SEARCH

SEARCH BY CITATION

Keywords:

  • Ptc1+/ mice;
  • Apoptosis;
  • DNA repair;
  • p21;
  • Medulloblastoma;
  • Carcinogenesis;
  • Neural stem cells;
  • Progenitor cells

Abstract

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Neural stem cells are highly susceptible to radiogenic DNA damage, however, little is known about their mechanisms of DNA damage response (DDR) and the long-term consequences of genotoxic exposure. Patched1 heterozygous mice (Ptc1+/−) provide a powerful model of medulloblastoma (MB), a frequent pediatric tumor of the cerebellum. Irradiation of newborn Ptc1+/− mice dramatically increases the frequency and shortens the latency of MB. In this model, we investigated the mechanisms through which multipotent neural progenitors (NSCs) and fate-restricted progenitor cells (PCs) of the cerebellum respond to DNA damage induced by radiation, and the long-term developmental and oncogenic consequences. These responses were assessed in mice exposed to low (0.25 Gy) or high (3 Gy) radiation doses at embryonic day 13.5 (E13.5), when NSCs giving rise to the cerebellum are specified but the external granule layer (EGL) has not yet formed, or at E16.5, during the expansion of granule PCs to form the EGL. We found crucial differences in DDR and apoptosis between NSCs and fate-restricted PCs, including lack of p21 expression in NSCs. NSCs also appear to be resistant to oncogenesis from low-dose radiation exposure but more vulnerable at higher doses. In addition, the pathway to DNA repair and the pattern of oncogenic alterations were strongly dependent on age at exposure, highlighting a differentiation-stage specificity of DNA repair pathways in NSCs and PCs. These findings shed light on the mechanisms used by NSCs and PCs to maintain genome integrity during neurogenesis and may have important implications for radiation risk assessment and for development of targeted therapies against brain tumors. Stem Cells 2013;31:2506–2516


Introduction

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Stem cells (SCs) are essential during embryonic development for specification and morphogenesis and over the lifetime to maintain tissue and organism homeostasis. To insure the maintenance of genomic integrity, SC compartments need to be specifically equipped with a defense capacity against DNA damage deriving from endogenous and exogenous insult [1, 2].

The mechanisms of SC responses to DNA damage vary greatly among different tissues [3]. Common protective mechanisms to limit massive apoptosis following DNA damage to preserve tissue function have been identified in tissue SCs from blood and hair follicle [4, 5]. On the other hand, it has been proposed that SCs might prefer to die following DNA damage rather than attempt repair [6]. For instance, intestinal SCs undergo extensive apoptosis after DNA damage [7, 8]. Differential DNA damage response (DDR) has even been reported among SCs of the same tissues at different stages of ontogeny. SCs in the mouse bone marrow, for instance, are more resistant to radiation-induced apoptosis than more differentiated progenitor cells (PCs) [4, 9].

Maintaining genome stability during neural development is of paramount importance, and defects in DNA damage signaling are frequently associated with neurodegeneration, neurodevelopmental disease, and brain tumors. Double-strand breaks (DSBs), the most biologically significant DNA lesions, arise from oxidative damage, replication, and exogenous sources, including chemicals in the environment or ionizing radiations. The increasing use of medical radiation procedures such as computed tomography (CT) scans raises health concerns on potential cancer risks in bone marrow and brain of children [10], more radiosensitive compared to adults. SCs or stem-like cells are a target for radiation carcinogenesis. However, little is known about the molecular, cellular, and oncogenic responses to DNA damage of multipotent neural progenitors (named NSCs throughout the text) and fate-restricted PCs in their natural setting in vivo.

Irradiation of the cerebellum has multiple effects on its development, including reduction of the overall size, foliar/lobular malformations [11, 12], and functional deficiencies such as tremor, ataxia, hypoactivity, and cognitive deficits [13-17]. Populations of NSCs orchestrate the specification and patterning of the cerebellum. The cerebellar cells are born and migrate sequentially from two germinal zones. The first one is the ventricular zone (VZ), which contains NSCs that give rise to most of the neurons and glia. A subset of these cells move tangentially to the upper rhombic lip (RL), the second germinal zone, generating fate-restricted granule PCs. An important aspect of cerebellar neurogenesis in mammals is that distinct populations of NSCs and fate-restricted PCs are born during temporal windows of specification. In the mouse, starting at about embryonic day (E) 13–14, granule PCs leave the RL and migrate rostrally over the cerebellar anlage to form an external granule layer (EGL) of cells. The EGL starts to form at E15, and granule PCs remain mitotically active until the second postnatal week.

Medulloblastomas (MB) are malignant tumors of the cerebellum, considered pediatric tumors as less than 30% occur in individuals older than 16 years. Neural precursors of the VZ and the EGL are considered the cells of origin of MB. Origins of MB are deeply related to development of the normal cerebellum, a process in which the Sonic hedgehog (Shh) pathway plays a pivotal role. In the canonical Shh signaling, upon binding of Shh to its receptor Patched1 (Ptc1), Smoothened (Smo) is released from inhibition and activates downstream Shh targets genes. Aberrant activation of the Shh pathway has been unambiguously linked to MB in human and mice. Mice in which a copy of the Ptc1 gene has been knocked out (Ptc1+/) are developmentally nearly normal in a large number of instances but show a marked predisposition to tumor development, including MB [18, 19]. Our previous work has shown that irradiation of Ptc1+/ mice during postnatal days 1–4, when granule PCs are highly proliferative, dramatically increases the frequency of MB [20, 21]. Here, we investigate the mechanisms through which multipotent NCSs and fate-restricted PCs respond to DNA damage induced by radiation, the effect of radiation exposure on cerebellar patterning and their likelihood of oncogenic transformation. These responses were assessed in Ptc1+/ mice exposed in utero to high (3 Gy) or low (0.25 Gy) radiation doses at E13.5, when fate specification of NSCs generating the cerebellum has already occurred but the EGL is not formed, or at E16.5, during the burst of proliferation of granule PCs to form the EGL. The effects of homologous recombination (HR) and nonhomologous end joining (NHEJ) deficiency on apoptotic response of NSCs and granule PCs were also investigated using mouse strains with inactivated Rad54 or DNA-PKcs DNA repair genes.

Materials and Methods

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Animal Breeding

Mice lacking one Ptc1 allele (Ptc1neo67/+, named Ptc1+/ throughout the text) were generated through disruption of exons 6 and 7 in 129/Sv embryonic stem cells [19] and maintained on CD1 background. Genotyping for Ptc1 was performed as described previously [19]. Ptc1+/ mice were also crossed with Rad54/ [22], and DNA-PKcs/ [23] mice were maintained on C57Bl/6 background. F1 mice of the desired genotypes (DNA-PKcs+//Ptc1+/ × DNA-PKcs+//Ptc1+/+ and Rad54+//Ptc1+/ × Rad54+//Ptc1+/+) were intercrossed to produce F2 populations. Genotyping of mice was performed as described previously [24].

Animal Treatment and Irradiation

Mice were housed under conventional conditions with food and water available ad libitum and a 12-hour light cycle. Mice were irradiated in utero at E13.5 or E16.5 with 0.25 or 3 Gy of x-rays. Irradiation was performed using a Gilardoni CHF 320 G x-ray generator (Gilardoni S.p.A., Mandello del Lario, Italy) operated at 250 kVp, 5 mA for 0.25 Gy, and 15 mA for 3 Gy, half-value layer = 1.6 mm Cu (additional filtration of 2.0 mm Al and 0.5 mm Cu). Experimental protocols were reviewed by the ENEA Institutional Animal Care and Use Committee.

Tissue Collection

For DDR analysis, Ptc1+/ mice irradiated at E13.5 or E16.5 with 3 Gy and age-matching controls were sacrificed at 0.5 or 6 hours post-irradiation. For morphometric analysis, brains from Ptc1+/ mice irradiated at E13.5 or E16.5 with 0.25 or 3 Gy and age-matching controls were collected at P1, P7, and P14. For the analysis of DNA repair pathways to cell survival, brains from Ptc1+/, Ptc1+//DNA-PKcs−/−, or Ptc1+//Rad54−/− embryos irradiated at E13.5 and E16.5 were collected at 6 hours post-irradiation. All samples were fixed in 10% buffered formalin. For tumorigenesis, mice were observed daily for their lifespan, and upon decline of health (i.e., severe weight loss, paralysis, ruffling of fur, or inactivity), they were killed and autopsied. Normally appearing and tumor-bearing brains were fixed in 10% buffered formalin and partly snap frozen. Samples were processed for histological analysis using standard methods.

Morphometric Analysis

The cerebella and internal granule layer (IGL) total cross-sectional areas were measured in at least three cerebella for time point. To determine the EGL thickness at P14, 10 individual measurements of the EGL in folia VIII were taken for each cerebellum. The imaging software NIS-Elements BR 4.00.05 (Nikon, Instruments Europe B.V., Italy) was used for morphometric analyses.

Immunohistochemical Analysis

Paraffin sections were cut at 4-μm thickness for immunohistochemical analysis. Before incubation in 3% H2O2 in methanol for 10 minutes, sections were dewaxed, rehydrated, and heated in unmasking buffer. Antibodies used were: rabbit polyclonal antibody against 53BP1 and Sox2 (Abcam, Cambridge, U.K., http://www.abcam.com), cleaved caspase-3 (Cell Signaling Technology, Danvers MA, http://www.cellsignal.com), p21, B-cell lymphoma-2 (Bcl-2) (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), and p53 (Novocastra Laboratory, Newcastle, U.K., http://www.novocastra.co.uk).

Antibody–antigen complexes were visualized using a horseradish peroxidase-conjugated secondary antibody (Dako, North America, Inc., Carpinteria, CA, http://www.dako.com) and the 3,3′-Diaminobenzidine (DAB) chromogen system (Dako, North America) or a rabbit biotinylated-conjugated secondary antibody; after incubation with avidin-biotin immunoperoxidase, staining was visualized with Vector NovaRED Substrate Kit (Vector Laboratories, Inc., Burlingame, CA, http://www.vectorlabs.com). Immunohistochemistry analysis of monoclonal antibody against NeuN (Millipore, Billerica, MA, http://www.millipore.com) and Pax6 (Millipore) was performed using the HistoMouse MAX Kit (Invitrogen Corporation, Camarillo, CA, http://www.invitrogen.com) according to manufacturer's instructions.

Caspase-3 Quantification

Three to five tissue sections from RL/EGL of Ptc1+/−, Ptc1+/−/DNA-PKcs/, and Ptc1+/−/Rad54/ mice were immunostained for cleaved caspase-3 at 6 hours after irradiation and imaged by HistoFAXS software (TissueGnostics, Austria, http://www.tissuegnostics.com) at ×40 magnification. Specific regions of interest (dashed lines in Fig. 6B–6L) were analyzed with HistoQuest software (TissueGnostics) for automatic color separation and quantification.

Tumor Phenotyping

For tumor phenotyping, four digital images at ×40 magnification of tumors from Ptc1+/− mice irradiated with 3 Gy at E13.5 (n = 6) or E16.5 (n = 10), randomly selected among those that were not snap frozen, were collected by NIS-Elements F 3.2 software (Nikon Instruments Europe B.V.) and analyzed for Sox2 and NeuN. The numbers of Sox2 (brown stain) and NeuN (red stain) positive cells were calculated relative to area (mm2) using the imaging software NIS-Elements BR 4.00.05 (Nikon Instruments Europe B.V.).

Microsatellite Analysis at the Ptc1 Locus in MB

DNA from MBs and normal tissue was extracted using Wizard SV Genomic DNA Purification System (Promega Corporation, Madison, WI, http://www.promega.com). Polymerase chain reaction amplifications were performed as described previously [25]. About 17 microsatellites spanning the length of chromosome (chr) 13 were used to examine tumor DNA in comparison with the corresponding genomic DNA (Fig. 6A).

Statistical Analysis

Analyses were performed using GraphPad Prism version 4.02 for Windows (GraphPad Software, San Diego, CA). Statistical comparisons were made using Student's t test. Kaplan-Meier survival curves were compared, and log-rank test p values were calculated. p values <.05 were considered statistically significant.

Results

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Characterization of NSCs and Granule PCs at E13.5 and E16.5

Panels A and B of Figure 1 illustrate the embryonic cerebellar primordium at the time of irradiation: at E13.5, when only NSCs giving rise to the cerebellum are specified (A), and at E16.5, when the RL cells are migrating rostrally over the dorsal surface of the cerebellar anlage to form the EGL (B). We first examined the expression of Sox2, an embryonic transcription factor and SC marker, in unexposed Ptc1+/ embryos at E13.5 and E16.5. At E13.5, we detected a strong Sox2 expression in the whole RL (Fig. 1C), while local differences were observed at E16.5. In fact, Sox2 was markedly expressed only in the ventral RL region but not in the dorsal RL and the EGL (Fig. 1D, 1E). By contrast, Pax6, which has a critical role for granule cells differentiation and migration, was not expressed in either the RL at E13.5 or the ventral RL at E16.5 (Fig. 1F, 1G), but showed a robust expression in the rostral RL and EGL at E16.5 (Fig. 1G, 1H), consistent with the differentiation program of NSCs toward lineage-restricted PCs. As lengthening of the G1 phase in neural progenitors is associated with control of the switch from expansion to differentiation [26, 27], we examined expression of the p21 G1 checkpoint protein in unexposed Ptc1+/ embryos at E13.5 and E16.5. We show that p21 displayed a constitutive expression pattern complementary to that of Sox2 and overlying Pax6 expression (Fig. 1I–1K), therefore associating p21 with differentiation of PCs.

image

Figure 1. Differential expression pattern of stemness and differentiation markers in multipotent neural stem cells and fate-restricted progenitor cells (PCs) from embryos at E13.5 and E16.5 (A, B). Hematoxylin and eosin-stained sagittal sections of embryonic cerebellum at E13.5 and E16.5. Box in (A) represents higher magnification of the RL region at E13.5 in the panels below. Solid and dashed boxes in (B) show a higher magnification of the RL and EGL regions at E16.5 in the panels below. Dashed arrow in (B) indicates the path of migration of PCs over the cerebellum anlage to form the EGL. Immunohistochemical analyses for Sox2 (C–E), Pax-6 (F–H), p21 (I–K). Dashed lines in (D, G, and J) indicate the rostral RL and in (E, H, and K) the EGL. Scale bar = 100 μm (A, B), 10 μm (C–K). Abbreviations: E, embryonic; EGL, external granule layer; RL, rhombic lip.

Download figure to PowerPoint

DDR in NSCs and Granule PCs After In Vivo Irradiation with 3 Gy of X-rays at E13.5 or E16.5

We next investigated DDR signaling and radiosensitivity in NSCs at E13.5 and their descendants at E16.5. The DDR pathways are a complex network of sensors, mediators, and effectors that provide two main critical functions, DNA repair and activations of cell-cycle checkpoints that cause cell cycle arrest. We examined the in vivo response of multipotent NSCs and fate-restricted PCs to DNA damage in Ptc1+/− mouse embryos irradiated with 3 Gy of x-rays at E13.5 or at E16.5 by immunohistochemistry for 53BP1. We found that this protein, a DNA damage sensing protein used as surrogate marker for DNA DSBs, is expressed endogenously (Supporting Information Fig. 1A–1C) but relocates to form foci after irradiation (Fig. 2A–2C). Whereas at E13.5, 30 minutes after irradiation, we observed a strong induction of 53BP1 foci through the whole RL area (Fig. 2A), at E16.5 regional differences were observed in focal accumulation of 53BP1. In the ventral RL, where Sox2-positive cells reside, the 53BP1 response resembled that at E13.5. In contrast, lower levels of 53BP1 were detected in the dorsal region of the RL and in the EGL (Fig. 2B, 2C). This difference in 53BP1 staining is likely to reflect differential activation of DDR in multipotent NSCs and fate-restricted PCs in their natural setting in vivo.

image

Figure 2. Multipotent neural stem cells from embryos at E13.5 and fate-restricted progenitor cells from embryos at E16.5 show marked differences in DDR after irradiation with 3 Gy of x-rays. Immunohistochemical analyses for 53BP1 (A–C), p53 (D–F), p21 (G–I), cleaved caspase-3 (J–L), Bcl-2 (M–O). Insets in (B) and (C) higher magnification of 53BP1 foci. Scale bar = 10 μm (A–O). Abbreviations: Bcl-2, B-cell lymphoma-2; DDR, DNA damage response; E, embryonic; EGL, external granule layer; RL, rhombic lip.

Download figure to PowerPoint

We then evaluated the activation of p53-p21 DNA damage signaling pathway in mouse embryos at 6 hours after irradiation. While similarly strong p53 responses were detected in embryos at E13.5 and E16.5 (Fig. 2D–2F), marked differences in p21 induction were observed between embryos at E13.5 and E16.5 (Fig. 2G–2I). Consistent with the absence of p21 in mouse embryo SCs even after challenge with radiations [28], we found virtually no expression of the p21 G1 checkpoint protein in multipotent NSCs from the RL in embryos at E13.5 (Fig. 2G). By contrast, a marked p21 expression was found in the lineage-restricted granule PCs of the dorsal RL and in the EGL layer itself but not in the ventral RL in embryos at E16.5 (Fig. 2H, 2I). The different radiation-induced expression pattern in multipotent NSCs and restricted PCs was qualitatively similar to that of unirradiated embryos, although more intense, suggesting a meaningful correlation of the strength of p21 response to endogenous/exogenous damage on cell differentiation status.

To further explore the differential sensitivity of NSCs and PCs, we examined the activation of caspase-3, a well-established apoptotic marker, 6 hours after irradiation. We hypothesized that, in the absence of p21-dependent G1 arrest, damaged cells enter S-phase where replication exacerbates DNA damage resulting in cell death. However, despite dramatic caspase-3 expression in the VZ, virtually no caspase-3 expression or any other sign of apoptosis, such as nuclear pyknosis, were found in the RL at E13.5, suggesting that multipotent NSCs are highly resistant to DNA damage-induced apoptosis (Fig. 2J). Conversely, at E16.5 we found a marked caspase-3 expression and extensive nuclear pyknosis in both dorsal RL and EGL, with few apoptotic figures in the ventral RL region expressing Sox2 (Fig. 2K, 2L).

To determine whether the mechanism of resistance to radiation-induced apoptosis in NSCs involves enhanced expression of prosurvival genes, we analyzed the expression of the Bcl-2 protein in mouse embryos at 6 hours after irradiation. At E13.5, we detected a strong Bcl-2 expression in the whole RL (Fig. 2M). In contrast, region-specific differences in the staining pattern were again observed at E16.5. While NSCs in the ventral RL stained positively, the dorsal RL and the EGL layer, where the lineage-restricted granule PCs are located, showed little or no Bcl-2 immunostaining, respectively (Fig. 2N, 2O).

Effects of in Utero Irradiation with 3 Gy on Cerebellum Development and Tumorigenesis

The relatively few well-defined cell types within the cerebellum and the stereotypical foliation pattern make the cerebellum particularly amenable to the study of factors affecting morphogenesis. Here, we evaluated the effects of irradiation in the cerebellum of Ptc1+/ mice exposed to 3 Gy at E13.5 or at E16.5 relative to control mice. Animals were sacrificed at P1, P7, and P14, and the cross-sectional area of the cerebellum was determined. Compared to unirradiated embryos, irradiation at E13.5 caused a shallow principal fissure at P1 (asterisks in Fig. 3A, 3B) but did not significantly affect the size of the cerebellum at P1–14 (Fig. 3D, 3H, 3L) or cause gross defects in the foliation pattern (Fig. 3E, 3F, 3I, 3J). In contrast, the size of the cerebellum was significantly reduced both at P1 and P7 (p = .0003 and p = .0258) after irradiation at E16.5 (Fig. 3C, 3D, 3G, 3H). In addition, alterations of the principal fissures caused a clear agenesis of the folia I–V and defects in folia VI, as well as a marked sparseness of the IGL in the anterior cerebellar lobes at P14 (Fig. 3I, 3K). Therefore, irradiation of embryos at E16.5 impaired cerebellum patterning more severely compared to irradiation at E13.5.

image

Figure 3. Cerebellum morphogenesis after embryo irradiation. Exposure to 3 Gy of x-rays at E16.5 impaired the cerebellum patterning more severely compared to irradiation at E13.5. Hematoxylin and eosin-stained section of unirradiated (A), irradiated at E13.5 (B), or at E16.5 (C) cerebella at P1. (D): Quantification of the cerebellum size at P1 showing a significant decrease after irradiation at E16.5 (p = .0003). (E–G): Sections of (E) unirradiated, (F) irradiated at E13.5, or (G) at E16.5 P7-cerebella. (H): Quantification of the cerebellum size at P7 showing a significant decrease after irradiation at E16.5 (p = .0258). (I–K): Sections of (I) unirradiated, (J) irradiated at E13.5, or (K) at E16.5 P14-cerebella, in which a marked agenesis of the folia 1–5 and defects of folia 6 were observed. (L): Quantification of the cerebellum size at P14. Asterisks indicate principal fissures that were shallow either after irradiation at E13.5 or at E16.5 compared to unexposed embryos. *, p ≤ .05; ***, p ≤ .001. Scale bar = 400 μm (A–C, E–G, I–K). Abbreviations: E, embryonic; P1, postnatal day 1; P7, postnatal day 7; P14, postnatal day 14.

Download figure to PowerPoint

Cerebellum tumorigenesis was assessed in groups of Ptc1+/ mice prenatally exposed to 3 Gy of x-rays (E13.5, n = 22; E16.5, n = 24) and monitored for 40 weeks. Irradiation of Ptc1+/ mouse embryos at E13.5 led to rapid tumor formation, significantly accelerating MB development and mortality compared with mice irradiated at E16.5 (median survival 12 vs.19 weeks; p = .0058) (Fig. 4A). Results of this experiment indicate that MB can be initiated by irradiation of both multipotent NSCs and fate-restricted granule PCs of the developing cerebellum, however, exposure of NSCs results in enhanced oncogenesis compared to irradiation of fate-restricted PCs.

image

Figure 4. Effects of in utero irradiation with 3 Gy of x-rays on cerebellum tumorigenesis. (A): Kaplan-Meier kinetic analysis of medulloblastoma (MB) after irradiation at E13.5 or E16.5 showing significant increase in MB mortality at E13.5 (p = .0058). (B, C): Immunostaining for the stem cell marker Sox2 and (E, F) the mature neuron marker NeuN. Quantitative representation of Sox2 (D) and NeuN (G) immunostaining showing that tumors from embryos irradiated at E13.5 show increased expression of Sox2 and decreased expression of NeuN compared to tumors from embryos irradiated at E16.5 (p = .0354; p = .0334). *, p ≤ .05; **, p ≤ .01. Scale bar = 30 μm (B, C, E, and F). Abbreviation: E, embryonic.

Download figure to PowerPoint

To determine whether the differences in tumor onset between MBs developed in embryos irradiated at E13.5 or E16.5 reflect differences in tumor phenotype, we analyzed the expression of markers of NSCs (Sox2) (Fig. 4B–4D) and mature neurons (NeuN) (Fig. 3E–3G) in tumors from embryos irradiated at E13.5 (n = 6) or E16.5 (n = 10). Compared to MBs from E16.5-irradiated embryos, tumors from mice irradiated at E13.5 were more largely undifferentiated, showing a significant increase in the expression of Sox2 (49%; p = .0354) and a significant decrease in the expression of NeuN (55%; p = .0334). These results show that MBs from mice irradiated at E13.5 more closely resemble multipotent NSCs rather than committed granule PCs based on marker expression.

Effects of in Utero Irradiation with 0.25 Gy on Cerebellum Development and Tumorigenesis

DDR response mechanisms after low-dose irradiation are qualitatively different from those at high-dose [29]. We evaluated the effects of 0.25 Gy of x-rays on cerebellum morphogenesis in the Ptc1+/ mice exposed at E13.5 and E16.5. Mice were sacrificed at P1, P7, and P14, and the cross-sectional area of the cerebellum was determined. Neither significant size variation nor other morphologic abnormalities were observed in P1- and P7-cerebella from embryos irradiated at E13.5 and E16.5 (data not shown). Interestingly, however, we observed a significant increase of 30% (p = .0345) and 50% (p = .0090) in the cross-sectional area of P14-cerebellum from embryos irradiated at E13.5 and E16.5, respectively, compared to unexposed cerebellum (Fig. 5A–5D). Postnatal development of the cerebellum occurs through rapid proliferation of granule PCs in the EGL, followed by cell cycle exit and final inward migration to the IGL, causing progressive EGL disappearance and formation of the IGL. Therefore, to monitor cerebellum development in the different groups we measured the thickness of the EGL in folia VIII (Fig. 5E, 5F) and the area of the IGL in cerebellum at P14 (Fig. 5G, 5H). Compared to unirradiated cerebellum, irradiation at E13.5 had no significant effects on EGL thickness, while a significant increase in the IGL area was observed (p = .0120). By contrast, irradiation at E16.5 significantly decreased the thickness of EGL (p = .0107) and increased the IGL area (p = .0062). Taken together, these findings indicate that low-dose irradiation may preferentially affect differentiation in fate-restricted PCs (E16.5) compared to NSCs (E13.5).

image

Figure 5. Developmental and oncogenic effects of in utero irradiation with 0.25 Gy of x-rays. (A–C): Hematoxylin and eosin-stained sagittal sections of (A) unexposed, (B) irradiated at E13.5, or (C) at E16.5 postnatal day 14 (P14)-cerebella. (D): Quantification of the cerebellum area at P14, showing significant increases after irradiation at E13.5 or E16.5 (p = .0345; p = .0090). (E): Representative section of folia VIII from P14-cerebellum. (F): Quantification of the EGL thickness in folia VIII at P14, showing significant decrease after irradiation at E16.5 (p = .0107). (G): Schematic representation of the IGL area at P14. (H): Quantification of the IGL area, showing significant increases after irradiation at E13.5 (p = .0120) and E16.5 (p = .0062). (I): Kaplan-Meier kinetic analysis of medulloblastoma (MB) after irradiation at E13.5 and E16.5 showing significant increase in MB mortality at E16.5 (p = .0084). *, p ≤ .05; **, p ≤ .01. Scale bar = 400 μm (A–C, G), 100 μm (E). Data relative to unirradiated mice are reported from Pazzaglia et al. [30]. Abbreviations: E, embryonic; EGL, external granule layer; IGL, internal granule layer.

Download figure to PowerPoint

image

Figure 6. Ptc1 LOH patterns in MB and age effects of inactivation of HR and NHEJ. (A): Analysis of chr 13 LOH in MBs from Ptc1+/ mice exposed to 0.25 or 3 Gy of x-rays at E13.5 or E16.5. The distance of microsatellite markers (D13Mit) from the centromere is given in cM, with values taken from the Genetic and Physical Maps of the Mouse Genome (1999). Black circles indicate no loss of polymerase chain reaction (PCR) signal from either allele; open circles denote loss of signal from one allele (LOH); gray circles indicate not informative markers. PCR primers were purchased from Research Genetics (Huntsville, AL). (B–L): Detection of cleaved caspase-3 at 6 hours after irradiation with 3 Gy of x-rays in sections of cerebellar primordium from Ptc1+/− mouse embryos at E13.5 (B), or at E16.5 (F, J), from HR-deficient (Ptc1+/−/Rad54/) mouse embryos at E13.5 (C) or at E16.5 (G, K), and from NHEJ-deficient (Ptc1+/−/DNA-PKcs/) mouse embryos at E13.5 (D) or at E16.5 (H, L). Quantification of active caspase-3, expressed as stained area per μm2, in the RL at E13.5 (E) and in the RL and EGL at E16.5 (I) from DNA repair proficient (Ptc1+/−), HR-deficient (Ptc1+/−/Rad54/) and NHEJ-deficient (Ptc1+/−/DNA-PKcs/) mouse embryos. (E): Impairment of HR significantly increased apoptosis in the RL of embryos irradiated at E13.5 compared to DNA-repair proficient (p = .0013) and NHEJ-deficient embryos (p = .0244), (B–D). (I): Significant increase of apoptotic response in the RL of NHEJ or deficient HR embryos irradiated at E16.5 compared to DNA-repair proficient mice (p = .0067; p = .0068; F–H). Impairment of NHEJ significantly increased apoptosis in the EGL of embryos irradiated at E16.5 compared to DNA repair proficient (p = .0026) and HR-deficient embryos (p = .00197; J, L). Dashed lines delineate the RL in (B–G) and the EGL in (H–J). *, p ≤ .05; **, p ≤ .01. Scale bar = 20 μm. Abbreviations: cM, centiMorgan; chr, chromosome; E, embryonic; EGL, external granule layer; HR, homologous recombination; LOH, loss of heterozygosity; MB, medulloblastoma; NHEJ, nonhomologous end joining; RL, rhombic lip; VZ, ventricular zone.

Download figure to PowerPoint

We next evaluated cerebellum tumorigenesis in Ptc1+/ mouse embryos exposed to a low radiation dose (0.25 Gy) at E13.5 (n = 41) or at E16.5 (n = 45). We observed a significant increase in MB mortality in mice irradiated at E16.5 compared to mice irradiated at E13.5 (p = .0084), in which no excess MB mortality was found over unexposed controls (Fig. 5I). These findings show that fate-restricted PCs represent a preferential oncogenic target after low-dose exposure compared to multipotent NSCs.

Ptc1 Loss of Heterozygosity Patterns in MB

Loss of heterozygosity (LOH) is the most common type of genetic alteration in human cancer that contributes to carcinogenesis by altering gene dosage. In Ptc1+/− mice, MBs are almost invariably characterized by loss of the normal remaining Ptc1 allele, occurring through interstitial deletions on chr 13 where the Ptc1 gene is located. To seek for possible mechanistic differences in the genesis of oncogenic alterations, we compared LOH patterns in MB developed after irradiation with 0.25 or 3 Gy of x-rays at E13.5 (n = 7) and E16.5 (n = 9). To this aim, we analyzed tumor DNA with a set of microsatellite markers spanning the length of chr13 (Fig. 6A). Regardless of radiation dose, we found a marked heterogeneity of LOH patterns in MBs developed after embryonic irradiation (E13.5 and E16.5) compared to tumors developed in mice irradiated as newborns (typically showing large chr13 interstitial deletions [31]). In mice irradiated at E13.5 only 43% (3/7) of the tumors showed typical interstitial deletions, while 28% (2/7) displayed large terminal losses and 28% (2/7) showed loss of the whole chr13. The LOH pattern of MBs from mice irradiated at E16.5 shows that 67% (6/9) of tumors had interstitial chr13 deletions, while 11% (1/9) showed upper terminal deletions and 22% (2/9) retention of heterozygosity for all chr13 markers analyzed.

The different LOH patterns in tumors initiated in multipotent NSCs (E13.5) and early (E16.5) PCs versus late (P1) lineage-restricted PCs clearly indicate that induction of oncogenic genetic alterations depends on age at exposure and may imply that distinct DNA repair mechanisms are acting along the differentiation process.

Effects of HR and NHEJ Deficiency on Apoptotic Response of NSCs and PCs

To explore the possibility that DNA repair mechanisms are differentiation/maturation-stage specific, we examined caspase-3 activation in the cerebellum of NHEJ (DNA-PKcs/) and HR (Rad54/)-deficient mice at E13.5 and E16.5, 6 hours after irradiation with 3 Gy of x-rays. In embryos irradiated at E13.5, impairment of HR dramatically increased apoptosis in the RL area (p = .0013; p = .0244; Fig. 6B–6E) compared to DNA repair-proficient and NHEJ-deficient mice. At E16.5, deficiency of either NHEJ or HR significantly increased apoptotic response in the RL (8.5–10-fold) compared to DNA repair-proficient mice (p = .0067; p = .0068; Fig. 6F–6I). Notably, fate-restricted PCs of the EGL were highly susceptible to radiation-induced apoptosis when NHEJ was impaired (p = .0026 vs. DNA repair proficient mice; Fig. 6J, 6L, 6I; p = .00197 vs. HR-deficient mice; Fig. 6K, 6L, 6I), indicating a major contribution of NHEJ in more differentiated progenitors. Taken together, these results clearly indicate that differentiation stage-specific DNA damage repair pathways are specifically acting in NSCs and fate-restricted PCs lineages.

Discussion

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Despite the increased usage of CT scanning and frequency of flying, knowledge on SC-specific radiation effects on tissue growth and long-term consequences on genome stability are sparse, except for hematopoietic tissue [4, 32].

To clarify the mechanisms used by NSCs and PCs for maintenance of genome integrity during neurogenesis, we have combined for the first time a comparative analysis of DDR activation and radiosensitivity in multipotent NSCs versus PCs, with evaluation of the long-term developmental and oncogenic consequences of irradiation in the cerebellum in vivo. We have also addressed the mechanisms of oncogenic events in MBs from E13.5- and E16.5-irradiated mice and dissected the contribution of HR and NHEJ DNA repair pathways upon induction of DNA damage in NSCs and their differentiating progeny.

Characterization of NSCs and Granule PCs in Embryonic Cerebellum at E13.5 and E16.5

During nervous system development, maintaining the balance between SC self-renewal and neurogenesis is essential. Too much self-renewal would result in a brain with too few neurons and abnormal circuitry; too much neurogenesis would quickly deplete the NSCs pool, resulting in a small brain and neurological abnormalities. In the developing cerebellum, Sox2 is expressed at high levels in NCSs and is required for their self-renewal, while Pax6 is critical for proper differentiation and migration of granule cells. Here, we show a complementary expression pattern of Sox2 and Pax6 at E13.5 and E16.5, suggesting that the Sox2/Pax6 ratio is involved in the regulation of cell specification of fate-restricted PCs in the developing cerebellum. An analogous mechanism has been described for proper regionalization of the optic cup during eye development [33]. On the other hand, p21 has been recently shown as a major player in the control of expansion of NSCs through negative regulation of the Sox2 gene [34]. Consistent with these findings, our results show p21 expression in fate-restricted progenitors, suggesting a possible link with Sox2 downregulation and granule lineage specification. Of note, p21 was previously shown to be highly expressed in replicating hematopoietic progenitor and precursor cells, indicating an important regulatory function in differentiation [35-37].

DDR Activation After Irradiation of NSCs and PCs

When comparing radiation-induced DDR signaling in NSCs versus fate-restricted PCs in the developing cerebellum in vivo, we detected several differences in marker expression, such as higher expression of 53BP1, complete lack of p21 induction, enhanced expression of the prosurvival protein Bcl-2, and marked resistance to radiation-induced apoptosis, which highlighted a dramatic influence of differentiation stage on the regulation of DDR signaling.

Lack of the p21-dependent cell cycle arrest at the G1/S transition after irradiation was recently detected in NSCs and PCs in the VZ of the developing mouse brain [38]. This study shows that this phenomenon extends to the NSCs of the cerebellum primordium. Although the mechanism through which p21 expression is repressed in NSCs remains to be clarified, two known mechanisms might be involved. One involves multiple miRNA families known as regulators of p21 mRNA, playing a key role in governing the G1/S transition and cell cycle checkpoint in undifferentiated human embryonic SCs [39]. The other is mediated by transcription factors such as Bmi1, FoxG1, and Olig2, strong repressors of p21, promoting amplification of NSCs and often overexpressed in brain tumors [40-44].

p21-induced G1 growth arrest is generally considered protective from apoptosis, and lack of the G1/S checkpoint in SCs is interpreted as a mechanism to eliminate damaged cells and preserve genomic integrity. Our results show that in the developing cerebellum lack of p21 is associated with a marked radioresistance in NSCs, while p21 expression correlates with a strong apoptotic response in fate-restricted PCs. These findings are consistent with reports highlighting a proapoptotic function of p21 under certain conditions in specific systems [45-47]. Our results also strongly involve the antiapoptotic Bcl-2 signaling pathway in resistance of NSCs to radiation-induced apoptosis. Consistently, Bcl-2 overexpression is a known mechanism used by hair-follicle-bulge and hematopoietic SCs to increase their resistance to DNA damage induced cell death [5, 48].

Collectively, by addressing DDR signaling and radiosensitivity of NSCs versus fate-restricted PCs in vivo, we show that these two distinct cell populations engage specific DDR programs differing in many respects, including the activation of robust antiapoptotic mechanisms in NSCs but not in fate-restricted PCs.

Alteration of Cerebellum Patterning After Irradiation of NSCs and PCs

Our results may also explain some aspects of the cellular events determining morphogenetic defects. Abnormal cerebellum patterning due to cell killing was predominant after high-dose irradiation (3 Gy). However, when comparing developmental effects in NCSs and PCs we found that cerebellum patterning was more severely impaired after irradiation at E16.5 compared to irradiation at E13.5, in accordance with the increased apoptotic response in fate-restricted PCs. An unexpected result of this study was the acceleration of differentiation observed after a low dose (0.25 Gy). Indeed, cell differentiation under stress conditions represents an alternative mechanism to cell killing to eliminate damaged cells from the SC pool. Accordingly, recent evidence suggests cell differentiation as a potential new outcome of the DDR in many cell types, including neuronal cells [49]. Notably, differentiation effects after low-dose radiation were more prominent in fate-restricted PCs compared to multipotent NSCs, suggesting that damage-related induction of p21 may negatively control Sox2, altering Sox2/Pax6 dosage and causing a switch from expansion to differentiation. Collectively, our data point to a general increased radiosensitivity of PCs compared to NSCs. From an evolutionary point of view the intrinsic radioresistance of NSCs, shared by SCs from other tissues such as blood and skin, may be interpreted as part of an efficient DDR that favors survival and lineage functionality aimed to preserve the overall tissue/organ function.

Effects of Irradiation of NSCs and PCs on MB Tumorigenesis

A further finding of this study is that a high radiation dose (3 Gy) can initiate MB in both multipotent NSCs and fate-restricted PCs, although significantly accelerated MB mortality was observed after irradiation at E13.5. This is probably due to the strong antiapoptotic response in NSCs, earlier occurrence of oncogenic mutations in the lineage, and/or to a decreased proliferation potential in granule PCs. Our results with radiation are in agreement with previous work using conditional Ptc1 knockout, showing that both SCs and PCs can serve as cells of origin for MB, and that the cell in which the tumor is initiated can have an important impact on the rate of tumor progression [50].

Remarkably, we show that after a low radiation dose (0.25 Gy) only granule PCs are able to initiate MB. Although unexpected, this result might imply that intrinsic DDR features in NSCs reduce the oncogenic consequences of low-dose exposure and/or that a larger pool of cells at risk (RL and EGL at E16.5, Fig. 1B) may be needed for genetic alterations and tumor effects to become manifest at low doses.

There is a general scarcity of in vivo data about DDR mechanisms acting in SCs and more differentiated lineages. Our results that the pattern of genetic oncogenic alterations depends on embryonic/postnatal age at exposure indicate that distinct DDR strategies can be adopted to limit the impact of radiation-induced DSBs along the differentiation process. Our data also show that different DNA repair pathways, as well as other DDR mechanisms, mainly cell cycle checkpoint arrest and apoptosis, may be acting in a balanced fashion to adjust survival versus genomic stability.

HR takes advantage of the homologous chromosome or sister chromatid for accurate repair and occurs in late S and G2 phases [51, 52]. Instead, error-prone NHEJ directly joins DSBs with little or no sequence homology between the broken ends and may operate in all cell cycle phases [52]. Our data clearly indicate a predominant role of HR deficiency in NSCs, consistent with lack of expression of G1 checkpoint protein p21, and a more prominent effect of NHEJ inactivation in fate-restricted PCs. In agreement, different DNA repair systems in SCs at different stages of ontogeny have also been detected in the VZ/subVZ of the mouse brain [53] and in the hematopoietic system [4, 32]. However, further research is needed to clarify the oncogenic consequences of selective inactivation of HR and NHEJ in NSCs and PCs.

Our phenotypic characterization of MB shows that these tumors express markers reminiscent of the differentiation status of the cell of origin. Since tumorigenesis can be considered as aberrant organogenesis carried on by cancer SCs (CSCs), clarification of the relationship between tumors and the tissue SCs may have important implications for therapy [54]. Phenomena of radioresistance have been evidenced in highly malignant glioma and MB and have been ascribed to the presence in the tumor tissues of cells expressing the SC marker CD133 [55-57]. Because CSCs may use DNA repair mechanisms used by SCs to mediate chemoresistance and radioresistance [1], signaling pathways activated by NSCs in response to radiation may be exploited to increase the efficacy of therapies against brain tumors.

Conclusions

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Our current in vivo study provides original insights into the molecular mechanisms underlying developmental and oncogenic effects of radiation exposure in NSCs and their differentiating progeny. This is relevant for understanding mechanisms of DDR and maintenance of genome integrity of NSCs, and how deregulation of these mechanisms leads to cancer. In particular, a central role for the cell cycle inhibitor p21 and its ability to influence cell-fate decision, DNA repair process, and apoptotic response has been highlighted. By defining distinct NSCs and PCs lineages in which exogenous damage initiates MB and systematically assessing their DDR for cancer and cancer-relevant endpoints, we provide better understanding of the mechanisms of response to radiation therapy in brain cancer and possible strategies to overcome radioresistance.

Acknowledgments

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

This work has been partially supported by grants from European Community's Seventh Framework Program (EURATOM) contract Fission-2011–295552 (CEREBRAD), contract Fission-2011–249689 (DoReMi), and Grant 10357 from the Associazione Italiana, Ricerca sul Cancro (AIRC).

References

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. A
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1485-sup-0001-suppfig1.tif1741KSupporting Information Figure 1

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.