Neural precursors (NPs) derived from ventral mesencephalon (VM) normally generate dopaminergic (DA) neurons in vivo but lose their potential to differentiate into DA neurons during mitogenic expansion in vitro, hampering their efficient use as a transplantable and experimental cell source. Because embryonic stem (ES) cell-derived NPs (ES NP) do not go through the same maturation process during in vitro expansion, we hypothesized that expanded ES NPs may maintain their potential to differentiate into DA neurons. To address this, we expanded NPs derived from mouse embryonic day-12.5 (E12.5) VM or ES cells and compared their developmental properties. Interestingly, expanded ES NPs fully sustain their ability to differentiate to the neuronal as well as to the DA fate. In sharp contrast, VM NPs almost completely lost their ability to become neurons and tyrosine hydroxylase-positive (TH+) neurons after expansion. Expanded ES NP-derived TH+ neurons coexpressed additional DA markers such as dopa decarboxylase and DAT (dopamine transporter). Furthermore, they also expressed other midbrain DA markers, including Nurr1 and Pitx3, and released significant amounts of DA. We also found that these ES NPs can be cryopreserved without losing their proliferative and developmental potential. Finally, we tested the in vivo characteristics of the expanded NPs derived from J1 ES cells with low passage number. When transplanted into the mouse striatum, the expanded NPs as well as control NPs efficiently generated DA neurons expressing mature DA markers, with approximately 10% tumor formation in both cases. We conclude that ES NPs maintain their developmental potential during in vitro expansion, whereas mouse E12.5 VM NPs do not.
Neural precursors (NPs) can be derived from various brain regions and stages during development and can generate all three cell types of the neural lineage (i.e., neurons, astrocytes, and oligodendrocytes) . Thus, these NP cells could be a useful resource of cells for various applications, including basic developmental studies, high-throughput drug screening, and cell replacement therapy of brain disorders. However, their proliferative and developmental potentials are poorly understood. In the case of NPs derived from the fetal brain, while proliferating in mitogen-containing media, they change their developmental property and generate more glia than neurons [2, 3]. Adult brain-derived neural stem cells also display limited ability for spontaneous neurogenesis upon mitogens removal even though they can be guided to generate more neurons by specific environmental cues . Fetal ventral mesencephalon (VM)-derived NPs, though they generate dopaminergic (DA) neurons in vivo and also without expansion in vitro, show limited ability for differentiation to specific neuronal phenotype such as the DA fate . Although modification of culture conditions such as ascorbic acid or low oxygen somewhat increased their differentiation to the DA phenotype [5–7], once significantly expanded in vitro, these stem cells from fetal and adult brain show very limited ability to maintain their developmental potential for neurogenesis and/or differentiation to the specific neuronal fate. Thus, it is a critical issue to determine whether NP cells derived from different sources can maintain their developmental potential during mitogenic expansion.
Recently, many laboratories have reported that embryonic stem (ES) cells can efficiently differentiate into NP cells and then into DA neurons via optimal culture conditions and/or genetic manipulation [8–12]. Thus, multipotent NP cells can be efficiently derived from ES cells. Because ES cell-derived NPs (ES NP) do not go through the same maturation process during in vitro expansion as fetal brain-derived NPs do , it is tempting to hypothesize that ES NPs maintain their potential to differentiate to the neuronal and/or DA fate during mitogenic expansion. To address this, we expanded NPs derived from embryonic day-12.5 (E12.5) VM or ES cells and compared their developmental properties in vitro. Here, we report that ES NPs, but not VM-derived NPs, maintain the potential to generate Tuj1+ neurons as well as functional DA neurons after extensive in vitro expansion. These ES NP cells maintained their differentiation potential through multiple freeze-thaw cycles.
Materials and Methods
ES Cell Culture and In Vitro Differentiation
Early passage J1 ES cells (with the passage number of 10) were obtained from Dr. Rudolf Jaenisch's laboratory (Massachusetts Institute of Technology, Lexington, MA). The mouse ES cell lines J1 and N2 were maintained as described previously . Briefly, undifferentiated ES cells were cultured on gelatin-coated dishes in Dulbecco's modified minimal essential medium (Life Technologies, Rockville, MD, http://www.invitrogen.com) supplemented with 2 mM glutamine (Life Technologies), 0.001% β-mercaptoethanol (Life Technologies), 1× nonessential amino acids (Life Technologies), 10% donor horse serum (Sigma, St. Louis, http://www.sigmaaldrich.com), and 2,000 U/ml human recombinant leukemia inhibitory factor (LIF; R&D Systems, Minneapolis, http://www.rndsystems.com).
ES cells were differentiated into embryoid bodies (EBs) on nonadherent bacterial dishes (FisherScientific, Pittsburgh, http://www1.fishersci.com) for 4 days in EB medium as described above without LIF and replacing horse serum with 10% fetal bovine serum (FBS; HyClone, Logan, UT, http://www.hyclone.com). EBs were then plated onto an adhesive tissue culture surface. After 24 hours of culture, selection of neuronal precursor cells was initiated in serum-free ITSFn (insulin, transferin, selenium, and fibronectin) medium . After selection for 7–10 days, cells were trypsinized and nestin+ neuronal precursors were plated on polyornithine (PLO; 15 μg/ml; Sigma)- and fibronectin (FN; 1 μg/ml; Sigma)-coated plates in NP media (N2 medium  supplemented with 1 ng/ml laminin [Sigma] and 10 ng/ml basic fibroblast growth factor [bFGF; R&D Systems]). For expansion of nestin+ neuronal precursors, cells were passaged once every week and plated onto PLO/FN-coated six-well plates at 7 × 105 cells per well in NP media. For differentiation of neuronal precursors, 1.5 × 105 cells were plated onto PLO/FN-coated 24 wells with coverslips, expanded in the presence or absence of 500 ng/ml Shh-N (R&D Systems), and 100 ng/ml fibroblast growth factor 8 (FGF-8; R&D Systems) for 4 days, and then bFGF was removed to induce differentiation to neuronal phenotypes in the absence or presence of 200 μM ascorbic acid (Sigma) [8, 12]. Cells were either fixed after 4 days of expansion for NP stage analysis or at 9 days after starting neuronal differentiation for analysis of neuronal phenotypes differentiation (ND) stage.
NP cells were frozen after trypsinization by suspension in freezing media (90% serum and 10% dimethyl sulfoxide) at 1 × 107 cells per ml and placed in a styrofoam container at −80°C to ensure a gradual decrease in temperature. After 24 hours, frozen cells were moved to a liquid nitrogen tank. Frozen NPs were thawed in a 37°C water bath followed by centrifugation to remove freezing media. The cells were then plated on PLO/FN-coated plates in N3bFGF media.
VM Dissection and Expansion
Timed-pregnant mice were purchased from Charles River Laboratories (Wilmington, MA, http://www.criver.com). On E12.5, the ventral-most part of mesencephalon was cut out and collected in NP media with 5% FBS. Mechanical trituration in phosphate-buffered saline (PBS) yielded single-cell suspension. Cells were plated on PLO/FN-coated plates and maintained in NP media. Expansion of VM precursors consisted of passaging cells once every week on to PLO/FN-coated plates in NP media. To differentiate VM precursors, 1.5 × 105 cells were plated onto PLO/FN-coated 24 wells with coverslips and expanded in the presence or absence of 500 ng/ml Shh-N (R&D Systems) and 100 ng/ml FGF-8 (R&D Systems) for 4 days. ND was induced by removal of bFGF in the absence or presence of 200 μM ascorbic acid (Sigma). Cells were fixed after 4 days of expansion for analysis of NP stage or after 9 days from the initiation of neuronal differentiation for analysis of ND stage.
For immunofluorescence staining, cells were fixed for 30 minutes in 4% formaldehyde (Electron Microscopy Sciences, Ft. Washington, PA, http://www.emsdiasum.com), rinsed with PBS, and then incubated with blocking buffer (PBS, 10% normal donkey serum [NDS]) for 10 minutes. Cells were then incubated overnight at 4°C with primary antibodies diluted in PBS containing 2% NDS. The following primary antibodies were used: mouse anti-nestin (Rat401, 1 μg/ml; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww), rabbit anti-β-tubulin (1:2,000; Covance, Princeton, NJ, http://www.covance.com), sheep anti-tyrosine hydroxylase (TH, 1:200; Pel-Freez, Rogers, AK, http://www.pel-freez.com), sheep anti-aromatic L-amino acid decarboxylase (1: 200; Chemicon, Temecula, CA, http://www.chemicon.com), and rat anti-dopamine transporter (DAT, 1:2,000; Chemicon). After additional rinsing in PBS, the coverslips were incubated in fluorescent-labeled secondary antibodies (Cy2- or Rhodamine Red-X-labeled donkey immunoglobulin G; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, http://www.jacksonimmuno.com) in PBS with 2% NDS for 30 minutes at room temperature. After rinsing 3 × 10 minutes in PBS, sections were counterstained using 1 μg/ml DAPI (4,6-diamidino-2-phenylindole) and then mounted onto slides in Gel/Mount (Biømeda Corp., Foster City, CA, http://www.biomeda.com). Coverslips were examined using a Leica TCS/NT confocal microscope (Leica, Heerbrugg, Switzerland, http://www.leica.com) equipped with krypton, krypton/argon, and helium lasers.
Cell Counting and Statistical Analysis
Cell density of DA neurons was determined by counting the numbers of TH+ cells and the numbers of β-tubulin+ cells per field at ×63 magnification using a Zeiss Axioplan I fluorescent microscope (Carl Zeiss, Thornwood, NY, http://www.zeiss.com). Ten visual fields were randomly selected and counted for each sample, and cell densities were calculated by dividing the numbers of TH+ cells by that of β-tubulin+ cells. Numbers presented in figures represent the average percentage and SEM of TH+ cells over β-tubulin+ cells from five samples per ES cell clone.
For statistical analysis, we used Statview software (SAS, Cary, NC, http://www.statview.com) and performed analysis of variance (ANOVA) with an alpha level of 0.01 to determine possible statistical differences between group means. When significant differences were found, post hoc analysis was performed using Fisher's PLSD (protected least significant difference; alpha = 0.05).
Semiquantitative Reverse Transcription-Polymerase Chain Reaction and Real-Time Polymerase Chain Reaction Analysis
Total RNA from plated cells at different stages of the differentiation protocol was prepared using TRIzol Reagent (Sigma) followed by treatment with DNase I (Ambion, Austin, TX, http://www.ambion.com). For reverse transcription-polymerase chain reaction (RT-PCR) analysis, 5 μg of total RNA was transcribed into cDNA using the SuperScript Preamplification Kit (Life Technologies) and oligo (dT) primers. The cDNA was then analyzed in a PCR assay using the following primers: β-actin: 5′-GGCATTGTGATGGACTCCGG-3′; 5′-TGCCA-CAGGATTCCATACCC-3′ (358 bp); nestin: 5′-GGAGT-GTCGCTTAGAGGTGC-3′; 5′-TCCAGAAAGCCAAGAGA-AGC-3′ (327 bp; ; Bmi1: 5′-TTGCTGCTGGGCATCGTAAG-3′; 5′-CCAATGGCTCCAATGAAGACC-3′ ; β-tubulin: 5′-AACTATGTAGGGGACTCAGACCTGC-3′; 5′-TCTCACACT-CTTTCCGCACGAC-3′ (274 bp); TH: 5′-TCCTGCACTCCCT-GTCAGAG-3′; 5′-CCAAGAGCAGCCCATCAAAGG-3′ (423 bp); DDC: 5′-CCTACTGGCTGCTCGGACTAA-3′; 5′-GCG-TACCAGGGACTCAAACTC-3′ (715 bp); DAT: 5′-CA-GAGAGGTGGAGCTCATC-3′; 5′-GGCAGATCTTCCAGA-CACC-3′ (328 bp); Pitx3: 5′-CTCTCTGAAGAAGA-AGCAGCG-3′; 5′-CCGAGGGCACCATGGAGGCAGC-3′ (491 bp); Nurr1: 5′-CATGGACCTCACCAACACTG-3′; 5′-GAGA-CAGGTGTCTTCCTCTG-3′ (383 bp).
PCR reactions were carried out with 1 × IN Reaction Buffer (Epicenter Technologies, Omaha, NE, http://www.epicentertechnology.com), 1.4 nM of each primer, and 2.5 units of Taq I DNA polymerase (Promega, Madison, WI, http://www.promega.com). Samples were amplified in an Eppendorf Thermocycler (Brinkmann Instruments, Westbury, NY, http://www.brinkmann.com) under the following conditions: denaturing step at 95°C, 40 seconds; annealing step at 60°C, 30 seconds; amplification step at 72°C, 1 minute for 20 to 28 cycles and a final amplification step at 72°C, 10 minutes. For semiquantitative PCR, cDNA templates were normalized by amplifying actin-specific transcripts, and levels of gene transcription were detected by adjusting PCR cycling and primer design in such a way that each primer set amplified its corresponding gene product at its detection threshold to avoid saturation effects.
For quantitative analysis of the expression level of mRNAs, real-time PCR analyses using SYBR green I were performed using DNA engine Opticon (MJ Research, Waltham, MA, http://www.mjr.com). To reduce nonspecific signals, oligonucleotides amplifying small amplicons were designed using MacVector software (Oxford Molecular Ltd., Burlington, MA, http://www.accelrys.com). We selected primer sets amplifying the specific product without nonspecific bands. The following are the primer sets used for real-time PCR analysis: TH: 5′-TTGGCTGAC-CGCACATTTG-3′; 5′-ACGAGAGGCATAGTTCCTGAGC-3′ (336 bp).
The oligonucleotides used for RT-PCR were used to detect β-actin, nestin, β-tubulin, DDC, and DAT mRNA. Amplifications were performed in 25 μl containing 0.5 μM of each primer, 0.5× SYBR Green I (Molecular Probes, Inc., Eugene, OR, http://probes.invitrogen.com), and 2 μl of fivefold diluted cDNA. Forty PCR cycles were performed with the temperature profile of 95°C for 30 seconds, 55°C for 30 seconds, 72°C for 30 seconds, and 79°C for 5 seconds. The dissociation curve of each PCR product was determined to test the specificity of the fluorescent signals. The melting temperatures (Tm) of the PCR products were 85°C, 85°C, 85°C, 85°C, 84°C, and 85°C for β-actin, nestin, β-tubulin, TH, DDC, and DAT, respectively. After each PCR cycle, fluorescence was detected at 79°C to melt primer dimers (the Tm of all primer dimers used in this study was <76°C). Plasmid DNAs containing the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene (from 102 to 107 molecules) were used to make a standard curve. The fluorescent signals from specific nestin, β-tubulin, TH, DDC, and DAT PCR product were normalized against that of the β-actin gene, and then relative values were calculated by setting the value of unexpanded VM as 100. Two replicates were done for each sample, and all reactions were repeated more than twice.
Analysis of Catecholamines
Differentiated VM NPs and ES NPs (ND stage day 9) in six-well plates were treated with 200 μl of N3 medium supplemented with 50 mM KCl, and the media were collected after 30 minutes and concentrated solutions of perchloric acid (PCA) were added to a final concentration of 0.1 M PCA/0.1 mM EDTA. For measurement of catecholamine cell contents, cells were harvested in 0.1 M PCA/0.1 mM EDTA. These deproteinated samples were centrifuged, and supernatants were kept at −80°C until further analysis. Samples were further purified by using a 0.22-μm nylon filter (Osmonics, Inc., Trevose, PA, http://www.osmonics.com) followed by determination of cate-cholamine content by reverse-phase high-performance liquid chromatography (HPLC) using a Velosep RP-18 column (100 × 3.2 mm; Brownlee Labs, Shelton, CT, http://las.perkinelmer.com) and an ESA Coulochem II electrochemical detector (ESA, Inc., Chelmsford, MA, http://www.esainc.com) equipped with a model 5014 analytical cell as we described . The mobile phase was composed of 0.1 M sodium phosphate buffer (pH 2.65), 0.1 mM EDTA, 0.4 mM sodium octyl sulfate, and 9% (vol/vol) methanol and used at a flow rate of 0.8 ml/minute. The potential of the guard cell was set at 330 mV. The potential of the first electrode in the analytical cell was set at 0 mV, the second at 310 mV. l-DOPA, dopamine, dihydroxyphenyl acetic acid, and homovanillic acid were identified by retention time and quantified based on peak height using an EZChrom Chromatography Data System (ESA, Inc., Chelmsford, MA, http:// www.esainc.com/). The limit of detection for all compounds was less than 1 pg. DA content of each sample was normalized with the amount of total cellular proteins. For protein measurement, after harvesting cells in 0.1 M PCA/0.1 mM EDTA, precipitates were suspended in 0.2% Triton-X 10 mM potassium phosphate buffer (pH 7) and sonicated. The protein content was measured by the Bradford method (Bio-Rad Assay; Bio-Rad Laboratories, Hercules, CA, http://www.bio-rad.com) .
J1 ES cell-derived NPs before and after 4 weeks of expansion were differentiated by removal of bFGF, trypsinized at day 3 of the differentiation stage [11, 12], and resuspended at a density of 200,000 cells per μl. One microliter of cell suspension was grafted into the right striatum (from the bregma: anterior-posterior +0.05, lateral −0.18, ventral −0.30, incisor bar 9) of C57/BL6 mice (n = 9) (Charles River Laboratories). Prior to surgery, mice received an i.p. injection of pre-anesthesia (acepromazine [3.3 mg/kg; PromAce, Fort Dodge, IA, http://www.bi-vetmedica.com/] and atropine sulfate [0.2 mg/kg; Phoenix Pharmaceuticals, Inc., St. Joseph, MO, http://www.phoenixpeptide.com]) followed by an i.p. injection of ketamine (60 mg/kg; PromAce) and xylazine (3 mg/kg; Phoenix Pharmaceuticals, Inc.). Transplantation was performed using a 22-gauge 10-μl Hamilton syringe (Hamilton Company, Reno, NV, http://www.hamiltoncompany.com) and a Kopf stereotaxic frame (David Kopf Instruments, Tujunga, CA, http://www.kopfinstruments.com). Buprenorphine (0.032 mg/kg, subcutaneous; Sigma) was given twice during 24 hours as postoperative analgesia. Four weeks after transplantation, mice were terminally anesthetized with an i.p. overdose of pentobarbital (150 mg/kg; Sigma). Mice were perfused intracardially with 100 ml of heparin saline (0.1% heparin in 0.9% saline) followed by 200 ml of paraformaldehyde (4% in PBS). Brains were postfixed for 8 hours, equilibrated in sucrose (20% in PBS), sectioned at 40 μm on a freezing microtome, and collected in PBS. For histological analysis, sections were stained with antibodies against TH (see above) and counterstained by Nissle staining. For measurement of graft volume and counting of total TH+ cell number within grafts, every sixth section was stained with antibodies against TH (see above) and counterstained by Nissle staining. The stained sections were subjected to stereological analysis using an integrated Axioskop two microscope (Carl Zeiss) and StereoInvestigator image capture equipment and software (Microbright Field, Williston, VT, http://www.microbrightfield.com). TH+ cell density was calculated by dividing the total TH+ cell number by the graft volume.
NP Cell Population Derived from VM and ES Cells Can Be Expanded In Vitro in the Presence of bFGF
Previously, we have shown that N2 ES cells overexpressing the transcription factor Nurr1 efficiently generate DA neurons . Thus, to obtain NPs that can readily generate DA neurons, we differentiated N2 ES cells to the NP stage according to the five-stage in vitro differentiation procedure [8, 12]. Furthermore, we used the wild-type ES cell line, J1, as well as VM cells dissected from E12.5 embryonic tissues from timed-pregnant mice, for the generation of NPs. These NP cells were expanded on adherent culture dishes in the presence of bFGF as described in Materials and Methods. ES cell-derived and VM-derived NPs showed different efficiencies of expansion under these conditions. Whereas the VM-derived NP cell number increased approximately 10-fold after 4 weeks of expansion, the ES-derived NP cells proliferated exponentially, reaching approximately 1,000-fold after expansion for the same period (Fig. 1A). Expanded cells from both VM- and ES-derived precursors were positive for the NP marker nestin (Fig. 1B–1E). In addition, semiquantitative RT-PCR analysis showed that expression of nestin and another NP marker, Bmi1, known to be essential for neural stem cell self-renewal , was comparable before and after expansion in both cell populations. Real-time PCR analysis confirmed that consistent expression levels of nestin mRNAs were maintained after expansion among each group (Fig. 1G). These results show that both ES NP and VM NP cells maintain at least some aspects of their NP state upon expansion.
NPs Derived from ES Cells but Not from VM Cells Efficiently Differentiate to Tuj1+ Neurons and TH+ DA Neurons after Mitogen-Induced Expansion
To analyze their developmental potential, we differentiated expanded NPs in vitro into mature neurons. As previously reported, NPs derived from both unexpanded ES cells and VM cells readily differentiated into neurons and TH+ neurons (Fig. 2A, 2C, 2G, 2I, 2M, 2O, 2S, 2U). After 4 weeks of expansion, VM-derived NPs almost completely lost their potential to differentiate into neurons (Fig. 2G, 2H) or into TH+ neurons (Fig. 2M, 2N). In sharp contrast, N2 ES cell-derived NPs differentiated into neurons (Fig. 2I, 2J) and TH+ neurons (Fig. 2O, 2P) with the same efficiency after 4 weeks of expansion. To test whether this dramatic difference was artificially affected by overexpression of Nurr1 by N2 ES cells, we tested the wild-type J1 ES cell line. As shown in Figure 2L and 2R, NPs derived from the wild-type J1 ES cells also maintained their developmental potential after the same 4 weeks of expansion. These results show that ES cell-derived NPs dramatically differ from VM-derived NPs in their developmental potential after extensive expansion although they appear similar before expansion.
For quantitative analysis, we performed cell counting as described in Materials and Methods (). To ensure an accurate quantitative analysis of cell numbers, we designed a cell-counting method using a grid of 10 random fields, which was applied to multiple coverslips per analysis. After in vitro differentiation, unexpanded VM NPs, expanded VM NPs, unexpanded N2-ES NPs, expanded N2-ES NPs, unexpanded J1-ES NPs, and expanded J1-ES NPs contained 218 ± 38, 48 ± 11, 321 ± 43, 335 ± 25, 350 ± 12, and 330 ± 10 β-tubulin+ neurons in 10 fields, respectively (average cell number ± SEM; Fig. 2Y). We also counted the number of TH+ neurons, and the above cells contained 5.9% ± 0.7%, 1.1% ± 1.1%, 24.7% ± 2.2%, 21.6% ± 1.1%, 22.7% ± 2.4%, and 20.0% ± 0.6% TH+/β-tubulin+ neurons, respectively, and 1.24% ± 0.24%, 0.04% ± 0.04%, 6.95% ± 1.78%, 6.45% ± 0.78%, 6.61% ± 1.07%, and 5.43% ± 0.87% TH+/total cells, respectively (average percentage ± SEM; Fig. 2Z).
In Vitro Expanded NPs Derived from ES Cells Differentiate into Functional Midbrain-Like DA Neurons
Next, we performed further immunocytochemistry analyses to test whether TH+ neurons generated from expanded NPs express other DA neuronal markers. After in vitro differentiation, both unexpanded (control) and 4-week expanded N2 cells expressed another DA neuronal marker, DDC, with almost a complete coexpression pattern with TH (Fig. 3A–3F), and some of them also expressed the late DA marker, DAT (Fig. 3G–3L). In addition, J1 ES cell-derived NPs prominently generated DA neurons, many of which coexpress DDC (Fig. 3M–3R) and DAT (Fig. 3S–3X) before and after expansion.
We next analyzed DA marker gene expression by semi-quantitative RT-PCR. mRNA expression of the neuronal marker β-tubulin gene was comparable before and after expansion of N2 and J1 cells (Fig. 4A). In contrast, expression of β-tubulin mRNA was almost undetectable in the case of expanded VM NP cells, which is consistent with immunocytochemistry analysis (Fig. 2). mRNA expression of midbrain DA markers such as TH, DDC, DAT, Pitx3, and Nurr1 was well maintained after expansion of N2 and J1 cells, whereas VM cells almost completely lost their expression after expansion (Fig. 4A). As shown in Figure 5B, all of these gene expression data were quantitatively confirmed by real-time PCR analyses (Fig. 4B).
An important physiological aspect of authentic DA neuron phenotypes is the ability to synthesize DA and release it in response to membrane depolarization. Having observed specific midbrain DA marker expression after in vitro differentiation of expanded N2 NP cells, we next tested whether DA could be released after membrane depolarization. In vitro differentiated cells (day 9 of the five-stage protocol) from unexpanded and expanded NP cells were treated with 50 mM KCl, and the released DA in the media was analyzed by HPLC. In response to membrane depolarization from unexpanded VM, expanded VM, unexpanded N2, and expanded N2 cells, 2.70 ± 0.22, 0.00 ± 0.00, 3.59 ± 0.41, and 3.29 ± 0.42 pg DA/μg cellular proteins were released, respectively (Fig. 5). Taken together, neurons derived from expanded N2 cells maintain the ability to differentiate into midbrain-like DA neurons and to produce and release DA in response to membrane depolarization whereas VM NP cells completely lost these properties after expansion.
Expanded ES Cell-Derived NPs Efficiently Generate DA Neurons After In Vivo Transplantation into Mouse Striatum
To characterize the developmental potential of ES cell-derived NP cells in vivo, we differentiated unexpanded and expanded J1 ES NPs in vitro, trypsinized them at day 3 of the stage V protocol, and transplanted them into normal mouse striatum. Early passage J1 ES cells were used for transplantation analysis to avoid high-frequency tumor formation from high-passage N2 ES cells . At day 3 of the stage V protocol, the induction of DA marker genes was not evident in vitro (data not shown; ). Four weeks after transplantation, the animals were sacrificed and the grafts were analyzed for survival and expression of phenotypic markers. J1 ES cell-derived NPs generated mostly well-contained TH+ grafts, but occasional tumor formation was observed in approximately 10% of transplanted mice per each group. This result suggests that 4-week expansion does not apparently increase teratoma/tumor formation for the low-passage J1 ES cells. As shown in Figure 6A and 6B, both unexpanded and expanded J1 NP cells efficiently generated abundant TH+ neurons in the host brain. TH+ cell bodies appear to be derived from transplanted cells, because sham-treated side (contralateral to the transplanted side) shows only TH fibers projecting from substantia nigra but never TH cell bodies as in the grafts. No significant differences in the density of TH+ cells in the grafts were observed between unexpanded and expanded NP cell grafts (Fig. 6C; 3,196.5 ± 739.7 and 3,305.5 ± 453.1, respectively). We next examined whether these TH+ neurons generated in vivo also express other markers of DA neurons. Triple immunocytochemistry analysis demonstrated that virtually all of TH+ neurons coexpressed DAT and DDC in grafts of unexpanded (Fig. 6D–6G) and expanded (Fig. 6H–6K) NP cells.
Long-Term Storage and Freeze-Thaw Cycle Do Not Abrogate the Developmental Potential of Expanded ES-Derived NP Cells
The above data demonstrate that expanded NP cells from ES cells do not lose their developmental potential for neurogenesis and lineage-specific differentiation to DA neurons. These results suggest that these cells may serve as a convenient and unlimited cell source for many basic and therapeutic approaches. One prerequisite for this application is that these expanded cells can be stably stored in liquid nitrogen without losing their developmental and proliferative potential. We tested whether expanded NPs from ES cells maintain their differentiation potential after a freeze-thaw cycle. We froze N2 ES-derived NP cells in liquid nitrogen after a 4-week expansion and subsequently thawed and subjected them to an additional 4 days of expansion in the presence of bFGF followed by in vitro differentiation. As shown in Figure 7, generation of Tuj1+ cells (Fig. 7B, 7F) and TH+ neurons (Fig. 7C, 7G) was as efficient as those without the freeze-thaw cycle. These thawed NP cells could also be expanded further, demonstrating that they also fully sustain the proliferative capacity. The same pattern was observed after the freeze-thaw cycle of expanded NP cells from J1 ES cells. Furthermore, we found that these ES NP cells could be stored up to 1 year in liquid nitrogen without losing their differentiation potential to TuJ1+ cells and TH+ neurons (data not shown).
We systematically analyzed and compared the developmental potential of fetal VM- and ES cell-derived NPs before and after 4-week expansion in vitro in the presence of mitogens. To perform quantitative comparisons between NPs from different sources, we used an adherent culture system. NPs can be cultured in vitro using different methods, such as adherent culture or neurosphere cultures, which have pros and cons. Neurosphere culture can preserve cell-cell interaction better, which may play an important role in the behavior of NPs . However, it is not easy to efficiently quantify all the cells within the spheres even after attaching to the substrate. Thus, we compared ES-derived NPs and fetal VM-derived NPs using the same culture method, one that is easier for quantification and could clearly demonstrate differences in terms of their differentiation potential during mitogenic expansion. Our results demonstrate that ES NPs do not lose their potential to differentiate into mature neurons, including DA neurons, after subsequent expansion by passaging in vitro. This was in sharp contrast to VM-derived NPs, which almost completely lost their potential for both neurogenesis and DA differentiation after expansion in vitro. Salient features about in vitro expanded NP cells derived from ES cells are as follows. First, NP cells derived from ES cells (both the wild-type J1 and the genetically engineered N2) could be exponentially expanded in vitro with a 1,000-fold increase in numbers after 4 weeks, whereas VM-derived NP cells were expanded at a much lower efficiency (10-fold increase after 4 weeks). Second, NP cells from both J1 and N2 ES cells could efficiently generate neurons and TH+ neurons after expansion. In addition, ES NPs spontaneously differentiated into neurons after expansion by removal of bFGF with high efficiency, which is again different from fetal tissue-derived NPs. Expanded ES NPs also showed maintenance of differentiation potential in vivo after transplantation into mice striatum. Furthermore, ES NPs proliferate robustly in the presence of mitogen and can be easily kept by cryopreservation without losing their developmental potential, which make them an even more convenient cell source.
Importantly, our in vivo transplantation studies showed that expansion of NPs of the low-passage J1 ES cells did not apparently increase the incidence of tumor formation after 4 weeks of expansion. However, it has been also reported that long-term-passaged fetal tissue-derived NPs showed increased tumor formation accompanied by changes in cell behavior such as faster proliferation kinetics . Thus, it would be interesting to see whether further expansion of ES NPs (e.g., more than several months) may lead to increased tumor formation. If that is the case, even though expansion of ES NPs could provide unlimited number of NPs with stable developmental potential, changes in cellular and molecular characteristics should be carefully monitored, if long-term expansion is attempted for cell replacement therapy.
In line with our current observation, there are previous reports of the intrinsic differences between NPs derived from fetal tissue or ES cells. Hitoshi et al.  reported that ES NPs do not change their progenitor status over passage in vitro unlike fetal tissue-derived NPs. For example, ES NPs respond only to bFGF but do not become responsive to EGF even after passaging ([13, 22] and S. Chung, O. Isacson, and K.-S. Kim, unpublished observation). In contrast, fetal tissue-derived NPs mature to respond to EGF over the course of culture in vitro in the presence of mitogen [13, 23]. Also, the expression of the primitive ectodermal marker FGF5 is not downregulated during expansion for ES NPs, whereas embryo-derived NPs lost their expression in culture . In addition to the intrinsic differences between fetal tissue-derived versus ES-derived NPs, environmental difference between these two cell preparations may have changed the differentiation potential of these NPs. In line with this, there have been reports that endothelial cells or astrocytes can alter the proliferation and phenotypes of NPs [4, 24, 25]. The identification of the intrinsic or environmental factors/ pathways that are involved in controlling the phenotype determination of DA NPs will make it possible to further understand the mechanism of DA differentiation, as well as to maintain and generate transplantable cells for neuronal replacement therapy for Parkinson's Disease [26–28]. Notably, low oxygen condition was not necessary to maintain DA phenotype of ES NPs during expansion, even though low oxygen condition has been shown to be important for maintenance of VM-derived NPs [7, 29]. Thus, it is possible that ES NPs may have intrinsic mechanism(s) to counteract the stress caused by high oxygen condition compared with VM NPs.
In summary, this study shows that ES NPs, but not VM-derived NPs, can be greatly expanded as a population (1,000-fold in 4 weeks) into transplantable cells without compromising their developmental potential. Expanded NPs from J1 ES cells with low passage number did not increase tumor/teratoma formation and maintained the developmental potential to generate midbrain-like DA neurons and to produce DA in vitro as well as in vivo. Furthermore, these ES-derived NPs can be easily kept by cryopreservation without losing their developmental potential, which make them a very convenient and readily available cell source for drug screening, developmental stem cell studies, and transplantation studies.
By using the same culture and in vitro differentiation methods, we compared the differentiation potential of fetal VM-derived NPs and ES-derived NPs during mitogen-induced expansion. ES NPs fully maintained their potential to spontaneously generate neurons, including DA neurons, whereas VM NPs did not. ES NPs also maintain their proliferative and developmental potential during cryopreservation and the freeze-thaw cycle, thus providing a useful source for cell replacement therapy for neurodegenerative disorders as well as biological and drug discovery studies.
This work was supported by Udall Parkinson's Disease Center of Excellence grants P50 NS39793, MH48866, DAMD-17-01-1-0762, and DAMD-17-01-1-0763 and the postdoctoral fellowship program of Korea Science & Engineering Foundation (B.-S.S.).