The evidence for a role of pericytes in neural tissue repair was obtained from animal studies with experimentally induced brain ischemia, demonstrating that pericytes and pericyte-like cells, termed adventitial cells, originate neurons in the subgranular zone (SGZ) and glial cells in the dentate gyrus of hippocampus in monkeys (1). Moreover, the expression of stromal-derived factor-1 and angiopoietin-1 in the vascular niche close to the infarct areas is related with the guidance and survival of tissue-specific intrinsic progenitors in the subventricular zone (SVZ) of mice after focal cortical stroke (2). Pericytes surrounding brain blood vessels constitutively express nestin, a marker of neural progenitor cells (3), and, therefore, might participate in neuronal regeneration (4). The existence of mesenchymal stem cells (MSC) in perivascular niches supports the hypothesis that vessels in different tissues participate in originating such stem cells (SCs) (5–8). These findings do not agree with the idea of distinct developmental origins of all MSC, such as from neural crest (9, 10) and other mesodermal derivatives (11). The ability of these cells to differentiate into several cell types in vitro probably reflects the efficacy of the perivascular niche in the maintenance of their stemness. Evidence collected so far indicates that the pericyte in the perivascular compartment functions as a SC in the postnatal organism (7).
Neurogenesis in the brain occurs in close proximity to blood vessels and may be associated with angiogenesis (8). A web of growth and morphogenetic factors, including sonic hedgehog, bone morphogenetic proteins, Ephs/ephrins, Notch and fibroblast growth factor, nitric oxide, and erythropoietin present in adult neural SC niches (SVZ and SGZ of the hippocampal zone), participate in regulation of angiogenesis (12–15). In addition, blood vessels are conduits for the delivery of paracrine factors, such as hormones (sexual hormones, glucocorticoids, and prolactin) and cytokines, from distant sources. These “long-distance” cues may act directly on neural stem and progenitor cells and endothelial cells to regulate angiogenesis and neurogenesis (16, 17). The leptomeninges (pia mater), closely associated with blood vessels and also with microvascular pericytes/perivascular cells throughout the central nervous system (CNS), reveal neural stem/progenitor cell activity in response to ischemia and can generate neurons (18).
The therapeutic potential of SCs has already been recognized. Pericytes were able to regenerate skeletal muscle and promote functional recovery of the diseased heart and kidney (19). Actually, the pericyte is considered to be a pluripotent SC and, therefore, should be able to differentiate into tissues originated from the three germ layers (20). Recent data have shed light on the role of the CNS capillary pericytes as regulatory cells with SC capability (18, 21, 22). Brain pericytes originate from pluripotent neural crest cells like the neurons, supporting the hypothesis of one common cell lineage origin (23). CNS pericytes may be a source for purified viable SC with the potential of directed neurogenesis and could be important for future therapeutic strategies (21, 22). Roles for neural pericytes in brain remodeling after injury are unquestionable; however, the question whether outer CNS pericytes have the same SC properties needs yet to be resolved. Therefore, we now show in the present work that non-CNS perivascular pericytes obtained from rat aorta ex-plants and induced to neural differentiation express stage-specific embryonic antigen (SSEA-1), suggesting that they pass through a pluripotent stage and then differentiate into neural phenotypes in the presence of appropriate stimuli.
Materials and Methods
Reagents were purchased in cell-culture grade and highest available quality from Sigma, unless otherwise indicated. Male rats Sprague–Dawley 270–350 g (75–90 days old) from the animal facility of the Instituto de Química (Universidade de São Paulo, Brazil) were sacrificed by decapitation and used for extraction of aorta explants according to procedures approved by the local Ethics' Committee.
Isolation and Neural Differentiation of Pericyte SCs
The entire aorta was quickly dissected aseptically and washed for the removal of blood cells. Fat tissue was stripped off and vessels of vasa vasorum were prepared under a microscope in a laminar flow cabinet. Small ex-plants were transferred into 35-mm culture plates by using a scalpel followed by careful addition of culture medium in order to prevent tissue floating. The cell-culture medium consisted of DMEM low glucose supplemented with penicillin (100 U/ml), streptomycin (100 μg/ml), L-glutamine (2 mM), sodium pyruvate (2 mM), and FBS (10%). Pericytic cells spread from the explants and invaded the culture plate reaching at 70% cell confluence in culture medium supplemented with FBS after 1 week. Then, primary pericytes cells were collected by trypsinization and transferred into a new culture flask (75 cm2) or 24 well cell-culture plate (TPP, St. Louis, MO) on poly-L-lysine-coated glass coverslips for immunocytochemistry.
Neural differentiation was induced by addition of all-trans retinoic acid (RA; 1 μM) in serum-free defined media (DMEM) supplemented with insulin (5 μg/ml), apotransferrine (30 μg/ml), ethanolamine (20 μM), sodium selenite (30 nM), nonessential amino acids (1×), and sodium piruvate (2 mM). Pericyte spheres were obtained as previously described (19) in the presence of RA for 2 days in defined serum-free medium, then collected, and transferred into adherent culture dishes. Then, cultures were kept in defined medium in the absence or presence of the inhibitor of phosphodiesterases 3-isobutyl-1-methylxanthine (IBMX; 500 μM), and nerve growth factor (NGF; 400 ng/ml), brain-derived neural factor (BDNF; 10 ng/ml), supplemented with B27 (2%), neurobasal medium N2 (1%) in a humidified incubator at 5% CO2 and 37°C. After 8 days, immunophenotypic protein expression was analyzed, and transient elevations of cytosolic calcium concentration ([Ca2+]i) were determined in response to the stimulation by KCl (1 mM) or N-methyl-D-aspartate (NMDA; 100 μM).
Nondifferentiated and neural-differentiated cells following induction with RA plated in 35 mm dishes were incubated with Fluo3-AM (4 μM) in DMSO (0.5%) and 0.1% of the nonionic surfactant pluronic acid F-127 for 30 min at 37°C in HEPES (10 mM), pH 7.4, containing NaCl (140 mM), KCl (3 mM), MgCl2 (1 mM), CaCl2 (2.5 mM), and glucose (10 mM). After loading with Fluo3-AM, the cells were washed with incubation buffer and incubated for a further 20 min at 37°C. Calcium imaging was performed with a Inverted Research Microscope ECLIPSE-TiS (Nikon, Melville, NY) equipped with a 14 bit high-resolution CCD camera CoolSNAP HQ2 (Photometrics, Tucson, AZ) and analyzed with NIS-Element software (Nikon) using image acquisition rates of one frame per second. Fluo-3 fluorescence was excited with a xenon lamp at 488 nm, and the emitted light was detected using a bandpass filter at 515–530 nm. For the calculation of free cytosolic calcium concentration ([Ca2+]i), ionophore (4-Br-A23187; 5 μM) followed by EGTA (10 mM) were used to determine Fmax and Fmin fluorescence values, respectively. [Ca2+]i values were calculated as described previously (24, 25). At least 10 individual cells of cultures treated with RA alone or with RA and the differentiation cocktail were analyzed for their variations in [Ca2+]i.
For immunocytochemical reactions, cells grown on rounded coverslips (1 cm diameter) were fixed with paraformaldehyde (4%; Sigma) in phosphate-buffered saline (PBS) overnight at 4°C, washed three times with PBS, permeabilized with Triton-X-100 (0.05%) for 5 min, and incubated for 30 min in a blocking solution containing Tween-20 (0.1%) and heat-inactivated FBS (2%). The cells were incubated overnight at 4°C with rabbit polyclonal primary antibodies at 1:200 dilution against neurofilament (NF)-200 (polyclonal rabbit IgG, Sigma), glial fibrillary acidic protein (GFAP; polyclonal rabbit IgG, Dako, Carpinteria, CA), β3-tubulin (polyclonal mouse IgG, Sigma,) and SSEA-1 (monoclonal mouse IgM, Chemicon, Temecula, CA). The slides were washed three times with Tween-20 (0.1%) and FBS (0.2%) and then incubated for 1 h at room temperature with 1:200 dilutions of Alexa Fluor® 546 IgG goat anti-rabbit, Alexa Fluor® 488 IgG/M goat anti-mouse, or Alexa Fluor® 546 IgM goat anti-mouse secondary antibodies (all from Invitrogen), respectively. Primary antibodies were omitted in control reactions. Counterstaining of cell nuclei was done with 4′,6-diamidino-2-phenylindole (DAPI; 0.1%). Following another washing step, coverslips were mounted on slides by using Vectashield (Vector Laboratories, Burlingame, CA). Slides were examined on a confocal microscope (LSM 510-Meta, Zeiss, Jena, Germany) (26).
Enrichment of Pericytes in Aortic Explants and Induction to Neural Differentiation
Pericytic cells spread from aortic explants were taken into tissue culture. Cells revealed large cytoplasms with rounded nucleus, with a typical myo-endothelial morphology given by cytoplasmatic stress fibers observed by phase-contrast microscopy (Fig. 1A; Supporting Information Fig. S1A). These cells expressed smooth muscle α-actin (Supporting Information Fig. S1B) being in agreement with previous reports that pericytes may express smooth muscle actin in small and large vessels and are a source for MSC, which can differentiate into adipogenic, chondrogenic, osteogenic, and smooth muscle lineages (27, 28). Most of the cells expressed the pericyte and MSC marker Thy-1/CD90 as well as αSMA and platelet-derived growth factor (PDGF) receptors α and β, characteristic for pericyte cells as well as nestin, a marker of neural stem and progenitor cells (Supporting Information Fig. S1C–S1E). Pericyte spheres were obtained by floating cell culture in nonadherent agarose-treated plates in the presence of RA for 2 days in defined serum-free medium (Fig. 1B). The spheres were collected and transferred into adherent culture dishes, in which pericytes initiated displayed cell projections following 2 days of culture and under defined media (Fig. 1C, day 4). After the fourth day, the defined media was supplemented with a cocktail of NGF, BNDF, B27, N2, and the phosphodiesterase inhibitor IBMX for the consolidation of neural differentiation (Figs. 1D–1F).
Expression of Embryonic and Neurogenic Markers
We investigated whether aortic pericyte-derived could differentiate into the major cell populations present in the CNS. Coverslips were prepared for immunocytochemistry and stained for the detection of SSEA-1, astrocytic, and neuronal marker expression. Aortic pericytes expressed smooth muscle actin (Supporting Information Fig. S1B), but were negative to immunostaining against the endogenous neural markers β3-tubulin, GFAP, and NF-200 (data not shown). Although RA-treated pericytes did not express the pluripotency marker SSEA-1 (Fig. 2A), following induction to differentiation in the presence of RA, NGF, BNDF, B27, N2, and the phosphodiesterase inhibitor IBMX, pericytic cells reflected phenotypes of embryonic neuroectodermal cells by expressing SSEA-1 (Fig. 2B). Neural phenoypes were obtained, when cells were induced to differentiation for two days in the presence of RA, then kept for another 2 days in defined serum-free medium, and finally exposed for 4 more days to growth factor and IBMX supplemented media (differentiation cocktail). Astroglial cells were identified by the expression of GFAP (Fig. 2D), which was not detected when cells had been differentiated the absence of the growth factors and IBMX (Fig. 2C). Staining for the marker proteins β3-tubulin, characteristic for immature neurons (26), was visible following RA treatment and floating culture (Fig. 2E), and maintained during neural differentiation (Fig. 2F). The progress of differentiation into neuron-like cells (Figs. 2G and 2J) was detected by the expression of NF-200 (Fig. 2I), which was not expressed when cells were induced to differentiation in the absence of growth factors and IBMX (Fig. 2H). Aorta pericytes differentiated in the absence of NGF, BNDF, B27, N2, and the phosphodiesterase inhibitor IBMX did not reveal neuroectodermal-specific gene expression. Moreover, spheres treated for 2 days as floating culture in the presence of RA and afterward maintained as adherent culture for 6 days more in defined medium returned to initial pericytic morphology as shown in Figures 2A, 2C, 2E, and 2H, indicating that the above-mentioned morphogenetic factors were essential for neural differentiation of pericytes.
Functional Neural-Like Excitable Phenotypes and Response of Neural-Differentiated Pericytes to Stimulation by Neurotransmitters
In addition to neural-specific morphology and protein expression, one would expect that neural-differentiated pericytes are excitable and express functional neurotransmitter receptors. Single-cell calcium imaging was used to determine basal [Ca2+]i levels of pericytes exposed only to RA and of neural-differentiated pericytes in the presence of RA and differentiation cocktail (Figs. 3A and 3B) as well as changes in [Ca2+]i following application of KCl (100 mM) (Figs. 3C and 3D) or NMDA (100 μM; Figs. 3E and 3F). Maximal and minimal fluorescence (Fmax) were obtained in the presence of the Ca2+ ionophore A23187 (5 μM) and the Ca2+ chelator EGTA (10 mM) respectively (Figs. 3G–J). Single-cell calcium imaging of neural-differentiated pericytes revealed that application of a depolarizing stimulus KCl (100 mM) induced transient elevations of [Ca2+]i due to activation of voltage-operated calcium channels. KCl addition to the extracellular medium did not provoke any significant responses in pericytes induced to differentiation in the absence of growth factors and IBMX (Figs. 3C, 3K, 4A, and 4B; see also the Supporting Information Video 1). NMDA-evoked responses were not observed in pericytes, which had been treated with RA only during differentiation (Figs. 3E, 3L, 4A and 4B), suggesting that differentiation to functional neurons did not happen. However, differentiated cells in the presence of growth factors and IBMX revealed [Ca2+]i transients mediated by NMDA-glutamate receptors, specific for neurons and glial cells (Figs. 3F, 3L, 4C and 4D, as well as following depolarization with KCl (Figs. 3D, 3K, 4C and 4D). All 10 individually analyzed cells with neural morphology, differentiated in the presence of RA, and the differentiation cocktail, responded with [Ca2+]i transients following stimulation by KCl or NMDA (Fig. 4C), suggesting the presence of a population of neural-differentiated cells (40% of the total cell population). Time kinetics of KCl- and NMDA-evoked [Ca2+]i transients are shown in Figures 4B and 4D (see also Supporting Information Video 2). In summary, RA-induction was essential for the induction of neural differentiation, while the differentiation cocktail was necessary for the completion of differentiation into phenotypes with functional neuronal properties.
Recent research efforts have largely focused on the detection, phenotypic expression characterization, and in vitro replication and differentiation of SCs from human adult tissue. The determination of the origin and identity of SCs together with their niches in adult tissue also provides important information on their participation in endogenous tissue regeneration and their possible applications as pluri- or multipotent cells in cellular regeneration therapy (6, 7, 27–33). Evidence accumulated in the last few years show that the adult macro- and microvessels contain multi- or pluripotent SCs, including MSCs, and/or pericytes, as well as hemopoietic SCs, and lineage committed progenitors, such as vascular walls endothelial progenitor and Sca-1+ smooth muscle progenitor cells (6, 7, 23, 29–31, 33–36). Pericyte SCs have been isolated from different vascular tissues including abdominal adipose, bone marrow, dental pulp, and umbilical cord (6, 7, 23, 29–31, 33–35). Actual discussions of the role of pericytes as SCs propose that pericytes are a source for MSC (29). It has yet not been resolved how MSCs/pericytes contribute to the formation, maturation, and homeostasis of all vascularized tissues (5, 7). Because blood vessels and with them pericytes are part of all tissues and organs, there would be many therapeutic applications if these cells could be induced to differentiate into defined phenotypes (37). In view of that, we have studied the capability of these cells to differentiate into neural phenotypes.
During embryonic development, neural crest cells surround aortic vessels providing all components including pericytes and musculature-connective tissue except the endothelial cells (23). Brain microvascular pericytes show a pluripotential SC activity, express neural markers as nestin, GFAP, NF1, and oligodendrocyte O4 antigen when induced to differentiation (22), supporting the idea that pericytes are pluripotent cells in the blood–brain barrier. The neural crest cell origin of pericytes, in addition to a high similarity with MSCs and other sources of SCs, support the idea that SCs have an embryonic neuroectodermal/epiblast common origin that will give rise to cells of the three germ layers: ectoderm, mesoderm, and endoderm (1, 23, 38). This assumption would also explain why undifferentiated SCs express neural genes (39). It is unquestionable that CNS pericytes have a role in the brain remodeling after injury; however, whether non-CNS pericytes possess SC properties, such as pluri- or multipotency, is yet being discussed. Therefore, we have developed an in vitro protocol for differentiation of pericytes into neural phenotypes.
Pericyte isolation and culture did not depend on FACS selection or other complex equipment for cell separation. With the described protocol, 75% of cells expanded from the aorta explants express Thy-1/CD90 (Supporting Information) and pericyte markers (αSMA, PDGFRs). Our protocol reveals advantages over previously described methods for the differentiation of CNS microvascular pericytes (21, 22). Aortic pericytes required 8 days of differentiation for the acquisition of neural morphology, excitability, and expression of functional NMDA-glutamate receptors. Besides neural stem and progenitor cells that are intensively studied for the potential in regeneration therapy of neurodegenerative diseases (38), pericyte SCs might be an alternative SC source based on the easiness of isolation of these cells. In the near future, in vivo studies will reveal the possible therapeutic potential of these cells for the treatment of neurodegenerative diseases. Experimental support has been already obtained for the common regulation of angiogenesis and neurogenesis during developmental processes and regeneration following ischemic stroke (40, 41), being in agreement with our observation of pericytic cells with angiogenic potential being involved in neuronal regeneration.
The medium for the induction of neural differentiation contained BDNF and RA, factors known to induce neural differentiation of embryonic cells (42), of neural stem cells (43), bone-marrow MSCs, and adipose-derived mesenchymal cells with pericyte-specific markers, producing functional response profiles to stimulation by neurotransmitters (33, 34, 44, 45). In agreement with the characteristics of excitable neuronal cells, differentiated pericytes responded with [Ca2+]i transients to membrane depolarization by KCl indicating the presence of voltage-operated ion channels. Moreover, differentiated cells expressed functional NMDA-glutamate receptors such as in neural-differentiated adipose-derived SC where the presence of NMDA receptors served as an indicator and marker of the efficiency of neuronal differentiation (46). However, the presence of RA alone is not enough for maintaining pericytes in their neural-differentiated stage, as evaluated by their morphology, expression of neural marker proteins and excitability. However, the presence of the used morphogenetic factors (NGF, BNDF, B27, and N2) and IBMX maintained neural phenotypes of differentiated pericytes, but no excitability could be observed in the presence of RA alone. Consequently, RA and the cocktail of morphogenetic factors contributed to differentiation induction in independent manners by activation of different cell signaling pathways. RA activates nuclear receptors, while the used growth factors and IBMX mediate membrane receptor dependent down-stream signaling pathways.
In summary, pericyte cells are pluripotent SCs that can be induced to neural differentiation. Expression of SSEA-1 suggests that cells pass through a pluripotent stage, followed by expression of β3-tubulin, GFAP and NF200 such as observed during differentiation of neural crest cells. Differentiated cells express a protein characteristic for functional NMDA-glutamate receptors and voltage-gated ion channels. Following studies will elucidate whether neural-differentiated pericytes can be integrated in neuronal networks and participate in synaptic transmission. Besides being important for understanding mechanisms of neuronal differentiation, pericytes may provide a novel source of SCs for cell regeneration therapy in the CNS.