Novel lines of Pax6 -/- embryonic stem cells exhibit reduced neurogenic capacity without loss of viability
© Quinn et al. 2010
Received: 4 August 2009
Accepted: 24 February 2010
Published: 24 February 2010
Skip to main content
© Quinn et al. 2010
Received: 4 August 2009
Accepted: 24 February 2010
Published: 24 February 2010
Embryonic stem (ES) cells can differentiate into all cell types and have been used extensively to study factors affecting neuronal differentiation. ES cells containing mutations in known genes have the potential to provide useful in vitro models for the study of gene function during neuronal differentiation. Recently, mouse ES cell lines lacking the neurogenic transcription factor Pax6 were reported; neurons derived from these Pax6 -/- ES cells died rapidly after neuronal differentiation in vitro.
Here we report the derivation of new lines of Pax6 -/- ES cells and the assessment of their ability to survive and differentiate both in vitro and in vivo. Neurons derived from our new Pax6 -/- lines were viable and continued to elaborate processes in culture under conditions that resulted in the death of neurons derived from previously reported Pax6 -/- ES cell lines. The new lines of Pax6 -/-ES cells showed reduced neurogenic potential, mimicking the effects of loss of Pax6 in vivo. We used our new lines to generate Pax6 -/- ↔ Pax6 +/+ chimeras in which the mutant cells survived and displayed the same phenotypes as Pax6 -/- cells in Pax6 -/- ↔ Pax6 +/+ chimeras made by embryo aggregation.
We suggest that loss of Pax6 from ES cells reduces their neurogenic capacity but does not necessarily result in the death of derived neurons. We offer these new lines as additional tools for those interested in the generation of chimeras and the analysis of in vitro ES cell models of Pax6 function during neuronal differentiation, embryonic and postnatal development.
Pax6 is a highly-conserved transcription factor whose main sites of expression are in the developing eye and central nervous system [1–5]. Homozygous loss-of-function mutations of Pax6 cause failure of eye morphogenesis and severe abnormalities of brain development [6, 7]. Pax6 plays an important role promoting neurogenesis; in vivo, loss of Pax6 results in neural progenitors having reduced neurogenic potential [8, 9] whereas its over-expression in vitro pushes cells towards a neuronal fate [8, 10, 11]. The mechanisms by which Pax6 directly promotes neurogenesis are not yet known.
Pluripotent embryonic stem (ES) cell lines have provided a means to exploit gene targeting for the analysis of gene function in vivo. In addition, since ES cell lines can be differentiated into a variety of cell types in culture they provide an opportunity to study gene function by comparing the phenotypes of ES-derived cells in vitro. ES cells have been used systematically as a model system for examining factors controlling differentiation pathways in vitro. In particular, differentiation protocols have been reported which recapitulate differentiation of various neuronal cell types, or their precursors, in vitro [12–17].
Recently, ES cell lines derived from mouse embryos lacking Pax6 (Pax6 -/-) were studied in culture . Neurons derived from these Pax6 -/- ES cells died rapidly after induction of neuronal differentiation, with almost none surviving beyond about a week after plating. This is surprising as Pax6 -/-neurons can survive into postnatal life following transplantation into wild-type embryos  and following early embryonic neural tissue-specific conditional deletion of Pax6  suggesting that loss of Pax6 alone is not necessarily sufficient to abolish neuronal viability.
We derived new lines of Pax6 -/- ES cells and assessed their ability to survive and differentiate both in vitro and in vivo. These new lines (which are labelled with a developmentally neutral nuclear transgenic marker) were used to generate Pax6 -/- ↔ Pax6 +/+ chimeras in which the mutant cells survived and displayed the same phenotype as Pax6 -/- cells in Pax6 -/- ↔ Pax6 +/+ chimeras made by embryo aggregation [9, 20–22]. In vitro, we found that the new lines of Pax6 -/- ES cells showed reduced neurogenic potential, mimicking the effects of loss of Pax6 in vivo. Contrary to previous reports that ES-derived cells lacking Pax6 show reduced process development as little as 4 days after plating and die 2-4 days later , cells derived from these new Pax6 -/- lines were viable and continued to elaborate their processes after comparable culture times under comparable conditions. We offer these lines as a validated tool for investigation of gene function during neuronal differentiation in vitro and the generation of chimeric animals for in vivo analysis of development of the central nervous system.
Animals in this study were bred in-house following Home Office (UK) regulations and were approved by the University of Edinburgh Ethical Review Panel.
ES cells that were either wild type (Pax6 +/+) or Pax6 SeyEd/SeyEd(designated here as Pax6 -/-;  were derived using the following protocol. Female mice [129Sv(Ola); Pax6 +/-] were superovulated and mated with males [129Sv(Ola)] that were Pax6 +/- and homozygous for a reiterated β-globin repeat transgene (Tg/Tg; , which acts as a developmentally neutral marker detectable by DNA-DNA in situ hybridization [9, 20, 22, 25]. Delayed implantation (diapause) was induced at 2.5 days post coitum (dpc) by intraperitoneal injection of Tamoxifen (Sigma, UK; 10 μg/animal) and subcutaneous injection of Depo-Provera (Sigma, UK; 1-3 mg/animal). At 7.5 dpc delayed blastocysts were flushed from the uterus, transferred to a gelatinised well containing N2B27 medium (Stem Cell Sciences, UK) with 10 ng/ml Leukaemia Inhibitory Factor (LIF) and cultured at 37°C in 5% CO2. After approximately 5 days in culture, the inner cell mass was detached from the bottom of each well using a fine pulled Pasteur mouth pipette before being disaggregated in trypsin (0.025% for 2-3 minutes at 37°C) and disrupted to give individual clusters of between 1-5 cells. Cell clusters were then transferred to a fresh gelatinised well containing N2B27 media containing LIF (1000 U/ml) with the addition of BMP4 (10 ng/ml). Primary colonies of ES cells were visible after approximately 5 days in culture. ES cell lines were passaged at least twice in feeder-free conditions in BHK-21 Glasgow MEM (GMEM; Sigma, UK) with 15% fetal bovine serum (FBS) and LIF (1000 U/ml). Cell lines were genotyped as described previously . All cell lines were karyotyped to confirm normal chromosome complement.
Chimeric embryos were produced by blastocyst injection and processed histologically, by in situ hybridization and immunohistochemistry, to detect cells carrying the Tg marker and to analyse their expression of transcription factors, as described before .
The neural differentiation protocol followed a previous description . Briefly, ES cells maintained in feeder-free conditions were plated in non-adherent dishes in GMEM +15% FBS in the absence of LIF (designated Day 0, or day of LIF withdrawal). After 4 days, all-trans retinoic acid (5 μM, Sigma, UK) was added. The developing embryoid bodies (EBs) were allowed to develop for a further 2 or 4 days (to Days 6 or 8) before dissociation using Tryple Express™ (Invitrogen, UK). EBs taken on Day 6 and some of those taken on Day 8 were exposed to 40 μg/ml bromodeoxyuridine (BrdU) for 30 minutes before being dissociated and plated for two hours in 8-well glass chamber slides, fixed with paraformaldehyde and processed for immunohistochemistry. Other cells dissociated on Day 8 were plated at 2.0 × 105 cells/cm2and cultured in serum-free GMEM/F12:Neurobasal medium (1:1) supplemented with N2 and B27 (Invitrogen, UK) for 6 or 8 days (to Days 14 or 16 after LIF withdrawal) before fixation and processing.
The cortices of E11.5 embryos generated from timed matings of Pax6 +/- animals were isolated. Each cortex was kept separate, its cells were dissociated into separate wells and were allowed to grow for 7 days in DMEM:F12 1:1 culture media supplemented with 100 μg/ml transferrin, 25 μg/ml insulin, 20 nm progesterone, 30 nm sodium selenite, 60 μg/ml putrescine, 20 ng/ml EGF and bFGF (all Sigma, U.K). The genotypes of individual embryos were discovered by PCR analysis as described previously . Primary neurospheres were dissociated on Day 7 and stained with propidium iodide for analysis using flow cytometry.
Antibodies used in this study were against β-III-tubulin (Sigma, UK), microtubule associated protein 2 (MAP2; Sigma), UK, Pax6 (Developmental Studies Hybridoma Bank, USA), glial fibrillary acid protein (GFAP; DAKO, UK) and BrdU (Beckton Dickinson, UK). Visualisation was achieved using Alexafluor-488 or Alexafluor-568 (Invitrogen, UK) conjugated secondary antibodies and cell nuclei were counterstained using TOPRO3 (Molecular Probes, USA). For counting, 3-8 areas were selected randomly per well. Apoptotic cells were identified by their dense chromatin condensation visible with TOPRO3 staining. Chromatin condensation was easily recognized, is a defining feature of apoptosis  and has been shown previously to coincide with other indicators of apoptosis such as terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labelling (TUNEL) .
For flow cytometry, cells were stained with propidium iodide to allow discrimination of single cells and analysis of DNA content. Staining reactions were carried out in duplicate. Cells were analyzed on a Beckman-Coulter XL flow cytometer with Expo32 software. 8,000-20,000 cells were analyzed per sample.
Sub-confluent Pax6 -/- ES cells were dissociated using Tryple Express™ and plated in GMEM media containing LIF and FBS. Twenty-four hours later they were co-transfected using Fugene (Invitrogen, UK) with pCMV-Script plasmid containing the full length Pax6 cDNA  and plasmid pEGFPN1 (Clontech, USA) expressing green fluorescent protein (GFP). Transfected cells were selected by culture in medium containing G-418 (500 μg/μl). After 7 days in selection medium individual colonies were picked and expanded by routine passage in GMEM+FBS+LIF. Stably transfected cell lines were subjected to neural induction as described above and cultures fixed for immunohistochemistry at Day 14.
We established 9 lines of Pax6 +/+ ES cells (wtMM1-9) and three lines of Pax6 -/- ES cells (SeyD1-3) containing the Tg transgene . Karyotyping of these ES cell lines showed that each had a normal chromosome complement . Pax6 -/- ES cell lines displayed no abnormalities of growth during routine passage; two (SeyD1 and SeyD2) were analysed in detail.
Embryoid bodies (EBs) from previously-derived Pax6 -/- ES cells contained normal complements of neuronal progenitors . In agreement with this, we found that proportions of BrdU-labelled proliferating cells were no different between wtMM4, SeyD1 and SeyD2 EBs (Figure 3). We also compared proportions of cells positive for the progenitor markers RC2 and Nestin in EBs on Day 8 and found no evidence of differences between the proportions of progenitors in EBs from our new wild type and Pax6 -/- lines, in agreement with a similar comparison by Nikoletopoulou et al.  of their wild-type and mutant lines.
Proportions of BrdU-labelled proliferating cells declined between Day 6 and Day 8 after LIF withdrawal, coinciding with the onset of differentiation (Figure 3). Small proportions of β-III-tubulin-expressing cells were detected on Day 8 after LIF withdrawal and small proportions of cells labelled with BrdU on Day 8 were also β-III-tubulin-positive several hours later. There were no differences in the proportions of early differentiated neurons between cultures from wild type and Pax6 -/- lines at Day 8 (Figure 3). Overall, these data indicate that neuronal differentiation is just beginning on Day 8 and the time of onset is not perturbed by the loss of Pax6 expression in our ES cells.
Proportions of apoptotic cells on Day 6 and Day 8 after LIF withdrawal were low with no significant differences between the wild-type and the Pax6 -/- ES cells (Day 6: Pax6 +/+ mean percentage of apoptotic cells = 5.42% ± 1.82 s.e.m., n = 5 cultures; Pax6 -/- mean = 7.64% ± 3.05, n = 5; Day 8: Pax6 +/+ mean = 2.44% ± 1.27, n = 5; Pax6 -/- mean = 1.62% ± 0.65, n = 5).
As at Day 14, many Pax6 -/- SeyD1 and D2 cells remained negative for β-III-tubulin at Day 16 but the extensive processes surrounding all cells in the cultures made it impossible to count proportions of these β-III-tubulin-negative cells accurately. We found that some cells in wtMM4 and SeyD1 and SeyD2 lines were positive for GFAP, a marker of cells of glial identity (Figure 6).
To test whether primary neural progenitors derived from Pax6 -/- embryos showed increased cell death in culture, neural progenitors were isolated from wild type and Pax6 mutant cortex at E11.5 and expanded in vitro for 7 days. Neurospheres were dissociated and proportions of dead cells were determined from the size of the hypodiploid (subG1) population of cells by flow cytometry as described previously . No significant difference was observed in the proportions of hypodiploid cells between the wild-type and mutant cells (mean percentages of cells in subG1 were: Pax6 +/+, 12.6% ± 2.10 s.e.m.; Pax6 -/-, 7.8% ± 2.85, n = 3 cultures each). These data indicate that loss of Pax6 in primary neural progenitors does not increase cell death.
In many respects, the in vitro development of our new Pax6 -/- ES cells resembles that of previously derived Pax6 -/- ES cells  and that of primary Pax6 -/- progenitors taken from mutant embryos . Firstly, our new wild-type and Pax6 -/- mutant lines produced EBs containing proliferating progenitors in large proportions that were similar to those in EBs of wild type ES cells and of previously derived Pax6 -/- ES cells. This agrees with a previous conclusion that the generation of early neurogenic progenitors from ES cells is not perturbed in the absence of Pax6 . Secondly, Nikoletopoulou et al.  showed that abnormally high proportions of these progenitors expressed the ventral telencephalic marker Mash1; similarly, we found abnormal ectopic expression of Mash1 in Pax6 -/- ES-derived cells in the dorsal telencephalon of chimeras. Thirdly, previous studies have shown that primary Pax6 -/- progenitors from mutant cortex generate reduced proportions of neurons in culture ; we found the same trend in cells differentiated from our lines of Pax6 -/- ES cells in vitro. Together these data suggest that our novel Pax6 -/- ES cell lines are mimicking the in vivo and in vitro characteristics of neuronal progenitors derived from primary tissue of the Pax6 -/- embryo.
In one respect, however, our new Pax6 -/- ES cell lines developed quite differently to Pax6 -/- ES cell lines reported previously. Our Pax6 -/- ES cell-derived neurons showed no signs of early failure of process development and the extensive cell death previously reported; on the contrary, they were continuing to develop ever more elaborate processes at an age when virtually all of the previously described cells had died .
Since the culture protocol we used is similar to that used to differentiate previously derived Pax6 -/- ES cells, possible explanations for this difference in viability must centre of the nature of the ES cells themselves. One possibility is variation in genetic background. Our ES cells were derived from Pax6 mutant mice inbred on a 129Sv(Ola) background whereas those used by Nikoletopoulou et al.  were not (N.D. Allen, personal communication). It is possible that genetic background is critical for the manifestation of abnormally high levels of cell death observed in the previous study. This may not be surprising as genetic background has been shown to play a role in determining the exact nature of the eye and craniofacial defects associated with the Pax6 -/- phenotype .
Nikoletopoulou et al.  did not report the phenotype of Pax6 -/- cells derived from their mutant ES cells in mouse chimeras. They did, however, describe the introduction of Pax6 -/- ES cell derived progenitors into chick telencephalon: some of the neurons derived from these cells survived in the chick telencephalon, suggesting that even in these cells the absence of Pax6 does not necessarily cause extensive cell death given an appropriate supporting environment. Therefore induction of Pax6 genotype-dependent neuronal cell death might depend on both the genetic background of the mutant cells and the environment in which they are placed. It seems quite conceivable that loss of Pax6 from ES cells might cause the rapid death of all derived cells with some genetic backgrounds in vitro but not in vivo, whereas loss of Pax6 from ES cells with other genetic backgrounds might generate cells that are viable both in vitro and in vivo, as is the case in our novel Pax6 mutant cell lines.
Although Nikoletopoulou et al.  suggested that the rapid death of all Pax6 -/- ES cell derived neurons in culture reflects an increased incidence of cell death in the cerebral cortex of Pax6 -/- mice, other studies did not report an increase in apoptotic cells in the Pax6 -/- cortex in vivo or among Pax6 -/- primary cells in culture [8, 9, 30]. Data included in this present study on cultured cells from embryos support previous evidence against an increase in cell death in the absence of Pax6. On the contrary, Pax6 -/- neurons have been shown to survive for extended periods, even into postnatal life following transplantation into wild type embryos  and following early embryonic neural tissue-specific deletion of Pax6 . In the present study Pax6 -/- ES-derived neurons in chimeras also survived into postnatal life. Overall, most published studies point to the importance of Pax6 in the early regulation of neurogenesis [8, 9, 30–33], whereas the evidence that Pax6 has a prominent role in cell death remains controversial.
In conclusion, we have successfully derived Pax6 -/- ES cells that develop in ways predicted from previous work on Pax6 -/- progenitors and neurons in vivo and in vitro. Neurons derived from our new lines show enhanced viability compared to neurons derived from previously studied Pax6 -/- ES cell lines. Our new lines offer a good model for the study of Pax6 gene function in vivo and in vitro.
We thank Jenny Nichols, Sally Lowell, Rowena Smith, Ian Simpson, Katy Gillies, Ailsa Travers and Julie Thompson for their expert assistance. Research in the authors' laboratory is supported by the Medical Research Council (MRC), The Wellcome Trust and the Biotechnology and Biological Sciences Research Council (BBSRC).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.