μ- and κ-Opioids Induce the Differentiation of Embryonic Stem Cells to Neural Progenitors*

Growth factors, hormones, and neurotransmitters have been implicated in the regulation of stem cell fate. Since various neural precursors express functional neurotransmitter receptors, which include G protein-coupled receptors, it is anticipated that they are involved in cell fate decisions. We detected μ-opioid receptor (MOR-1) and κ-opioid receptor (KOR-1) expression and immunoreactivity in embryonic stem (ES) cells and in retinoic acid-induced ES cell-derived, nestin-positive, neural progenitors. Moreover, these G protein-coupled receptors are functional, since [d-Ala2,MePhe4,Gly-ol5]enkephalin, a MOR-selective agonist, and U69,593, a KOR-selective agonist, induce a sustained activation of extracellular signal-regulated kinase (ERK) signaling throughout a 24-h treatment period in undifferentiated, self-renewing ES cells. Both opioids promote limited proliferation of undifferentiated ES cells via the ERK/MAP kinase signaling pathway. Importantly, biochemical and immunofluorescence data suggest that [d-Ala2,MePhe4,Gly-ol5]enkephalin and U69,593 divert ES cells from self-renewal and coax the cells to differentiate. In retinoic acid-differentiated ES cells, opioid-induced signaling features a biphasic ERK activation profile and an opioid-induced, ERK-independent inhibition of proliferation in these neural progenitors. Collectively, the data suggest that opioids may have opposite effects on ES cell self-renewal and ES cell differentiation and that ERK activation is only required by the latter. Finally, opioid modulation of ERK activity may play an important role in ES cell fate decisions by directing the cells to specific lineages.

The maintenance of ES cells in an undifferentiated state in vitro is dependent on self-renewing cell division in the presence of leukemia inhibitory factor (LIF), 2 which signals through var-ious receptor complexes (reviewed in Refs. 1-3; see also Refs. 4 and 5). Mouse ES cells can be induced to differentiate into neural cells in the presence of retinoic acid (RA) in vitro (reviewed in Refs. 6 and 7). During this induction, ES cells undergo a series of steps that resemble key stages in the early mouse embryo, supporting the hypothesis that the in vitro pathway represents the normal developmental pathway (8,9).
Considerable effort has been recently devoted to characterizing intrinsic factors, extrinsic factors, and signaling pathways regulating proliferation and differentiation of stem cells. The ERK/MAP kinase signaling pathway has been implicated in both proliferation and differentiation of many cell types, including stem cells (10). In several studies, ERK activation had a negative influence on self-renewal/proliferation in murine ES cells (4,(11)(12)(13)(14). Moreover, LIF-dependent activation of STAT3 was not mediated by ERK activity (15). However, a dual role for the Ras/ERK pathway was proposed for ES cells, affecting both their division and differentiation (16). Ras activation down-regulated levels of Nanog, a protein that is normally expressed in high amounts in self-renewing ES cells (17). A functional role for ERK was proposed in the phosphatidylinositol 3-kinase-dependent regulation of ES cell self-renewal (18). The importance of ERK activation and its duration for cell differentiation has been investigated in various cells (19 -22). Some evidence suggests that RA-induced inhibition of LIF signaling in ES cells is ERK-independent (23). However, a specific requirement of the ERK pathway was reported for RA-dependent commitment of murine ES cells (24). Cross-talk may occur by noncanonical actions of RA with the MAP kinase phosphorylation cascade. RA-induced differentiation of ES cells may be achieved by restricting nuclear entry of activated ERK (25). ERK activation is required for regulation of cyclin D1 levels, the increase of which parallels ES cell differentiation (26,27).
Neurotransmitters as well as growth factors and hormones can influence cell division and differentiation of self-renewing stem cells and neural progenitors (NPs) via ERK in some cases (28,29). For example, stimulation of the muscarinic acetylcholine receptor, a G protein-coupled receptor, induces DNA synthesis in rat cortical neuroepithelium progenitors via stimulation of phosphatidylinositol 3-kinase and ERK (30). The activated CB1 receptor inhibits neuronal progenitor differentiation via attenuation of ERK signaling in E17 cortical cultures and in adult dentate gyrus (31). This endocannabinoid system was also found to promote astroglial differentiation by acting on neural progenitor cells (32). Glutamate activates N-methyl D-aspartate (NMDA) receptors and promotes proliferation of E15 precursors derived from the germinal zone of the ventral telencephalon in vivo and in vitro (33). Haloperidol acting via dopamine D2 receptors increases the number of NPs, neurons, and glia in the adult rat brain (34). G protein ␤␥ subunits of heterotrimeric G proteins are required for proper mitotic-spindle orientation and asymmetric cell fate decisions of cerebral cortical progenitors (35). These studies were performed with brain-derived stem cells or NPs, but little is known about the consequences of G protein-coupled receptor-ERK cross-talk in blastocyst-derived ES cells or their NPs.
Since the plasticity of uncommitted stem cells has opened new perspectives in tissue regeneration, recent research has been directed to understanding the signaling mechanisms that control proliferation and/or differentiation of ES cells. Here, we detectedand -ORs in ES cells and in ES cell-derived NPs. More importantly, we found that both opioids induced ES cell differentiation via ERK and an ERK-independent attenuation of proliferation in RA-driven NPs.

Mouse ES Cells and Their Growth Conditions-D3 (ATCC)
and PRX-129/S6 # 7 ES cells (Primogenix, Inc.) were used. PRX-129/S6 # 7 ES cells were isolated from the inner cell mass of a day 3.5 129/S6/SvEv mouse blastocyst. Cells have a normal male karyotype and are specific pathogen-free. They were injected into blastocysts and produced chimeras at a high rate (first injection: 10/15 pups, no perinatal death). PRX-129/S6 # 7 cells were grown on mouse embryonic fibroblasts in the presence of LIF, passage 10 cells were propagated 2-3 times on gelatin-coated flasks, and these cells are currently used in the laboratory. Quality control and characterization of the D3 ES cells are documented by their depositor (36). D3 ES cells were propagated on irradiated STO cell feeders (ATCC) for several passages before they were transferred to gelatin-coated flasks.
D3 and PRX-129/S6 # 7 ES cells were maintained in Dulbecco's modified Eagles medium containing 10% fetal calf serum, LIF (100 ng/ml), and 0.1 mM 2-mercaptoethanol (ME) on gelatincoated dishes. These growth conditions are known to prevent ES cells from differentiating (6,37,38). Here, cells grown under these conditions will be described as "undifferentiated, self-renewing ES cells." The neural induction of ES cells entails the "4Ϫ/4ϩ protocol," which has two phases (8,39). In the first phase, cells are cultured in nonadhesive, bacterial grade Petri dishes in serumcontaining media (without LIF and ME) in the absence of RA for 4 days. During this phase, embryoid bodies of different sizes appear. In the second phase, cells are grown in RA (1 M) and serum-containing media for an additional 4 days. Differentiation of these cells is also a two-phase process: 1) early differentiation, during which nestin-positive cells develop, and 2) their terminal differentiation, the final step in ES cell growth. For this purpose, embryoid bodies are dissociated and cultured on tis-sue culture plates in serum containing Dulbecco's modified Eagle's medium for an additional 2-5 days (short term differentiation) or for 5-12 days (long term differentiation). The RAinduced, short term differentiated cells are referred to as NPs.
Opioid Treatments-ES cells or ES cell-derived progenitors are maintained in their corresponding growth media. Opioid agonists (0.1-1 M; DAMGO, a MOR-selective agonist, or U69,593, a KOR-selective agonist) were added for various times (up to 24 h). In some experiments, cells were preincubated with opioid antagonists (1 M, 60 min; CTAP (MOR-selective) or nor-BNI (KOR-selective)) and then treated with the corresponding opioid agonist (0.1 M) for a given time.
Measurement of ERK Activity-For these experiments, cells were grown in 6-well plates. In most cases, ES cells were maintained in serum-containing media upon opioid treatment. In contrast, ES cell-derived NPs were grown in media deprived of serum for 24 h before treatment with opioids. ERK phosphorylation was measured by performing immunoblotting experiments with a phospho-ERK Ab that recognizes the active form of ERK (40 -42). Cells were washed with PBS and lysed with buffer (20 mM HEPES, 10 mM EGTA, 40 mM ␤-glycerophosphate, 2.5 mM MgCl 2 , 2 mM sodium vanadate, 1% Nonidet-40, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, and 20 g/ml leupeptin) and spun at 14,000 ϫ g, and protein concentration of the supernatants was determined. Cell lysates (10 -20 g of protein/lane) were separated by 10% SDS-PAGE. Proteins were blotted on Immobilon membranes. Nonspecific sites were blocked with 5% milk in Tris-buffered saline plus 0.2% Tween 20. Blots were incubated with monoclonal phospho-ERK Ab (1:2000; Cell Signaling) for at least 15 h at 4°C, followed by incubation with a horseradish peroxidase-conjugated IgG (1:2000; Sigma) for 1 h at room temperature. Bands were visualized with an ECL chemiluminescence detection system (GE Healthcare) and exposure to Classic Blue sensitive x-ray film. Band intensities were determined by densitometry with an Eastman Kodak Co. DC120 digital camera, Kodak ds 1D version 3.0.2 (Scientific Imaging Systems), and NIH Image ImageJ version 1.32 software. ERK stimulation in opioidtreated cells was expressed as -fold change over basal levels in control cells.
OR and Nestin Immunofluorescence Microscopy-ES cells or ES cell-derived NPs were grown in glass chamber slides (Nunc). The cells were fixed in 4% PA for 20 min at room temperature and permeabilized in 0.1% Triton/PBS for 5 min. Cells were then incubated in PBS containing 0.5% bovine serum albumin and 0.1% Tween 20 for 30 min to reduce nonspecific binding, followed by overnight incubation at 4°C with the following rabbit polyclonal OR Abs: MOR-1 raised against C terminus (1:2500; Neuromics); MOR-1 raised against N terminus (1:50; Santa Cruz Biotechnology, Inc., Santa Cruz, CA); KOR-1 (1:50; Santa Cruz Biotechnology). In some cases, a mouse monoclonal nestin Ab (1:1000; Chemicon) was added together with the OR Abs. After washing, Alexa Fluor 594-conjugated (red) and/or Alexa Fluor 488-conjugated (green) secondary Abs (1:1000) were applied for 1 h at room temperature. DAPI (1:200) was added together with the secondary Abs. The slides were treated with anti-fade reagent (Molecular Probes, Inc., Eugene, OR) and examined for immunofluorescence with NIKON-OP-TIPHOT-2 or OLYMPUS AH3 microscopes with simultaneous recording of dual fluorescence label images.
Real Time Quantitative Reverse Transcriptase (qRT)-PCR-Total RNA from ES cells and RA-induced NPs was isolated using the Qiagen RNeasy minikit (Qiagen, Valencia, CA) according to the manufacturer's instructions. cDNA was generated from 1 g of total RNA using the High Capacity cDNA archive kit from Applied Biosystems (Foster City, CA) with random primers as described by the manufacturer. qRT-PCR was performed with SYBR green chemistry on an Applied Biosystems 7500 instrument. The primers were designed using Primer Express (Applied Biosystems) and supplied by Integrated DNA Technologies (Coralville, IA). Primer sequences for MOR-1 were as follows: forward, 5Ј-CCACTAGCACGCT-GCCCTT-3Ј; reverse, 5Ј-GCCACGTTCCCATCAGGTAG-3Ј. Primer sequences for KOR-1 were as follows: forward, 5Ј-AG-AGAGAGAAGCGGCAAGCA-3Ј; reverse, 5Ј-GCCAAGGCT-CACTAACTCCAA-3Ј. The cDNA templates for qRT-PCR were diluted 1:10, and the 50 l SYBR-green reaction consisted of 1ϫ SYBR-green Master Mix (Applied Biosystems), 300 nM forward and reverse primers, and 5 l of diluted cDNA. Four replicates of each qRT-PCR reaction were run on 96-well plates. The amplification efficiencies of MOR-1 and KOR-1 were consistent with those of the endogenous control, glyceraldehyde-3-phosphate dehydrogenase. Relative quantification measurements were made as described (43) by using the comparative C T method (⌬⌬C T method) in which the gene C T values are normalized by subtracting glyceraldehyde-3-phosphate dehydrogenase C T .
SOX-1 Immunostaining and Counting-D3 ES cells were grown in chamber slides, in serum/LIF/ME-containing media, and were treated with 1 M DAMGO or U69,593 for 24 h. In some cases, cells were pretreated for 1 h either with MOR antagonist (CTAP, 1 M) or KOR antagonist (nor-BNI, 1 M) before the addition of opioid agonists for 24 h. Control and opioid-treated ES cells were washed with PBS and fixed with 2% PA for 10 min at room temperature. Cells were then permeabilized with 0.4% Triton/PBS for 5 min and incubated in PBS containing 10% fetal calf serum and 0.4% Triton X-100 for 30 min to reduce nonspecific binding, followed by incubation with chicken polyclonal SOX-1 Ab (1:500; Chemicon) overnight at 4°C. After washing, Alexa Fluor 594 (red)-conjugated antichicken secondary Ab (1:1000; Molecular Probes) was applied for 1 h at room temperature. Cells were counterstained with DAPI (1:200) to visualize nuclei for cell counts. Chamber slides were treated with anti-fade reagent (Molecular Probes) and examined for immunofluorescence as described above. NIH Image J version 1.32 software was used to count cells. The percentage of SOX-1-positive cells was estimated from the ratio between the total number of cells (DAPI-stained) and the number of SOX-1-stained cells. Cells from 3-5 fields/slide were counted, and 3-4 slides/treatment group were used for the counting. The total number of counted cells was 500 -800/ treatment group.
Cell Proliferation Assays-For these studies, equal numbers of cells were seeded per well in 12-well dishes. Cell numbers were estimated by serial dilutions and counting cells (volume of 10 l) in several visual fields using a bright line hemacytometer and light microscopy. Basal levels or opioid-induced changes in DNA synthesis were assessed by measuring the rate of [methyl-3 H]thymidine incorporation into cells. Cells were grown in media with or without opioids for 24 h. In some experiments, the MEK-selective inhibitor U0126 (1 M) was present for 28 h, whereas opioid agonists were added for 24 h, and [methyl-3 H]thymidine (0.02 Ci/ml) was present for the last 4 h. [methyl-3 H]Thymidine incorporation was measured as described in our previous studies with some modifications (44). Briefly, cells were washed with PBS, followed by incubation with 5% trichloroacetic acid at 4°C for 30 min. Then cells were collected in 2% NaHCO 3 /0.1 N NaOH, and [methyl-3 H]thymidine incorporation was determined by liquid scintillation counting.
BrdUrd Labeling of Cells-D3 ES cells or RA-induced NPs were grown until they reached about 50% confluence. They were then treated with either 1 M DAMGO or U69,593 for 24 h. In some cases, cells were pretreated with inhibitors or opioid antagonists as described in the figure legends. At the end of the treatment, medium was replaced with BrdUrd labeling medium (10 M; BrdUrd detection kit I; Molecular Probes), and cells were incubated for 20 -30 min. After several washes, cultures were fixed with ethanol fixative for 20 min at Ϫ20°C. For nestin co-labeling, cells were incubated together with a monoclonal BrdUrd Ab (1:10; Roche Applied Science kit) and polyclonal nestin Ab (1:1000; Covance) at 37°C for 30 min. For Sox-1 co-labeling, cells were first incubated with a monoclonal BrdUrd Ab (1:10; Roche Applied Science kit) at 37°C for 30 min and then incubated in PBS containing 10% fetal calf serum and 0.4% Triton X-100 for 30 min to reduce nonspecific binding, followed by counterstaining with a chicken Sox-1 polyclonal Ab (1:500; Chemicon) overnight at 4°C. After washing, cells were treated with fluorescein-conjugated anti-mouse IgG (green, 1:10; Roche Applied Science kit) together with Alexa Fluor 594 (red)-conjugated anti-chicken or anti-rabbit secondary Abs (1:1000; Molecular Probes) for 1 h at room temperature. Slides were examined for immunofluorescence as described above.

Characterization of ES Cells and ES Cell-derived NPs-
The state of differentiation of D3 or PRX ES cells and RA-induced ES cell-derived NPs was characterized by immunocytochemistry and immunoblotting. For this purpose, D3 and PRX ES cells were maintained under serum/LIF/ME-induced self-renewal conditions, or they were differentiated in the presence of RA. For the immunocytochemistry experiments, cultures were stained with nestin (a neural progenitor marker) Ab and DAPI (a nuclear marker). As seen in double DAPI/nestin images (Fig.  1, D3 and PRX), there were no nestin (red)-stained cells among the undifferentiated ES cells, whereas RA-induced NPs show heavy staining with nestin ( Fig. 1, dD3 and dPRX).
Since the transcription factor Oct4 has been demonstrated to be vital for the formation of self-renewing pluripotent ES cells, cell lysates from undifferentiated D3 ES cells or RA-induced NPs were probed with an Oct4 Ab (1:100; Chemicon) as presented in Fig. 1B. The densitometry data from the Oct4 immunoblots suggest that there is about a 50% reduction in Oct4 protein levels in D3 ES cell-derived NPs (3767 Ϯ 85; n ϭ 3) in comparison with the levels of this transcription factor in undifferentiated ES cells (6954 Ϯ 463; n ϭ 4).

MOR-1 and KOR-1 Gene Expression and Immunoreactivity Was Detected in ES Cells and in ES Cell-derived NPs-Three
independent approaches were taken to determine the presence of MOR-1 and KOR-1 in undifferentiated D3 ES cells and RAinduced ES cell-derived NPs: qRT-PCR, immunoblot analysis of cell lysates, and immunofluorescence microscopy of cells at different stages of development. In qRT-PCR experiments, we found that in comparison with ES cells, NPs have 1.37-and 2-fold increases in MOR-1 and KOR-1 gene expression, respectively (n ϭ 3). As shown in Fig. 2, immunoblotting performed with polyclonal MOR-1 (C terminus; Neuromics)-, MOR-1 (N terminus; Santa Cruz Biotechnology)-, and KOR-1 (Santa Cruz Biotechnology)-specific Abs shows the presence of major 50 and 55 kDa bands that correspond to the expected molecular masses for MOR-1 and KOR-1, respectively. In addition to the 50 kDa band, the N terminus MOR-1 Ab reveals the existence of higher molecular mass bands (Fig. 2, lanes 6 and 7), which were suggested to correspond to the glycosylated form(s) of the receptor by several groups (45)(46)(47). Both MOR-1 and KOR-1 bands were detected in cell lysates, obtained from D3 ES cells maintained under self-renewal conditions (serum/LIF/ME). RA treatment does not appear to induce significant changes in MOR-1 or KOR-1 gene expression or protein levels ( Fig. 2) in D3 NPs. Brain homogenate was used as a positive control, and lysates from late passage immortalized rat astrocytes were used as negative controls, because these cells lose detectable amounts of ORs with passaging (42).
Immunofluorescence microscopic analysis further supports the occurrence of MOR-1 and KOR-1 in D3 ES cells and NPs (Fig. 3). Once again, the undifferentiated state of D3 ES cells is confirmed by the absence of nestin (green)-stained cells (D3; Fig. 3, E and G). In contrast, D3 ES cell-derived NPs (dD3) show heavy nestin staining (Fig. 3, F and H). The presence of MOR-1 (red) immunoreactive cells was confirmed by applying the two polyclonal MOR-1 Abs that were used for the immunoblotting experiments (Fig. 2). It appears that the N terminus MOR-1 Ab detects the receptor mainly on the cell surface of ES cells (Fig.  3B) and NPs (Fig. 3D). Fig. 3, E and F, shows MOR-1 immunostaining using the C terminus MOR-1 Ab. KOR-1 immunoreactive cells were detected in self-renewing D3 ES cells (Fig. 3G) and in NPs (Fig. 3H).  Opioid Regulation of ERK Phosphorylation in Undifferentiated ES Cells-Having found that MOR-1 and KOR-1 are present in both undifferentiated ES cells and in ES cell-derived NPs, we sought to determine the functionality of these receptors. Since MAP kinases have been implicated in regulation of cell proliferation and differentiation, we studied opioid regulation of ERK activation in ES cells and ES cell-derived NPs. MOR-or KOR-induced ERK activation was measured by performing immunoblotting experiments with a phospho-ERK Ab that recognizes the active form of ERK. Undifferentiated ES cells were grown in media devoid of serum/LIF/ME for at least 24 h. Thereupon, cells were treated with the selective opioid agonist DAMGO (0.1 M, 5 min) or the -selective opioid agonist U69,593 (0.1 M, 5 min). Both opioids stimulated ERK activity in D3 ES cells by about 2-fold (Fig. 4). Similar data were obtained with undifferentiated PRX ES cells (data not shown). The corresponding, selective MOR (CTAP) and KOR (nor-BNI) antagonists blocked agonist induced stimulation of ERK phosphorylation, supporting the notion that these are ORmediated effects.
Opioid effects on ERK signaling were also evaluated in cells grown in serum/LIF/ME-containing media, conditions that are expected to maintain D3 ES cells in an undifferentiated, selfrenewing state. Under these conditions, both DAMGO and U69,593 induced a sustained activation of ERK signaling throughout the 24-h treatment period (Fig. 5). For DAMGO, there was 2-fold or greater activation of ERK at all time points (Fig. 5B). Maximal phosphorylation was seen at 5 min and 12 h with each having greater than 4-fold stimulation of ERK. The ERK activation observed with U69,593 was also sustained for all time points (Fig. 5C). This activation was slightly less overall, although the greatest phosphorylation seen was over 6-fold at 5 min.
Although serum induces high basal levels of phosphorylated ERK in many cell types, ERK basal levels were low in ES cells grown in serum-containing media (Fig. 5A). The low basal levels of ERK phosphorylation in these cells support the notion that ERK activity may not be required for self-renewal of ES cells as recently reported (see Introduction).
Opioids Induce Limited Proliferation of Undifferentiated ES Cells via the ERK Pathway-To correlate the ERK data with cell proliferation, we studied opioid regulation of ES cell growth. In undifferentiated D3 ES cells, the findings are that DAMGO, U69,593, and morphine all induced limited proliferation. Specifically, DAMGO and morphine caused a 40% rise in proliferation over basal levels, whereas U69,593 caused a 30% increase (Fig. 6). The mitogen, epidermal growth factor (10 ng/ml), induced about a 100% increase in ES cell proliferation (n ϭ 3, p Ͻ 0.05). Epidermal growth factor stimulates ES cell proliferation via the ERK/MAP kinase signaling pathway (48). Since ERK mediates mitogen-induced cell proliferation, the MEKselective inhibitor U0126 was added to determine the role of ERK in opioid enhanced cell proliferation. Pretreatment of ES cells with U0126 blocked the opioid-induced increase of proliferation seen in the absence of the inhibitor, suggesting that opioids act through ERK/MAP kinase to increase proliferation in D3 ES cells (Fig. 6). In PRX ES cells, morphine also induced a 30% increase in cell proliferation over basal levels (data not shown). U0126 abolished this increase, suggesting that ERK is responsible for opioid-enhanced proliferation in undifferentiated PRX ES cells as well.
Although basal levels of proliferation in the presence and absence of U0126 were set at 1 to simplify the estimation of opioid effects (controls in Fig. 6), we found that U0126 pretreatment enhanced basal levels of D3 ES cell proliferation by about 40% (basal levels: 8726 Ϯ 1168 dpm versus basal in the presence of U0126: 12,005 Ϯ 1225 dpm). This finding supports the notion that undifferentiated ES cells (in the absence of opioids)   NOVEMBER 3, 2006 • VOLUME 281 • NUMBER 44 can proliferate even when ERK signaling is blocked and suggest that the ERK/MAP kinase signaling pathway may not be involved in ES cell self-renewal.

Opioids Induce ES Cell Differentiation
Evidence for Opioid-induced D3 ES Cell Differentiation-To determine whether the opioid-induced proliferation of ES cells is due to an increase in self-renewal or asymmetric cell division to form NPs, we examined the phenotype of the opioid-induced "newly formed" cells. Undifferentiated ES cells were treated with opioids and labeled with BrdUrd, followed by staining with nestin and BrdUrd Abs. The immunofluorescence analysis raises the possibility that DAMGO and U69,593 induce the appearance of nestin-positive cells that may be NPs derived from ES cells (Fig. 7A). The representative double images of control cells show mainly BrdUrd (green) staining and only a few nestin (red)-stained cells.
To determine further the nature of opioid-induced cells, we performed a quantitative analysis of Sox-1 immunostained, control, and opioid-treated ES cells (Fig. 7, B-D). Sox-1 is a selective marker of proliferating NPs; it belongs to a family of transcription factors that are expressed in ectodermal cells upon acquisition of neural progenitor identity (49,50). By counting cells (Ն1000 cells/treatment group), we estimated that only a few of the control cells (4 Ϯ 1.1%, n ϭ 8) were Sox-1-positive. Exposure of undifferentiated ES cells to 1 M MOR (Fig. 7B) or KOR (Fig. 7C) opioids for 24 h initiated the appearance of about 7-8-fold more Sox-1-positive cells than in control cells (Fig. 7, D and E). The corresponding, selective MOR (CTAP) and KOR (nor-BNI) antagonists blocked agonist-induced appearance of Sox-1-positive cells, suggesting that this is an OR-mediated process. CTAP (10 Ϯ 2.6%, n ϭ 4) and nor-BNI (2 Ϯ 0.5%, n ϭ 4), alone had no significant effect on the number of Sox-1-labeled cells.
Opioid Regulation of ERK Phosphorylation in RA-induced ES Cell-derived NPs-To investigate opioid modulation of ERK activity upon RA-induced differentiation of ES cells, RA-induced embryoid bodies were dissociated and plated in serum containing Dulbecco's modified Eagle's medium for an additional 2-5 days (short term differentiation) or for 5-12 days (long term differentiation). After growing in serum-deprived media for 24 h, these cells were treated with opioids for various times, and ERK phosphorylation was measured by immunoblotting (Fig. 8). The results support the following conclusions: 1) basal levels of ERK phosphorylation in long term differentiated cells were significantly higher (2.2 Ϯ 0.2, n ϭ 9, p Ͻ 0.05) than basal levels of ERK activation in short term differentiated cells (NPs), suggesting that terminal differentiation of these cells may require activated ERK; 2) MOR and KOR agonists elicit similar effects on ERK activation in short and long term differentiated cells (Figs. 8 and 9).
More detailed time course experiments with DAMGO and U69,593 were also conducted (Fig. 9). These data further confirm that the two opioids induce a similar biphasic ERK activation profile with peaks occurring at 2-15 min and at about 2-6 h in short and long term differentiated ES cells (Fig. 9, A-D). Thus, the opioid-induced sustained ERK activation seen in

undifferentiated ES cells changes in differentiated cells to a biphasic ERK activation profile.
Opioids Inhibit NP Proliferation-To evaluate opioid effects on RA-induced NP proliferation, we performed a quantitative analysis of Sox-1/BrdUrd double-labeled, control, and opioidtreated NPs (Fig. 10). Since Sox-1 is a selective marker of proliferating NPs, we counted the BrdUrd-labeled cells among the Sox-1-positive cells and thus estimated the changes in NP proliferation upon treatment with opioids. The data presented in Fig. 10 indicate that 1 M DAMGO or U69,593 decreased proliferation of NPs by 50 -60% in 24 h. The inhibitory effect of both opioid agonists was reversed by the corresponding MOR (CTAP) and KOR (nor-BNI) antagonists, indicating that DAMGO and U69,593 were acting via their respective receptors. Interestingly, nor-BNI alone induced a 27% increase (59 Ϯ 4.7%, n ϭ 4, p ϭ 0.0149) in cell proliferation over basal levels.
This finding may be due to the possibility that NPs secrete endogenous -opioid peptides that may have inhibitory effects on basal levels of cell division. Upon the addition of KOR antagonist, the effects of both endogenous and exogenous KOR ligands were suppressed, resulting in an increase in proliferation over basal levels. In contrast, in the presence of CTAP alone, NP proliferation was similar to basal levels (34 Ϯ 5.4%, n ϭ 4, p ϭ 0.168).
The involvement of the ERK signaling pathway in opioid regulation of NP proliferation was evaluated by treatment of the cells with the MEK inhibitor, U0126, for 1 h before opioid agonist exposure (Fig. 10). U0126 alone did not significantly affect basal levels of NP proliferation (44 Ϯ 3.2%, n ϭ 4, p ϭ 0.832 versus controls of 43 Ϯ 4.1%, n ϭ 6). Interestingly, administration of U0126 in the presence of DAMGO did not reverse the inhibitory effect induced by the MOR agonist alone, suggesting were grown and treated with opioids as described in A. In some cases, cells were pretreated for 1 h either with MOR antagonist (CTAP, 1 M; B) or KOR antagonist (nor-BNI, 1 M; C ) before the addition of opioid agonists for 24 h. After washing, cells were fixed with 2% PA for 10 min at room temperature. Cells were then permeabilized in 0.4% Triton/PBS for 5 min and incubated in PBS containing 10% fetal calf serum and 0.4% Triton X-100 for 30 min to reduce nonspecific binding, followed by incubation with chicken Sox-1 polyclonal Ab (1:500; Chemicon) overnight at 4°C. After washing, Alexa Fluor 594 (red)-conjugated anti-chicken secondary Ab (1:1000; Molecular Probes) was applied for 1 h at room temperature. DAPI (1:200) was added together with the secondary Ab. Representative double DAPI (blue)/Sox-1 (red) images are shown in control and opioid-treated cells. n ϭ 4 -8. Magnification was ϫ20 -40. Bar graphs (D and E) show the quantification of the Sox-1 counting data. The total number of cells was determined by counting DAPI-stained cells (Ն1000 cells/treatment group). Sox-1-positive cells are expressed as a percentage of the total number of cells. n ϭ 4 -8. *, significantly greater than controls; #, significantly less than agonist alone, p Ͻ 0.0001. that MOR regulation of NP proliferation is independent of the ERK signaling pathway. Furthermore, the inhibitory effect of U69,593 was only partially reversed upon blockade of the ERK signaling by U0126.

DISCUSSION
The analysis of our data reveals several important findings. 1) MOR-1 and KOR-1 gene expression and immunoreactivity was detected in undifferentiated ES cells and in ES cellderived NPs. This is the first evidence of the occurrence of MOR-1 and KOR-1 in blastocyst-derived ES cells. 2) MOR-1 and KOR-1 that occur in ES cells and NPs show functionality as established by the detection of opioid-induced temporal regulation of ERK/MAP kinase signaling. A sustained activation of ERK was characteristic for the opioid-treated undifferentiated ES cells, whereas the RA-induced NPs showed a biphasic profile of opioid-induced ERK activity.
3) The ERK signaling pathway mediatedand -opioid-induced limited proliferation in undifferentiated D3 ES cells, although in the absence of opioids, these cells may not require ERK activation for their self-renewal. More importantly, opioids induced ES cell asymmetric division to generate nestin/ SOX1-positive NPs and reduced the self-renewal of ES cells. Thus, a novel finding here is that opioid-induced sustained activation of ERK may play a role in the initial differentiation of ES cells. 4) Opioids attenuated proliferation of the population of NPs that were generated upon differentiation of ES cells with RA. This inhibitory effect is ERK-independent for DAMGO and partially for U69,593 and correlated with MOR and KOR agonist induction of a biphasic rather than a sustained ERK activation profile in NPs.
In prior studies, KOR was found in the P19 embryonic carcinoma cell line, which contains pluripotent stem cells (51)(52)(53). Although these cells are similar to ES cells derived from 4 -5day-old mouse embryos, they are transformed cells and therefore capable of overexpressing genes not expressed in their parent cell. RA promoted expression of the MOR gene, whereas the KOR gene was first suppressed and then reactivated in P19  cells (54 -56). KOR was also found on the cell surface and nuclear fractions of GTR1 ES cells (57)(58)(59). In addition, MOR and ␦-opioid receptor, but not KOR, are expressed in adult hippocampal progenitors, and opioid peptides, such as ␤-endorphin, can regulate proliferation of these progenitor cells via ERK (60). Reduced ERK signaling via MOR decreases proliferation of these progenitors, increases the number of in vitro generated neurons, and reduces the number of astrocytes and oligodendrocytes. Finally, opiates were found to inhibit neurogenesis in the adult rat hippocampus (61).
We have characterized the mechanism of ERK activation by ORs and demonstrated that opioids either enhanced or inhibited DNA synthesis in several types of primary cultures and cell culture model systems (40 -42, 44, 62, 63). In most of these cases, opioids inhibited mitogen-stimulated proliferation as seen here (Fig. 10). It was important to determine how opioids regulate the ERK signaling pathways in ES cells and in ES cellderived NPs. As discussed in the Introduction, several studies suggest that ERK activation may not be required for ES cell self-renewal (4,(11)(12)(13)(14). Our thymidine incorporation results further support this hypothesis. For example, the finding that MEK inhibitor U0126 alone enhanced basal levels of ES cell proliferation by about 40% supports the notion that undifferentiated ES cells can proliferate when ERK signaling is blocked and suggest that ERK activity may not be necessary for ES cell self-renewal. Moreover, our U0126 data on basal levels of ERK suggest that activation of this kinase may have inhibitory effects on ES cell self-renewal. In addition, under self-renewal conditions (serum/LIF/ME), ES cells maintain low basal ERK activity, data that further support the lack of requirement for ERK activ-ity by these cells (Fig. 5). In contrast, the results also suggest that opioid regulation of ES cell asymmetric cell division in undifferentiated ES cells is ERK-dependent. Therefore, we propose that in undifferentiated ES cells, two processes take place. There is an ongoing ERK-independent, symmetric cell division leading possibly to ES cell self-renewal and an opioid-induced ERK-dependent, asymmetric cell division leading to ES cell differentiation (Fig. 11). Thus, bothand -opioids promote ES asymmetric cell division at a slightly more rapid rate than that of serum/ME/LIF-induced self-renewal of ES cells.
The above findings raise the following questions. If ERK signaling is not the regulator of ES cell division/self-renewal, then what is the signaling pathway that modulates this process? It was proposed that mouse ES cells utilize phosphatidylinositol 3-kinase to progress through the G 1 phase and to avoid differentiation (18,27). What is the mechanism of opioid induction of NP appearance? It has been suggested that Sox proteins might contribute to the transcriptional activation of the MOR gene and that MOR could mediate some Sox-regulated developmental processes (64).
ERK signal duration impacts different cellular responses in many cells and may produce different outcomes in the same cells, such as proliferation, differentiation, or apoptosis (10,26,44,(65)(66)(67)(68). In some cells, sustained but not transient activation of ERK is required to initiate proliferation (41,65,69,70). 3 Sustained ERK activity appears to be required by many cells to pass the G 1 restriction point and to enter S phase, in which , followed by opioid agonist treatment. Cells were then incubated with BrdUrd and co-labeled with monoclonal BrdUrd and chicken Sox-1 polyclonal Abs as described under "Experimental Procedures." Representative double BrdUrd (green)/Sox-1 (red) images are shown in control and opioid-treated cells. Magnification was ϫ20. B, quantification of BrdUrd/Sox-1 counting data. NIH ImageJ version 1.32 software was used to count cells. Cells from Ͼ5 fields/well were counted (Ͼ1000 cells/treatment group). The total number of NPs was estimated by counting Sox-1-stained cells. BrdUrd-positive cells were expressed as a percentage of the total number of Sox-1-labeled NPs. n ϭ 3-7. *, significantly less than controls (p Յ 0.0004); #, significantly greater than agonist alone (p Յ 0.0281). cellular DNA is replicated (65,71). In contrast, other studies have shown that Ras/Raf mediated a transient epidermal growth factor-induced ERK activation that leads to proliferation, whereas Rap/Raf-mediated sustained activation of ERK is required for PC12 cell differentiation (22,72). Our data suggest that the opioid-induced sustained activation of ERK triggers ES asymmetric cell division, and the "newly" formed cells appear to be ES cell-derived NPs.
The important role of the ERK signaling pathway in neural differentiation has been well established. Whereas some reports indicate that ERK signaling inhibits differentiation of some types of cells (73,74), other studies support the idea that the ERK signaling may be a positive regulator of this process (75). A recent paper discusses the existence of a biphasic regulation of ERK activity during myogenic differentiation and suggests that the signaling pathway(s) may play a dual role in this multistep process, wherein the late phase is responsible for the formation of postmitotic myotubes (76). A similar study shows that an early stimulation and late inhibition of ERK activity by insulin-like growth factor mediates the switch in insulin-like growth factor action from inhibition to stimulation of skeletal muscle cell differentiation (77). Finally, it was found that the early stage of neuronal differentiation of mouse ES cell line P19 triggered by aggregation of the cells and RA treatment is accompanied by biphasic activation of ERK signaling: a transient phase and a second, sustained ERK activation that is maintained until the appearance of neural phenotype (78).
Here, we propose that opioid-induced sustained activation of ERK may be a required step for the development of ES cellderived NPs, and opioid-induced biphasic ERK phosphorylation may play a supporting role in the increased differentiation rate during the generation and maturation of NPs. This hypoth-esis is based on the following findings. As seen in Fig. 7, basal levels of ERK activity are higher in long term differentiated ES cells than in short term differentiated cells (NPs). In addition,and -opioids induce a moderate, biphasic activation of ERK signaling in short and long term differentiated ES cells. More importantly, we found that the biphasic stimulation of ERK phosphorylation by opioids accompanies the opioid-induced inhibition of proliferation in NPs. Although MOR regulation of NP proliferation was independent of the ERK signaling pathway, the inhibitory effect of U69,593 was only partially reversed upon blockade of the ERK signaling by U0126. This finding may be explained by the involvement of dual signaling pathways in KOR modulation of NP proliferation. In addition to ERK signaling, p38/ MAP kinase is a candidate for a second signaling pathway that could be required for KOR inhibition of NP proliferation.
The absence of a developmental phenotype in opioid receptor knock-out mice has been reported (79). This is not uncommon, given the functional redundancies that cells display after targeted gene disruption of critical signaling molecules, such as G protein-coupled receptors. There are numerous accounts in the literature regarding the absence of alteration of function after the loss of G protein-coupled receptor gene expression due to redundant or compensatory signaling mechanisms (80, 81). In some instances, developmental phenotypic changes are subtle after gene disruption and only seen when the organism is stressed as it is in its natural environment (82). Alternatively, MOR and/or KOR may be dispensable for development, and the observations made here may be only of relevance to maternal opiate abuse.
In conclusion, our studies suggest that MOR-1, KOR-1, and their exogenous ligands are able to modulate ES cell proliferation and differentiation. Thus, opioids and opioid-induced ERK signaling may play an important role in ES cell fate decisions by directing the cells to specific lineages.