Neurotrophic Features of Human Adipose Tissue-Derived Stromal Cells: In Vitro and In Vivo Studies

Due to its abundance, easy retrieval, and plasticity characteristics, adipose-tissue-derived stromal cells (ATSCs) present unquestionable advantages over other adult-tissue-derived stem cells. Based on the in silico analysis of our previous data reporting the ATSC-specific expression profiles, the present study attempted to clarify and validate at the functional level the expression of the neurospecific genes expressed by ATSC both in vitro and in vivo. This allowed evidencing that ATSCs express neuro-specific trophins, metabolic genes, and neuroprotective molecules. They were in fact able to induce neurite outgrowth in vitro, along with tissue-specific commitment along the neural lineage and the expression of the TRKA neurotrophin receptor in vivo. Our observation adds useful information to recent evidence proposing these cells as a suitable tool for cell-based applications in neuroregenerative medicine.


Introduction
Adipose-tissue-derived adult pluripotent cells, commonly known as adipose tissue stromal cells (ATSCs) are mesenchymal stem cells (MSCs) residing in the connective stroma of adipose tissue. They represent a valuable source of adult stem cells, being easily isolated from an abundant and accessible tissue [1][2][3]. Their plasticity along with the ease of in vitro culturing and propagation makes them the most used cell type in a wide range of tissue regeneration applications [4][5][6][7].
We have previously shown the ATSC-specific molecular properties, by comparatively analyzing the geno-mewide expression profiles of MSCs from different adult tissues [3]. The study allowed indicating the main molecular features which regulate the stemness maintenance of MSCs and a more extensive plasticity of ATSC in vitro. The complete result dataset of this previous study (available at the Gene Expression Omnibus (GEO) database, http://www.ncbi .nlm.nih.gov/gds, accession number GSE8954) also indicated that ATSCs specifically express neurospecific genes.
The purpose of this study is to extract the biologically significant genes from this dataset and validate the functional relevance of the neurotrophic genes expressed by ATSC both in vitro and in vivo.

In Silico Biological Analysis of the Microarray Dataset.
In order to identify the candidate genes involved in the neurotrophic properties of ATSCs, the gene list of ATSC-specific genes obtained through the microarray-based gene profiling of ATSC compared to bone-marrow-derived mesenchymal cells (BMSCs) and fibroblasts (http://www.ncbi.nlm.nih .gov/gds, accession number GSE8954) [3] underwent an ad hoc biological analysis, aimed at finding neurologically relevant genes. For this purpose, the list of 441 genes specifically upregulated in ATSC (P value <0.01), resulting from the statistical analysis (see [3] for statistical methods used in data analysis), were categorized according to the "biological function" annotations implemented from the Gene Ontology Annotation (GOA) database (http://www.ebi.ac .uk/GOA/). Specific neuroprotective, neurodevelopmental, and/or neurotrophic functions were further studied using the "Gene Reference Into Function" tool in GenBank (http://www.ncbi .nlm.nih.gov/gene/about-generif).

Patients and Specimens.
Adipose tissue (AT) specimens were obtained by lipoaspiration from healthy volunteers (mean age 40.2 ± 14.2 years) upon obtaining a written consent. A skin biopsy was obtained from the retroauricular region of an healthy male donor (aged 45) and served for the isolation of human dermal fibroblasts (HDF). Individuals data were handled confidentially and anonymously. All the procedures employed in this study were approved by the ethical committee of the Catholic University of Rome (Rome, Italy; number P552 (A.779)/CE2007).

Chemicals and Reagents.
Cell culture media and supplements were purchased from Lonza (Basel, Switzerland). Enzymes, growth factors, and all other chemicals used in this study were purchased from Sigma (Sigma-Aldrich, St Louis, Mo,USA), unless otherwise specified.

ATSC Isolation and Culture.
Mesenchymal stromal cells were isolated in primary culture from the lipoaspirates, as already described elsewhere [3]. Briefly, AT was extensively washed, mechanically fractionated, and digested using 0.1% collagenase type VIII. The lysed tissue was then filtered through a 100 μm mesh, and the cell suspension was centrifuged. The cell pellet was then plated in T75 tissue culture flasks using Dulbecco's Modified Eagle's Medium (DMEM), supplemented with 10% FBS, 100 units/mL penicillin, 100 μg/mL streptomycin, and 0.2 ng/mL fibroblast growth factor beta (bFGF). Cells were subcultured as previously described [3] and then used for in vitro and in vivo experiments, as detailed in the following paragraphs. ATSCs growth kinetics up to fifteen culture passages and their immunophenotype were assessed as already described elsewhere [6].

HDF Isolation and Culture.
Dermal fibroblast were isolated in primary culture from the skin biopsy and cultured as previously described [8]. These cells served as a mesodermalderived differentiated controls to produce the conditioned medium (HDF-CM) used in the in vitro experiments (see following paragraphs).

In Vitro Experimental Procedures: Neural Cell Line Cultures and Treatments
In order to assess the functional significance of the neurotrophic genes specifically expressed by ATSCs, LAN5 and PC12 cells were used as neural undifferentiated cell lines for the in vitro experiments. These cell lines are commonly employed as valuable models to study the neuronal differentiation and degeneration processes in vitro [9][10][11][12][13].

Cell Lines and Treatments.
The human LAN-5 dopaminergic cells were grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 2 mM glutamine, 40 μg/mL gentamicin, and 10% heat-inactivated fetal calf serum (FCS), according to standard protocols [13]. Cells were plated at a 10 4 /cm 2 seeding density in 24-well plates. The day after plating, ATSCs were seeded in the same wells using a 10 4 /cm 2 seeding density. In addition, separate wells of LAN5 cells were cultured in presence of ATSC-conditioned medium (ATSC-CM), which was obtained by filtering through a 0.2 μm cellulose acetate filter the supernatant medium of subconfluent ATSC cultures. Thereafter, both LAN5-ATSC cocultures and ATSC-CM-treated cells were grown for three days without changing the culture medium.
The rat PC12 noradrenergic cell line was seeded at a density of 5000 cells/cm 2 in RPMI 1640 medium containing 5% fetal calf serum and 10% horse serum and grown till 80% confluence, according to standard protocols [12]. Between the third and the fourth culture passage, cells were plated in 24-well plates, using a 10 4 /cm 2 seeding density. The day after plating, the PC12 culture medium was replaced by either ATSC-CM or HDF-CM. LAN5 and PC12 cells in standard culture medium served as controls in the experiments. As PC12 cells are known to differentiate along a functional neuronal phenotype upon NGF treatment, cells primed with 50 ng/mL of NGFβ were used as positive control of differentiation [9]. Cellular morphology was evaluated by an invertoscope up to four days of culture.

In Vivo Experimental Procedures:
Neonatal Rat Brain ATSC Inoculation

Adenoviral-Mediated Cell Transduction.
In order to make ATSC recognizable in living tissues, cells were transfected using a defective adenoviral vector carrying the enhanced green fluorescent protein (AdEGFP) as a reporter gene. AdEGFP stocks were kindly provided by the Vector Core Facility of the University of Pittsburgh (Pa, USA). Cells were plated at a 10 4 /cm 2 seeding density and treated with AdEGFP using a multiplicity of infection (MOI) of 100 plaque-forming units (pfu)/cell. The efficiency of cell transduction was assessed observing fluorescent cells 48 hours later using an invertoscope equipped with a fluorescent lamp. EGFP-expressing cells were then inoculated in neonatal rats, as further described.

Cell Transplantation.
Human ATSCs were transduced with Ad.eGFP 48 hours prior to in vivo transplantation. The surgery was performed on neonatal rats at postnatal day 1 (P1), after the induction of deep anesthesia by hypothermia. A small parietal hole was made into the skull above the frontal cortex, and cells were slowly injected into the lateral ventricle (1 mm posterior to the bregma, 1 mm lateral to the midline, and 2-2.5 mm ventral to the pial surface) using a glass micropipette coupled to a Hamilton microsyringe. For each animal treated, 5 × 10 4 ATSCs suspended in 1 μL of Puck's saline A (Invitrogen, Carlsbad, Ca) were used. Shamoperated animals were injected with the same volume of The colocalization of eGFP with the above-mentioned markers was examined with a Zeiss LSM 510 confocal laser scanning microscopy system.

ATSCs Express Neurospecific
Genes. Data extracted from previously published microarray data showed the selective upregulation of 441 genes (P < 0.01) in ATSC compared to BMSC and human fibroblast MRC5 cells (Figure 1). The in silico biological analysis of the microarray data (GEO dataset number GSE8954) allowed to identify a short list of biologically relevant genes, involved in neuroprotection, neural developmental processes, and neurotrophic functions (see Table 1). In particular, this 12-transcript list included genes, namely, nerve growth factor beta (NGFB), neuropilin 1 (NRP1), and GTP cyclohydrolase 1 (GCH1), encoding soluble neurotrophins which are known to mediate neuronal growth, differentiation, migration, and neuroprotection [9,14,15]. The neuronal cadherin CDH2 belongs to the major transmembranar signalling complex cadherin/catenin that plays a key role in neuronal processes during early development. It is activated during neural circuit formation and maturation to mediate axonal outgrowth and arborisation [16,17]. Moreover, nearly all genes in the list are implicated in developmental processes within the nervous system, such as neurogenesis, neuron differentiation, axonogenesis, axon guidance, nerve growth, and glia differentiation and migration ( Table 1). The phosphoribosyl pyrophosphate synthetase 1 (PRPS1) and the phosphoglycerate mutase 1 (PGAM1) genes are implicated in metabolic pathways which are essential in neuronal function and maintenance (see function details and references in Table 1).

ATSCs Induce Neurite Outgrowth in PC12 and LAN5
Cells. In order to evaluate the effects of the supposed neurotrophic properties of ATSC, the capability of inducing visible changes in cell morphology of neural cells was first assessed in vitro. For this purpose, LAN5 cells were either cultured in ATSC-CM or cocultured with human ATSC for three days. Both cells cultured in ATSC-CM (Figure 2(b)) and those in coculture (Figures 2(c)-2(d)) displayed evident changes in shape and morphology, compared to those grown in standard culture medium (Figure 2(a)). The morphological changes consisted in the formation and elongation of neurite-like processes observed in discrete loci of the culture plate. The outgrown neurites seemed to establish contacts with both neural cells and ATSC in culture (Figure 2(d)).
In addition, the adrenergic PC12 cell line was cultured in presence of ATSC-conditioned medium (ATSC-CM) for four days. PC12 primed with βNGF and PC12 cultured in HDF-CM was used as positive and negative neuro-differentiation controls, respectively. The morphological analysis showed the extensive outgrowth and extension of neurite-like structures in both βNGFand ATSC-CM-treated cells exhibiting essentially overlapping features (Figure 2(f)-2(g)), compared to cells cultured in standard medium (Figure 2(e)). Cells grown in HDF-CM showed clear morphological signs of distress, becoming small-rounded vacuolized cells, with a marked tendency to detach (Figure 2(h)).

In Vivo Analysis of ATSC-Specific Neurotrophic Features.
The functional significance of the ATSC-specific upregulation of genes involved in the neural lineage has been further investigated in vivo after transplantation of ATSCs in the neonatal rat brain. ATSCs were efficiently transduced with Ad.eGFP prior to in vivo transplantation (Figure 3).
Histological examination of ATSC-transplanted young rats sacrificed 7 days after transplantation showed clusters of eGFP-positive ATSCs, characterized by rounded morphology, localized in the wall of the lateral ventricle, near the needle tract, surrounded by GFAP positive astroglial endings (Figure 4(a) A-C). In particular, based on the results observed in vitro, we assessed the expression of the anti-NGF-β receptor, as to further investigate the significance of the NGF/TRKA signaling pathway. ATSCs exhibited immunopositivity for the TRKA antibody 7 days after transplantation (Figure 4(a) D-F). At this time point, no colocalization with neuronal (Doublecortin, NeuN), astroglial (GFAP), or oligodendroglial (O4) markers were observed (not shown). Histological examination of young rats sacrificed 15 days after transplantation confirmed the survival of ATSCs in the brain of injected animals. Grafted cells examined at this time point were mainly localized within the brain parenchyma, near the ventricular system and frequently in the hippocampus. They showed a bipolar or multipolar morphology with processes extending in various directions. Interestingly, confocal microscopy examination revealed that many of these eGFP-positive ATSCs coexpressed also the astroglial marker GFAP (Figure 4(b) A-C), while no colocalization between eGFAP and Doublecortin, NeuN, or O4 was found (not shown). Virtually all engrafted ATSCs expressed immunopositivity for anti-TRKA antibody (Figure 4(b) D-F)). Sham-operated animals exhibited only a mild GFAPstained glial reaction around the needle tract (not shown).

Discussion
Different evidences indicated that transplanted MSCs promote endogenous repair of neurologically damaged areas and neural differentiation, via the release of soluble trophic factors and cytokines [27].
In particular, recent studies indicated that ATSC culture medium should contain neurotrophic factors, which were able to induce neuritogenesis in PC12 cells in vitro and protect brain from both hypoxic damage and glutamate neurotoxicity [28][29][30]. Nonetheless, only selected molecules have been dosed in ATSCs as possible neurotrophic candidates [28][29][30][31], while the expression of a wider panel of neuro-specific molecules has not been assessed in ATSCs so far.
The possible complete list of neurotrophic/neuroprotective factors specifically expressed by ATSC is proposed in this study, as a result of the in silico analysis of differentially expressed genes in MSC isolated from different adult tissues [3]. This revealed that ATSCs strongly and specifically express at least three neurotrophins: NFGB, NRP1, and FGF2. These secreted molecules reasonably represent the molecular background of ATSC-neurotrophic features. The in vitro assays in this study demonstrated that ATSCs could in fact induce neurite outgrowth not only in PC12, but also in human neuroblasts (LAN5 cell line). The induction of neuronal differentiation should be the result of the demonstrated presence of soluble secreted factors in ATSC culture medium [28] along with cell-to-cell contacts with neural cells in vitro. Thus, this event could be reasonably mediated by both NGFB, which promotes neuronal differentiation [9], and NRP1 that guides axon growth [15]. Also the nonneurospecific growth factor FGF2 could play a role in this event, being able to promote neurogenesis [22]. In addition, the adhesion molecule CDH2 that is expressed on the plasma membrane and is involved in axonogenesis and synapse assembly [19] could play a role in ATSC-mediated neuronal differentiation of LAN-5 cells. Although, the possibility that other factors participate in mediating this effects cannot be excluded.
Our data could also suggest that ATSCs neurotrophic function resides in a sort of astrocyte-like phenotype, as they specifically express genes belonging to the glial phenotype, including VLDR, FGF2, and NDN, according to GOA annotations. To this end, the necdin homolog (NDN) gene, involved in the NGFB signalling pathway, is particularly relevant, as it drives glial migration during nervous system development and is expressed in the cell projections [23]. Although the neural transdifferentiation capacity of MSCs has been largely debated, many recent studies emphasise the possibility of both bone marrow-and adipose tissue derived-undifferentiated stromal cells to differentiate along the neuroectodermal lineage to neuronal-like cells of the ectodermal lineage, mainly in vitro [32][33][34][35][36][37][38][39][40][41][42].
Recent data indeed assess the importance of cell-cell interactions along with the release of growth factors from the host tissue in ATSCs neural transdifferentiation [43]. In line with these observations, the results obtained in vivo, following cell implantation in the neonatal rat brain, indicate that ATSCs survive, migrate, and essentially differentiate toward an astroglial fate. Taken together, our observations suggest that ATSCs show a predisposition to the neural fate as they express a molecular phenotype resembling neural commitment in vitro and transdifferentiate along the neural lineage in vivo.
Recent reports evidence the successful implantation and migration of ATSCs in vivo using experimental models of rat brain ischemia, where they were able to promote functional recovery [44][45][46][47]. In addition, different groups reported the neural transdifferentiation of ATSCs transplanted in the injured spinal cord [48,49], evidencing that, when detached from the physiological niche, they express ectoderm neural markers [50]. We may speculate that secreted soluble factors from neighbouring cells and physical reciprocal contacts with neural cells may cause/facilitate transdifferentiation processes, as also indicated by the expression of the NGF receptor TRKA by transplanted ATSCs. This evidence, reported in in vitro studies [51], could suggest a possible autocrine mechanism on ATSC, as they express NGFB in vitro, although the functional significance of this observation deserves further studies.
Taken together, the results obtained in this study seemed to indicate that ATSC neurotrophic features reside in their specific capability of expressing not only secreted neurotrophins/neuroprotective molecules, but also structural protein-coding genes, mimicking the astrocyte function in sustaining neurons metabolism and function in the central nervous system and being able to differentiate into astrocytes. These properties, along with their reported capacity to migrate in injured tissues, could suggest possible future applications of ATSCs in many diverse neurological contexts.