation of the interactions between mesenchymal stem cells and glioblastoma cells, also involving other nonmalignant stromal and immune cells within the tumor microenvironment. In the figure, the critical pathways that may support or impair tumor growth via a variety of mechanisms are highlighted gold standard in the study and use of MSCs, involving basic and clinical researchers, and even biotech and pharmaceutical companies, could allow a better sharing of information resulting in a real advancement in the field.

CONFLICT OF INTEREST
The authors declared no potential conflicts of interest.

AUTHOR CONTRIBUTIONS
A.B., F.B., T.F.: wrote the manuscript; all authors contributed to literature searching, analysis, and critical review of the article cited, and proofed and approved the manuscript.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.