Characterization and Therapeutic Potential of Human Amniotic Fluid Cells in Mediating Neuroprotection

Recently, human amniotic fluid (AF) cells have attracted a great deal of attention as an alternative cell source for transplantation and tissue engineering. AF contains a variety of cell types derived from fetal tissues, of which a small percentage is believed to represent stem cell sub-population(s). In contrast to human embryonic stem (ES) cells, AF cells are not subject to extensive legal or ethical considerations; nor are they limited by the lineage commitment characteristics of adult stem cells. However, to become therapeutically valuable, better protocols for the isolation of AF stem cell sub-populations need to be developed. This study was designed to examine the molecular components involved in selfrenewal, neural commitment and differentiation of AF cells obtained at different gestational ages. Our results showed that, although morphologically heterogeneous, AF cells derived from early gestational periods ubiquitously expressed KERATIN 8 (K8), suggesting that the majority of these cells may have an epithelial origin. In addition, AF cells expressed various components of NOTCH signaling (ligands, receptors and target genes), a pathway involved in stem cell maintenance, determination and differentiation. A sub-population of K8 positive cells (< 10%) co-expressed NESTIN, a marker detected in the neuroepithelium, neural stem cells and neural progenitors. Throughout the gestational periods, a much smaller AF cell sub-population (< 1%) expressed pluripotency markers, OCT4a, NANOG and SOX2, from which SOX2 positive AF cells could be isolated through single cell cloning. The SOX2 expressing AF clones had the capacity to give rise to a neuronal phenotype in culture, expressing neuronal markers such as MAP2, NFL and NSE. Taken together, our findings demonstrated the presence of fetal cells with stem cell characteristics in the amniotic fluid, highlighting the need for further research on their biology and clinical applications. Chapter 2 Probing stemness and neural commitment in AF cells 40 Introduction Stem cells can be derived from a variety of tissues during embryonic and fetal development as well as in adulthood; and hence may potentially be exploited for both tissue engineering and cell transplantation [1]. Pluripotent stem cells, derived from human embryonic stem (ES) cells, have the ability to give rise to all three embryonic germ layers, while maintaining their ability to proliferate indefinitely in culture. However, their application for therapy remains closely associated with ethical concerns related to both their origin and teratogenic potential [1-5]. Similarly, adult stem cells can be obtained from a number of tissues, but they generally display low proliferation rates and lineage restrictions [6-9]. Consequently, ongoing efforts continue to search for more practical sources of stem cells in humans. Recent studies have shown that human amniotic fluid (AF) contains a small sub-population of stem cells, enriched based on C-KIT expression, with self-renewal and differentiation potential [10]. However, despite the wide and well-established use of AF in routine prenatal diagnosis, the properties and origin of AF derived stem cells is poorly understood. It is believed that AF cells originate from several fetal tissues, including skin and the lining of the digestive, respiratory and urinary systems [11-13]. In addition, the nature and characteristics of AF cells may vary with gestational age. Thus, we have used several samples from the second and third trimesters of gestation to provide a comprehensive molecular characterization of AF cells, with regard to the genes involved in stem cell maintenance, neural fate commitment and differentiation, with an underlying focus on KERATIN expression and NOTCH signaling. Chapter 2 Probing stemness and neural commitment in AF cells 41 The evolutionarily conserved NOTCH signaling pathway plays diverse roles in cell-fate specification during embryogenesis and in adult tissues [14]. NOTCH comprises a family of highly conserved receptors, whose activation is induced by their specific ligands, DELTALIKE (DLL) and JAGGED (JAG), through cell–cell interactions [14]. Once activated, the NOTCH intracellular domain (NICD) is cleaved by -secretase, which leads to translocation of the NICD into the nucleus. Subsequently, NICD is associated with the transcription factor recombination signal binding protein for immunoglobulin kappa J region (RBP-J) to generate a transactivation complex, which initiates transcription of target genes such as the hairy/enhancer of split (HES) transcriptional repressors. Of particular interest is the fact that NOTCH signaling regulates stem cells in many different settings, including the skin and nervous system, among others (reviewed in [15]). During fetal skin development, the epidermis changes from a simple epithelium covered by periderm to stratified epithelium and finally to keratinized epidermis [16, 17], suggesting that epithelial cells likely contribute significantly to the cellular composition of the AF. During development, different compartments of the epidermis containing stem cell sub-populations are established that are maintained into adulthood [18]. Since NOTCH signaling has been implicated in both epithelial and epidermal stem cells [18, 19], we sought to determine whether we could identify the expression of keratin filament proteins in AF culture. We examined the expression of keratin intermediate filament protein, KERATIN 8 (K8), a marker of simple epithelia [20-22], since it is the first keratin to appear in embryogenesis, as early as in preimplantation embryos [23] and has been shown to be expressed in undifferentiated human ES cells [24, 25] and mouse epiblast stem cells [25, 26]. Lastly, we determined whether AF cells express the conventional markers of pluripotency: OCT4a, SOX2 and NANOG, which Chapter 2 Probing stemness and neural commitment in AF cells 42 could be exploited as a method of enriching for stem cell sub-populations by clonal analysis. A better understanding of the cellular and molecular profile of AF cells will provide a basis for the generation of novel lineages of primary human cell lines that are suitable for cellbased therapies. Chapter 2 Probing stemness and neural commitment in AF cells 43 Materials and Methods Cell Culture Samples of amniotic fluid were obtained from the Ottawa Hospital (Ottawa, Ontario, Canada) following routine amniocentesis carried out on pregnant women at 15 to 35 (AF15 AF35) weeks of gestation. All the procedures were performed following the guidelines established by the Ottawa Hospital and National Research Council Canada Research Ethics Boards; a written consent was obtained from each woman to use the amniotic fluid for both genetic analysis and research purposes. Genetic analysis revealed normal karyotypes and the pregnancies resulted in successful births. Shortly after amniocentesis, amniotic fluids were diluted with Phosphate Buffered Saline (PBS) and spun at 200 x g for 5 mins; pellets were re-suspended in the expansion media composed of Dulbecco’s Modified Eagle Medium (DMEM) (Invitrogen) supplemented with 20% inactivated fetal bovine serum (FBS, Hyclone). The cells isolated from each AF sample were plated in a 10 cm culture dish (Corning) and expanded. Cells were passaged by trypsinization (0.25% Trypsin, 0.1% EDTA, Wisent) and expanded serially with a split ratio of 1:3 for 5 -10 passages at 70% confluency and used for cellular and molecular analyses. AF cells were maintained in a humidified incubator under 5% CO2 at 37°C. The viability of AF cells was routinely checked with trypan blue and cell survival dye, carboxyfluorescein diacetate (CFDA). NT2/D1 (NT2) embryonal carcinoma (EC) cells (ATCC) were seeded at a density of 2 x 10 cells in HG-DMEM (Invitrogen) and 10% FBS (Hyclone) replenished every 2 days [27]. Human SH-SY5Y cells (ATCC) and E13 embryonic cortical cultures were prepared, as previously described [28, 29]. Chapter 2 Probing stemness and neural commitment in AF cells 44 To generate AF-derived single cell clones, a single cell suspension was prepared by gentle trypsinization and individual AF cells were deposited, by Fluorescence Activated Cell Sorting (FACS), with a cell density of 1 cell per well of a 96-well flat bottom plate (Nunc) in 100 μl of expansion medium. Once the cultures became 70% confluent, clones were subcultured first into 24-well plates and then, 6-well plates (Nunc) and eventually into 10 cm culture plates (Corning), using the above-mentioned conditions (Figure 8A). The clonal AF cells, were thereafter expanded serially with a split ratio of 1:3 and cultured in DMEM containing 0.5% FBS and N2 supplement for up to one week to generate neurons. RNA Extraction and RT/QPCR Total RNA was extracted from cell, using TriReagent (Molecular Research Centre), as per manufacturer’s instructions. The RNA was quantified with NanoDrop (Thermo Fisher Scientific), with a final A260/A280 ratio of 1.9-2.0. One microgram of total RNA was reverse transcribed, using Quantitect Reverse Transcriptase (Qiagen) for cDNA preparation. A no reverse transcriptase control was performed in parallel. PCR amplifications were carried out on a BIORAD DYNA Cycler with a 25 μl volume containing: 1X PCR Reaction Buffer, 200 μM dNTP mixture, 1.5 mM MgCl2, 0.05 units Taq polymerase, 5 μM sense and antisense oligonucleotide primers and 10 ng of cDNA. The PCR program consisted of a denaturing step for 3 mins at 94°C, followed by 30 secs at 94°C, 30 secs at 56-70°C and 30 secs at 72°C for 40 cycles. The PCR final extension period was 5 mins at 72°C. The reaction products were separated on a 2% agarose gel containing 0.2 μg/ml of ethidium bromide and visualized on a UV transilluminator. The gel images were captured with a FluorChem 8900 Imager (Alpha Innotech) and the amplicon size was determined by comparison with a 1 kb Chapter 2 Probing stemness and neural commitment in AF cells 45 Plus DNA Ladder (Invitrogen). β-ACTIN (ACTB), NT2 and SH-SY5Y cells were used as a normalizing housekeeping gene and positive controls for RT/QPCR expression analysis, respectively. Quantitative PCR (qPCR) was performed using Fast SYBR Green Master Mix (Bio-Rad) on 7500 Fast Real-Time PCR System (Applied Biosystems). Approximately, 10 ng of cDNA was used per individual reaction with primer concentrations of 5 μM. Quantitative PCR amplifications were performed using the following conditions: Initial denaturation at 94°C, 20 secs and 40 cycles at 94°C, 5 secs; 60°C, 30 secs. All data was normalized to ACTB, using the ΔΔCT method. To identify the expression of self-renewal, neural commitment and differentiation genes, specific primers were designed using publicly available Primer3 software [30] Chapter 2 Probing stemness and neural commitment in AF cells 46 Table 1. Sequence and annealing temperatures of primers for RT/QPCR Designation Sequence (5’-3’) Annealing Temp. (°C) Amplicon Size (bp) Self-renewal SOX2 F GGAGCTTTGCAGGAAGTTTG 60 460 SOX2 R GGAAAGTTGGGATCGAACAA qSOX2 F CACTGCCCCTCTCACACATG 60 81 qSOX2 R TCCCATTTCCCTCGTTTTTCT OCT4a F CTTCGGATTTCGCCTTCTCG 60 412 OCT4a R TGCAGAGCTTTGATGTCCTG qOCT4a F CTGAGGGCGAAGCAGGAGTC 60 170 qOCT4a R CTTGGCAAATTGCTCGAGTT OCT4b F CTTGCTGCAGAAGTGGGTGGAGGAA 70 169 OCT4b R CTGCAGTGTGGGTTTCGGCA qOCT4b F GCTCGAGAAGGATGTGGTCC 60 81 qOCT4b R CGTTGTGCATAGTCGCTGCT NANOG F GACTGAGCTGGTTGCCTCAT 56 276 NANOG R TTTCTTCAGGCCCACAAATC qNANOG F GCAGAAGGCCTCAGCACCTA 60 81 qNANOG R AGGTTCCCAGTCGGGTTCA Notch Signaling Pathway NOTCH1 F CACCCAGAACTGCGTGCA 58 726 NOTCH1 R GGCAGTCAAAGCCGTCGA NOTCH2 F GCTGATGCTGCCAAGCGT 58 475 NOTCH2 R CCGGGGAAGACGATCCAT DLL1 F GTCGACTCCTTCAGTCTGCC 60 413 DLL1 R CAGATCGGCTCTGTGCAGTA DLL3 F GACCCTCAGCGCTACCTTTT 60 249 DLL3 R TACATCTTCAGGGCGATTCC DLL4 F ACCGACCTCTCCACAGACAC 60 484 DLL4 R CGCTGATATCCGACACTCTG HES1 F AAAATGCCAGCTGATATAATGGAG 60 314 HES1 R GGTCTGTGCTCAGCGCAGCCGTCA HES5 F CAGAGTCCCTGCCGTTTTAG 60 265 HES5 R GAGTTCGGCCTTCACAAAAG JAG1 F ACACACCTGAAGGGGTGCGGTATA 60 608 JAG1 R AGGGCTGCAGTCATTGGTATTCTGA JAG2 F TGAACGGGTACCAGTGTGTG 60 275 JAG2 R TCTTGCCACCAAAGTCATCA Neural Markers NES F GCGTTGGAACAGAGGTTGGA 58 327 NES R TGGGAGCAAAGATCCAAGAC PAX6 F CAATCAAAACGTGTCCAACG 60 431 Chapter 2 Probing stemness and neural commitment in AF cells 47 PAX6 R TGGTATTCTCTCCCCCTCCT ASCL1 F GTCGAGTACATCCGCGCGCTG 60 220 ASCL1 R AGAACCAGTTGGTGAAGTCGA CD133 F CAGTCTGACCAGCGTGAAAA 60 200 CD133 R GGCCATCCAAATCTGTCCTA NSE F’ GAGCGGGCAGTGGAAGAAAAGG 56 292 NSE F’ GTCCGGCAAAGCGAGCTTCAT NFL F” TCCTACTACACCAGCCATGT 56 284 NFL R” TCCCCAGCACCTTCAACTTT MAP2 F TCAGAGGCAATGACCTTACC 56 321 MAP2 R GTGGTAGGCTCTTGGTCTTT β-ACTIN F TCACCCACACTGTGCCCATCTACGA 60 295 β-ACTIN R CAGCGGAACCGCTCATTGCCAATGG * q designates primers used for QPCR ’ designates primers from Van Buren et al., Clin Can Res 2007: 13(16) 4704-4712; ” primers from Lee et al., Plos One 2009: 4(5) e5586, Antibodies The following antibodies were used in this study: SOX2 (1:300, ICC, in house), OCT3/4 (1:200, WB, Santa Cruz), CD133 (1:200, WB, Santa Cruz), C-KIT (1:50, ICC, Santa Cruz), β-ACTIN (1:5000, WB, Sigma), PAX6 (1:1000, WB, ICC, Santa Cruz), JAG1 (1:100 ICC, Santa Cruz), RBP-J (1:100, ICC, Cosmo Bio Co., LTD.), MAP2a+2b (1:200, ICC, Sigma), Hoechst (1:1000, ICC), NOTCH1 (1:100, ICC, Upstate Biotech), NFL (1:200, ICC, Chemicon), Keratin 8 (1:2, Developmental Studies Hybridoma Bank), horseradish peroxidase (HRP) anti-rabbit and anti-mouse (1:5000, WB, Jackson Immunoresearch Laboratories) and immunocytochemistry fluorescence-conjugated (Alexa Fluor 488 antirabbit, goat or rat) secondary antibodies were used at a dilution of 1:500 (Molecular Probes, Invitrogen). Chapter 2 Probing stemness and neural commitment in AF cells 48 Protein Extraction and Western Blot Analysis Cells were washed with ice-cold TBS (25 mM Tris-HCL, pH 7.5, 150 mM NaCl) and lysed directly in the tissue culture plate on ice, in ice-cold lysis buffer (25 mM Tris-HCL, pH 7.6, 150 mM NaCl, 1% Triton-X, 1% Na.Deoxycholate) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and a protease inhibitor cocktail (Roche Diagnostics), as previously described [29]. Cell lysates were incubated on ice for 30 mins and clarified by centrifugation at 16 000 x g at 4°C for 15 mins. Total protein concentrations were determined, using the Bio-Rad Protein Assay (Bio-Rad). The supernatant aliquots containing equal amounts of total protein (40 μg) were diluted 1:2 with Laemmli sample buffer (Bio-Rad) for a total loading volume of 20 μl, run on a 12% SDS-PAGE (90 V for approximately 3 hrs and transferred to a nitrocellulose membrane (Amersham), using a wet transfer apparatus (Bio-Rad) at 20 V overnight at 4°C. Membranes were washed once with TBST (TBS containing 0.05% Tween-20, Sigma) and blocked with 5% skim milk in TBST for 2 hrs at room temperature. The membrane was incubated with primary antibodies (see the Antibodies section for specifics) overnight at 4°C with gentle rocking. Β-ACTIN (ACTB) (Bio-Rad, 1:5000) was used as a loading control. Membranes were washed five times for 5 mins per wash in TBST. All secondary antibodies, anti-rabbit and anti-mouse IgG-HRP conjugates (Bio-Rad, 1:5000), were diluted in 5% skim milk and the membranes were incubated for 1 hr at room temperature. Membranes were washed five times for five minutes per wash in TBST and consequently detected by the enhanced chemiluminescence system (New England Nuclear). The ECL signal was analyzed by FluorChem 8900 (Alpha Innotech). Rainbow markers (Amersham) were used to determine the protein sizes. Chapter 2 Probing stemness and neural commitment in AF cells 49 Immunocytochemistry Cells were rinsed three times with PBS, fixed with 65% ethanol containing 0.15 M NaCl for 20 mins and blocked with Protein Block Serum Free (Dako) for 30 mins. Primary antibodies were diluted in Antibody Diluent (Dako) and cells were incubated with the primary antibody for 1 hr at room temperature. Following washes with PBS, cells were incubated with the secondary antibody for 1 hr, washed and covered with the Vectashield mounting medium (Vector Laboratories). The immunoreactivity was examined under an Axiovert 200 M fluorescence microscope (Zeiss). Pictures were taken with AxioCam and Axiovision 4.7 (Zeiss) and processed with Adobe PhotoShop (Adobe Systems Incorporated). SOX2, OCT4a or KERATIN positive cells were quantified against the total number of Hoechst positive cells, using at least 15 random images at 20X magnification. The majority of images shown are representative of AF26, unless otherwise specified. DNA transfection and virus production The promoter reporter SOX2 construct (pAP2-PrSox2-EGFP) was generated by sub-cloning the SmaI-BamHI full-length Sox2 promoter fragment (-5255 to +279 CDS), derived from the p-PrSox2-EGFGP-IRES-PURO plasmid (a generous gift from Dr. Fred Gage, Salk Institute, La Jolla, CA, USA), into the pAP2-EGFP retroviral vector [31]. Briefly, 293 GPG cells [32] were plated the night before on a 10 cm dish in 5 ml DMEM supplemented with 10% inactivated FBS, 1 μg/ml tetracycline (Sigma), 2 μg/ml puromycin (Sigma) and 300 μg/ml G418 (Gibco). Twenty micrograms of the pAP2-PrSox2-EGFP was diluted in 500 μl of serum free medium (DMEM) and incubated at room temperature for 7 mins with 30 μl of Lipofectamine2000 (Invitrogen) also previously diluted into 500 μl of DMEM. The DNAChapter 2 Probing stemness and neural commitment in AF cells 50 Lipofectamine complex was added drop-wise to 293 GPG cells at 70% confluency. Medium was replaced 7 hrs post-transfection and the transfection efficiency was confirmed based on GFP expression 24 hrs later, using fluorescence microscopy (Zeiss). Virus conditioned medium was collected and replenished daily for a span of three days. The virus-conditioned medium was filtered through a 45 μm low protein binding filter (Millipore) and concentrated, using an Ultra Centrifugal Device (Ambion) [29], according to manufacturer’s instructions. The concentrated virus was either used directly to infect AF cells or stored at 80°C. AF cells were seeded the night before and infected at 70% confluence three times within 24 hrs, with 200 μl of concentrated virus and 8 μg/ml of polybrene in 1 ml of DMEM media. Three days following infection, the medium was replaced with DMEM, 0.5% FBS and N2 supplement to induce neuronal differentiation. In parallel experiments, the pAP2PrSox2-EGFP was also directly electroporated into AF cells, using the Neon Transfection System (Invitrogen). Briefly, 1 μg of pAP2-PrSox2-EGFP was diluted to a total volume of 10 μl Re-suspension Buffer E containing 1.0 X 10 dissociated AF cells. The Neon pipette containing the cell suspension and plasmid was electroporated inside the Neon 10 μl tip, using the following conditions: Voltage: 1350 V, Pulse Width: 15 ms and Pulse Number: 3. Following electroporation, the cells were seeded into a well of a 24-well plate containing 500 μl of DMEM + 20% FBS. Three days following transfection, the medium was replaced to induce neuronal differentiation. Flow Cytometry Analysis AF cells were washed with PBS, trypsinized and resuspended in 2 ml of cold 2% FBS/PBS. Dissociated single cells were stained with C-KIT antibody (Santa Cruz sc-5535) for 30 mins Chapter 2 Probing stemness and neural commitment in AF cells 51 at 4°C. Following the incubation period, the cells were washed twice with cold 2% FBS/PBS and incubated for 30 mins at 4°C with a secondary phycoerythrin (PE)-conjugated antibody. The cells were subsequently washed, resuspended in 2 ml of cold 2% FBS/PBS, filtered through a 70 μm filter and analyzed using a MoFlo Cell Sorter (Beckman Coulter). Karyotypic Analysis For karyotypic integrity of clonal AF-F5 cells, the cells were incubated for 4 hrs in medium containing 0.1 μg/ml Karyomax Colcemid solution (Invitrogen), harvested by trypsinization and spun at 175 x g for 5 mins at room temperature. The cells were incubated for 20 mins at room temperature in 10 ml of hypotonic solution (0.075 M KCl) and fixed with fresh methanol/acetic acid fixative (3:1 vol/vol). The fixed cell suspension was dropped onto a clean glass slide. The slide was subsequently incubated with Hoechst for 10 mins and mounted with Vectashield (Vector). Pictures of the metaphase spreads were taken with a 40X objective (Axiovert Zeiss). Statistical Analysis Results are analyzed using a one-way analysis of variance (ANOVA), followed by Newman-Keuls post-hoc test. Results are expressed as mean + standard error of mean (SEM) and considered significant at p <0.05. Chapter 2 Probing stemness and neural commitment in AF cells 52 Results Early AF cells express abundant levels of KERATIN 8 Based on their morphology and growth characteristics, AF cells can be classified into three cell types: epithelial-like, amniotic fluid-specific and fibroblast-like [33]. Since the composition and cellular and molecular properties of AF cells may vary with gestational age, we examined AF-derived cells ranging from 15 to 35 weeks of gestation. The cells isolated from the AF were plated in culture-grade dishes to reproducibly establish primary cultures. Ki67, a well-characterized cell growth marker, that excludes the resting (G0) phase, was used along with Hoechst staining to show proliferation (Figure 1A) and mitogenic configuration (Figure 1B, arrowheads); respectively. Furthermore, cell survival dye, CFDA, was used to ensure that AF cells maintain a viable and metabolically active state throughout culturing (Figure 1C). The doubling time of the established primary cultures was 21.5 hrs, with approximately 500 doublings observed. Given that fetal epithelial cells contribute largely to the composition of AF during gestation, we sought to investigate whether K8, a key marker of simple epithelia and early embryonic stages [23], is expressed in AF cells. Immunocytochemistry showed fibrillary cytoplasmic staining for K8 in AF cells (Figure 2A, bottom panel). Furthermore, K8 positive cells were abundant (over 97%) in early gestation periods (AF 16 and 26); whereas, they were significantly decreased (approximately 75% reduction) at later gestational periods (AF28 and AF35; Figure 2B), suggesting that the majority of earlier gestation cells in the AF are epithelial in origin. This notion is further established by expression of K7, K18 and K19 in AF cells (Supplementary Figure 1). Chapter 2 Probing stemness and neural commitment in AF cells 53 Figure 1. The proliferative activity of AF cells A. AF cells readily expressed Ki-67 (red), a marker commonly used for proliferation analysis. Ki-67 was detected at both interphase (arrowheads) and mitotic (arrow) stages. B. Similarly, Hoechst (blue) staining demonstrated proliferation in AF culture. Several mitotic figures (arrowheads) were present in any random field of view. C. Cell survival dye, CFDA (green), reveals that AF cells are viable and metabolically active throughout culturing. Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 54 Figure 2. Expression of KERATIN 8 in AF cells at different gestational periods A. Immunocytochemistry of AF cells from 16, 26, 28 and 35 weeks of gestation (AF16 – AF35) showed significantly higher KERATIN 8 (K8, green) positive cells in AF16 and AF26 cultures compared with AF28 and AF35 cultures. Hoechst (blue) staining was used to determine the total number of cells. B. Statistical analysis confirmed that more than 95% of AF16 and AF26 cells were K8; whereas, there was a significant reduction in expression of K8 in AF28 and AF35 cultures. Data (mean + SEM, n=3, *** p<0.0001 vs AF28/35) was obtained from 15 independent fields of view. Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 55 AF cells contain components of the Notch signaling pathway Since NOTCH signaling has been shown to play multiple roles in the regulation of epidermal development [34] and since several NOTCH receptors (NOTCH1-NOTCH4), ligands (DELTA-LIKE (DLL1 and DLL4) and JAGGED (JAG1 and JAG2) and effectors (HES1 and HES5, Figure 3A), are expressed in the developing skin [35, 36], we examined these components of the NOTCH signaling pathway to gain a better understanding of its critical role in AF cells. Our RT-PCR analysis detected multiple NOTCH receptor and ligand transcripts expressed in AF samples ranging from 15 to 35 weeks of gestation (Figure 3B). NOTCH1, NOTCH2 and JAG1 were all ubiquitously expressed and HES1 appears to be the effector gene uniformly expressed in all gestation samples examined, by comparison to HES5. Immunocytochemistry further highlighted the ubiquitous expression of NOTCH1, JAG1 and the presence of DLL1 and DLL4 in sub-populations of AF cells (Figure 3C). Ubiquitous RBP-J expression, a key mediator of NOTCH signaling, was also observed in the AF cultures (Figure 3C). When associated with some NOTCH proteins, RBP-J acts as a transcriptional activator that activates transcription of NOTCH target genes, such as HES1. The presence of the various components of NOTCH signaling suggests that AF cells can be used as a human cell model to study complex signaling pathways. AF cells express genes involved in neural cell fate lineage During fetal development, expression of NOTCH1 or its downstream regulators, such as HES1, inhibits neuronal differentiation and results in the maintenance of a progenitor state, as a precursor to neural differentiation. Hence, we used RT-PCR to examine the Chapter 2 Probing stemness and neural commitment in AF cells 56 Figure 3. Characterization of NOTCH signaling in AF cells A. Schematic diagram of the NOTCH signaling pathway. Upon binding of NOTCH ligands (JAG1, JAG2 and DLL1, DLL3, DLL4), NOTCH receptors (NOTCH1 and NOTCH2) are cleaved and the NOTCH intracellular domain (NOTCH1/2 ICD) translocates to the nucleus to associate with an RBP-J transcription factor to regulate gene expression of target genes, HES1 or HES5. B. RT-PCR analysis of the mRNA harvested from AF cells at 15 to 35 weeks (wks) of gestation for genes involved in the NOTCH signaling pathway. AF cells displayed a ubiquitous expression of NOTCH1 and NOTCH2, JAG1 and HES1, regardless of their gestational age. The other ligands (DLL1, DLL3 and JAG2) and HES5 appeared to be present only in some of the samples. β-ACTIN (ACTB) and undifferentiated human NT2 cells were used as internal and positive controls, respectively. NTC, No Template Control. C. Representative immunocytochemistry further confirming the presence of NOTCH1, DLL1, DLL4, JAG1 and RBP-J in AF cells. Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 57 Figure 4. Expression of neural markers in AF cells A. RT-PCR analysis of mRNA harvested from AF cells at 15 to 35 weeks (wks) of gestation for early neural markers. AF cells expressed low levels of PROM1 (CD133), NESTIN (NES), PAX6 and ASCL (MASH). β-ACTIN (ACTB) and undifferentiated human NT2 cells were used as internal and positive controls, respectively. NTC, No Template Control. B-C. Western blot analyses confirmed that both PROM1 and PAX6 proteins were expressed in AF cells. NT2 cells and ACTB were used as positive and internal controls, respectively. Chapter 2 Probing stemness and neural commitment in AF cells 58 expression of common early neural markers. Our data showed that AF cells ubiquitously express low levels of reported early neural genes: NESTIN (NES), PAX6, ASCL (MASH-1), along with PROMININ-1 (PROM1, also known as CD133) [37] (Figure 4A). Western blotting further confirmed that AF cells express PROM1 and PAX6 (Figure 4B-4C). These findings suggest that a distinct cell population, within the amniotic fluid cultures, may harbour progenitor cells with the potential to differentiate along the neural lineage, under the appropriate conditions. Since both RT-PCR and Western blot rely on cell lysates from the heterogeneous cell population, we further determined more specifically what percentage of K8 positive cells were also positive for NES expression, a marker of neuroepithelial stem cells (Figure 5A-5C). Our double staining data showed that NES is co-expressed with K8 in approximately 10% of the AF cells in early gestational periods (Figure 5D). The subcellular analysis of K8 and NES confirmed distinct localization of these intermediate filaments within the cytoplasm (Figure 5C). Small sub-populations of AF cells express pluripotency markers Although NESTIN has been used as a marker of neuroepithelial and neural stem cells, it has also been shown to be expressed in early progenitor cells in other tissues such as limbal epithelia and the bulge region of the hair follicle [38-40]. Therefore, we employed a number of pluripotency markers, SOX2, OCT4 and NANOG, to determine whether AF cells express the genes known to be critical for self-renewal of embryonic stem cells. We observed very low transcript levels of SOX2 and OCT4a expression by QPCR (Figure 6A), which fell outside of the linear range, reflecting the small subset of AF cells positive for nuclear SOX2 (approximately 1%) and OCT4a expression, using immunocytochemistry (Figure 6B). Chapter 2 Probing stemness and neural commitment in AF cells 59 Figure 5. Co-expression of NESTIN and KERATIN 8 in AF cells A. Representative immunocytochemistry of AF cells from 26 weeks of gestation showing that a sub-population of K8 positive cells (green) also expressed NESTIN (red) with Hoechst nuclear counterstain (blue). B. The corresponding phase contrast image of A. C. Higher magnification detailing co-expression of K8 and NES immunofluorescence (yellow) in AF cells. D. Statistical analysis showed that there was a significant difference in the percentage of cells positive for KERATIN 8 (K8) and both KERATIN 8 and NESTIN (K8NES) at different gestational ages (AF16 – AF35). Data (mean + SEM, n=3, ** p<0.0001 vs K8 AF28/35) were obtained from 15 independent fields of view. Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 60 Notably, based on a recent publication by Liedtke et al. [41], we discriminated the expression of OCT4 by its two known isoforms: OCT4a and OCT4b, only the former of which is ascribed as a pluripotency marker [41]. To distinguish OCT4a from OCT4b, RT/QPCR primers for OCT4a were designed to flank Exon1 and Exon2, whereas those for OCT4b were designed to flank Exon3 and Exon4 (lacking Exon1). Functionally, the two OCT4 human isoforms show different expression patterns and differ in their ability to confer self-renewal [42, 43]. In accordance, we used an OCT4 antibody (SC-5279), which recognizes a single epitope localized at amino acid 1-134 of the OCT4 protein, which excludes the concomitant recognition of the OCT4b isoform [41]. Approximately 0.1% of AF cells expressed OCT4a (Figure 6B). Similar to the levels observed for SOX2 and OCT4a, we found very low NANOG expression by QPCR analysis (Figure 6A). NT2 cells were positive for the expression of all genes examined, characteristic of an embryonal carcinoma cell line. Although very few AF cells express SOX2, OCT4a and NANOG, we were able to enrich for the expression of these genes by purifying a population of cells that expressed the cell surface antigen C-KIT (CD117), the receptor for stem cell factor (SCF) [44], using fluorescence activated cell sorting (FACS), as previously reported [10]. A total of 1.05% ± 0.3% of AF cells were C-KIT positive, resulting in approximately a 1-3 fold increase in gene expression of the SOX2, OCT4a and NANOG, compared to C-KIT negative cells (see Supplementary Figure 2A-D). SOX2 expressing AF cells have the capacity to differentiate into neuronal cells Since AF cultures contain a few cells that express SOX2, we infected AF cells with pAP2PrSox2-EGFP, a retroviral vector that contains the SOX2 promoter upstream of the EGFP Chapter 2 Probing stemness and neural commitment in AF cells 61 Figure 6. Presence of pluripotent stem cell markers in a small subset of AF cells A. QPCR analysis of AF cells ranging between 15 to 35 weeks of gestation (AF15 – AF35) revealed low levels of transcripts for SOX2, OCT4a, OCT4b and NANOG. Fold expression levels were normalized against β-ACTIN (ACTB) (mean + SEM, n=3). Undifferentiated NT2 cells were used as positive control. B. Immunocytochemistry confirmed that only a small number of AF cells express SOX2 and OCT4, compared to undifferentiated NT2 cells. Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 62 reporter gene (Supplementary Figure 3A) to identify promoter activity and hence SOX2 expression. Following retroviral infection, we monitored the fate of a sub-population of SOX2-EGFP positive cells, mainly found in clusters (Supplementary Figure 3B). Since SOX2 expression is found in neural stem cells, each cluster was carefully marked after which cultures were treated with DMEM, 0.5% FBS and N2 supplement for up to one week to induce neuronal differentiation. Neuronal morphology was monitored within each cluster, using live fluoresence and Hoffman modulation contrast imaging, followed by MAP2 immunocytochemistry (Figure 7E-7F). In contrast, such a sub-population of cells was not observed in either uninfected control cultures (Figure 7A-7B) or those infected with EGFP retrovirus (Figure 7C-7D). Further analysis by both RT-PCR and immunocytochemistry revealed that these SOX2-positive clusters were also positive for other neuronal specific markers, neurofilament light chain (NFL)/NF68 and neuron-specific enolase (NSE) (Figure 7G and 7I, also see Supplementary Figure 3C). Since AF26 cells displayed a high proliferative potential, with a doubling time of 21.5 hrs, expressed many of the markers involved in stem cell maintenance, neural commitment and differentiation, we used AF26 cells to generate single cell derived clones to enrich for SOX2 expression (Figure 8A). We characterized the top three clones AF-C2, AF-F5 and C-KIT positive AF-C12 based on the expression of SOX2, OCT4a, NANOG and NESTIN (Figure 8B). AF-F5 and C-KIT positive AF-C12 clones expressed SOX2, NANOG and NESTIN to a higher level that AF-C2 clones. Given the heterogeneity of AF cells, the generation of a homogenous cell population, through single cell cloning, is important in the interpretation of molecular analyses and functional tests. Specifically, of the SOX2 expressing clones, AF-F5 Chapter 2 Probing stemness and neural commitment in AF cells 63 Chapter 2 Probing stemness and neural commitment in AF cells 64 Figure 7. Neuronal differentiation of SOX2 positive AF cells A and C. Immunocytochemistry showed no MAP2 positive cells in either uninfected (A) or pAP2-EGFP infected (C) AF cultures. E. Clusters of AF cells infected with pAP2-PrSox2EGFP differentiated into MAP2 positive neuronal cells. B, D and F. Phase contrast images of A, C and E; respectively. Only cells in panels E/F displayed neuronal morphological features. G. RT-PCR analysis of the prSox2-EGFP AF26 cells in neuronal induction media revealed the presence of transcripts for NFL, MAP2 and NSE. SH-SY5Y cells and ACTB were used as positive and internal controls; respectively. H, I. Immunocytochemistry confirmed the expression of NFL (red) in mouse embryonic cortical cultures (H) and AF infected cultures (I) with Hoechst nuclear counterstain (blue). Scale bar: 50 μm. Chapter 2 Probing stemness and neural commitment in AF cells 65 Figure 8. Generation and karyotypic integrity of AF derived single cell clones A. AF cells were harvested from donors and expanded in culture in 10 cm plates. For the generation of single cell clones, a single cell suspension was prepared and individual AF cells were deposited, by Fluorescence Activated Cell Sorting (FACS), with a cell density of one cell per well of a 96-well plate. Resulting clones were sub-cultured first into 24-well plates and thereafter expanded serially to 10 cm plates and subsequently used for a number of cellular and molecular purposes. B. RT-PCR analysis of the top three clones: AF-C2, AFF5 and C-KIT positive AF-C12 for the expression of SOX2, OCT4a, NANOG and NESTIN. C. AF-F5 clones were shown to preserve their karyotypic integrity after multiple passages in culture. Chapter 2 Probing stemness and neural commitment in AF cells 66 cells were able to survive multiple freeze/thaw cycles and SOX2 expression was found in approximately 70% of the cells, in contrast to the cells found in the heterogenic culture (Figure 9A-9B). The SOX2 expression in the clones was validated by Western blotting and immunocytochemistry (Figure 9C-9D) and represents a clonal cell line that can be used to examine the neural differentiation potential of SOX2 positive AF cells for future cell-based therapies. Accordingly, we used AF-F5 cells in subsequent studies in Chapter 2 and 3 to evaluate the neuroprotective potential of AF cells in the injured brain. Importantly, AF-F5 clones were able to preserve their karyotypic integrity after multiple passages (500 doublings) and freeze/thaw cycles in culture (Figure 8C). Chapter 2 Probing stemness and neural commitment in AF cells 67 Figure 9. SOX2 enriched AF single cell clones A. SOX2 was not detected in AF cells, using conventional RT-PCR analysis. B. In contrast, SOX2 was readily detected in the single cell clones obtained from AF-F5 (AF) cells. βACTIN (ACTB) and undifferentiated human NT2 cells were used as internal and positive controls, respectively. NTC, No Template Control. C. Immunocytochemistry verified the expression of SOX2 protein in the same sub-population. Right panel: Phase contrast image of the SOX2 positive AF clone. D. Similarly, western blot analysis demonstrated the presence of SOX2 in the AF single cell clone. Undifferentiated NT2 cells: positive control. ACTB: loading control. Scale bar: 25 μm. Chapter 2 Probing stemness and neural commitment in AF cells 68 Discussion Human AF cells harbour a high proliferative capacity, without forming teratomas in vivo [45], a typical problem associated with human ES cells. By examining their growth phase, we found that AF cells have a doubling time (21.5 hrs) similar to that of AF membrane epithelial cells (18 -24 hrs, depending on passage number [46]) and shorter than that observed for human ES cells (32 hrs [47]) and bone marrow-derived mesenchymal stromal cells (36-72 hrs [48]). However, a major underlying issue remains surrounding their heterogeneity, based on AF composition, which varies with gestation and hence the source of putative stem cell sub-populations remains unknown [49, 50]. This necessitates the generation of clonally-derived AF cell populations for future studies. Despite the heterogeneity of AF-derived cells, we found that the majority of the cells obtained from earlier gestations (AF16-26) were positive for K7, K8, K18 and K19. The presence of these early type I (K18 and K19) and type II (K8 and K7) heteropolymeric keratin pairs suggests that AF cells mostly resemble simple epithelial cells [20-22]. The epithelium includes continuous sheets of tightly linked cells that constitute the surfaces (such as epidermis and corneal epithelium) and linings (such as the digestive, respiratory and uro-genital epithelia) of the body [16, 19], and are readily shed into the AF by the developing fetus [49]. Interestingly, the expression of K8 and K18 pair has been recently shown to be characteristic of the epithelial nature of undifferentiated human ES cells and epiblast stem cells [24, 25]. Thus, since K8 is preferentially expressed in early gestational periods, putative stem cell sub-populations found within the native AF may originate either from the fetal skin, urothelium [51, 52] or cells lining the lungs and oral-nasal cavities of the fetus [49]. Indeed, the majority of cells derived from the skin at the time of amniocentesis (16-17 weeks of Chapter 2 Probing stemness and neural commitment in AF cells 69 gestation) are periderm cells [52] and both K8 and K18 are found in the periderm cells of the developing skin [16, 53]. During embryonic and fetal development, the epidermis changes from a simple epithelium, covered by the periderm, to stratified epithelium and finally to the keratinized epidermis [16]. The change from fetal periderm to definitive epiderm occurs at about 21 weeks of gestation [52]. At this time, the periderm cells have shed and the five cell layers typical of adult epidermis are present [16]. Interestingly, early epidermal cells also express high levels of K8 during development [54]. The subsequent keratinization of the fetal skin, including follicular and interfollicular keratinization, occurs around 22-25 weeks of gestation [49], which would explain why fewer K8 cells are present in the later gestational periods. Although the cells found within the AF have been categorized into three main types: epithelial (E-type), fibroblast-like (F-Type) and amniotic fluids (AF-Type) cells [51, 52], previous studies have demonstrated that keratin staining is restricted to mainly epithelioid (Type-E) and some amniotic fluid (TypeAF) type cells; fibroblast-like (Type-F) cells have been shown to be negative for keratin suggesting a more mesenchymal origin [55]. Our results certainly support these findings. Since NOTCH signaling is one of the most important pathways implicated in both epithelial and epidermal stem cells (reviewed in [18, 19]), we investigated the expression of multiple components of NOTCH signaling in AF cells. Our results showed that various components of the NOTCH signaling pathway were expressed in AF cells. Among them, we did find that NOTCH1, NOTCH2 and JAG1 were ubiquitously expressed in all samples and that HES1 represents the primary target of NOTCH-RBP-J signaling in AF cells. Interestingly, HES1 also represents one of the main target genes in the skin epidermis [34]. In fact, multiple Chapter 2 Probing stemness and neural commitment in AF cells 70 NOTCH receptors are expressed in the skin and when all canonical NOTCH signaling is ablated in embryonic epidermis, proliferation is reduced and differentiation is impaired, suggesting that NOTCH signaling also acts as a switch directing epidermal stem cells to initiate differentiation [34]. However, NOTCH signaling is complex and depends on the epidermal compartment [56], and since the AF contains a mixture of cell types from various fetal tissues, including cells partially or fully committed to a particular fate, the exact role of NOTCH signaling in AF cells remains elusive. Since NOTCH is important in all tissues and signaling through NOTCH ligands may involve many responses in cells from different systems, a more homogenous population of AF cells would certainly aid in better defining the role of NOTCH during fetal development. Similarly, it remains to be determined whether NOTCH signaling is involved in neural commitment of AF cells. AF cells expressed low levels of neural markers, NESTIN and PAX6, with trace expression of ASCL. PROM1, a proposed marker of progenitor cells, was ubiquitously expressed in all gestational periods examined. In particular, PROM1 expression has recently been proposed as a means of isolating neural progenitor cells from human ES cells [37]. However, it is important to note that PROM1 negative cells have also been shown to give rise to clonogenic neural stem cells [57], bringing into question the redundancy of PROM1 marker expression. Hence, in addition to PROM1, we also examined NESTIN, a marker used in a number of stem and progenitor cell types, such as neural stem and progenitors cells as well as in cells of human limbal epithelia [40] and the bulge region of the hair follicle [38, 39], the latter of which may also contain resident epithelial stem cells [58]. Interestingly, the NESTIN-positive cells in the bulge region were shown to be able to Chapter 2 Probing stemness and neural commitment in AF cells 71 differentiate into neuronal cells, when cultivated under clonal conditions [38] and NESTIN expression was shown to co-localize with K5, K8 and K15 in the hair follicle bulge cells, outer-root sheath cells and basal cells of the sebaceous glands [39]. Similarly, we also observed NESTIN co-expression in K8 positive AF cells (approximately 10%), suggesting that the K8-NESTIN positive cells may represent a potential progenitor cell population within fetal epithelial cells. In fact, NESTIN positive cells were also found in the basal layer of the epidermis [59], which further corroborates our findings. Our results suggest that NESTIN expression defines a sub-population of epithelial-like AF cells whose potential as a putative stem cell population is currently under investigation. Initial efforts have focused on the use of cell surface antigen C-KIT for the isolation of stem cell sub-populations. In fact, clonal AF stem cells, originating from a sub-population of CKIT (receptor for stem cell factor) positive cells, have been shown to give rise to cell lineages inclusive of all three germ layers [10] – a hallmark which confers a greater advantage over most known adult stem cells sources. Despite these promising results, it is still unclear as to the origin of C-KIT derived AF stem cells, which is further complicated by the developmental distribution of C-KIT expression. This receptor protein is present in human ES cells [60], primordial germ cells, hematopoietic stem cells and some somatic cells. Multipotent C-KIT positive cells isolated from the dermis can acquire neural, hepatic and renal properties [61], and those of the bone marrow have been differentiated into bone, cartilage, fat, neural and pancreatic cells [48], making it challenging to truly define the origin of C-KIT positive AF cells. Therefore, a more specific marker for the isolation of AF stem cell sub-populations is required. To resolve this issue, we determined the degree to Chapter 2 Probing stemness and neural commitment in AF cells 72 which AF cells express the hallmarks of pluripotency: OCT4a, SOX2 and NANOG. Although our data showed that all three genes were enriched 1-3 fold in C-KIT positive cells, it is possible that other important factors are also enriched by C-KIT sorting [62] that could contribute to stemness. In addition, the low number of OCT4 and SOX2 expressing cells prior to cloning can be an explanation of why no detectable levels of the transcript or protein were found, as also recently reported by Li et al. [45]. This is not surprising since when ES cells are triggered to differentiate, the expression of OCT4a and SOX2 is downregulated, confirming that these genes are required for self-renewal and maintenance of pluripotency. Indeed, elevated levels of SOX2, OCT4a and NANOG expression were recently shown in induced Pluripotent Stem (iPS) cells generated from AF cells via ectopic expression of four human factors: OCT4/SOX2/KLF4/C-MYC [45]. Infection of AF cultures with a SOX2 promoter EGFP retroviral vector identified small clusters of SOX2 positive cells, which were differentiated into NFL, NSE and MAP2 positive cells, markers which are expressed in neurons. NFL and MAP2 provide structural support to cell shape and facilitate the transport of particles and organelles within the cytoplasm [63]. NSE (also known as enolase gamma subunit) is a glycolytic isoenzyme, which is expressed in the central and peripheral nervous systems. Together, the expression of these proteins was found in sporadic clusters of AF cells, very similar to the pattern and distribution observed for SOX2 positive cells. Supporting our observations, the expression of SOX2 in the brain has been shown to be restricted to neural stem and progenitor cells, glial precursors and proliferating astrocytes [29]. Notably, SOX2 has been shown to employ BRN2 as a partner (rather than OCT4) for targeting NES expression in neural stem and Chapter 2 Probing stemness and neural commitment in AF cells 73 progenitor cells [64, 65] and seems to function upstream of C-KIT in ES cells, based on recent microarray and chromatic immunoprecipitation data [66]. Whether, SOX2 serves the same function in AF cells remains to be elucidated. Since single cell assays, performed by limiting dilution (a direct measure of clonality), are an essential and rigorous approach to examine the cell fate of any given stem cell, we employed this method to enrich for AF clones that maintained SOX2 expression following multiple passages and freeze/ thaw cycles. SOX2 expression was validated with RT-PCR, Western blotting and immunocytochemistry in two of the three clones generated. Given the heterogeneity of AF cells and the ability to isolate clonal sub-populations, the characterization and standardization of AF cells is crucial before AF cell-based therapies can be translated to future clinical applications. In support of this goal, we have generated clonal cell lines of AF cells based on single cell cloning protocols as well as SOX2 and C-KIT expression for subsequent studies. This is the first report, to our knowledge, describing the NOTCH signaling pathway in AF cells and showing that pure SOX2 positive clones can be enriched from AF culture and used for neuronal induction. Our data is in agreement with others reporting the presence of neural-derived AF cells [10, 67]. Chapter 2 Probing stemness and neural commitment in AF cells 74