Earlier studies reported that neural stem (NS) cells injected into blastocysts appeared to be pluripotent, differentiating into cells of all three germ layers. In this study, we followed in vitro green fluorescent protein (GFP)–labeled NS and embryonic stem (ES) cells injected into blastocysts. Forty-eight hours after injection, significantly fewer blastocysts contained GFP-NS cells than GFP-ES cells. By 96 hours, very few GFP-NS cells remained in blastocysts compared with ES cells. Moreover, 48 hours after injection, GFP-NS cells in blastocysts extended long cellular processes, ceased expressing the NS cell marker nestin, and instead expressed the astrocytic marker glial fibrillary acidic protein. GFP-ES cells in blastocysts remained morphologically undifferentiated, continuing to express the pluripotent marker stage-specific embryonic antigen-1. Selecting cells from the NS cell population that preferentially formed neurospheres for injection into blastocysts resulted in identical results. Consistent with this in vitro behavior, none of almost 80 mice resulting from NS cell–injected blastocysts replaced into recipient mothers were chimeric. These results strongly support the idea that NS cells cannot participate in chimera formation because of their rapid differentiation into glia-like cells. Thus, these results raise doubts concerning the pluripotency properties of NS cells.
In the brain, the subventricular zone (SVZ) is a source of proliferating cells that replenish olfactory interneurons and glial cells [1,2]. Cells from the SVZ have several characteristics that distinguish them as neural stem (NS) cells, including their capacity to self-renew indefinitely in defined media supplemented with either epidermal growth factor (EGF) or basic fibroblast growth factor (FGF) and the capacity to differentiate into neurons, astrocytes, and oligodendrocytes in vitro upon cytokine exposure [3,4].
Recent studies have suggested that under certain conditions, NS cells are not only multipotent, i.e., able to produce more than one neural cell type, but also able to differentiate into other cell types, such as blood  and muscle . In fact, when NS cells are placed in embryos at the blastocyst stage (wherein differentiation between trophoblast and fetus has begun but before germ layer formation occurs) and examined later in development, pluripotency, or the ability to produce all cell types of the embryo proper, has been noted, suggesting that these cells can be reprogrammed to assume a state analogous to the embryonic stem (ES) cell [7,8]. However, although cellular markers from NS cells were found in various organs of these blastocyst-injected animals, cellular reprogramming has to be reassessed in light of new data reporting potential fusion events between NS cells and ES cells [9,10]. Thus, the pluripotent potential of NS cells must be further tested.
Because of our interest in NS cell pluripotency after blastocyst injection, we developed an in vitro model system to observe these cells for several days after injection. Our results indicate that these cells rapidly differentiate into glial fibrillary acidic protein (GFAP)–positive, nestin-negative cells and then regress. This behavior is strikingly different from that of ES cells but is consistent with our own experience that these cells preferentially form glia when transplanted into the central nervous system (CNS) . Furthermore, unlike other reports [7,8], we were unable to demonstrate any chimeric animals after NS cell injection into blastocysts. Because these cells consistently become astrocytes after injection, the pluripotent property of these cells under these conditions is questionable.
Materials and Methods
Cell Isolation and Maintenance
Green fluorescent protein (GFP)–expressing NS cells were obtained from the brains of transgenic mice (Swiss Webster) expressing GFP under the control of the mouse prion promoter and a chicken β-actin–cytomegalovirus (CMV) immediate early enhancer . Bacterial β-galactosidase–expressing NS cells were obtained from Rosa26 mice (B6,129-Tgr-TgR[ROSA26]26Sor; Jackson Laboratory, Bar Harbor, ME, http://www.jax.org), which express a reporter gene encoding β-galactosidase in all tissues . The SVZ/striatal areas from day-16 heterozygote embryos were derived and maintained as described elsewhere [14,15]. Briefly, the SVZs were dissected, triturated to a single-cell suspension, and seeded in culture medium consisting of DMEM/F12, 0.3% glucose, 23 μg/ml insulin, 92 μg/ml transferrin, 55 μM putrescine, 50 U/ml penicillin-streptomycin, 27.5 nM sodium selenite, 20 nM progesterone (Sigma, St. Louis, http://www.sigma-aldrich.com), and 20 ng/ml EGF (Becton Dickinson Labware, Bedford, MA). ES cells expressing enhanced GFP (clone B5) driven by a CMV promoter with chicken β-actin enhancer randomly integrated into R1 ES cells were obtained from Dr. A. Nagy from The Samuel Lunenefeld Research Institute (Mount Sinai Hospital, Toronto, Ontario, Canada) and were maintained in a leukemia inhibitory factor–supplemented medium.
Blastocyst Injection for In Vitro Studies
Blastocyst injections were performed in the Transgenic Core Facility of the University of Massachusetts Medical School. Day 3.5 blastocysts were extracted from the uteri of C57Bl/6J pregnant females. Injections were performed with 10 to 15 GFP-ES cells or 10 to 15 GFP–single NS cells (SC, passages 1–8), which were passaged and filtered on the day of injection to ensure single-cell suspensions or with a single day 2–GFP neurosphere (D2NS, passages 1–8), that were in culture for 48 hours after passaging. Injected blastocysts were placed one per well on eight-well chamber slides (Nalgen Nunc International, Naperville, IL) coated with gelatin, and their development was observed by fluorescence microscopy (Leica DMIRB microscope, Wetzlar, Germany, http://www.leica-microsystems.com) for 48 to 96 hours.
Blastocyst Injection for In Vivo Studies
Day-3.5 blastocysts were injected with 10 to 15 GFP-SCs (passages 1–5 or 53), 1 to 12 small neurospheres cultured for only 24 hours after passaging, or a single D2NS. Injected blastocysts were implanted in uteri of Swiss Webster pseudo-pregnant females. For some experiments, neonates (postnatal day 1) were euthanized, and histological or polymerase chain reaction (PCR) analyses were conducted on brains, intestines, livers, and kidneys of pups to detect GFP-labeled cells. For some neonates, histological and PCR studies were simultaneously conducted on organs. Neonates were perfused through the heart with phosphate buffered saline (PBS). Half of each organ was either placed in 4% paraformaldehyde for 48 hours or used for DNA extraction and PCR. Fixed organs were cut on a freezing microtome (35-μm-thick sections); the sections were mounted, coverslipped, and observed with a fluorescent microscope. For other neonates or embryos (between days 11.5 and 15.5) used only for PCR analyses, DNA was extracted from organs.
PCR for GFP Detection
Genomic DNA from neonates or embryos was extracted using a Genomic DNA isolation kit (Lamda Biotech, St. Louis). Primers used for detection of GFP gene were as follows: forward 5′-TAAACG GCC ACAAGTTCAGC-3′ and reverse 5′-TGT TCT GCT GGTAGT GGT CG-3′ (predicted length of PCR product, 472 bp; Sigma Genosys, The Woodlands, TX, http://www.genosys.com) . Primers for the housekeeping gene gapdh, forward 5′-CGG AGT CAACGG ATT TGG TCG TAT-3′ and reverse 5′-AGC CTT CTC CAT GGT GGT GAAGAC-3′ (predicted length of PCR product, 309 bp), were used to confirm the presence of template DNA in the reactions. The PCR reactions were performed in 50 μl with 50 ng of DNA, each with 10 mM Tris, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, and 200 μM dNTPs each, 0.4 μM each of forward and reverse primer, and 1.25 U AmpliTaq DNA polymerase (Applied Biosystems, Roche Molecular System, Branchburg, NJ, http://www.pebio.com). The PCR reactions were performed as follows: loading at 95°C for 2 minutes, denaturation at 95°C for 30 seconds, annealing at 55°C for 30 seconds, and elongation at 72°C for 1 minute for 30 cycles. For positive controls, DNA templates from brains of GFP heterozygotic mice were used. Anegative control with only water was performed. In some cases, to detect the presence of GFP DNA, nested PCR was performed as follows: 20 cycles at 55°C for annealing with the primers described above, followed by 30 cycles at 63°C for annealing with the forward oligomer 5′-CAG CCG CTACCC CGACCACA-3′ and the reverse primer 5′-CGC TGC CGT CCT CGATGT TG-3′ on 5 μl of DNA from the first PCR reaction (predicted length of PCR product, 314 bp). All of the PCR reactions were analyzed on 2% agarose gels.
Real-Time PCR forGFP Detection
Quantitative real-time PCR analysis was performed with a DNA Engine Opticon (MJ Research) using the Quantitect SYBR Green PCR kit (Qiagen, Valencia, CA, http://www.qiagen.com). Primers used were either the first set of GFP primers (melting point of product, 87°C) or the primers used for the nested PCR (melting point of product, 85°C). The PCR program consisted of 30 cycles with a denaturation at 94°C for 30 seconds, annealing at the above-specified temperature for 30 seconds, and elongation at 72°C for 30 seconds. Fluorescent DNA melting curves were taken from 55.0°C to 90.1°C and read every 0.2°C with a hold of 1 second between reads. The Opticon Fluorescent Detector's sensitivity to small quantities of chimeric DNA in a background was assessed using genomic DNA from GFP-expressing mouse serially diluted with wild-type mouse genomic DNA. A negative control with only water was also performed.
Reversed Transcription PCR for oct 4 Detection
To evaluate oct 4 expression, poly A+ mRNA was extracted from ES cells, SCs, or D2NS using a QuickPrep mRNA kit (Amersham Biosciences, Buckinghamshire, U.K.) and reverse transcribed in cDNA using the Superscript First Strand Synthesis kit (Invitrogen Life Technologies, Carls-bad, CA, http://www.invitrogen.com) using the following primers: forward 5′-AAG TGG GTG GAG GAAGCC GAC AAC-3′ and reverse 5′-TGG GGG CAG AGG AAA GGA TAC AGC-3′ (predicted length of PCR product, 333 bp). The PCR amplification of the oct4 cDNA was performed with an annealing temperature of 60°C as described above. A negative control with only water was also performed, and detection of the housekeeping gene gapdh was performed to confirm presence of templates in the reactions. The PCR reactions were analyzed on 2% agarose gels.
Blastocysts were fixed in 4% paraformaldehyde for 10 minutes and then rinsed three times in PBS for 5 minutes each and incubated in blocking solution containing 0.4% Triton X-100 (Amersham Biosciences) and 10% normal goat serum in PBS for 1 hour at room temperature. Blastocysts were then incubated in a cocktail of primary antibodies made in 2% normal goat serum and 0.4 % Triton X-100 in PBS for 2 hours at room temperature. The primary antibodies used were a combination of the rabbit anti-human GFAP, a marker for astrocytes (1/1000; ICN Pharmaceuticals Cappel, Aurora, OH), with the mouse anti-rat βIII-tubulin, a marker for early neurons (1/1000, TUJ1 clone, Covance BAbCo, Richmond, CA, http://www.babco.com), the mouse anti-rat nestin, a marker for NS cells (1/1000; BD Biosciences Pharmingen, San Diego, http://www.bdbiosciences.com), or the mouse anti-human stage-specific embryonic antigen-1 (SSEA-1), a marker for ES cell pluripotency (1/50, Chemicon International, Temecula, CA, http://www.chemicon.com). Samples were then rinsed in PBS and incubated for 2 hours in a cocktail of Alexa 594 goat anti-rabbit (1/500; Molecular Probes, Eugene, OR, http://www.probes.com) and CY5 goat anti-mouse (1/500; Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) secondary antibodies. Samples were rinsed in PBS and mounted with Prolong Antifade reagent (Molecular Probes). A similar procedure was used for characterizing cultured cells, except that only one primary antibody and its corresponding secondary antibody were used at a time and the nucleus of the stained cells were labeled with 4′6-diamidino-2-phenylindole. Immunofluorescent staining was observed using confocal or regular microscopy.
Characterization of Cultured Cells
As previously noted, NS cells isolated from E16 embryos grow as nonadherent spheres when exposed to EGF in absence of any coating substrate [3,4]. All cells from SC and D2NS preparations were nestinimmunoreactive, with only a few of these NS cells occasionally positive for GFAP or βIII-tubulin. Almost all ES cells were SSEA-1–positive (Fig. 1A) and nonimmunoreactive for GFAP- and βIII-tubulin (data not shown). Only a few ES cells (<8%) were nestin-immunoreactive. Additionally, GFP-ES cells expressed oct 4 mRNA (Fig. 1B). In contrast, neither SSEA-1 nor oct 4 mRNA expression was observed in SCs or D2NS cells (Fig. 1).
Assessment of In Vitro Chimerism after Blastocyst Injection of GFP-ES or GFP-NS Cells
Forty-eight hours after blastocyst injection, GFP-ES cells could be detected in 93% of the inner masses of hatched blastocysts (52 blastocysts with GFP cells per 56 blastocysts injected), a rate that was significantly higher than that for either GFP-SCs (62%, 40/65) or GFP-D2NS (64%, 44/69) (p < .001, χ2 test, Fig. 2). At 72 and 96 hours after injection, the detection rate of GFP-ES cells decreased to 92% and 55.5%, respectively, compared with 43% and 10% for GFP-SC and 42% and 13% for GFP-D2NS. The detection rates at 72 hours (p = .009) and 96 hours (p < .001) were significantly higher for GFP-ES cells than for GFP-SC or GFP-D2NS (χ2 test, Fig. 2). Furthermore, by 48 hours, both the singly injected NS cells and the small neurospheres extended long cellular processes (Fig. 3A). By comparison, the ES cells maintained an undifferentiated appearance. Moreover, by 72 hours, blastocysts injected with GFP-NS cells presented only small round GFP bodies (Fig. 3B2), whereas blastocysts injected with GFP-ES cells showed a greater number of GFP cells than when initially injected (Fig. 3B1). Similar observations were made when we increased the number (up to 40 cells) of inoculated GFP-labeled or β-galactosidase–expressing SCs into blastocysts (data not shown).
Immunocytochemical Characterization of GFP Cells Present in Blastocysts
Double immunocytochemistry for GFAP and bIII-tubulin, nestin, or SSEA-1 was performed on blastocysts injected with GFP cells (n = 22). GFP-ES cells in the blastocysts (n = 6) were negative for GFAP, bIII-tubulin, and nestin but were positive for SSEA-1 (Fig. 4). However, cells in the blastocysts injected with GFP-SC (n = 9 blastocysts) or GFP-D2NS (n = 7 blastocysts) were GFAP-positive and negative for nestin, βIII-tubulin, and SSEA-1 (Fig. 4). These results demonstrated an astrocytic fate for NS cells injected into blastocysts.
Assessment of In Vivo Chimerism in Neonates and Embryos after Blastocyst Injection of NS Cells
To extend our in vitro results, we also injected NS cells into blastocysts, placed the injected blastocysts into pseudo-pregnant females, and characterized the resulting pups. Blastocysts (n = 74) were injected with 10 to 15 GFP-SCs, and 42 neonates were obtained. No difference in survival was noted whether early (1–5) or late (53) passage cells were used. All neonates were screened by histology, PCR/nested PCR, or nested PCR only for the presence of GFP-expressing cells. None of the 42 pups were chimeric in the heart, liver, intestines, kidney, skin, or brain (Table 1).
Table Table 1.. Summary of in vivo results after injections of neural stem cells in blastocysts
Abbreviations: D1NS, day 1–GFP neurosphere; D2NS, day 2–GFP neurosphere; GFP, green fluorescent protein; PCR; polymerase chain reaction; RT, reverse transcriptase; SC, single neural stem cell.
Given our lack of success detecting chimerism in neonates, we decided to examine embryos. Moreover, because of the possibility that the pluripotent cells represented only a small proportion of NS cells, additional injections were performed with small neurospheres selected 1 day after seeding single-cell suspensions. To ensure that we were selecting cells that would grow into mature neurospheres, we seeded one small neurosphere per well and followed its development in a separate experiment. These small neurospheres gave rise 75% of the time to mature neurospheres when singly seeded (data not shown).
Fifty-four blastocysts were injected with 1 to 12 small neurospheres (passages 1–5), and 11 embryos at day 11.5 were obtained. No chimerism was detected by histology or by real-time PCR in organs of the embryos examined. However, serial dilution of genomic DNA from GFP-expressing mice with wild-type mouse DNA showed that a quantity of 0.001 ng GFP mouse DNA (1:100,000 dilutions) resulted in a detectable amplification signal, and a quantity of 0.01 ng GFP DNA (1:10,000 dilutions) yielded a signal cleanly above background noise (Fig. 5). None of the 11 embryos showed any detectable GFP DNA signal in either heads (Fig. 5) or whole bodies (Table 1).
To check for the possibility that a more developed neurosphere might better incorporate into blastocysts, we chose to inject D2NS when the spheres were more developed but still small in diameter. Blastocysts (n = 44) were injected with one neurosphere (passages 1–5), and 26 embryos between E 13.5 and E 15.5 were obtained. No chimerism was detected in any organs of the 26 embryos by real-time PCR (Table 1). Overall, we used multiple methods to examine 82 embryos or neonates and found no evidences for chimerism.
Several studies have reported that various stem cells show pluripotency after injection into blastocysts [7, 8, 16, 17]. Two of these used NS cells and showed incorporation into all three germ layers [7,8], supporting the view that NS cells could be induced to express a true pluripotent phenotype. In the present study we demonstrate, however, that NS cells injected into blastocysts cannot form chimeras because of their rapid differentiation into glial-like cells within the blastocyst.
To understand NS cell pluripotency after injection into blastocysts, we chose to follow in vitro the incorporation and development of NS and ES cells into blastocysts. Moreover, we assessed whether pluripotent capacity resided in only a small population of NS cells. The optimal way would be to select for pluripotent stem cell markers, but EGF-grown NS cells (harvested at E16) did not express either the pluripotent marker SSEA-1 (Fig. 1A) or the transcription factor Oct 4 (Fig. 1B), which is present only in germ line and some pluripotent cells . Instead, because only a small percentage of NS cells display self-renewal properties based on clonal analyses [3,4], we chose to inject individual NS cells shortly after passaging or small neurospheres as early as 24 hours after passaging. Although long-term passaging typically provides a large pool of NS cells , early passages (below 10) were used in this study to replicate experimental procedures used by others  and to avoid major changes in NS cell growth and gene expression that have been reported to occur after long-term culturing and passaging of neurospheres .
Blastocysts can be maintained for up to 96 hours in vitro, during which the integrity of the inner cell mass (ICM) is maintained. We noted that compared with blastocysts injected with GFP-ES cells, significantly fewer blastocysts contained GFP-NS cells by 48 hours, a difference that was more pronounced by 96 hours (Fig. 2). Moreover, GFP-NS cells in blastocysts adopted a differentiated morphology that was not observed for GFP-ES cells (Fig. 3A). Furthermore, 96 hours after injection of GFP NS cells, residual GFP appeared in a very small cellular compartment. Similar results were obtained with small neurospheres or SC suspensions. Because injection of β-galactosidase–labeled NS cells into blastocysts yielded a very low rate of integration and also differentiated (data not shown), this appears to be a characteristic of EGF-grown NS cells and is not a reflection of the reporter expression used to label the cells. In contrast, GFP-ES cells in blastocysts incorporated in the ICM and increased in numbers (Fig. 3B). Thus, regardless of whether NS cells are selected for self-renewal and proliferation properties, they do not incorporate and proliferate like ES cells in the blastocyst environment.
In terms of differentiation markers, by 48 hours, GFP-NS cells in blastocysts had differentiated into GFAP-positive cells that no longer expressed the stem cell marker nestin (Fig. 4). A similar differentiation into a nestin-negative, GFAP-positive cell was noted after injections of NS cells selected for sphere-forming capability (Fig. 4). This pattern of expression clearly differs from the one observed in undifferentiated NS cells, in which GFAP and nestin are coexpressed in vivo as well as in vitro [19,20]. Such expression of differentiation markers was never seen after ES cells were implanted; these cells retain cellular markers of pluripotency, such as the SSEA-1 (Fig. 4), and continue to divide in vitro. Thus, because implantation of NS cells into blastocysts leads to a rapid differentiation into nestin-negative, glia-like cells, it seems very unlikely that they can participate in tissue development.
Such results further suggest that survival of NS cells until gastrulation is unlikely. Consistent with this is the fact that we did not detect chimerism in brains and other tissues of more than 80 experimental animals after in vivo implantation of NS cell–injected blastocysts (Table 1 and Fig. 5). Tissues from embryos and neonates were assessed by various techniques, including a very sensitive real-time quantitative PCR assay that can identify as little as 0.001 ng GFP DNA (Fig. 5). Because injections of ES cells into blastocysts yield a very high rate of chimeric animals (S. Jones, personal observations), these results reinforce the idea that NS and ES cells do not display equivalent pluripotent abilities.
It is possible that the blastocyst may not represent the optimal environment to test NS cell pluripotency. Indeed, a timely and specific cellular level of transcription factors such as Oct 4, Nanog, and Stat3 is crucial for the expression of pluripotency in the ICM of blastocyst and for embryo formation [21,22]. However, such markers are generally not encountered after very early development  and are absent in NS cells (Fig. 1). Nevertheless, extrinsic environmental stimuli and the highly proliferative properties of the blastocyst should be permissive to NS cell implantation and development , providing they are pluripotent through Oct 4–independent mechanisms.
Other studies have reported chimerism in multiple organs of experimental animals after blastocyst injections with either EGF-grown adult or EGF/FGF-grown apoptosis-resistant embryonic NS cells [7,8]. It is unlikely that differences in methods are sufficient to explain the discrepancies between our study and these reports. Indeed, except during aging, there is no evidence of major differences in cell properties between EGF-grown embryonic and adult NS cells . Adult NS cells proliferate in vivo under EGF stimulation and display similar in vitro self-renewal and multipotent properties as embryonic NS cells [25,26]. Furthermore, although EGF- and FGF-stimulated embryonic NS cells might represent different types of multipotent stem cells in the CNS [27,28], EGF- and EGF/FGF-grown NS cells are similar in their self-renewal and differentiating properties, with only a small percentage of EGF-grown NS cells differentiating into neurons . Thus, the apparent “chimeric” animals in the previous studies of blastocyst injections might be the result of other cellular processes, such as cell fusion. Recent studies have, indeed, shown that coculture of ES cells and somatic cells can lead to the formation of hybrids in which pluripotent ES cell–like properties are dominant. This phenomenon has now been observed with cells such as fully differentiated thymocytes  or stem cells of neural or hematopoietic origin [9, 10, 30] and has been reported to lead to the formation of “chimeric” animals . Therefore, formation of hybrid ICM-NS cells may be one mechanism by which “chimeric” animals were obtained in previous blastocyst injection studies [7,8]. In the present study, we have not tested for the formation of tetraploid cells; however, the frequency of hybrid formation in vitro can be low (between 10−4 and 10−5, for example ), and, thus, fusion events might not have been detected in the present experimental sample. Nevertheless, the consistent, early differentiation of NS cells into glia provides an explanation why chimerism occurs rarely, if ever, after blastocyst injection.
Previous CNS transplant studies [11,31] indicate that EGF-grown NS cells preferentially differentiate into glial cells in the absence of differentiating stimuli. In the present study, placement of the NS cells in a primitive, undifferentiated environment resulted in a similar glial fate, suggesting that this differentiation pattern is a default pathway of these cells in vivo. This additionally suggests that injection of NS cells into blastocysts is unlikely to ever yield chimerism. Moreover, this consistent differentiation and cessation of proliferation of NS cells after blastocyst injections raises the question of whether NS cells are ever pluripotent.
We want to thank Dr. Anthony van den Pol for the gift of the GFP transgenic mice and Drs. Todd Savarese and William Schwartz for their helpful comments on the writing of this manuscript. This work was supported in part by NS43879 from the National Institute of Neurological Disorders and Stroke (L.R.).