Nestin-cre transgenic mice have been widely used to direct recombination to neural stem cells (NSCs) and intermediate neural progenitor cells (NPCs). Here we report that a readily utilized, and the only commercially available, Nestin-cre line is insufficient for directing recombination in early embryonic NSCs and NPCs. Analysis of recombination efficiency in multiple cre-dependent reporters and a genetic mosaic line revealed consistent temporal and spatial patterns of recombination in NSCs and NPCs. For comparison we utilized a knock-in Emx1cre line and found robust recombination in NSCs and NPCs in ventricular and subventricular zones of the cerebral cortices as early as embryonic day 12.5. In addition we found that the rate of Nestin-cre driven recombination only reaches sufficiently high levels in NSCs and NPCs during late embryonic and early postnatal periods. These findings are important when commercially available cre lines are considered for directing recombination to embryonic NSCs and NPCs.
Cre mediated recombination is a powerful and broadly utilized genetic approach to conditionally delete or express selective genes, and is extensively used to lineage trace distinct progenitor populations during development (reviewed by Branda and Dymecki, 2004; Bouvier and Cheng, 2009). Transgenic Nestin-cre lines of mice have been broadly utilized over the past decade to direct recombination to neural stem cells (NSCs) and intermediate neural progenitor cells (NPCs) (Gavériaux-Ruff and Kieffer, 2007). Several independent Nestin-cre lines have been generated and published (e.g. Isaka et al., 1999; Trumpp et al., 1999; Petersen et al., 2002; Dubois et al., 2006), but only one (Tronche et al., 1999) is available through the Jackson laboratories and has been used in numerous studies. Based on early reports by the originating investigator it has been assumed that this line is highly efficient in directing cre-mediated recombination to neural progenitors. Here we tested whether or not the only commercially available line of Nestin-cre mice is sufficient for recombination in NSCs and NPCs of the forebrain during embryonic and early postnatal development. For comparison we utilized a knock-in Emx1cre line that is also broadly used for cre-mediated recombination in NSCs and NPCs of the dorsal telencephalon.
Materials and Methods
Animals were used under Institutional Animal Care and Use Committee and North Carolina State University regulations, and were housed in the Laboratory Animal Research facility at the College of Veterinary Medicine. Nestin-cre (B6.Cg-Tg (Nes-cre)1Kln/J; Jackson Lab, #003771; on a mixed C57BL/6/SJL background) (Tronche et al., 1999) and knock-in Emx1cre mice (B6.129S2-Emx1tm1(cre)Krj/J; Jackson Lab, #005628; on a 129S2/SvPas background) (Gorski et al., 2002) were crossed to mice expressing the reporter genes tdTomato (tdTom) (B6.129S6-Gt (ROSA) 26Sortm14 (CAG-tdTomato) Hze/J; Jackson Lab, #007908; on a mixed 129S6/SvEvTac/C57BL/6NCr background) (Madisen et al., 2010) or LacZ (B6.129S4-Gt (ROSA) 26Sortm1Sor/J; Jackson Lab, #003309, on a 129S4/SvJaeSor background) (Soriano, 1999), which are conditionally expressed under a knock-in pCAG promoter into the Rosa locus to track cre-mediated recombination.
For mosaic analyses, MADM11-TG/GT:Nestin-cre or MADM11-TG/GT:Emx1cre mice were generated using breeding schemes previously described (Hippenmeyer et al., 2010). In brief, Nestin-cre or Emx1cre mice were first crossed to the MADM11-GT line (on a 129X1/SvJ background) and the resulting MADM11-GT:Nestin-cre or MADM11-GT: Emx1cre mice were then crossed to the MADM11-TG+ mice (also on a 129X1/SvJ background) to generate the MADM11-TG/GT:Nestin-cre or MADM11-TG/GT:Emx1cre mice for mosaic analyses.
For embryonic analyses, the morning of observed vaginal plug was designated as embryonic day 0.5 (E0.5). Embryos were harvested from time-pregnant females following Avertin overdose (7.5 mg/g body weight), their brain removed and fixed with 4% paraformaldehyde for a minimum of 24 hours prior to processing. New born (P0) mice were anesthetized by hypothermia followed by transcardial perfusion with 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS). All mice received a pulse of 5-bromo-2′-deoxyuridine (BrdU, Aldrich; 100 mg/kg body weight via intraperitoneal injections to pregnant females or P0 pups) one hour prior to harvestation.
Tissue processing and imaging
Following perfusion, brains were removed and sectioned in the sagittal plane at 50 µm on a vibratome (Leica VT 1000 S). Floating sections were immunohistochemically processed using standard procedures with rabbit anti-RFP (Abcam; 1:1000), chicken anti-GFP (Abcam; 1:1000), mouse anti-BrdU (BD Bioscience; 13:1000), rabbit anti-cre (Covance; 1:500), and rabbit anti-PH3 (Millipore; 1:500) antibodies. Immunoreactivity was visualized using Alexafluor conjugatged secondary antibodies (Invitrogen). When needed, sections were counterstained with TO-PRO-3 Iodide (Invitrogen; 1:2000) or DAPI (VECTASHIELD mounting medium, Vector labs) for cytoarchitectonic characterization of sections.
Staining of whole embryos for LacZ was as described previously (Nagy et al., 2007). For tissue preparation, embryos were cryopreserved within 30% sucrose in 0.1 M PBS overnight at 4°C followed by overnight freezing in tissue freezing medium (Triangle Biomedical Sciences) at −80°C. Embryos were then sectioned at 20 µm on a cryostat (Leica).
Percentages of tdTom+ recombined cells in cycling progenitors in various regions/tissues were calculated by dividing numbers of either BrdU+/tdTom+ pulse labeled progenitors, or PH3+/tdTom+ mitotic progenitors, by total number of counted BrdU+ or PH3+ nuclei, respectively (n = 3 animals; 5 sections per animal; a total of 225 and 217 PH3+ cells at E12.5 and E14.5 respectively; a total of 789 BrdU+ cells at P0). Significance was determined using Student's t-test and all values were expressed as mean ± standard error of the mean (SEM).
Results and Discussion
The efficiency of recombination in the central nervous system (CNS) was assessed in multiple reporter lines using the only Nestin-cre line of mice that was commercially available at the time of this report (Fig. 1A). Similar to past reports on the same line, we found that Nestin-cre mediated recombination commences in the CNS during embryogenesis (Fig. 1B). However, much to our surprise, in mice on a tdTomato (tdTom) reporter background the recombination rate was extremely insufficient in the ventricular (VZ) and subventricular (SVZ) zones of the neocortex during early (E12.5) and mid stages of forebrain development (E14.5) (Fig. 1B). Additional expression analyses using cre-specific antibodies confirmed the observed patterns at the various stages of development (Fig. 1C). High-magnification imaging revealed an extremely low recombination rate in PH3 or BrdU immunoreactive cycling NSCs and NPCs in the E14.5 lateral ganglionic eminence (LGE), olfactory ventricle (OV) and the cerebral cortex (Ctx) (Fig. 1D) (1.3±0.7% at E12.5 and 7.4±4% at E14.5 of PH3+ cells in the VZ and SVZ, respectively; n = 3 mice/group). Consistent with these findings, Nestin-cre on a different reporter line (Rosa-lacZ) also yielded low levels of reporter activity in the VZ of E12.5 cerebral cortices (Fig. 1D).
Insufficient rate of recombination in NSCs and NPCs of Nestin-cre transgenic mice.
In both reporter backgrounds, the majority of recombined cells were situated in the preplate layer (PPL) of the E12.5 cortex as well as intermediate zone (IZ) and cortical plate (CP) of the E14.5 cortex, suggesting that recombination became sufficient during postmitotic stages of neuronal differentiation (Fig. 1B–D). In contrast, the rate of recombination in NSCs and NPCs increased dramatically in the subependymal zone (SEZ) and the rostral migratory stream (RMS) of newborn (P0) mice (Fig. 1B,E). Nearly every BrdU-labeled NSC and NPC expressed tdTom (93±5% of BrdU+) (Fig. 1E), suggesting sufficient Nestin-cre mediated recombination in these cells. These findings suggest a gradual increase in the rate of Nestin-cre mediated recombination during perinatal development in NSCs and NPCs.
For comparison we utilized a knock-in Emx1cre line of mice (Fig. 2A) that has been reported to drive sufficient recombination in dorsal telencephalic NSCs and NPCs (Gorski et al., 2002). In contrast to the Nestin-cre line, Emx1cre mediated recombination was sufficient in the VZ/SVZ of the developing cortex at both E12.5 and E14.5 (Fig. 2B), which we confirmed again using antibody staining for cre (Fig. 2C). In addition, robust recombination was found in the majority (83±8%) of PH3+ progenitors in the E12.5 VZ and SVZ of the cortex (Fig. 2D). The efficiency increased to nearly 100% by E14.5 (Fig. 2D). Thus, the low degree of recombination in Nestin-cre brains is not due to potential issues associated with cre-loxP interactions in embryonic progenitors, but is most likely related to the transgene in this specific strain.
Sufficient rate of recombination in embryonic NSCs and NPCs of Emx1cre knock-in mice.
To conclusively confirm our surprising findings on the developmental patterns of cre-mediated recombination in NSCs and NPCs in the Nestin-cre lines, and to rule out possible recombinogenic issues related to the Rosa locus in the reporter lines in the first part of the study, we crossed both cre lines into the Mosaic Analysis with Double Markers on chromosome 11 (MADM-11) line of mice (Hippenmeyer et al., 2010). In this line, cre-dependent mitotic recombination results in expression of GFP or tdTom independently or together resulting in various cells with three potential expression patterns (Fig. 3A, red, green, yellow). Recombination can also occur in postmitotic cells during the G0 phase, but in this case only cells expressing both GFP and tdTom (yellow cells) will be obtained (Fig. 3A). Using this strategy we found recombined cells with all three colors (red, green, yellow) throughout all layers of the cerebral cortices in Emx1cre MADM11 brains at E12.5 (Fig. 3B). In addition, recombined NSCs and NPCs were co-labeled with the mitosis marker PH3 in the VZ region (Fig. 3B). Thus, the Emx1cre knock-in allele is sufficient for driving mitotic recombination in NSCs and NPCs as early as E12.5.
Analysis of mitotic and post-mitotic specificity of recombination in the cerebral cortices of Emx1cre and Nestin-cre mice using the MADM-11 genetic system.
In contrast to the Emx1cre line, Nestin-cre mediated recombination in the MADM-11 background was extremely infrequent in the cerebral cortices at E12.5, and the few recombined cells failed to co-label with PH3 in the VZ. Moreover, the few MADM-11 cells were predominantly situated in the preplate (PPL) and all of these cells expressed both GFP and tdTom, suggesting that recombination had occurred during postmitotic stages of their development (Fig. 3B). At E14.5, a few more recombined cells could be found in Nestin-cre:MADM-11 cerebral cortices, but fewer clones could be detected in the VZ/SVZ compared to IZ/CP situated cells. Moreover, the majority of VZ/SVZ recombined cells were largely devoid of PH3 immunostaining (Fig. 3C). By E17.5, recombination became sufficient in the entire cerebral cortex and PH3+ recombined cells in both VZ and SVZ regions could be readily detected (Fig. 3C, arrows). These data confirm our finding that mitotic recombination (i.e., recombination in NSCs and NPCs) only becomes sufficient during perinatal development in the specific Nestin-cre line we obtained from the Jackson laboratory.
In summary, our study unraveled an important technical issue that has significant implications for neurodevelopmental investigators using the cre-lox technology as a genetic approach. We found that a readily utilized, and the only commercially available, Nestin-cre line of mice is a poor cre-driver for directing recombination to embryonic neural stem- and progenitor- cells. Recombination commences in the embryonic brain as reported in the past, but is largely confined to postmitotic neurons such as those occupying the preplate and cortical plate of the developing cortex. We think this discrepancy previously led many investigators to assume that the high rate of recombination includes progenitors in the VZ and SVZ of the developing central nervous system. However, we find that the rate of Nestin-cre driven recombination reaches nearly 100% in neural and glial progenitors during perinatal development. Whether or not this temporal gradient of recombination efficiency is a transgenic issue in the specific line of Nestin-cre mice used in our study, or if the endogenous Nestin promoter also exhibits a similar expression pattern remains to be determined. Temporal and spatial patterns of transgene expression have inherent problems including uncontrollable random copy number and integration site of the transgene (reviewed by Haruyama et al., 2009). For example, a transgene may be integrated into a locus that contains repressor sites or one with epigenetic modifications that do not favor transcription. It is also known that dynamic alterations occur in both transcriptional and epigenetic machinery in maturing NSCs (Juliandi et al., 2010). Hence, randomly integrated transgenes may be transcriptionally repressed in stem cells, but become active upon differentiation, or vice versa, as may be the case in the Nestin-cre line tested in this study. Thus, investigators planning to use the only Nestin-cre line of transgenic mice distributed by the Jackson Laboratory should take this technical limitation into consideration prior to utilizing it in their studies.
This work was supported by NIH grant R01NS062182, a grant from the American Federation for Aging Research, and generous institutional funds to H.T.G.
The authors have no competing interests to declare.