|Home | About | Journals | Submit | Contact Us | Français|
Repopulation of brain circuits by neural precursors is a potential therapeutic strategy for neurodegenerative disorders; however, choice of cell is critical. Previously, we introduced a two-step culture system that generates a high yield of neural precursors from small samples of adult canine skin. Here, we probe their gene and protein expression profiles in comparison with dermal fibroblasts and brain-derived neural stem cells and characterize their neuronal potential. To date, we have produced >50 skin-derived neural precursor (SKN) lines. SKNs can be cultured in a highly replicable fashion and uniformly express a panel of identifying markers. Upon differentiation, they self-upregulate neural specification genes, generating neurons with basic electrophysiological functionality. This unique population of neural precursors, derived from mature skin, overcomes many of the practical issues that have limited clinical translation of alternative cell types. Easily accessible, neuronally committed, and patient specific, SKNs may have potential for the treatment of brain disorders.
Profound neuronal and synaptic loss is a hallmark of neurodegenerative disorders and is the best pathological correlate of cognitive impairment in Alzheimer's disease (AD) (Terry et al., 1991, Scheff and Price, 2006). Restoration of depleted neuronal populations with stem cell therapy is a promising treatment strategy (Daley, 2012, Lindvall et al., 2012, Duncan and Valenzuela, 2017), but the inaccessibility of endogenous neural stem cells within the brain's neurogenic niches means that alternative cell sources are required, preferably of autologous origin, to avoid long-term immunosuppression, which is harmful to both graft and host.
Reprogramming technology transformed the field, converting fibroblasts (FBTs) and other somatic cells into induced pluripotent stem cells (iPSCs) (Takahashi and Yamanaka, 2006) capable of generating functional neurons in vivo (Wernig et al., 2008). Alternatively, somatic cells can be converted directly into induced neurons (iNs), in effect bypassing the pluripotent state by induced upregulation of neuronal specification genes ASCL1, NEUROD, MYTL, and BRN2 (Pang et al., 2011, Pfisterer et al., 2011). However, clinical translation of reprogramming technology remains limited for a number of reasons. First, conversion efficiency is typically very low, <1% of initial cells for iPSCs (Liao et al., 2008) and <6% for iNs (Pang et al., 2011). Second, both cell types suffer from unacceptably high line-to-line (and even clone-to-clone) variability (Truong et al., 2016), a factor that precludes clinical translation. Third, the unlimited capacity of iPSCs to self-renew presents an inherent risk of uncontrolled cell growth in vivo, with high (60%) tumor rates reported (Ring et al., 2012). Finally, genetic modification raises significant safety concerns, including activating oncogenes and the risk of unanticipated mutagenesis.
The discovery of niches within adult skin that continue to harbor multipotent stem cells raised the prospect of overcoming some of these limitations. Cell populations within these niches exhibit stem cell-like properties, distinct from mesenchymal or hematopoietic stem cells, and retain a neurogenic developmental potency without genetic manipulation, including expression of neural stem cell-related proteins such as NESTIN and SOX2 in situ (Toma et al., 2001, Fernandes et al., 2004). Using a neurosphere propagation method, these native stem cell-like cells, termed skin-derived precursors (SKPs), can be expanded for multiple passages (>50), and mature into neural cell types when exposed to neurodifferentiation factors (Toma et al., 2001, Biernaskie et al., 2006, Lavoie et al., 2009). However, the final neuronal yield achieved by this approach has been very low (2%–10% across studies), and the propensity for glial cell differentiation (Toma et al., 2005, Hunt et al., 2008) has generally excluded translation of SKP cells to neuronal therapeutic application.
Responding to this, we previously reported enhanced cellular homogeneity and neurogenic potential using a two-step neurosphere-adherent culture system that begins with mature adult canine skin (Valenzuela et al., 2008). The complex three-dimensional growth environment and uneven exposure to growth signals inherent in the neurosphere assay permits and promotes heterogeneous cell growth (Bez et al., 2003, Babu et al., 2007), with neural stem cells reported to represent less than 1% of this population (Reynolds and Rietze, 2005). By contrast, we used neurospheres solely as a primary selection step, further expanding the resultant cells as an exclusively epidermal growth factor (EGF)/basic fibroblast growth factor (bFGF)-dependent adherent monolayer culture. Adherent expansion of human SKPs has been reported previously (Joannides et al., 2004), and while better neuronal yields were achieved using serum and astrocyte-conditioned medium, glial and mesenchymal cell types remained commonplace. Adherent culture systems have also been employed to expand brain-derived neural stem cells, producing more homogeneous cell populations biased toward GABAergic and glutamatergic neurons (Conti et al., 2005, Pollard et al., 2006, Goffredo et al., 2008). Combining these approaching in our two-step serum-free culture system, a unique population of skin-derived neural precursors (SKNs) can be routinely generated from adult canine skin, maturing to produce greater than 90% neuronal yields in vitro without genetic manipulation (Valenzuela et al., 2008). Accordingly, SKNs represent a promising candidate for autologous neural cell therapy.
Our choice of studying canine skin was intentional because of the poor history of translation of rodent research into effective human neurodegenerative treatment. Rodents do not naturally develop AD pathology or neurobehavioural signs in late life, and transgenic models have failed to predict outcomes in human clinical trials (Cummings et al., 2014, Breitner, 2015). By contrast, canine cognitive dysfunction (CCD) is a naturally occurring analogue of human AD; affected dogs display a progressive amnestic syndrome (Cummings et al., 1996, Salvin et al., 2011) as well as AD pathology (Cummings et al., 1996), and, as in humans, prevalence accelerates exponentially in old age (Salvin et al., 2010). CCD may therefore be an ideal translational model to test regenerative therapies. Yet prior to this, any candidate cell type needs thorough characterization. Here, we therefore assess the line-to-line replicability and neurogenic potential of canine SKNs in comparison with both canine dermal FBTs and brain-derived neural precursor cells (NPCs) isolated from the canine neurogenic niche.
Our protocol for the culture of canine SKNs combines initial neurosphere selection with passage as an adherent monolayer for cellular homogenization (Figure 1A). From an approximately 6 cm2 abdominal skin sample, under proliferative conditions, floating neurospheres (≈100 μm in diameter) formed in culture within 7 days of isolation (Figure 1B). Following selection of these neurospheres, and their enzymatic dissociation, the resulting cells were cultured as an adherent monolayer. A large amount of cell death occurred within the first 2 days of this adherent culture, but a selective population consisting of exclusively EGF/bFGF-maintained SKNs survived. Following first passage, approximately 1,000,000 SKNs could be routinely generated from over 50 individual skin donors (85% donor-wise success rate). The clonal ability of SKNs was demonstrated by seeding at single-cell density in a collagen matrix that precluded cell fusion, with spheres approximately 50 μm in diameter forming after 7 days of culture (Figure 1C).
Under our adherent culture conditions, a homogeneous SKN population could be routinely expanded for up to five passages. At first and second passage, 80% confluence (passage density) was reached within 7 days, undergoing an average of 2.7 population doublings, defined as [log(final cell count/starting cell count)]/log(2) (Greenwood et al., 2004). Following third passage, proliferative rates began to decline, with only 1.4 population doublings occurring within an equivalent time frame. This decline was also quantified by 5-ethynyl-2′-deoxyuridine labeling of actively dividing cells. Initially comparable levels of active proliferation in SKNs and brain-derived NPCs significantly declined in the SKN cultures at passage 3 and beyond (Figure 1D). This observation was also consistent with SOX2 gene expression, a regulator for stem cell proliferation, where expression in SKNs was significantly lower than in NPCs (p = 0.01; Figure 1E). Under adherent proliferative conditions, SKNs are therefore limited in terms of their proliferative capacity.
Global transcriptome analysis was carried out on canine dermal FBTs, SKNs, and NPCs, in order to better understand the neural potential of SKNs and the general relationship between these three cell types. Principle-component analysis of 32,000 transcripts accounted for 60.1% of total transcriptomic variance, with three cell populations observed, each with distinctive expression profiles (Figure 2A). Cell lines within cell types were highly correlated (r ≥ 0.98 SKN, r ≥ 0.98 NPC, r ≥ 0.98 FBT), while globally SKN and FBT lines were more closely correlated (r range: 0.96–0.97) than SKN and NPC lines (r range: 0.92–0.94). Correlation coefficients for individual cell lines are displayed in Table S1.
To elucidate whether underlying molecular profiles were similar between these three cell types, ANOVA was performed at a false discovery rate of <0.05 to identify genes with a greater than 2-fold difference in normalized expression between cell types. Of the 32,000 genes analyzed, 291 genes were differentially expressed between SKNs and FBTs, while 2,157 genes were differentially expressed between SKNs and NPCs (Figure 2B). These univariate results indicate that SKNs are closer in global gene expression to FBTs than NPCs. A multivariate method of hierarchical clustering on differentially expressed genes also placed SKNs in a closer relationship to FBTs than NPCs (Figure 2C). Nevertheless, qPCR examination of specific neural stem and precursor markers showed that SKNs retain a distinct gene expression profile to FBTs, with NES, P75NTR, DCX, and TUBB3 all significantly upregulated (Figure 2D). Moreover, from the Venn diagram we can infer that there are at least 40 genes specific to SKNs (detailed in Table S2). Many of these genes are associated with cell proliferation, adhesion, migration, and cytoskeletal organization. Of particular note were genes coding for protocadherins. These cell-adhesion proteins are predominantly expressed by the developing CNS (Sano et al., 1993), and have key roles in establishing neuronal connectivity and dendrite arborization (Schalm et al., 2010, Lefebvre et al., 2012).
Transcript lists containing the most commonly cited genes for pluripotency, neurogenesis, SKPs, mesenchymal stem cells (MSCs), growth factors, and growth factor receptors were compared for expression intensity between SKNs and NPCs (Figure 3). A full list of the individual gene expression fold changes between SKNs and NPCs can be found in Table S3. The iPSC pluripotency reprogramming factor KLF4 (Takahashi and Yamanaka, 2006, Takahashi et al., 2007), and the TBX3 gene implicated in embryonic stem cell proliferation and neuroepithelial differentiation (Esmailpour and Huang, 2012), were found to be upregulated in SKNs. Pluripotency markers BUB1, TCL1A, and PRC1 were all upregulated in NPCs, confirming their greater proliferative capacity, as reported in Figure 1D. As expected, NPCs also showed stronger expression of neurogenesis markers. However, early neurogenesis markers MKI67, PAX6, NCAM1, NES, and SOX2 were also strongly expressed in SKNs, indicative of a largely undifferentiated population of neural precursors. SKNs expressed the majority of SKP and MSC markers more intensely than NPCs, with significant upregulation of PDGFRA, HOXA5, HOXC4, HOXC6, TWIST2, EYA1, MAB21L1, MSX1, RHOBTB3, and AP2A1.
Growth factor and receptor gene expression were explored in order to gain insight into factors that SKNs may produce, or be responsive to, if transplanted in the brain. Interestingly, genes for insulin-like growth factors and their receptors were highly expressed in SKNs compared with NPCs. Furthermore, expression of IL7 and VEGFC were upregulated in SKNs relative to NPCs. These growth factors, while expressed in adult skin, also have key roles in both the adult and developing brain, including neurogenesis and synaptogenesis, and as migratory cues for neuronal precursors (Michaelson et al., 1996, D'Ercole et al., 2002, Rosenstein et al., 2010). As such, they may facilitate SKN survival and engraftment following transplantation.
Pathway enrichment was then carried out on the 2,157 differentially expressed gene list (SKNs versus NPCs) and scored based on Fisher's exact test (p < 0.01) to identify functional groups with more than two over-represented genes from a specific pathway. Genes were referenced against the Kyoto Encyclopedia of Genes and Genomes C. familiaris genome database (Kanehisa et al., 2004). The aim was to characterize and compare the biological pathways best represented in both cell types. Focal adhesion, extracellular matrix (ECM)-receptor interaction and PI3-AKT signaling pathway were the three most enriched pathways (Table 1). The relative gene expression in SKNs and NPCs for components of these pathways and others have been mapped in Figures S1–S6. The focal adhesion pathway (Figure S1), incorporating upregulated expression of p21-activated kinase proteins in SKNs, has an important role in the regulation of the actin cytoskeleton and is implicated in cell proliferation and migration, mediating filopodia formation, and cell motility (Xia et al., 2008). Interlinked with the focal adhesion pathway, phosphatidylinositol 3-kinase-Akt signaling (Figure S2) was also highly enriched in SKNs. An important regulator of the cell cycle, this pathway has been implicated in promoting proliferation and inhibiting differentiation in adult hippocampal NPCs (Peltier et al., 2007). Axon guidance was also identified as an enriched pathway between NPCs and SKNs (Figure S3). Ephrins, upregulated in SKNs, have been shown to mediate growth cone survival and collapse, and play a large role in the neural crest migration during gastrulation. EPH-A in particular has also been shown to interact with P75 neurotrophin receptor (P75NTR) in axon guidance and mapping (Lim et al., 2008). In addition, the differentially expressed gene list also contained a number of upregulated genes associated with glutamatergic and GABAergic synapse (Figure S4) pathways, including AMP(A) and GABA(A) postsynaptic receptors, which were both upregulated in SKNs compared with NPCs. Nevertheless, these synaptic pathways in their entirety were not significantly enriched, likely due to the cells' undifferentiated neural precursor phenotype.
As expected, embryonic pluripotency marker NANOG was not expressed. KI67 and SOX2 proliferation markers were present in a majority of SKNs (51% and 72%, respectively), and while OCT4 was highly expressed, staining was perinuclear, consistent with reported expression in dermal cells (Li et al., 2010) (Figures 4A and 4B). SKNs were homogeneous for stem cell-like marker CD133, neural stem/crest cell-like marker NESTIN, and neural crest cell marker P75NTR, a regulatory molecule implicated in both skin stem cell biology (Adly et al., 2009) and neurotrophin signaling (Roux and Barker, 2002) (Figures 4A and 4B). In fact, across these three distinct markers, line-to-line variance was remarkably low: fluorescence-activated cell sorting of three independent SKN lines (in triplicate) revealed >97% expression for each marker and a between-line coefficient of variance of <1.8%. This excellent level of line-to-line phenotypic replicability is therefore consistent with therapeutic usage.
SKNs also highly expressed an array of neural precursor markers: polysialylated neuronal cell-adhesion molecule (PSA-NCAM), DOUBLECORTIN, and immature neuronal BETAIII-TUBULIN were present in over 85% of cells. Of note, the astrocyte marker glial fibrillary acidic protein (GFAP) and oligodendrocyte marker O4 were observed in less than 1% of cells (Figure 4A). Expression of BETA1-INTEGRIN, combined with the absence of ALPHA4-INTEGRIN expression, may indicate a propensity of SKNs to preferentially bind laminins. This supports upregulated laminin-associated genes observed within the ECM-receptor-enriched pathway (Figure S5). Mesenchymal markers, platelet-derived growth factor receptor alpha (PDGFRA) and CD44 were found to be expressed in SKNs (Figures 4A and 4B). These proteins were found to be co-expressed with NSC markers CD133 and P75NTR within single cells (Figure S6). Of note, PDGFRA, while classically a mesenchymal marker, is also found in neuroepithelial cells (Andræ et al., 2001) and adult neural precursors. Moreover, adult human PDGFRA-expressing neural precursors of the periventricular region do not co-express GFAP (Chojnacki and Weiss, 2004), and so SKNs may share this non-glial propensity. Expression of mesenchymal myofibroblast marker smooth muscle actin was minimal. Transferrin receptor, traditionally considered a haematopoetic stem cell marker, was also observed in SKNs, although this receptor is also associated with neural cell types throughout the CNS (Erickson et al., 2008).
Following passage 3, the proliferative rate of SKNs declines rapidly and their morphology increasingly comes to resemble differentiating neurons. This spontaneous neuronal differentiation was enhanced through the removal of mitogens from the culture medium and supplementation with brain-derived neurotrophic factor (BDNF), known to support neuronal differentiation. Over a 28-day differentiation period, SKNs remained viable, and developed increasingly more complex elongated processes and mature neuronal morphology (Figure 5A), significantly more frequent than in NPCs (p < 0.0001; Figure 5B). Post-differentiation SKNs were also a more homogeneous population than NPCs. SKNs expressed neuronal marker genes TUBB3 and MAP2 but, unlike differentiated NPCs, glial cell GFAP was not expressed (Figure 5C). Mature neuronal gene ENO2 was expressed in both cultures, while GABAergic gene GAD67 was only present in differentiated SKNs. Protein expression corroborated these findings (Figure 5D). Mature neuronal proteins NEUN and neurofilament were seen in differentiated SKNs. BETAIII-TUBULIN and MAP2 co-expression can also be seen in morphologically immature (Figure 5D; left) and mature stage differentiated SKNs (Figure 5D; right), while peripheral nervous system (PNS) neuron markers peripherin and P75NTR (when co-expressed with a neuronal cytoskeletal marker, Fernandes et al., 2004) were not present. PNS glial protein MBP (Schwann cells), and CNS glial proteins GFAP (astrocytes) and OLIG2 (oligodendrocytes) were also absent. Positive control stains for these proteins can be seen in Figure S7. Presynaptic vesicular protein BASSOON and punctate accumulation of the FM4-64FX membrane-affinity dye was also detected along the length of the cell processes as well as internalized within the soma, indicative of synaptic vesicle turnover (Figure 5E).
qPCR analysis following differentiation revealed downregulation of proliferative markers NES (3-fold; p = 0.0001) and P75NTR (4.9-fold; p = 0.001), complemented by upregulation of the p21 gene CDKN1A (4.8-fold; p = 0.007), signifying terminal differentiation (Figure 5F). Notably, significant upregulation of neural specification gene ASCL1 was observed (27.8-fold; p = 0.04), as well as trends toward increased expression for POU3F2, MYT1L, and NEUROD. SKN differentiation was also accompanied by upregulation of oligodendrocyte repressor gene DLX2, and of several neurotransmitter-associated genes including glutamate transporter SLC17A7, monoamine transporter SLC18A2, GABA receptor GABBR2 (4.1-fold; p = 0.01), and dopamine transporter DAT1. Significant upregulation of the BASSOON gene BSN (5-fold; p = 0.03), and a trend toward upregulation of the voltage-gated sodium channel gene SCN8A, was also indicative of neuronal maturation (Figure 5F). Overall, these protein and gene expression patterns suggest that canine SKNs are highly neuronally committed.
Over the differentiation period, SKNs displayed a steady hyperpolarization of resting membrane potential, with the majority of cells exhibiting resting membrane potentials analogous to mature functional neurons (≥−55 mV) at 25–28 days in vitro (Figure 6A). Functional voltage-gated ion channels were observed under voltage-clamp conditions (Figure 6B), and sub-threshold depolarization and graded regenerative responses were observed via current-clamp recordings as early as 14 days in vitro (Figure 6C). However, under these conditions mature action potentials were not recorded, and as such SKNs remained in an electrophysiologically immature neuronal state.
Calcium (Ca2+) influx via voltage-gated channels is known to accompany action potentials. Fura-2-acetoxymethyl ester calcium-affinity dye was therefore used as a higher-throughput method for detection of voltage-gated events. Calcium entry was observed by change in the 340/380 florescence signal ratio over time (Figure 6D) and focal shifts in pseudocolorized image intensity (Figure 6E), when normalized to baseline T0. Differentiated SKNs exposed to a depolarization solution of 129 mM KCl and 2 mM Ca2+ show rapid Ca2+ internalization (+). Interestingly, depolarization solution containing 100 μM ATP (++) induced an increased influx of Ca2+, with repeated spikes of calcium entry observed consistent with repeated voltage-gated events.
Adult skin is extraordinarily complex, comprised of several interactive stem cell niches that originate from ectodermal, mesenchymal, and neural crest cell lineages. Here, we show that the canine dermis can be used to isolate and propagate precursor cells that display an interesting combination of neural crest, neural stem, and mesenchymal markers. Further, SKN cell lines can be reliably produced from mature dogs, display a high degree of line-to-line stability, are homogeneous across multiple phenotypic markers, and in vitro are highly neuronally committed. SKNs are therefore a promising candidate for neural cell therapeutic applications.
Embryologically, brain and skin originate from the same ectodermal lineage. Many adult tissues, including skin, continue to harbor populations of multipotent stem cells. Rather than being necessarily restricted to the tissue-specific lineages of origin, an increasing number of studies suggest that their differentiation potential is diverse and strongly influenced by environmental stimuli. For example, under defined culture conditions, dental pulp stem cells are capable of generating CNS neurons, yet are neural crest-derived and accordingly developmentally restricted to the PNS (Arthur et al., 2008). Similar plasticity is seen in dermal-derived precursor cells. When SKPs (closely related to SKNs) are isolated from neural crest-derived skin tissue they generate not only PNS neural cell types but also CNS neurons (De Kock et al., 2011), and even mesenchymal osteogenic cells (Lavoie et al., 2009). The converse is also possible: the dermal stem cell niche in ventral abdominal skin is non-neural crest in origin yet, despite this, SKPs derived from this niche are capable of generating neural crest-derived cell types (Jinno et al., 2010). SKNs in our study were isolated from ventral skin and therefore are likely to originate from the same non-neural crest-derived multipotent stem cells as ventral SKP cells.
Despite their similar origins, SKNs have several advantages over SKPs from a therapeutic perspective, including reliable homogeneous culture and strong neurogenic bias. Adherent culture is one potential reason for this increased uniformity, as observed in studies of neurogenic cells (Joannides et al., 2004, Conti et al., 2005, Pollard et al., 2006, Goffredo et al., 2008). In our case, uniformity was seen both within and across lines. We observed less than 2% coefficient of variance for three key phenotypic markers (CD133, NESTIN, and P75NTR) across multiple cell SKN lines and have, to date, produced more than 50 such lines from different dogs with a greater than 85% success rate. Furthermore, for several markers (P75NTR, CD133, and NESTIN), the level of positive expression was >97%, confirming that our protocol produces a reliably homogeneous SKN culture as required for clinical studies.
The pattern of SKN cell marker expression was unique, reflecting both their NPC-like properties (NESTIN and CD133) and dermal origins (PDGFRA and CD44). This pattern has some precedence, with human MSCs shown to spontaneously express neural linage markers SOX2, NESTIN, BETAIII-TUBULIN, and NEUROFILAMENT in vitro (Deng et al., 2006, Okolicsanyi et al., 2015). More generally, SKNs also exhibit upregulated and even unique expression of several genes associated with cell proliferation, cell adhesion, and cell migration. SKNs are hence an enriched population of neurogenic cells that may be well suited for survival and migration in vivo, a key factor that now requires empirical testing. Importantly, the absence of any genetic manipulation and observation that cell doubling tends to self-limit after four to five passages alongside downregulation of SOX2 further supports the safe use of these cells.
We found that SKNs are fate restricted by virtue of almost exclusive neuronal differentiation. Differentiated cells expressed a range of mature neuronal markers including synaptic and neurotransmitter-associated proteins. While these protein markers can also be found in PNS neurons, the absence of specific peripheral neuronal and glial proteins suggests that SKNs can mature in vitro into neurons compatible with the CNS. This neuronal fate bias may be linked to the endogenous upregulation of oligodendrocyte repressor gene, DLX2 (Petryniak et al., 2007), and spontaneous upregulation of several neural specification genes in response to a differentiation stimuli (high-dose BDNF). Indeed, these specification genes encode for the same four transcription factors (ASCL1, BRN2, MYTL1, and NEUROD) necessary for direct conversion of human somatic cells to iNs (Pang et al., 2011).
Differentiated SKNs also possessed basic neuronal functionality including voltage-gated ion channels and ATP-potentiated calcium transients, as seen in CNS neurons (Lin et al., 2007). However, mature action potentials remained elusive, suggesting that extended differentiation may be necessary for complete functional maturation into definitive neuronal subtypes. Indeed, embryonic stem cell-derived neural precursors have reportedly required up to 7 weeks to develop mature action potentials (Johnson et al., 2007). Nevertheless, global and focal gene expression data suggested that SKNs may possess a GABAergic and glutamatergic neuronal differentiation bias. This is in line with reports that similar adherent expansion of neural stem cells induces a bias toward GABAergic neurons (Goffredo et al., 2008), an intriguing finding that can be followed up in future work. In general, we hope that the strong neurogenic potential of canine SKN cells can be exploited to overcome one of the more pressing deficits in the cell replacement field: poor neuronal yields and an in vivo bias toward glial differentiation (Hofstetter et al., 2005, Blurton-Jones et al., 2009).
In summary, SKNs are capable of generating rich and uniform yields of mature neurons comparable with CNS neurons and may therefore have a role in replacement of lost brain cells. Having thoroughly characterized canine SKNs in vitro, we are now well placed to determine whether they are affected by disease-related signals and capable of generating structurally and functionally mature neurons in vivo.
Mature dogs up to 6 years of age had abdominal skin tissue (≈6 cm2) harvested following owner consent. Abdominal skin has been shown to harbor neurogenic stem cells (Toma et al., 2001, Fernandes et al., 2004, Valenzuela et al., 2008, Jinno et al., 2010) and is a practical region for future human clinical study as it is not frequently exposed to potentially mutagenic sunlight (UV radiation), and would represent a less-invasive harvest procedure for aged patients than alternatives such as facial skin (Fernandes et al., 2004). Canine tissue use was approved by the Animal Care and Ethics Committee of the University of New South Wales.
Skin biopsies were divided into 1–2 mm2 pieces and digested in 0.1% Trypsin (Thermo Fisher Scientific Australia Pty Ltd, Scorseby, VIC, Australia) for 40 min at 37°C, followed by 0.1% DNAse (Roche Applied Science, Castle Hill, NSW, Australia) for 1 min at room temperature. The tissue was then mechanically dissociated and passed through a 40 μm cell strainer (BD Bioscience, Sydney, NSW, Australia), and centrifuged at 180 × g for 5 min.
SKNs were then established using our published protocol (Valenzuela et al., 2008). In brief, the dissociated cells were then re-suspended in serum-free complete medium consisting of 3:1 DMEM/Ham's F-12 Nutrient Mixture (DMEM/F12; Thermo Fisher Scientific), 1% Penicillin Streptomycin (Thermo Fisher Scientific), 20 ng/mL EGF (BD Bioscience, Sydney, NSW, Australia), 40 ng/mL bFGF (Thermo Fisher Scientific), and 2% B-27 supplement (Thermo Fisher Scientific). The cells were then seeded at a density of 100,000 cells/cm2. When resulting neurospheres reached 50–100 μm in diameter, they were dissociated using TrypLE Select (Thermo Fisher Scientific) for 5 min at 37°C. Cells were then re-suspended in complete medium and grown as an adherent monolayer by seeding at 10,000 cells/cm2 on 0.1% gelatin-coated flasks, and passaged when 80% confluent.
For FBT culture, the dissociated cells were re-suspended in 3:1 DMEM/F12 medium containing 10% fetal bovine serum (Thermo Fisher Scientific), 1% GlutaMAX (Thermo Fisher Scientific), and 1% penicillin streptomycin. Hereafter referred to as FDMEM, and seeded at a density of 100,000 cells/cm2.
NPCs were derived from subventricular zone brain isolates, following previously published protocols (Duncan et al., 2016).
RNA from three individual donor lines each of SKNs, NPCs, and FBTs were subjected to genome-wide transcriptomic analysis using an Affymetrix Canine Genechip Gene 1.0 ST array (Affymetrix, Santa Clara, CA, USA). The complete microarray data are available online at https://www.ncbi.nlm.nih.gov/geo/ (accession number GEO: GSE74714).
Cells fixed with 4% paraformaldehyde were permeabilized in 0.5% Triton X-100 for 30 min, then blocked using 10% normal donkey serum (Sigma-Aldrich, Castle Hill, NSW, Australia) for 1 hr prior to overnight incubation in diluted primary antibodies. Alexo Fluor 488 or 594 secondary antibody (Thermo Fisher Scientific) incubation, and DAPI counterstaining was then carried out prior to imaging on a Leica DMI3000B Microscope, or quantification using a MACSQuant Analyzer flow cytometer (Miltenyi Biotec Pty Ltd, Sydney, NSW, Australia).
RNA was extracted using an illustra RNAspin Mini Isolation Kit (GE Healthcare, Little Chalfont, Buckinghamshire, UK) according to the manufacturer's instructions. cDNA was synthesized using Superscript III RT First Strand Synthesis System (Thermo Fisher Scientific). PCR amplification was then performed using Platinum Taq DNA Polymerase (Thermo Fisher Scientific). qRT-PCR was carried out using the CFX96 Real-Time PCR/C1000 thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) using SsoFast EvaGreen Supermix (Bio-Rad Laboratories). Relative gene expression was calculated using endogenous GAPDH housekeeping gene.
Whole-cell patch-clamp recordings were performed on differentiated SKN cells at room temperature (22.5°C ± 0.3°C). All membrane potentials were corrected for the liquid junction potential of 15.4 mV. Recordings were performed using an Axopatch-1D amplifier and digitized using pClamp 10.4 software and a Digidata 1440 interface (Molecular Devices, Sunnyvale, CA, USA) at 10 kHz, and filtered with an 8 pole Bessel filter at 2 kHz.
Fura2-AM dye (Thermo Fisher Scientific) was used as a ratiometric calcium indicator. Fluorescence intensity was recorded while perfused in a 5 mM KCl, 0 mM Ca2+ baseline solution, followed by 129 mM KCl, 2 mM Ca2+, then 129 mM KCl, 2 mM Ca2+, and 100 μM ATP solutions.
Repeat experiments were conducted on biologically independent cell lines. Statistical analysis was performed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA), analyzed by Student’s t test or ANOVA for multiple comparisons and considered significant at p < 0.05.
For further details of experimental procedures see Supplemental Experimental Procedures.
T.D. collected data, analyzed and interpreted results, and prepared the manuscript. A.L. and R.C.Y.L. designed the experiments, collected data, analyzed and interpreted results, and prepared the manuscript. K.S. and P.S. designed the experiments and analyzed and interpreted results. T.L., V.S., and M.V. designed the experiments, provided research material/equipment, and analyzed and interpreted results.
This work was supported by the Rebecca L. Cooper Medical Research Foundation, the ANZ Mason Foundation, the National Health and Medical Research Council (NHMRC) of Australia Program Grant (no. 568969), and philanthropic gifts from Ron Sinclair and Alzheimer's Australia. We appreciate the expert contributions of Dr Sandra Fok, Professor Mirjana Maletic-Savatic, and Professor Perry Bartlett.
Published: August 8, 2017
Supplemental Information includes Supplemental Experimental Procedures, seven figures, and three tables and can be found with this article online at http://dx.doi.org/10.1016/j.stemcr.2017.07.008.