Genetic deletion of genes in the cerebellar rhombic lip lineage can stimulate compensation through adaptive reprogramming of ventricular zone-derived progenitors
The cerebellum is a foliated posterior brain structure involved in coordination of motor movements and cognition. The cerebellum undergoes rapid growth postnataly due to Sonic Hedgehog (SHH) signaling-dependent proliferation of ATOH1+ granule cell precursors (GCPs) in the external granule cell layer (EGL), a key step for generating cerebellar foliation and the correct number of granule cells. Due to its late development, the cerebellum is particularly vulnerable to injury from preterm birth and stress around birth. We recently uncovered an intrinsic capacity of the developing cerebellum to replenish ablated GCPs via adaptive reprogramming of Nestin-expressing progenitors (NEPs). However, whether this compensation mechanism occurs in mouse mutants affecting the developing cerebellum and could lead to mis-interpretation of phenotypes was not known.
We used two different approaches to remove the main SHH signaling activator GLI2 in GCPs: 1) Our mosaic mutant analysis with spatial and temporal control of recombination (MASTR) technique to delete Gli2 in a small subset of GCPs; 2) An Atoh1-Cre transgene to delete Gli2 in most of the EGL. Genetic Inducible Fate Mapping (GIFM) and live imaging were used to analyze the behavior of NEPs after Gli2 deletion.
Mosaic analysis demonstrated that SHH-GLI2 signaling is critical for generating the correct pool of granule cells by maintaining GCPs in an undifferentiated proliferative state and promoting their survival. Despite this, inactivation of GLI2 in a large proportion of GCPs in the embryo did not lead to the expected dramatic reduction in the size of the adult cerebellum. GIFM uncovered that NEPs do indeed replenish GCPs in Gli2 conditional mutants, and then expand and partially restore the production of granule cells. Furthermore, the SHH signaling-dependent NEP compensation requires Gli2, demonstrating that the activator side of the pathway is involved.
We demonstrate that a mouse conditional mutation that results in loss of SHH signaling in GCPs is not sufficient to induce long term severe cerebellum hypoplasia. The ability of the neonatal cerebellum to regenerate after loss of cells via a response by NEPs must therefore be considered when interpreting the phenotypes of Atoh1-Cre conditional mutants affecting GCPs.
KeywordsCerebellum SHH signaling GLI2 Nestin-expressing progenitors Neurogenesis Atoh1-Cre Regeneration
External Granule Layer
Granule Cell Precursor
Green Fluorescent Protein
Genetic Inducible Fate Mapping
Internal granule cell layer
In situ hybridization
Mosaic mutant analysis with spatial and temporal control of recombination
Purkinje cell layer
Tandem Dimeric derivative of DsRed
Terminal deoxynucleotidyl transferase dUTP nick end labeling
The cerebellum (CB) consists of 80% of the neurons in the human brain  (60% in mouse ), and is involved in balance and motor coordination, but also modulates language, reasoning and social processes via its connections throughout the forebrain [3, 4, 5, 6, 7]. The CB undergoes its major growth in the third trimester and infant stage in humans, and the first 2 weeks after birth in mice, primarily due to expansion of the granule cell precursor (GCP) pool in the external granule cell layer (EGL) [8, 9, 10]. Given the late development of the CB compared to other brain regions, the CB is particularly sensitive to environmental and clinical factors that impact on growth (or cause injury) around birth. Furthermore, CB hypoplasia and prenatal injury is the second leading factor associated with autism . It is therefore important to identify genes that regulate cerebellum development. Many of the genes have been identified based on motor defects in homozygous null mutant mice, or in conditional mutants that remove genes in specific cell lineages. Intrinsic growth compensation mechanisms involving lineages where the gene does not function could however, obscure the normal function of a gene in cerebellar growth.
The CB develops from two germinal zones. The ventricular zone (VZ) gives rise to all the inhibitory neurons, including Purkinje cells (PCs)  as well as Nestin-expressing progenitors (NEPs) that expand in the cerebellar cortex after birth to produce astrocytes, including specialized Bergmann glia, and late born interneurons of the molecular layer [13, 14]. Ptf1aCre mice have been used to delete genes in inhibitory neurons and some glia . Excitatory neurons including granule cells (GCs) originate from the upper rhombic lip [16, 17, 18]. In mice, the EGL is established between embryonic day (E) 13.5 and E15.5. Atoh1-expressing GCPs then proliferate and expand in the EGL until ~postnatal day (P) 15 in response to Sonic Hedgehog (SHH) secreted by PCs [19, 20, 21]. When GCPs become postmitotic, they migrate down Bergmann glial fibers to form the internal granule cell layer (IGL). Interestingly, in rodent models the developing CB has been found to have a remarkable ability to recover from some injuries [22, 23, 24]. Indeed, we recently found that proliferating cerebellar GCPs can be replaced via adaptive reprogramming of NEPs after an acute depletion of the perinatal EGL due to irradiation [25, 26, 27]. Thus, NEPs in the neonatal CB have highly plastic behaviors. However, whether NEPs are harnessed to replenish cells lost in developmental mutants that lack key factors required for expansion and survival of GCPs has not been addressed.
One of the major pathways driving CB development is HH signaling. There are three hedgehog (Hh) genes in mammals, Indian (Ihh), Desert (Dhh) and Shh [28, 29]. Shh, the most widely expressed Hh gene, is required for development of most organs  by regulating a variety of cell behaviors including cell death, proliferation, specification and axon guidance. The cellular context (i.e. tissue, developmental stage, convergence of other signaling pathways) and concentration of SHH are thought to determine the particular response of a cell to SHH. HH signal transduction is mediated by the receptors Patched1 (PTCH1) and Smoothened (SMO) [28, 29, 30]. In the absence of HH signaling, PTCH1 constitutively represses SMO activity, whereas HH binding relieves this inhibition, in part by allowing accumulation of SMO in cilia . The GLI/Ci transcription factors are the effectors of the HH pathway. In mammals, the transcriptional activator (A) and repressor (R) functions of the GLIs have been divided between the three proteins . A general rule is that high levels of HH signaling induce the formation of a GLI2 activator (GLI2A) and this leads to transcription and translation of an addition activator, GLI1A, while a reduction or absence of the ligand allows for the formation of a GLI3 repressor (GLI3R). Importantly, we demonstrated that Gli1 expression is dependent on GLI2/3A, and thus is only expressed in cells receiving a high level of HH signaling [33, 34]. The three Gli genes, Shh, Smo, Ptch1 and Ptch2 are expressed in the CB and all but Ptch2 are required for CB development [20, 21, 35, 36, 37]. In particular, we have shown that SHH functions by inducing GLI1A/2A and is required for expansion of GCPs, primarily after birth [20, 38], whereas Gli3 is not required in the cerebellum after E12.5 . In addition to the crucial role of SHH in generating the pool of GCs, expansion of NEPs and thus production of NEP-derived interneurons and astroglia (astrocytes and Bergmann glia) also require SHH-signaling [13, 25, 39]. Furthermore, HH-signaling in NEPs is crucial for expansion of NEPs, recovery of the EGL and scaling of interneuron numbers after injury to the EGL .
Here we report that deletion of Gli2 in the vast majority of the GCPs is not sufficient to induce major cerebellar hypoplasia. Using our MASTR technique  in a mosaic mutant analysis of the effect of deleting Gli2 in scattered GCPs, we found that HH/GLI2-signaling is indeed necessary to maintain GCPs in an undifferentiated and proliferative state and to promote their survival. However, and similar to when the EGL is depleted using irradiation, we uncovered that NEPs are harnessed to repopulate the EGL and then wild type progenitors differentiate into GCs when Gli2 is deleted in most GCPs using an Atoh1-driven constitutive Cre . Our results not only provide more evidence for the unusual ability of the CB to recover from perinatal stress, but also reveal that NEP-dependent compensation should be taken into account when studying genes implicated in GCP development or survival and when using the Atoh1-Cre transgene.
The following mouse lines were used: Gli2flox/flox , Atoh1-Cre , Atoh-FlpoER, Nestin-FlpoER (a transgene similar to that described in ) and Rosa26MASTR(frt-STOP-frt-GFPcre) , Atoh1-GFP , Nes-CFP , Rosa26FRT-STOP-FRT-TDTom (Jackson Laboratory, 021875). The Atoh-FlpoER line, was made using the FLPoER1 cDNA described in  by subcloning it into the Atoh1 expression construct described in . All mouse lines were maintained on an outbred Swiss Webster background and both sexes were used for the analysis. Animals were housed on a 12 h light/dark cycle and were given access to food and water ad libitum. All experiments were performed using mice ages P0–P30.
Tamoxifen (Tm, Sigma-Aldrich) was dissolved in corn oil (Sigma-Aldrich) at 20 mg/ml. P2 Atoh1-FlpoER/+; R26MASTR/+; Gli2flox/flox, Atoh1-FlpoER/+; R26MASTR/+; Gli2flox/flox and P0 Nes-FlpoER/+; R26FSF-TDTom/+, Nes-FlpoER/+; R26FSF-TDTom/+; Atoh-GFP/+ mice as well as Nestin-FlpoER/+; R26MASTR/+; Gli2flox/flox, Atoh1-Cre/+; Gli2flox/flox and Nestin-FlpoER; R26MASTR/+; Atoh1-Cre/+; Gli2flox/flox mice and littermate controls received one 200 μg/g dose of Tm via subcutaneous injection. 50 μg/g 5-ethynyl-2_-deoxyuridine (EdU; Invitrogen) was administered via intraperitoneal injection (10 mg/ml in sterile saline) one hour before the animals were sacrificed.
Tissue processing, immunohistochemistry (IHC) and transcript detection
For animals younger than P4, they were anaesthetized by cooling and brains were dissected out and fixed in 4% paraformaldehyde overnight at 4 °C. Animals P4–30 received 50 μl intraperitoneal injections of ketamine and received ice-cold PBS via transcardial perfusion followed by 4% paraformaldehyde. Brains were collected and submersion fixed in 4% paraformaldehyde overnight at 4 °C. Tissues were processed for frozen embedding in optimal cutting temperature (OCT) compound and sectioned in the parasagittal plane on a Leica cryostat at 12 μm. For IHC, sections were incubated overnight at 4 °C with the following primary antibodies: rabbit anti-Ki67 (Thermo Scientific, RM-9106-S0), mouse anti-P27 (BD Pharmigen, 610,241), rabbit anti-PAX6 (Millipore, AB2237), goat anti-GLI2 (R&D System, AF3635), Goat anti-SOX2 (R&D System, AF2018), rabbit anti-GFP (Life Technologies, A11122), rat anti-GFP (Nacalai Tesque, 04404–84), mouse anti-NeuN (Millipore, MAB377) diluted in PBS with 5% BSA (Sigma-Aldrich) and 0.3% Triton X-100 (Fisher Scientific). Sections were then exposed for 2 h at room temperature to secondary species-specific antibodies conjugated with the appropriate Alexa Fluor (1:500; Invitrogen). EdU was detected using a commercial kit (Life Technologies) after the IHC reactions. TUNEL staining and in situ hybridization were performed according to standard protocols. Cre and Gli1 cDNAs were used as the template for synthesizing digoxygenin-labeled riboprobes. Images were collected on a DM6000 Leica microscope and processed using Photoshop software.
Ex vivo cerebellar slice culture was done as previously described . Briefly, P8 cerebella were embedded in 2.5% low-melting point agarose and saggitally sliced at 250 μM on a Vibratome. Slices were immediately taken to either a Leica TCS SP8 or SP5 confocal microscope platform. Slices were maintained in Eagle’s Basal Medium with 2 mM L-glutamine, 0.5% glucose, 50 U/ml Penicillin-streptomycin, 1xB27 and 1xN2 supplements at 37 °C and 5% CO2. Image stacks were acquired every 5 min for ~ 4 h. Cell tracking was performed using Imaris software. The autoregressive tracking function was employed with a spot size of 6 μM and a step size of 7 μM. Manual correction was performed.
Quantifications and statistical analyses
ImageJ software was used to measure the area (mm2) of cerebellar sections near the midline. For all IHC staining, cell counts were obtained using ImageJ and Neurolucida Software. For each developmental stage, three sections were analyzed per animal and ≥ 3 animals. Statistical analyses were performed using Prism software (GraphPad) and significance was determined at P < 0.05. All statistical analyses were two-tailed. For two-group comparisons with equal variance as determined by the F-test, an unpaired Student’s t test was used. Welch’s correction was used for unpaired t-tests of normally distributed data with unequal variance. P values are indicated in the figures. No statistical methods were used to predetermine the sample size, but our sample sizes are similar to those generally employed in the field. No randomization was used. Data collection and analysis were not performed blind to the conditions of the experiments.
Mosaic analysis reveals SHH-GLI2 signaling is critical for maintaining GCPs in an undifferentiated proliferative state and promoting their survival
As an alternative approach to a mosaic mutant analysis, we deleted Gli2 in the vast majority of GCPs (Atoh1-Cre/+; Gli2flox/flox or Atoh1-Gli2 CKOs). Consistent with previous studies using whole cerebellum Cre transgenes and our mosaic analysis, at P0 the anterior vermis of Atoh1-Gli2 CKOs (n = 5) appeared consistently smaller than controls (Gli2flox/flox) and the EGL was greatly diminished (Fig. 1g-j). SHH-GLI2 signaling loss was confirmed by the lack of Gli1 expression in the EGL of Atoh1-Gli2 CKO cerebella (Fig. 1g-h). Moreover, proliferation (Ki67) in the outer EGL and differentiation (P27 marking post mitotic cells in the inner EGL) were disrupted in the mutant EGL since two distinct EGL layers were not present (Fig. 1i-j). Interestingly, we observed an apparent increase of Gli1 expression in the Purkinje cell layer (PCL) suggesting that deletion of Gli2 in the EGL induced a cell non-autonomous up-regulation of HH-signaling in this layer (star in Fig. 1g-h). The lack of a phenotype in the posterior vermis is likely explained by low expression of Cre  in this region and thus low recombination  (Additional file 2: Figure S2).
All together, these results confirm a major role played by SHH-signaling through GLI2 to promote the expansion of the EGL and thus ensure the generation of the correct number of GCs.
The size of the Atoh1-Gli2 CKO cerebellum progressively recovers after birth
In summary, we found that depletion of the EGL at P0 by removing Gli2 from embryonic GCPs is not sufficient to induce consistent major hypoplasia of the vermis at P30. This raised the possibility of a compensation mechanism that allows partial recovery of the developing CB after genetic depletion of the EGL.
Wild type cells replenish the anterior EGL of Atoh1-Gli2 CKOs
NEPs switch their fate to become GCPs and produce GCs in Atoh1-Gli2 CKO cerebella
A subset of proliferating PCL NEP-derived cells migrate to the IGL in Atoh1-Gli2 CKO cerebella
Additional file 5: Video S1. P8 WT cerebellum shows no obvious movement of NEPs towards the EGL. Detection of native CFP fluorescence on sagittal slices of the vermis (lobule 2/3) of P8 Nes-CFP/+ mice showing displacement of CFP+ cells. Image stacks were acquired every 5 min for 4 h. (MOV 3481 kb)
Gli2 CKO in NEPs inhibits the recovery of the EGL in Atoh1-Gli2 CKOs
These results demonstrate that SHH-signaling through GLI2 plays a crucial role during NEP-mediated cerebellar recovery from loss of GCPs.
In this study we developed a conditional mutant strategy to delete Gli2 (the gene encoding the major effector of SHH signaling) in most GCPs in the anterior cerebellum using an Atoh1-Cre transgene used in many studies (e.g. [41, 45, 46, 47, 48, 49, 50, 51, 52]). Although we show using a mosaic analysis that SHH-GLI2 signaling is crucial for generating the correct pool of GCs by promoting GCP viability and proliferation, deletion of Gli2 in the EGL using this transgene is not sufficient to induce a major hypoplasia of the adult cerebellum in most mutants. We discovered that although the GCP pool is greatly diminished in Atoh1-Gli2 neonates, it is subsequently replenished by a cellular mechanism that includes adaptive reprogramming of WT NEPs to become GCPs. Importantly, since the transgene does not turn on in many of the newly generated GCPs, the EGL recovers and generates GCs. Rare GCPs that never express Cre could also contribute to replenishment of the GCPs if they can undergo more rounds of cell division than normal.
A question raised by this and previous studies is what signals induce the NEPs that reside in the PCL to change their fate and become GCPs. It was previously shown that when irradiation is used to kill most GCPs at P2–3, NEPs contribute to replenishment of the EGL . It is therefore possible that in Atoh1-Gli2 mutants a signal associated with the cell death we observed triggers NEPs to change their fate and generate GCPs. An alternative mechanism is that the depletion of the EGL results in a change in Purkinje cell signaling, possibly because of a reduction in excitatory input since less granule cells are generated. In turn, SHH might preferentially accumulate in the cell bodies of Purkinje cells, which lie close to the NEPs, thus increasing HH signaling to NEPs . Consistent with a role for GCs in regulating NEP behaviors, the Atoh1-Cre transgene is first expressed in GCPs at E13.5 , but the replenishment of the EGL in Atoh1-Gli2 CKOs only occurs several days after the EGL is depleted and when the IGL normally first becomes apparent (P3-P4). Thus, a possible involvement of NEPs in a compensation processes should be considered for conditional mutants that alter not only GCP proliferation and survival, but also genes involved in differentiation of GCs.
It might be expected that Atoh1-Smo CKOs would have a similar phenotype to Atoh1-Gli2 CKOs, given that both genes are required for SHH signaling. However, Atoh1-CreER/+; SmoloxP/Δ mice in which one allele of the Smoothened gene is deleted in the germline and deletion of the floxed allele is dependent on tamoxifen (Tm) administration have a severe cerebellum hypoplasia . A possible explanation for the phenotype in such mutants is that reprogramming of NEPs in Smo heterozygous mutants is partially compromised since HH signaling is crucial for the expansion of PCL NEPs and their migration to the EGL . In addition, since Tm diminishes cerebellum recovery after EGL depletion , the combination of a lower level of SMO protein in NEPs and administration of Tm to Atoh1-CreER/+; SmoloxP/Δ mice might blunt the response of NEPs to Smo-dependent depletion of the EGL leading to severe hypoplasia.
We observed a large variability in the vermal hypoplasia of Atoh1-Gli2 CKO adults, suggesting that only some mutant cerebella can efficiently recover. The variability in recovery is likely because the degree to which the Atoh1-Cre transgene is turned on in newly formed WT GCPs varies between mice. Since loss of Gli2 leads to cell death, the TUNEL staining observed in the EGL of P8 Atoh1-Gli2 CKO cerebella is consistent with Gli2 being deleted after P6 in some new wild type GCPs that are either derived from NEPs or rare GCPs that had not previously expressed Atoh1-Cre. In addition, our experiments indicate that some NEPs that enter the EGL and turn on PAX6 migrate back towards the IGL before becoming postmitotic and turning off Nes-CFP. We hypothesize that some NEP-derived GCPs that undergo deletion of Gli2 after entering the EGL survive but are unable to fully reprogram into GCPs. An interesting gene that might not be properly turned on is CxcR4, since SDF1 expressed by meninges signals through CXCR4 to maintain GCPs in the outer EGL and to enhance SHH-dependent proliferation . Furthermore, SHH-GLI1 signaling induces the transcription of Cxcr4 and Cxcr7 . We propose that in Atoh1-Gli2 CKOs a subset of PCL-derived NEPs express CRE after entering the EGL, and the subsequent loss of GLI2 protein reduces CXCR4, leading to migration of proliferating GCP-like cells back into the cerebellar cortex. Thus, the variable extent of growth of the cerebellum in Atoh1-Gli2 CKOs likely results from the amount of Atoh1-Cre induced after P4 in GCPs, and the resulting balance of cell death and premature migration from the EGL.
The cerebellum is broadly divided along the medio-lateral axis into a central vermis and two lateral hemispheres . Although recovery from depletion of the EGL at P0 occurs in both regions of Atoh1-Gli2 CKOs, the recovery is more robust in the hemispheres. Curiously, the hemispheres of Atoh1-Gli2 CKOs exhibit extra folds at P8 (arrow in Fig. 3g), suggesting a differential recovery response along the medio-lateral axis. The vermis and hemispheres are molecularly and functionally distinct [19, 55], and hemispheric GCPs have a higher sensitivity to high-level SHH-signaling than those in the vermis . We propose that hemispheric NEP-derived GCPs in the EGL maintain a higher level of SHH signaling and therefore expand more rapidly and efficiently than those in the vermis.
In this study, we show that the ability of NEPs to compensate for postnatal cerebellar damage must be considered in the interpretation of any mutant phenotype where genes involved in EGL cell proliferation/differentiation and survival have been disrupted. This is particularly the case if the Atoh1-Cre transgene utilized in this study  is employed to generate conditional mutants. Compensation for a loss of GCPs is most likely to occur for genes that are required after birth, once NEPs are present. Finally, our findings raise the question of whether similar recovery phenomena occur in other regions of the brain, and depending on the transgene used could complicate interpretation of mutant phenotypes.
We thank the past and present members of the laboratory for helpful discussions during the course of our study.
This work was supported by grants from the Brain Tumor Center at MSKCC and from the Philippe Foundation (to A.W), and from the NIH (R37 MH085726 and R01 NS092096 to A.L.J and F32 NS086163 to A.K.L.) and a National Cancer Institute Cancer Center Support Grant (P30 CA008748–48).
Availability of data and materials
All mouse lines are available from the Joyner lab or Jackson laboratories. All data generated or analyzed for this study are included in this published article.
AW and ALJ conceived the project; AW, AKL and ALJ designed the research; AW, MM, AKL and DNS performed the experiments; AW, MM, AKL and ALJ analyzed the data and all authors discussed the data; AW and ALJ wrote the manuscript with contributions from all authors. All authors read and approved the final manuscript.
Ethics approval and consent to participate
All animal procedures were performed according to a protocol (07–01-001) approved by the Memorial Sloan Kettering Cancer Center’s Institutional Animal Care and Use Committee.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 8.Altman J, Bayer SA. Development of the cerebellar system in relation to its evolution, structure, and functions. Boca Raton: CRC Press; 1997.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.