Author contributions: W.Z.: conception and design, collection and assembly of the data, data analysis and interpretation, writing of the manuscript; Y.M.L. and V.C.G.: collection and assembly of the data; C.R.F.: conception and design, data analysis and interpretation, writing of the manuscript.
First published online in STEM CELLS EXPRESS August 31, 2009.
Disclosure of potential conflicts of interest is found at the end of this article.
Parkinson's disease (PD) is characterized by the selective loss of midbrain dopamine neurons. Neural transplantation with fetal dopamine neurons can be an effective therapy for patients with PD, but recovery of human fetal cells is difficult. Scarcity of tissue has limited clinical application to a small number of research subjects worldwide. Selective differentiation of embryonic stem cells (ESCs) to dopamine neurons could lead to an unlimited supply of cells for expanded clinical transplantation. To facilitate the differentiation and purification of dopamine neurons, the green fluorescent protein (GFP) gene was inserted into the dopamine transporter (DAT) locus in mouse ESCs using homologous recombination. From these DAT-GFP ESCs, dopamine neurons expressing GFP were successfully produced by in vitro differentiation. The DAT-GFP ESCs were used to generate DAT-GFP knock-in mice. We have found that GFP was colocalized with DAT, Pitx3, Engrailed-1, and tyrosine hydroxylase-positive cells in midbrain, hypothalamus, and olfactory bulb but not in noradrenergic cell regions or other ectopic sites. The GFP-positive dopamine neurons could be isolated from embryonic day-15 ventral midbrain by fluorescence activated cell sorting. These purified dopamine neurons survived reculture and expressed tyrosine hydroxylase and DAT when cocultured with mouse astrocytes or striatal cells. Animals homozygous for DAT-GFP were hyperactive because they had no functional DAT protein. These DAT-GFP knock-in ESCs and mice provide unique tools for purifying dopamine neurons to study their physiology, pharmacology, and genetic profiles. STEM CELLS 2009;27:2952–2961
Parkinson's disease (PD) is a common neurodegeneration disease affecting approximately 1% of people older than 65 years [1–3]. Replacement of dopamine by the precursor L-DOPA remains the most effective treatment [4, 5]. However, after 5-10 years of L-DOPA therapy, patients develop motor complications including peak dose dyskinesias [6, 7]. In the past two decades, novel treatments for PD have emerged including fetal dopamine neuron transplantation [8–16], deep brain stimulation [17, 18], gene therapy [19, 20], and neuroprotective agents [21, 22]. Clinical trials have demonstrated that dopamine cell replacement therapy can duplicate the best effects of L-DOPA [12, 23]. Grafted dopamine neurons can survive indefinitely without immunosuppression [12, 23–25]. Because of the difficulty of acquiring fetal tissue, large-scale application of cell transplantation is impossible without a new cell source.
Embryonic stem cells (ESCs) may provide an unlimited supply of transplantable dopamine neurons. Dopamine neurons have been successfully differentiated from ESCs and can survive after transplantation in animal brain [26–31]. However, these ESC-derived dopamine neurons are a small fraction in a mixture of large numbers of nondopaminergic cells and stem cells [28–33]. Residual undifferentiated stem cells have led to tumor formation after transplantation [31, 33]. Purification of differentiated neurons is required for safe transplantation therapy .
One strategy to purify dopamine neurons is to identify a gene expressed in dopamine neurons and insert a fluorescent gene downstream from the promoter for that gene. Kessler et al. have generated transgenic mice with green fluorescent protein (GFP) expressed under control of the tyrosine hydroxylase promoter . The plasmid containing the TH promoter-GFP gene was randomly inserted in the genome. Although GFP is expressed in midbrain dopamine neurons, it is also expressed in noradrenergic cells and even in noncatecholamine cells.
Because transgene constructs lack the full regulatory environment of the native gene, transgenic methods do not have the maximal physiologic precision needed for isolation of authentic dopamine neurons generated either in vivo or in vitro. A knock-in strategy offers more accurate regulation of gene expression. Since tyrosine hydroxylase is expressed in many nondopaminergic cells, a gene more specific for dopamine neurons, such as the dopamine transporter (DAT), can narrow the selection of cells. In this report, we have generated DAT-GFP knock-in ESCs and used them to create DAT-GFP knock-in mice. We show that dopamine neurons recovered from embryonic mouse brain can be purified by fluorescence-activated cell sorting (FACS) and can survive replating in tissue culture.
MATERIALS AND METHODS
DAT-GFP Knock-In Targeting Vector
The DAT-GFP knock-in targeting vector was generated by inserting a GFP-pGK-Neo cassette into a mouse DAT genomic DNA construct (kindly provided by Dr. Xiaoxi Zhuang, Ph.D., University of Chicago, Chicago). The DAT genomic vector contained 5.1 kb of upstream sequence from the ATG start codon and 1.9 kb of downstream sequence from the ATG site (Fig. 1A). The GFP-pGK-Neo cassette was ligated into the unique NotI site between two homologous arms. The targeting vector was verified by restriction analysis and polymerase chain reaction (PCR).
ESC Culture, Gene Targeting, and Screening
Mouse ESCs (CJ7) were cultured on mouse embryonic fibroblast feeder layers in a medium consisting of Dulbecco's modified Eagle's medium (DMEM), 15% ESC-quality fetal bovine serum, nonessential amino acids (1X), 2 mM L-glutamine, 100 units/ml penicillin/streptomycin, 100 mM β-mercaptoethanol, and 100 units/ml leukemia inhibitory factor. For electroporation, 30 μg of AatII linearized targeting vector was electroporated into 2 × 107 CJ7 ESCs (500 microfarad, 0.24 kilovolts). Mouse ESCs were plated onto mouse embryonic fibroblast feeder cells in two 100-mm plates. One day after plating, G418 (300 μg/ml; Invitrogen, Carlsbad, CA, http://www.invitrogen.com) was added to the medium. After 7-10 days in culture, drug-resistant colonies were picked up and expanded in 96-well plates. We first used PCR screening to identify potential homologous recombinant clones with primer pairs indicated in Figure 1A. Positive ES clones were then expanded in 6-well plates and evaluated by Southern blotting. Three micrograms of genomic DNA was digested with EcoR1 and separated in 1% agarose gel, followed by DNA transfer to nylon membrane according to standard procedures. Southern blotting was performed using digoxigen-labeled probe (Fig. 1A), according to the protocol provided by the manufacturer (Roche, Basel, Switzerland, http://www.roche-applied-science.com). Two clones designated D8 and G4 were confirmed to have homologous recombination.
Generation of Knock-In Mice
The G4 line of DAT-GFP knock-in ESCs was used to generate knock-in mice in the Transgenic Mouse Core facility at University of Colorado Denver. The ESCs were injected into blastocysts isolated from C57BL/6 mice and transferred to surrogate mothers. From 20 pups, 6 chimeric male mice were identified by coat color. Chimeric males were mated with wild-type (WT) C57BL/6 females. Germ-line transmission was verified by PCR and Southern blotting using tail DNA from the resultant progeny. Breeding heterozygous male and female mice generated homozygous DAT-GFP mice. For PCR genotyping, we used the primer pairs of GACGTAAACGGCCACAAGTT and GAACTCCAGCAGGACCATGT under the following conditions: 95°C for 5 minutes, followed by 35 cycles of 95°C for 30 seconds, 55°C for 30 seconds, 72°C for 30 seconds, and a final extension of 72°C for 10 minutes.
In Vitro Differentiation of Knock-In ESCs
In vitro differentiation of ESCs to dopamine neurons was performed using both the PA6 cell method [26, 29] and the embryoid body method [30, 33]. Briefly, PA6 stromal cells were plated on gelatin-coated 12-well plates at 1 × 105 cells per well. One day later, DAT-GFP ESCs were dissociated to single cells and plated onto the PA6 cell layer at 300 cells per well in a differentiation medium (DMEM, 15% knockout serum replacement, nonessential amino acids, 2 mM L-glutamine, 100 units/ml penicillin/streptomycin, 100 mM β-mercaptoethanol). From days 7-20, the medium was replaced with DMEM/F12 supplemented with N2, B27 (Invitrogen), sonic hedgehog (Shh; 50 ng/ml), and fibroblast growth factor 8 (FGF8; 25 ng/ml) (R&D Systems, Minneapolis, http://www.rndsystems.com). Medium with trophic factors was replenished every 2 days. For the embryoid body method, undifferentiated ESCs were dissociated into single cells and plated onto nonadherent Petri dishes in basic differentiation medium that consisted of DMEM/F12 supplemented with N2 and B27. Embryoid bodies were formed after 6 days in suspension culture. The embryoid bodies were dissociated into single cells using 0.05% trypsin and plated onto poly-D-lysine (100 μg/ml)- and laminin (4 μg/ml)-coated dishes for another 10-14 days in the basic differentiation medium. During the adherent culture, FGF8 (5 ng/ml) and Shh (100 ng/ml) were added to the differentiation medium from days 2-6, followed by FGF8 (100 ng/ml) and Shh (400 ng/ml) treatment from days 6-10. Ascorbic acid (200 μM) was added from days 10-14. Cells were fixed with 4% paraformaldehyde and stained with antibodies to tyrosine hydroxylase (TH; see below for detail).
GFP Expression During Development and Postnatal Stages in DAT-GFP Knock-In Mouse
We examined GFP expression during mouse embryonic development and postnatal stages. Heterozygous embryos (by mating wild-type females with homozygous males) and homozygous embryos (by mating homozygous females with homozygous males) were collected at embryonic day 13 (E13), E15, and E18. Dissected mesencephalons were examined for GFP expression. Postnatal brains were recovered from 4- to 16-week-old mice that received an overdose of anesthetic followed by cardiac perfusion with cold saline followed by 4% paraformaldehyde. Whole brains were cryoprotected with 30% sucrose and sectioned at 40 μm. Sections were examined for GFP expression under an Olympus dissection microscope (MVX10; Olympus Microscope American, Center Valley, PA, http://www.olympusamerica.com) or Nikon Diaphot fluorescence microscope (Nikon Instruments, Inc., Melville, NY, http://www.nikoninstruments.com).
Colabeling of GFP-Positive Cells with Dopaminergic Markers in Differentiated Cells and Brain Sections
For dual immunostaining, cultures and brain sections were rinsed three times in phosphate-buffered saline (PBS). Samples were incubated for 30 minutes in blocking buffer of 5% normal goat serum, 1% bovine serum albumin, and 0.2% Triton X-100 in PBS, followed by overnight incubation at 4°C with the following antibodies: rabbit anti-TH (1:200; Pel Freez, Rogers, AK, http://www.pelfreez-bio.com), mouse anti-GFP (1:1000; Invitrogen), rat anti-DAT (1:200; Millipore, Billerica, MA, http://www.millipore.com), rabbit anti-Engrailed-1 (En-1; 1:200; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), and rabbit anti-Pitx3 (1:200; Invitrogen) . Samples were washed three times with PBS then incubated for 1 hour with the following secondary antibodies: fluorescence-conjugated goat anti-mouse (for GFP, 1: 200; EMD Biosciences, Gibbstown, NJ, http://www.emdbiosciences.com), Texas Red-conjugated goat anti-rabbit (for TH, En-1, and Pitx3, 1:200; EMD Biosciences), or cyanin 5-conjugated donkey anti-rat (for DAT, 1:200; Jackson ImmunoResearch, West Grove, PA, http://www.jacksonimmuno.com). After washing three times with PBS, samples were analyzed with a Nikon Diaphot fluorescence microscope (Nikon Instruments, Inc.).
Quantification of Dual-Labeled Cells
For in vitro differentiation experiments, we analyzed 200-500 cells for GFP, DAT, and TH colocalization in each condition from multiple wells of three independent experiments. For mouse brain sections, every fourth section of the entire brain from olfactory bulb to spinal cord was examined for GFP expression. For double-labeling experiments, we examined 200-600 GFP-positive cells for double staining under ×20 objectives. Multiple sections from three animals were examined.
Western Blotting Analysis of DAT-GFP Knock-In Mice
Mouse striatal and mesencephalic tissues were dissected from 4-, 8-, and 12-week-old WT, heterozygous, and homozygous DAT-GFP mice . The tissues were dissociated in a buffer composed of 50 mM Tris-HCl, pH 7.4, 10 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, and a cocktail of protease inhibitors (Roche). Thirty micrograms of soluble protein was separated on 12% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. After blocking nonspecific binding, membranes were incubated with mouse anti-GFP antibody (1-2000; Invitrogen), rat anti-DAT antibody (1-1000; Millipore), and mouse anti-β-actin (1-4000; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Horseradish peroxidase-conjugated goat anti-mouse or donkey anti-rat IgG (1-10,000; Jackson ImmunoResearch) was used as the secondary antibody, followed by chemiluminescent detection (Perkin Elmer Life Sciences, Waltham, MA, http://www.perkinelmer.com) [36, 37]. Three animals were used at each specified age. For quantification, Western blots were scanned into the computer, and ImageJ software (National Institutes of Health, http://rsb.info.nih.gov/ij) was used for band intensity measurement.
Open-Field Activity Test
Mice (five animals per group) were tested in 50 × 50-cm square boxes with clear Plexiglas walls and floors, evenly illuminated by white overhead fluorescent lighting. Mice were individually placed in the center of the open field and left to freely explore for 30 minutes. Activity was measured by a computer-assisted Photobeam Activity System with Flexfield (San Diego Instruments, San Diego, http://www.sandiegoinstruments.com). The total distance traveled and the total movement times were recorded in 5-minute intervals. Data were analyzed using multivariate analysis of variance and the Fisher's least square difference post hoc test. Significance was set at p < .05. Values are shown as mean ± SEM.
Primary Dopamine Neuron Culture
Primary cultures derived from E15 mouse ventral mesencephalon were prepared in Ca2+/Mg2+-free Hanks' balanced salt solution by mechanically dispersing tissue pieces with a 1-ml pipet as previously described. Subsequently, cells were centrifuged at 200g for 5 minutes and resuspended in F12 medium containing 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 μg/ml streptomycin, 100 U/ml penicillin, and 50 μg/ml ascorbic acid. Cells were seeded on polyethylenimine-coated (1 mg/ml; Sigma-Aldrich) 12-well plates in 1 ml of medium at 100,000 viable cells per square centimeter. Culture medium was changed every 2 days.
Dopamine and DOPAC Measurement by High-Performance Liquid Chromatography
To detect dopamine (DA) and 3,4-dihydroxyphenylacetic acid (DOPAC) in the culture medium, culture medium was collected and stored at −70°C. This medium contained ascorbic acid (50 μg/ml). After 4 days of culture, 22 μl of stock perchloric acid (final concentration of 0.2 M) was added to 1 ml of combined culture medium and centrifuged at 15,000g for 15 minutes at 4°C. An aliquot (20 μl) of the supernatant solution was analyzed by high-performance liquid chromatography (HPLC) equipped with an electrochemical detector (CoulArray system ESA Model 5600; ESA, Boston, http://www.esainc.com), a pump (ESA Model 580) set at 1.5 ml/minute, an automatic injector (ESA Model 540), and a catecholamine column (3 μm, 100 × 4.6 mm; Waters, Milford, MA, http://www.waters.com). The mobile phase was composed of 100 mm citric acid, 2% methanol, 1 mm EDTA, and 5 mg/L sodium octyl sulfate (pH 3.0). Electrochemical detector potentials were set at 0/50/200/300 mV (vs. Palladium). The results were expressed as picomoles of catecholamine per milliliter of medium.
Dopamine Neuron Purification from Embryonic Brain by FACS
We set up breeding using DAT-GFP homozygous males and C57BL/6 WT females, so that all embryos are heterozygous for DAT-GFP. Ventral mesencephalon was dissected from E15 fetal brain as described previously [38–40]. The dissected tissues were mechanically dissociated in Hanks' buffer using a 1-ml pipet. After centrifugation at 200g for 5 minutes, the supernatant was discarded, and the cell pellet was dissolved in Hanks' buffer and filtered through a 40-μm strainer (BD Bioscience, San Jose, CA, http://www.bdbiosciences.com). The cells were adjusted to a final concentration of 5 × 106 cells per milliliter. Cells were sorted based on endogenous GFP expression without immunostaining. FACS was conducted with FACSAria (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bdbiosciences.com) at the Clinical Immunology Flow Cytometry Core at University of Colorado Denver. FACSDiva software (Becton, Dickinson and Company) was used for acquiring samples. After initial analysis of approximately 5,000 cells, scatter gate, side scatter gate, and forward scatter gate were set to include just healthy cells. Gate P1 was set in fluorescein isothiocyanate channel (excitation wavelength at 488 nm) with fluorescent intensity greater than 2 log. GFP-positive cells were collected in cold DMEM supplemented with 10% FBS.
Coculture of Purified Dopamine Neurons with Mouse Striatal Cells
Two days before FACS separation, tissue culture plates were prepared for dopamine cell coculture on a supporting striatal cell base. Mouse striatum was dissected from E15 WT embryos [39, 40]. Tissues were mechanically dissociated and seeded on polyethylenimine (1 μg/ml; Sigma-Aldrich)-coated 8-well chamber slide in a medium consisting of F12, 15% FBS, 2 mM L-glutamine, and 100 units/ml penicillin/streptomycin. Immediately after FACS isolation, purified GFP-positive dopamine neurons were seeded onto the mouse striatal cell base. Dopamine neurons were cocultured with striatal cells for 7 days and fixed with 4% paraformaldehyde. Cultures were immunostained for TH and DAT antibodies using the procedure described above.
Generation of DAT-GFP ESCs and Knock-In Mice
The DAT-GFP knock-in ESC clones were identified by PCR screening and confirmed by Southern blotting. Homologous recombinant clones produced a 2.2-kb PCR product using primer pairs shown in Figure 1A (P1 and P2). Three clones were found positive by PCR (data not shown). With a 3′ external probe, we identified two positive clones (D8 and G4) by Southern blot (Fig. 1B). The G4 DAT-GFP ESCs were used to generate knock-in mice by blastocyst injections. Germ-line transmission of DAT-GFP was confirmed by PCR and Southern blotting from tail DNA of progeny. For genotyping, PCR was used to detect the presence of the GFP gene (800 base pairs). A typical PCR result is shown in Figure 1C. Breeding between heterozygous male and female mice generated homozygous DAT-GFP knock-in mice. Homozygous males were fertile so that uniformly heterozygous litters could be generated by breeding with WT C57BL/6 females.
In Vitro Differentiation of DAT-GFP Knock-In ESCs
We first used the PA6 cell method to induce dopamine neurons from DAT-GFP ESCs. By days 10-12, GFP-positive cells were seen in differentiating colonies (Fig. 2A–2C). By days 14-16, we observed many GFP-positive neurons with neurites growing out from the colonies (Fig. 2D). With TH immunostaining, we found that 92% of these GFP-positive cells were also TH positive (n = 425; Fig. 2E, merged in 2F). By days 18-20, we found clusters of GFP-positive neurons showing long neurites (Fig. 2G). With DAT immunostaining, approximately 95% of these GFP-positive cells were colocalized with DAT (n = 386; Fig. 2H, merged in 2I). Both D8 and G4 DAT-ESC lines showed similar results. We next used the embryoid body method to generate a monolayer of differentiated cells from the G4 DAT-ESC line. The intrinsic GFP signal was weak after 16 days of differentiation (data not shown). After immunostaining with an antibody directed against GFP, large numbers of GFP-positive cells were seen throughout the monolayer culture (Fig. 2J). When we double stained for TH, we found 80% of GFP-positive cells were also positive for TH (n = 291; Fig. 2K, merged in 2L). These results show that the DAT-GFP knock-in ESCs can produce GFP-positive, tyrosine hydroxylase-positive dopamine neurons in vitro.
Expression of GFP in DAT-GFP Knock-In Mouse Brain
We examined GFP expression during brain development in heterozygous DAT-GFP knock-in mouse. At E13, we could not see GFP fluorescence in mesencephalon (data not shown). At E15, we could see weak green fluorescence on both sides of dissected ventral mesencephalon (supporting information Fig. 1). At E18, we could see stronger green fluorescence in dissected mesencephalon (data not shown). In postnatal DAT-GFP knock-in mouse brain (heterozygous and homozygous mice, at 4, 8, and 16 weeks old), we examined brain sections for GFP expression. In coronal sections of midbrain, we observed large numbers of GFP-positive cells located in substantia nigra. A sample from 8-week heterozygous mouse brain is shown in Figure 3A. There were no differences in GFP intensity from mice of different ages. We found that mice homozygous for DAT-GFP showed much more intense GFP signal in midbrain dopamine neurons than heterozygous mice. After GFP antibody staining, the distribution of GFP-positive cells precisely matched the expected anatomic distribution of ventral midbrain dopamine neurons (Fig. 3B). In sagittal sections of DAT-GFP knock-in mouse brain, we found large number of GFP-positive neurons in substantia nigra (Fig. 3C). Interestingly, the forebrain bundles projecting from substantia nigra to striatum were also positive for GFP (arrows in Fig. 3D).
GFP-Positive Cells Were Colabeled with Dopamine Neuron Markers
We examined whether GFP-positive cells are truly authentic dopamine neurons. Brain sections from DAT-GFP heterozygous knock-in mice (at 4, 8, and 16 weeks old) were dual immunostained for GFP, DAT, Pitx3, En-1, and TH. In substantia nigra, we observed 98% of GFP-positive cells were also positive for DAT (n = 525; Fig. 4A–4C). Approximately 93% of GFP-positive cells had coexpression of Pitx3 (n = 458; Fig. 4D–4F). We also found that 90% of GFP-positive cells were colabeled with En-1 (n = 472; Fig. 4G, 4I). In addition, more than 92% of GFP-positive cells were stained for TH (n = 512; Fig. 4J–4L). In posterior hypothalamus, we found that GFP-positive cells were colocalized with TH in the dopamine neurons of the arcuate nucleus (Fig. 4G–4I). In septum, noradrenergic neurons are present, whereas dopaminergic neurons are absent. We observed that septal noradrenergic neurons were TH positive but were GFP negative, as expected for neurons that lack the dopamine transporter (Fig. 4P, 4Q). In a second noradrenergic nucleus, the locus coeruleus, TH-positive neurons were identified that did not express GFP (data not shown). The results demonstrate that GFP is faithfully expressed in dopaminergic neurons but not in noradrenergic neurons of DAT-GFP mice.
Expression of GFP in Olfactory Bulb
Because olfactory bulb also has dopamine neurons, we examined GFP expression in olfactory bulb sections. In wild-type mouse olfactory bulb, TH-positive neurons were clearly identified as periglomerular cells (Fig. 5A). In sections of olfactory bulb from heterozygous DAT-GFP knock-in mouse, GFP-positive cells were observed surrounding the glomerular layer (Fig. 5B, ×20 objective, and Fig. 5C, ×40 objective). With GFP and DAT dual immunostaining, approximately 97% of GFP-positive cells were colabeled with DAT (n = 232; Fig. 5D–5F). In addition, we found that 96% of GFP-positive cells were colocalized with TH (n = 275; Fig. 5G–5I). These results indicate that GFP is expressed in dopamine neurons of olfactory bulb.
Heterozygous DAT-GFP Mice Have Normal DAT Function and Homozygous DAT-GFP Mice Show Hyperactivity and Upregulation of GFP
To test whether heterozygous mice have normal DAT function, we cultured mesencephalon from E15 heterozygous and homozygous DAT-GFP knock-in as well as wild-type embryos. After 4 days of culture, medium was collected for DA and DOPAC measurement by HPLC, and cells were fixed for immunostaining. In heterozygous DAT-GFP knock-in mesencephalic cultures, we found GFP-positive cells with long neurites (Fig. 6A). Nearly 100% of GFP-positive cells were colabeled with DAT (n = 366; Fig. 6B, merged in 6C). We found that DA and DOPAC levels were similar between heterozygous DAT-GFP and wild-type dopamine neurons (p = .23; Fig. 6D), whereas DA levels, but not DOPAC, were significantly increased in homozygous DAT-GFP dopamine neurons compared with heterozygous and WT dopamine neurons (*, p < .05 for both comparisons, Fig. 6D). Results indicate heterozygous DAT-GFP dopamine neurons have normal dopamine transporter function.
The insertion of GFP leads to the disruption of DAT function in knock-in mice so that the homozygous DAT-GFP knock-in mice are equivalent to DAT-null mice. As expected for animals that cannot remove dopamine from the synaptic cleft, the homozygous mice showed hyperactivity in open-field testing, traveling twice the distance and twice the duration of WT and heterozygous mice (GFP +/−) (Fig. 6E, 6F, **, p < .01 for comparison between homozygous and heterozygous or homozygous and WT). Loss of DAT expression in the striatum of homozygous mice (GFP/GFP) was confirmed by Western blotting for DAT (Fig. 6G). The DAT protein level in heterozygous mice (GFP +/−) was similar to WT (Fig. 6G, 6H, p = .41). Western blot analysis in mouse mesencephalon showed that GFP expression was more than doubled in homozygous (GFP +/+) compared with heterozygous (GFP +/−minus]) mice, (Fig. 6G, 6I, **, p < .01). Although part of the increase in GFP signal can be accounted for by expression of two copies of the gene in homozygous animals and only one in heterozygotes, the loss of DAT function in homozygous knock-in mice may have led to feedback-induced overexpression at the DAT promoter, resulting in increased expression of GFP.
Dopamine Neurons Can Be Purified by FACS and Survive in Tissue Culture
To demonstrate the value of DAT-GFP mice for dopamine neuron purification, we dissected E15 midbrain tissues from DAT-GFP heterozygous embryos. The dissociated cells were analyzed and sorted by FACSAria (Fig. 7A). Approximately 2% of analyzed cells were positive for GFP as expected for the fraction of ventral midbrain cells that are dopaminergic (Fig. 7B). When the collected cells were resorted, approximately 90% of cells were positive for GFP (Fig. 7C). We tested whether FACS-purified dopamine neurons can survive in standard primary culture conditions. FACS-purified dopamine neurons were cultured on polyethylenimine-coated plates (1 μg/ml; Sigma-Aldrich) in dopamine neuron culture medium (see Materials and Methods) supplemented with 10 ng/ml of glial cell line-derived neurotrophic factor. However, by day 3, more than 95% of cells died and lifted from the surface (data not shown). When FACS-purified dopamine neurons were seeded on glial cell layers derived from any brain areas (cortex, ventral mesencephalon, or striatum of E15 WT embryos), dopamine neurons survived well and grew with long neurites, similar to results in Figure 6A–6C. We also cultured FACS-purified dopamine neurons with striatal cells derived from E15 WT embryos. After 7 days in culture, we observed many GFP-positive neurons showing long neurites (Fig. 7D, 7G). With dual immunostaining for TH (Fig. 7E) and DAT (Fig. 7H), we confirmed that these GFP-positive cells coexpressed TH (Fig. 7F) and DAT (Fig. 7I). These results indicate that dopamine neurons can be FACS purified from DAT-GFP knock-in mice and can survive if cocultured with astrocytes or striatal cells.
We have knocked GFP into the DAT locus of mouse ESCs and have shown that those cells can be differentiated to authentic dopamine neurons in vitro. The DAT-GFP ESCs were used to create knock-in mice with specific expression of GFP in dopaminergic neurons. In both heterozygous and homozygous DAT-GFP knock-in mice, the GFP gene expression pattern is highly specific to DAT-expressing neurons. GFP is present in midbrain dopamine neurons in A9 (substantia nigra), A10 (ventral tegmental area), A11 (posterior hypothalamus arcuate nucleus), and A16 (olfactory bulb), but not in TH-positive noradrenergic neurons in septum and locus coeruleus. Our mouse model is an improvement over transgenic mice in which GFP is under control of the TH promoter. In those transgenic mice, many cells express GFP nonspecifically and in ectopic sites [35, 41]. In a different knock-in mouse with GFP knocked into the Pitx3 locus, GFP expression was present in midbrain dopamine neurons but also in lens, head mesenchyme, muscles in the cranial facial region, and tongue, as expected for this broadly expressed gene . The high fidelity of GFP expression in dopamine neurons in our DAT-GFP knock-in mouse makes it a unique model.
We have performed in vitro differentiation of knock-in ESCs using both PA6 cell and embryoid body methods. With PA6 cell induction, we found that colonies expressed GFP after 14-16 days of differentiation. When we tried to dissociate the colonies and isolate GFP-positive cells, we found that the colonies were densely packed and were difficult to break down into single cells. Therefore, we tested the alternative method of embryoid body formation followed by dissociation and differentiation on attachment cultures. This method creates a relatively uniform layer of differentiated cells. Unfortunately, GFP intensity in attachment cultures was weaker than in PA6 cell cultures and was not bright enough for FACS separation (data not shown). Nonetheless, we were able to demonstrate GFP-positive cells in these cultures by amplifying signal after immunostaining with an antibody to GFP. Weak GFP expression using the embryoid body method indicates that the resultant dopamine neurons have low expression levels of DAT. It is possible that these dopamine neurons are immature cells, since DAT expression increases as cells mature [43, 44]. As we showed with dopamine neurons isolated from DAT-GFP mouse mesencephalon, coculturing with striatal cells may provide a more supportive environment for dopamine neuron survival and maturation as others have reported [45, 46].
To demonstrate the usefulness of DAT-GFP knock-in mice, we successfully purified embryonic dopamine neuron by FACS. At embryonic day 15, the GFP-positive cells were present in ventral mesencephalon, and the GFP intensity was bright enough for cell sorting. Importantly, these purified dopamine neurons could survive and develop a mature phenotype when cocultured on astrocytes or striatal cell layers. These FACS-purified dopamine neurons could not survive alone in tissue culture, probably because of a lack of trophic factor support provided by astrocytes or target cells. Similar dependence on astrocyte coculture has been reported after FACS sorting of pTH-GFP cells .
Our DAT-GFP knock-in mice make it possible to isolate pure populations of dopamine neurons without the complications of ectopic GFP expression. Because GFP-positive dopamine neurons can be visualized directly in live tissue slices and cultures, DAT-GFP mice will allow study of the physiology and pharmacology of dopamine neurons in living preparations. With high confidence that only dopamine neurons are being analyzed, gene microarray analysis of DAT-GFP-positive cells should yield precise characterization of genes uniquely expressed in dopamine neurons.
Human dopamine neurons created from human ESCs must be separated from residual stem cells if they are to be transplanted safely into patients with Parkinson's disease [31, 33]. Because these human dopamine neurons will not have a fluorescent label like GFP, other cell purification strategies must be developed. Antibodies that can bind to unique cell surface proteins could be used to separate cells, if novel cell surface proteins can be found on dopamine neurons. Although DAT is such a protein, our previous studies have shown that extracellular glycosylation of DAT prevents development of effective antibodies . Gene microarrays of our DAT-GFP dopamine neurons may identify new cell surface proteins unique to dopamine neurons and accessible to extracellular antibodies.
Homozygous DAT-GFP knock-in mice are equivalent to DAT knockout mice. We observed hyperactivity in homozygous DAT-GFP knock-in mice, which has been reported in DAT-null mice [49–51]. In heterozygous DAT-GFP knock-in mice, the DAT level is similar to that in WT mice. We have shown that DAT function is normal in heterozygous dopamine neurons. Interestingly, we found upregulation of GFP in homozygous DAT-GFP knock-in mice compared with GFP levels in heterozygous mice to 237% of heterozygote levels. Because homozygous mice have two copies of the GFP gene, much of the increased expression is accounted for by the second gene copy, although some fraction may be due to upregulation of DAT-GFP synthesis in homozygous mice. These results indicate that there may be feedback upregulation of DAT expression in the absence of DAT, similar to previous reports that cocaine can upregulate DAT .
DAT-GFP ESCs and the DAT-GFP knock-in mice created from them provide unique tools for dopamine neuron identification in live cells that can be purified by FACS. These GFP-labeled dopamine neurons will be useful for gene expression profiling as well as physiologic and pharmacologic studies.
The mouse DAT genomic DNA construct was kindly provided by Xiaoxi Zhuang, Ph.D., University of Chicago. We thank Kim Ellison for technical assistance, Dr. Liping Liang for HPLC measurement of dopamine and DOPAC, and the University of Colorado Denver Transgenic Core for producing knock-in mice. The study was supported by the American Parkinson Disease Association and the Leopold Korn and Michael Korn Professorship in Parkinson's Disease.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.