Generation of transgenic Dcx-CreERT2 mice
We previously demonstrated that a 3509-bp DCX genomic fragment could properly drive expression of reporter genes in neuronal precursors and immature neurons in vitro
and in vivo
]. Therefore, the CreERT2 encoding sequences were subcloned downstream of this DCX regulatory fragment (Additional file 1
). Two male founders carrying the CreERT2 transgene were obtained after pronuclear injection. Both founders transmitted their transgene to the F1 generation and Southern blot analysis suggested that only 1 copy of the transgene was integrated into the host genome (Additional file 1
Cre-recombinase activity was assessed on the F1 generation of both founder-derived lines following mating with Rosa26lacZ
reporter mice bearing a lacZ expression cassette activated following recombination [23
]. Two month-old DCX-CreERT2:Rosa26lacZ
mice were perfused two weeks after a tamoxifen (TAM) or vehicle injection and stained for β-gal activity. Both DCX-CreERT2 transgenic lines exhibited the expected TAM-induced β-gal expression in the adult neurogenic regions, i.e. SVZ and dentate gyrus (Additional file 1
). In contrast, no β-gal activity was observed following vehicle injections in the progeny derived from founder 2. However, in mice derived from founder 1, numerous β-gal positive profiles could be detected after vehicle injection indicating unspecific recombination events (data not shown). Therefore, only the transgenic DCX-CreERT2 founder 2 line was expanded and employed for the following experiments.
Expression and sub-cellular localization of CreERT2 in DCX positive (DCX+) cells
To validate that CreERT2 expression coincides with endogenous DCX expression in the DCX-CreERT2 transgenic mice, we investigated their respective expression patterns. At the cellular level, CreERT2 was detected in virtually all DCX+ cells of the developing CNS (E15.5). Furthermore, one day after injection of TAM, the CreERT2 has been translocated to the nucleus (Figure ). Similarly, in the adult brain, CreERT2 expression was, one day after TAM injection, strictly restricted to the nucleus of DCX+ cells (Figure ). Nuclear localization is induced by TAM administration and is a prerequisite for CreERT2 activity [24
Figure 1 CreERT2 cellular sub-localization in DCX-expressing cells. a) Immunodetection showing subcellular localization of CreERT2 (green) within DCX-expressing cells (red) in the cortex of an E15.5 embryo injected with TAM at E14.5. b to e) Cre subcellular localization (more ...)
To determine the time window in which CreERT2 exerts its function in the nucleus after the TAM injection, DCX-CreERT2 adult mice were perfused at different time points post-injection and the sub-cellular localization of the CreERT2 was assessed. Seven days after TAM injection, the nuclear localization was dramatically decreased as compared to the first day. At this time point, CreERT2 expression was still co-localized with DCX, but its distribution returned to be mostly cytoplasmic (Figure ). Furthermore, two weeks after TAM injection, CreERT2 was exclusively localized in the cytoplasm (Figure and ). Taken together, our results indicate that the CreERT2 nuclear localization rapidly recedes after the last TAM administration, indicating that CreERT2 activity was transient and virtually ceased after 7 days.
Assessment of CreERT2 activity in neuronal precursor cells
Having confirmed the correct co-localization of CreERT2 with DCX+ cells, we then analyzed recombination activity and specificity. To this end, we mated DCX-CreERT2 mice with Rosa26lacZ or CAG-CAT-EGFP reporter mice, which allow to monitor the activation of the respective reporter gene expression, following successful excision of the loxP-flanked cassette. The fate of DCX-expressing cells can then be followed by analyzing reporter gene expression at various time points following recombination.
The CreERT2 activity at embryonic stages was analyzed first. Twenty-four hours after a single TAM injection performed on E14.5, X-gal staining revealed that β-gal expression was restricted to the developing CNS and dorsal root ganglia (DRGs) (Figure ). This distribution coincided with the pattern of endogenous DCX expression at this stage (Figure ). Activation of CreERT2 at E17.5 by a single TAM-injection led to a wide distribution of EGFP reporter expression in the adult brain (Figure to and Additional file 2
). EGFP+ cells were detected in most regions of the brain parenchyma, such as in the granular cell layer of dentate gyrus, striatum, cortex, thalamus, Ammon's horn (CA1), etc. according to DCX expression pattern at E17.5.
Figure 2 Induction of reporter gene expression by TAM injections at different developmental stages. a) Immunodetection of DCX expression in an E14.5 mouse embryo. Detection of β-gal activity at E15.5 in a DCX-CreERT2: Rosa26lacZ embryos injected with (b) (more ...)
Noteworthy, virtually all EGFP+ cells expressed NeuN, and thus had become mature neurons (Figure to ). EGFP expression was neither detected in DCX-positive cells in the SVZ (Figure and ), nor in the rostral migrating stream (RMS) (data not shown) and subgranular zone (SGZ) of the dentate gyrus (Figure and ). A few EGFP+ cells could be found in or in close association to the ependymal layer of the lateral ventricles (Figure , arrows). These EGFP+ cells, however, did neither co-express NeuN nor DCX and their nature remains to be elucidated.
Due to the lower numbers of neurons continuously generated in the adult CNS, adult mice were injected with TAM on 5 consecutive days and then perfused for analysis 4 weeks after the last injection. At this time point, SVZ-generated EGFP+ cells reached the OB and were distributed mainly within the granular cell layer (GrO) (Figure and ) and to a lower extent in the periglomerular cell layer (pGl). The EGFP+ cells present in the OB were found to express the mature neuronal marker NeuN, whereas no co-expression of DCX could be detected (Figure and ). Within the rostral RMS in contrast, a few scattered EGFP-expressing cells appeared to have retained expression of DCX. Moreover, only rare EGFP+ cells could be found in the SVZ (data not shown). Similarly to cells detected in the OB, four weeks after TAM administration, EGFP+ cells in the dentate gyrus expressed NeuN and were devoid of DCX (Figure and ).
The efficiency and the specificity of the recombination event in DCX-expressing cells were evaluated in the adult SVZ and SGZ. The percentage of all DCX+ cells expressing EGFP was defined as the efficiency of recombination, and the percentage of all EGFP+ cells expressing DCX was defined as the specificity. Since levels of accumulated EGFP within cells were relatively low at the earliest time point studied, EGFP signals were amplified using an anti-EGFP antibody in all following experiments. In the SVZ, we observed that approximately 94% DCX+ cells were co-expressing EGFP two days after the last TAM injection (Figure and ). At the same time point in the SGZ of the dentate gyrus, EGFP could be detected in approximately 77% of DCX-expressing cells (Figure and ). Moreover, approximately 96% EGFP+ cells in SVZ and 90% EGFP+ cells in SGZ (Figure ) were co-expressing DCX, demonstrating that the CreERT2 activity was highly efficient and specific.
Figure 3 Co-localization of EGFP and DCX expression following TAM administration. Immunodetection of DCX (red) and EGFP (green) in the neurogenic regions of adult DCX-CreERT2:CAG-CAT-EGFP mice 2 days (a-d), 8 days (e-h) and 15 days (i-l) following last TAM administration. (more ...)
In order to characterize further the kinetics of the DCX+ cells' emigration from their site of birth to their target structures, the distribution of EGFP-expressing cells following recombination was investigated over time. To this end, adult mice were sacrificed eight, fifteen or twenty-nine days after the last TAM injection (D8, D15 and D29, respectively). Co-localization of EGFP expression with DCX and NeuN was analyzed within the SVZ-RMS-OB axis and within the dentate gyrus at these time points.
In contrast to the high percentage of EGFP-expression within DCX+ cells two days after the last TAM injection, the percentage of EGFP-expressing DCX+ neurons in the SGZ decreased to 41% at D8, and declined gradually to roughly 25% of all DCX+ cells at D15 (Figure ). Concomitantly, the frequency of co-localization in the SVZ decreased at D8 to 26.7%, further diminished to 12.5% at D15 (Figure ). Finally, at D29, only rare EGFP+ cells remained within the SVZ and were found to express DCX, whereas no co-localization could be detected in the SGZ at this time point. Taken together, our data indicate that the main emigration wave of EGFP-labeled neurons departed away from the SVZ within the first 15 days (Figure ).
Neuronal phenotypes of EGFP+ cells integrated into adult neurogenic regions
EGFP+ cells migrating along the RMS from D2 to D15 were found to maintain an immature neuronal morphology and only rare co-localization with NeuN could be documented (Figure ). Over the next 4 weeks, expression of NeuN within EGFP+ cells was broadly induced as cells reached the GrO or the pGl of the olfactory bulb (Figure and , Figure and ). Notably, in the GrO, we found only weak expression of DCX in EGFP+ cells. In contrast, DCX was still strongly expressed in the cytoplasm of EGFP+ located at the anterior end of the RMS (Figure , arrow), revealing that the expression of DCX decreased gradually as EGFP+ cells migrated into their target areas (data not shown). These observations reveal that DCX expression in cells migrating to the olfactory bulb is regionally regulated. In a similar fashion, EGFP-expressing cells in the dentate gyrus integrated in the inner granular layer over time and gradually induced the expression of NeuN. Quantitative analyses revealed that 15 days after the last TAM administration, more than 80% of EGFP+ cells found in the dentate gyrus (Figure ) and virtually all EGFP+ cells found in the olfactory bulb (data not shown) expressed the mature neuronal marker NeuN.
Figure 4 Co-localization of EGFP and NeuN expression following TAM administration. Immunodetection of NeuN (red) and EGFP (green) in the neurogenic regions of adult DCX-CreERT2:CAG-CAT-EGFP mice 2 days (a-d), 8 days (e-h) and 15 days (i-l) following last TAM administration. (more ...)
To further characterize the neuronal phenotypes of EGFP-labeled neurons, the presence of neurotransmitter-specific markers and calcium-binding proteins was investigated by immunohistology at D29 (Figure ). In agreement with previous studies [25
], expression of GAD65, a marker found in GABAergic neurons, could be detected at this time point in EGFP-expressing cells located in the OB. Moreover, a sub-population of periglomerular EGFP+ cells was found to co-express TH, a marker specific for dopaminergic neurons (Figure ).
Figure 5 Co-localization of EGFP and neuronal subtype-specific markers. Twenty-nine days after the last TAM administration, the EGFP-expressing cells (green) were examined for the expression of specific markers (red). (a) co-localization with the dopaminergic (more ...)
On the other side, VGLUT2, a commonly used marker for glutamatergic terminations, could be detected in the granular layer of the dentate gyrus and surrounded EGFP+ cells at D29, revealing that EGFP-expressing cells received glutamatergic inputs (Figure ). In addition, we scrutinized for the expression of the calcium-binding proteins calbindin-D28K, calretinin and parvalbumin in the EGFP-labeled granule neurons at this time point. The calbindin-D28K, which is expressed in mature granule neurons, could be detected in most EGFP+ cells of the dentate gyrus (Figure ). In contrast, no parvalbumin and only weak expression of calretinin could be detected in EGFP+ cells of the dentate gyrus at this time point, although cells expressing high levels of parvalbumin or calretinin, the latter been specifically found in newly generated granule cells, could be detected in the vicinity (see for example arrow in Figure ).
Proliferative capacity of EGFP+ cells in the adult neurogenic regions
DCX expression takes place in a relatively heterogeneous population of neuronal precursors and young neurons of various maturation stages and proliferative statuses. Administration of BrdU to label proliferative cells was performed at various time points following recombination. At the earliest time point investigated, D2, approximately 51.1% of EGFP+ cells in the SVZ, but only 7.7% of EGFP+ cells in the SGZ incorporated BrdU respectively, confirming that a fraction of the cells was still proliferating (Figure ). The higher proportion of mitotically active cells in the SVZ might be explained by the emigration of post-mitotic young neurons out of the SVZ towards the OB, leaving the most immature cells behind. Furthermore, no more BrdU incorporation in EGFP+ cells was observed in SGZ at D15 (Figure ), underscoring the temporally limited proliferative capacity of DCX+ cells. In contrast, although no more BrdU incorporation in EGFP+ cells was observed in SGZ at D15, a few BrdU+/EGFP+ co-labeling cells could be found in RMS (data not shown), suggesting that some EGFP+ cells in SVZ/RMS/OB axis could keep proliferative capacity until at least D15.
Figure 6 Proliferative potential of EGFP-labeled cells. One day or 14 days after the last TAM injection, dividing cells were labeled via a single BrdU injection and co-localization of EGFP with BrdU was examined at day 2 (a and c) and day 15 (b and d), respectively, (more ...)
EGFP+ cells outside of adult neurogenic regions
Following administration of TAM to adult DCX-CreERT2:CAG-CAT-EGFP mice, EGFP+ cells could be detected outside of the described neurogenic regions. In this respect, scattered DCX-expressing cells have previously been reported in adult cerebral cortex of rodents, cats and primates [7
]. To validate that EGFP expression in cells located outside of the neurogenic regions ensued from concomitant expression of DCX, we scrutinized by immunohistochemistry DCX expression pattern in the whole adult brain in respect to the activation of the EGFP expression.
Low to moderate levels of DCX expression could be detected in cells dispersed throughout the cerebral cortex (Figure ). Some weak DCX expression could also be perceived in corpus callosum, as well as around the 3rd ventricle and hypothalamus, the molecular cell layer (MCL) and the granular cell layer (GCL) of cerebellum (data not show). Four weeks after TAM injection, EGFP+ cells also appeared in these regions (Figure ). The reporter expression levels however were significantly higher than the endogenous DCX expression levels, since upon recombination, expression of the reporter gene was under the control of a strong promoter. Importantly, EGFP+ cells located outside of the neurogenic regions were not proliferative, as demonstrated by the absence of BrdU incorporation. Therefore, the nature and fate of these immature neurons remains to be deciphered.
Figure 7 Expression of DCX and EGFP outside of neurogenic areas. Low levels of DCX expression can be detected outside of neurogenic areas. For example, DCX-expressing cells were detected by immunohistochemistry in the piriform cortex (a) and in the corpus callosum (more ...)