Involvement of Atm and Trp53 in neural cell loss due to Terf2 inactivation during mouse brain development
- 955 Downloads
Maintenance of genomic integrity is one of the critical features for proper neurodevelopment and inhibition of neurological diseases. The signals from both ATM and ATR to TP53 are well-known mechanisms to remove neural cells with DNA damage during neurogenesis. Here we examined the involvement of Atm and Atr in genomic instability due to Terf2 inactivation during mouse brain development. Selective inactivation of Terf2 in neural progenitors induced apoptosis, resulting in a complete loss of the brain structure. This neural loss was rescued partially in both Atm and Trp53 deficiency, but not in an Atr-deficient background in the mouse. Atm inactivation resulted in incomplete brain structures, whereas p53 deficiency led to the formation of multinucleated giant neural cells and the disruption of the brain structure. These giant neural cells disappeared in Lig4 deficiency. These data demonstrate ATM and TP53 are important for the maintenance of telomere homeostasis and the surveillance of telomere dysfunction during neurogenesis.
KeywordsAtm DNA damage Apoptosis Brain development
DNA double-strand breaks
DNA damage response
Genomic instability resulting from DNA damage induced by either endogenous or exogenous insults could lead to defective neurodevelopment and neurological diseases (Lee et al. 2016). Ataxia telangiectasia mutated (ATM) is one of the early responders to DNA damage, particularly to DNA double-strand breaks (DSBs). ATM mutations cause Ataxia Telangiectasia (A-T) characterized by ataxia due to loss of Purkinje and granule cells in the cerebellum (McKinnon 2012, 2013). Similarly, Ataxia-telangiectasia and RAD3-related (ATR) recognizes single-stranded DNA resulting from replication stress. Seckel syndrome 1 (SCKL1) due to hypomorphic mutations in the ATR gene is characterized by microcephaly and mental retardation (Nam and Cortez 2011; McKinnon 2013). Once activated by DNA damage, ATM and ATR phosphorylate several overlapping substrates including tumor protein p53 (TP53/Trp53 in mice) to regulate DNA damage repair, apoptosis, and cell cycle arrest (Lee et al. 2001, 2012b; Lovejoy and Cortez 2009). Although it is well known that ATM and ATR are required to maintain genomic integrity, the precise roles of ATM and ATR during brain development are not fully understood, especially related to neuropathology such as ataxia, neurodegeneration, and microcephaly observed in human patients.
Telomere dysfunction is one of the endogenous sources to induce DNA damage response (DDR), since the unprotected telomere ends could be recognized as DNA strand breaks (de Lange 2005). To protect this DNA region, the telomere ends are coated with the Shelterin complex, which is composed of several proteins including telomeric repeat binding factor 2 (TERF2) and protection of telomeres 1 (POT1), to prevent telomere attrition and inappropriate DDR induction (de Lange 2005; Palm and de Lange 2008). TERF2 homodimers bind to double-stranded regions of telomeres and inactivation of TERF2 triggers ATM-dependent DDR signals, whereas POT1 protects the 3′ single-stranded overhang and its inactivation initiates ATR-dependent DDR signals at the cellular level (Karlseder et al. 1999; Zhang et al. 2006, 2007; Denchi and de Lange 2007; Sfeir and de Lange 2012). The exposed telomere ends as a result of Terf2 inactivation were processed by DNA ligase IV (Lig4) which is the ligase for the canonical Non-homologous end-joining repair (NHEJ) pathway for DSBs (Celli and de Lange 2005; Lee et al. 2016; Smogorzewska et al. 2002). The involvement of ATM and ATR in DDR induced by telomere dysfunction was mutually exclusive in this context.
However, selective deletion of the Pot1a gene during mouse brain development resulted in Atm-dependent neurological phenotypes including cerebellar defects, suggesting that telomere dysfunction might induce DDR in a tissue-specific manner during neurogenesis (Lee et al. 2014). For the current study, we asked whether telomere dysfunction due to Terf2 inactivation also induces neuro-specific DDR associated with Atm or Atr signaling pathway during neurogenesis.
Materials and methods
Floxed Terf2 animals (Karlseder et al. 1999; Celli and de Lange 2005) were purchased from the Jackson Laboratory (JAX #006568). Germline deletion of the Terf2 gene caused embryonic lethality at mid-gestation in the mouse (Celli and de Lange 2005). In order to restrict inactivation of the Terf2 gene in the nervous system during development, Terf2 LoxP/+ animals were interbred with Nestin-Cre animals (JAX #003771) or Emx1-Cre animals (JAX #005628). Cre recombinase expression driven by the Nestin promoter is active around embryonic day (E) 11 throughout the neural progenitors, whereas the Emx1 promoter is activated only in the dorsal telencephalon and the hippocampal progenitors around E10 during mouse embryogenesis (Gorski et al. 2002; Graus-Porta et al. 2001; Tronche et al. 1999). Terf2 LoxP/LoxP ;Nestin-Cre and Terf2 LoxP/LoxP ;Emx1-Cre mice were obtained through a proper breeding scheme. Conditional knockout animals could not be used for breeding. Terf2 LoxP/+ ;Nestin-Cre or Emx1-Cre animals did not show any discernable defects or shortened life span, and they were fertile. So these animals were included in control groups. Mutant alleles of Atm, Atr LoxP , Trp53 LoxP , Lig4 LoxP , and Pot1a LoxP genes and the polymerase chain reaction (PCR) conditions for genotyping were as previously described (Lee et al. 2001, 2012b, 2014; Shull et al. 2009). All animals were maintained in a mixed strain of C57BL/6 X 129 genetic background. Since we could not find any gender difference of DDR in the brain before (Lee et al. 2001, 2012a, 2014), we analyzed the experimental materials regardless of gender for the current study. Also Nestin-Cre expression was maintained in females for breeding to minimize any ectopic Cre recombinase activity outside of the nervous system. Genotypes of genetically engineered animals were determined by a routine PCR method using the following primers: Terf2 (Forward: 5′ ccaaccagggatacacagtga, Reverse: 5′ atccgtagttcctcttgtgtctg), Pot1a (Forward: 5′ ctcgaattccatctcctcccagtactctctcag, Reverse: 5 ggaactggtacgtatcagtgtgtgtgg), Atm wildtype ( WT) allele (Forward: 5′ gcctgtatcttctatgtgcaccgtcttcgc, Reverse: 5′ ggtgcggtgtggatgggactggagg), Atm targeted allele (Forward: 5′ gtgatgacctgagacaagatgctgtc, Reverse: 5′ gggaagacaatagcaggcatgc), Atr (Forward: 5 tacattttagtcatagttgcataacac, Reverse: 5 cttctaatcttcctccagaattgtaaaagg), Lig4 (Forward: 5′ atcgctcttgtcccagtacacctgc, Reverse: 5 gtgcattaaatggagtgctgtgc), Trp53 (Forward: 5′ cacaaaaacaggttaaacccag, Reverse: 5′ agcacataggaggcagagac).
PCR products for Terf2 genotyping were amplified for 35 cycles of 94 °C for 30 s, 60 °C for 45 s, and 72 °C for 45 s. PCR products for WT and floxed alleles (LoxP) of the Terf2 gene were 233 and 300 bp, respectively.
All of animal materials for experiments were courteously provided by Dr. Peter McKinnon (St. Jude Children’s Research Hospital, USA). The presence of a vaginal plug was indicated as E.5 and the day of birth as postnatal day (P) 0. All animals were housed in an AAALAC accredited facility and were maintained in accordance with the National Institutes of Health ‘Guide for the Care and Use of Laboratory Animals’. All procedures for animal use were approved by the Institutional Animal Care and Use Committee.
Histopathological procedures were performed as previously described (Lee et al. 2012a, 2014). In brief, both embryos and brains were collected at indicated time points after the fixation step using 4% phosphate-buffered paraformaldehyde. Cryosections of embryos or brains at 10 μm were collected using an HM500M (Microm) or an MEV (SLEE medical GmbH) cryostat for pathological analysis. Hematoxylin and Eosin (H/E), and Nissl staining were carried out in a routine procedure. Immunoreactivity was visualized by either colorimetric detection using the VIP substrate kit (Vector Labs) reactive with biotinylated secondary antibodies or fluorometric detection using FITC/CY3 conjugated secondary antibodies (Jackson Immunologicals) after incubation with primary antibodies. Counterstaining was done with 0.1% methyl green in 0.1 M sodium acetate buffer for colorimetric staining followed by mounting with DPX (Sigma) or DAPI/PI counterstaining mounting medium (Vector Laboratories) for fluorometric staining.
Antibodies used for this study were Calbindin (mouse, 1:2000, Sigma-Aldrich), Ctip2 (rat, 1:100, Abcam), Cux1 (rabbit, 1:100, Santa Cruz), Foxp2 (rabbit, 1:500, Abcam), GFAP (mouse, 1:500, Sigma-Aldrich), γ-H2AX (rabbit, 1:100, Cell signaling), γ-tubulin (mouse, 1:1000, Sigma-Aldrich), Histone H3phosphoS10 (rabbit, 1:2000, Cell signaling), MBP (myelin basic protein, rabbit, 1:200, Abcam), Myelin-PLP (rabbit, 1:200, Abcam), NeuN (mouse, 1:500, Millipore), Neuronal Class III β tubulin (clone TUJ1, mouse, 1:1000, Covance Research), Tbr1 (rabbit, 1:500, Abcam). Depending on primary antibodies, the citric acid-based antigen retrieval method was applied to enhance immunoreactive signals.
TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assay using Apoptag (Chemicon) was applied to measure apoptosis. Multiple histological slides were examined and imaged using an Axio Imager A1 microscope (Zeiss) or B600TiFL (Optika). Microscopic images were captured and processed using Photoshop (v.CS6.5, Adobe).
Telomere in situ using telomere-PNA probe
To detect telomere in situ, tissue sections were dehydrated in alcohol series, denatured at 80 °C for 4 min, followed by incubation for 4 h in the dark at room temperature with CY3-CCCTAACCCTAACCCTAA peptide nucleic acid (PNA) telomere probe (0.7 μg/ml, Panagene) in 70% formamide, 10 mM Tris–HCl pH 7.5, 5% MgCl2 buffer (82 mM Na2HPO4, 9 mM citric acid, 25 mM MgCl2), and 1% blocking reagent (Roche). Then, sections were washed with Solution I (70% formamide, 10 mM Tris–HCl pH 7.2, 0.1% BSA) and Solution II (0.05 M Tris–HCl pH 7.2, 0.15 M NaCl, 0.05% Tween 20) several times. After washing, slides were mounted with DAPI-containing aqueous mounting medium (Vector). Images were captured and analyzed as described in the “Histology” method section.
Terf2 was essential for brain development
It has been reported that germline deletion of the Terf2 gene caused embryonic lethality at mid-gestation in the mouse (Celli and de Lange 2005). So, we restricted Terf2 inactivation to the nervous system by cross-breeding of the Terf2 floxed animal model with either Nestin-Cre (Terf2 LoxP/LoxP ;Nestin-Cre, hereafter Terf2 Nes-Cre ) or Emx1-Cre (Terf2 LoxP/LoxP ;Emx1-Cre, hereafter Terf2 Emx1-Cre ) animal lines during development.
Terf2 inactivation induced massive neural apoptosis during development
Atm and Trp53, but not Atr, inactivation rescued neural apoptosis resulting from Terf2 deficiency
Previously, it has been demonstrated that DNA damage triggered by Terf2 inactivation in vitro induced Atm-dependent Atr-independent DDR (Karlseder et al. 1999; Celli and de Lange 2005; Zhang et al. 2006, 2007; Denchi and de Lange 2007). So, next we tested the in vivo involvement of Atm and Atr, as well as Trp53, a common downstream substrate, in lethality and massive neural apoptosis resulting from Terf2 inactivation during neurogenesis. Floxed Terf2 animals were cross-bred with Atr (conditional inactivation: Atr Nes-Cre and Atr Emx1-Cre ), Atm (germline inactivation: Atm −/−) or Trp53 (conditional inactivation; Trp53 Nes-Cre and Trp53 Emx1-Cre ) animals. Similar to the in vitro situation (Denchi and de Lange 2007; Karlseder et al. 1999), Atr inactivation did not have any influence on the phenotypes observed in both Terf2 Nes-Cre and Terf2 Emx1-Cre animals (Fig. 2). Terf2 Emx1-Cre ;Atm −/− and (Terf2;Trp53) Emx1-Cre animals also did not show any discernible improvement in gross phenotypes including smaller body size and shorter life span observed in the Terf2 Emx1-Cre animals (data not shown).
Although both Atm and Trp53 inactivation could reduce apoptosis dramatically in the Terf2-null brains (Fig. 2), there was a difference between these two genetic backgrounds. Neural cell death in both Terf2 Nes-Cre and Terf2 Emx1-Cre embryos had disappeared by E15.5 in a Trp53-null background, whereas Atm inactivation could not stall most of the apoptosis in the ventricular zone (VZ) where neural progenitor cells are located in (Fig. 2d).
The multiple defects were found in the Terf2-null brains
Although neuron-specific class III β-tubulin (Tuj1), which is a marker for postmitotic neurons, showed a normal distribution in the developing brains of the Terf2/Atm and Terf2/p53 double null embryos (Fig. 2b), the restored portion in the mature brains exhibited several faulty features. Trp53 inactivation led to better restoration of the missing parts in the Terf2-null brain than Atm deficiency did. However, the restoration was still incomplete (Fig. 1b). The hippocampus was not restored in the Terf2/Atm and Terf2/Trp53 double null brains (Fig. 1b).
Next, we analyzed the six-layered cerebral cortical structure created in an inside-out manner during development (Greig et al. 2013). The Foxp2, Ctip2, and Tbr1 immuno-positive neurons, which are generated during an early stage of cortical development and localized in the lower layers of the cerebral cortex in normal development, were found in the upper cortical part of the (Terf2;Trp53) Nes-Cre and (Terf2;Trp53) Emx1-Cre brains (Fig. 3b). Similarly, the Terf2 Emx1-Cre ;Atm −/− cortex showed mis-localization of Ctip2 and Tbr1 immuno-positive neurons. Furthermore, there were no proper upper layers formed (layers I, II, and III), which were Cux1 immuno-positive, in both the Terf2/Atm and Terf2/Trp53 double null cortices, suggesting that the Terf2-null developing brain could not generate cortical neurons at the later stage during development in both Atm- and Trp53-deficient backgrounds (Fig. 3b).
Trp53 inactivation resulted in multinucleated giant neural cells in the Terf2-null brain
Also, the considerable amounts of DNA damage visualized by phosphorylated H2AX (γ-H2AX) foci formation, which is a good marker for DNA strand breaks, were detected in all of the Terf2-null brains, but not in the control brain (Fig. 5a). In addition, telomere in situ visualization using a telomere-PNA probe showed its conjugated CY3 signals in the Terf2-null brains (Fig. 5a). These telomere-PNA signals were not noticeable in the control brain, suggesting that telomeres in Terf2-null neural cells were readily accessible by the telomere-PNA probe since they were not well protected.
Next we examined whether multinucleated giant neural cells were formed during neurogenesis. Certainly, γ-H2AX foci formation in the Terf2-null brains, particularly in the VZ of the developing brains, was dramatically increased in both Atm- and Trp53-deficient backgrounds (Fig. 5b). However, the chromosome mass at metaphase visualized by phosphorylated H3 immunoreactivity (H3pS10) showed a significant increase only in the Terf2/Trp53 double null neural progenitors (Fig. 5b). This giant chromosomal mass was not observed in an Atm-null background. It is possible that this giant chromosomal aggregation might partially result from incomplete cell division, since multiple centrosomes detected by γ-tubulin immunoreactivity in a single cell were observed only in the Terf2/Trp53 double null neural cells at metaphase (Fig. 5b). Even though DNA damage detected by γ-H2AX foci was evident, neural progenitor cells with giant chromosomal mass were not observed in the Pot1a Nes-Cre embryonic brains (Fig. 5b); consequently, there were no multinucleated giant neural cells found in the Pot1a Nes-Cre mature brains (Lee et al. 2014).
DNA ligase IV plays a role in neural abnormalities due to Terf2 deficiency during neurogenesis
Previously it was reported that Lig4 plays a role in abnormal fusions of telomere ends in Terf2-deficient cells (Celli and de Lange 2005; Smogorzewska et al. 2002). So we tested whether Lig4 inactivation during neurogenesis could alleviate neural abnormalities observed in the (Terf2/Trp53) Emx1-Cre brain. As described before (Lee et al. 2012a), the (Lig4;Trp53) Emx1-Cre brain did not show any neural phenotypes related to telomere dysfunction. The triple conditional knockout animals (hereafter (Terf2;Lig4;Trp53) Emx1-Cre ) were born alive, but the (Terf2;Lig4;Trp53) Emx1-Cre animals died around 1 month of age similar to the (Terf2;Trp53) Emx1-Cre animals.
Neuropathological analysis revealed that the neurological abnormalities were moderately corrected in the (Terf2;Lig4;Trp53) Emx1-Cre brain. During embryogenesis, the reduction of neural cell death in the (Terf2;Lig4;Trp53) Emx1-Cre embryos was the same as the (Terf2;Trp53) Emx1-Cre embryos, most likely due to the effect of Trp53 inactivation (Fig. 2d). However, proliferating cells at metaphase in the (Terf2;Lig4;Trp53) Emx1-Cre VZ were similar to those with two centrosomes found in control embryos (Fig. 5b). Consequently, the most obvious correction was disappearance of multinucleated giant cells in the triple conditional null mature brain (Figs. 4a, 5a). In addition, the upper layers of the cerebral cortex, which are Cux1 immuno-positive, were formed in the triple null brain (Fig. 3b). However, this recovery by Lig4 inactivation was not complete, such as partial reconstruction of the hippocampus and CC compared to the control groups (Figs. 1b, 4).
Telomere homeostasis is essential for brain development in the mouse
The role of TERF2 and POT1 as components of the Shelterin complex is to protect the telomere ends so that genomic integrity is maintained (Celli and de Lange 2005; Denchi and de Lange 2007; Karlseder et al. 1999). Previously the Terf2 gene of the mouse was conditionally targeted only in the liver and skin (Bojovic et al. 2013; Lazzerini Denchi et al. 2006; Martinez et al. 2014). Terf2 inactivation in the basal layer of epidermis using a K14-Cre animal line did not cause embryonic lethality (Bojovic et al. 2013), while Terf2 inactivation in epidermal stem cells using a K5-Cre animal line led to partial embryonic lethality and impaired skin development (Martinez et al. 2014). Similarly, here we demonstrated that a selective inactivation of the Terf2 gene in the neural progenitor cells resulted in embryonic fatality and loss of brain structures, providing another example that the protective role of TERF2 is important to maintain cellular and organismal viability. This result is clearly different from the animal model for selective inactivation of the Terf2 gene in the liver that showed normal liver function and regeneration through endoreduplication without cell division (Lazzerini Denchi et al. 2006). The status of cells, such as proliferating vs. resting cells, most likely contributes to this difference in cellular viability as suggested before (Martinez et al. 2014).
Interestingly, there was no similarity of the neural phenotypes between the Terf2 Nes-Cre and Pot1a Nes-Cre animals (Lee et al. 2014), even though germline deletion of the Terf2 or Pot1a gene in the mouse resulted in embryonic lethality (Celli and de Lange 2005; Hockemeyer et al. 2006; Wu et al. 2006), Pot1a inactivation in neural progenitor cells did not induce massive neural apoptosis during embryogenesis, and the neurological defects resulting from Pot1a inactivation were restricted mainly to the cerebellum (Lee et al. 2014). Furthermore, Pot1a Nes-Cre ;Trp53 −/− animals did not show any sign of ataxia. On the contrary, Terf2 inactivation in the neural progenitor cells induced more global effects throughout the central nervous system. Also it appeared that there was a differentiation defect of late-born neurons originated from Terf2-null neural progenitor cells which underwent more cell divisions in a Trp53-deficient background.
As illustrated in Fig. 5c, it is possible that the Shelterin complex with all components including POT1 presents only at the end of telomeres where the 3′ single-stranded GC rich overhang is exposed (de Lange 2005; Gramatges and Bertuch 2013; Palm and de Lange 2008). For this reason, the impact resulting from Terf2 inactivation in the neural progenitor cells was more severe than that of Pot1a inactivation in the murine nervous system.
ATM and Trp53 are key signaling mediators in telomere dysfunction due to Terf2 inactivation during brain development in the mouse
One of the interesting observations was the entity of multinucleated giant neural cells only in the Terf2/Trp53 conditional knockout mature brain, possibly resulting from endoreduplication without proper cell division as suggested in other Terf2 conditional knockout animal models (Lazzerini Denchi et al. 2006; Martinez et al. 2014; Ullah et al. 2009). It had been demonstrated that endoreduplication and mitotic failure in a Trp53-dependent manner during telomere crisis lead to polyploid cell accumulation (Davoli and de Lange 2012; Pampalona et al. 2012; Davoli et al. 2010). This defect of multinucleated giant cells was resolved by Lig4 inactivation, suggesting that the ligation function of LIG4 plays a role in this particular neural defect due to telomere dysfunction resulting from Terf2 inactivation during brain development.
Trp53 inactivation suppressed neural apoptosis triggered by Terf2 inactivation in the entire developing nervous system including the VZ in which big metaphase cells could be formed, whereas Atm deficiency could not inhibit programmed cell death in the VZ of the Terf2-null embryonic brain. Therefore, multinucleated giant neural cells were not observed in the Terf2/Atm knockout brain. Also this result suggests that Trp53 activation to induce neural apoptosis in the Terf2-null VZ was Atm-independent. This observation was consistent with the previous reports demonstrating that ATM is involved in neural apoptosis via Trp53 activation in the postmitotic zone, not in the VZ (Lee et al. 2000, 2001; Orii et al. 2006). Apparently, ATR is not the key kinase to activate Trp53 signaling in the Terf2-null VZ, since Atr inactivation did not have any effect on the neural defects in the Terf2-null brain, similar to the in vitro situation (Denchi and de Lange 2007). In addition, there was likely no cross-talk between Atm and Atr signalings responding to telomere dysfunction resulting from Terf2 inactivation during brain development, in contrast to the situation of Pot1a deficiency in the developing brain (Lee et al. 2014). Alternatively, DNA-dependent protein kinase (DNA-PK) might be involved in the Trp53 signaling in the VZ upon DDR induced by Terf2 deficiency (Rybanska-Spaeder et al. 2014).
Taken all together in combination with our previous report (Lee et al. 2014), we demonstrated that TERF2 is more critical to maintain telomere homeostasis than POT1a in the developing mouse brain. Telomere dysfunction by Terf2 inactivation induces the ATM-Trp53 signaling axis to trigger neural apoptosis as a part of the mechanisms to maintain genomic integrity during neurogenesis as illustrated in Fig. 5c. ATM is also activated via ATR signaling by telomere dysfunction due to Pot1a inactivation in the same physiological context.
We are deeply grateful to Dr. Peter McKinnon at St. Jude Children’s Research Hospital, USA for providing experimental materials and vital discussion. We thank the Hartwell Center for biotech support, the cytogenetics core for chromosome analysis and the Animal Resource Center for animal husbandry. We also thank Jingfeng Zhao, Yang Li and Dr. Helen Russell for technical support as well as Stuart Horwitz for English editing. YSL was supported by the NRF of Korea grant funded by the MSIP (Nos. 2011-0030043 and NRF-2014R1A1A2056224). YSL is also supported by the new faculty research fund of Ajou University School of Medicine.
Compliance with ethical standards
Conflict of interest
We declare no conflict of interest regarding this paper.
All procedures using animal models were approved by the IACUC.
Supplementary material 1 (MOV 3414 kb)
- Wu L, Multani AS, He H, Cosme-Blanco W, Deng Y, Deng JM, Bachilo O, Pathak S, Tahara H, Bailey SM, Deng Y, Behringer RR, Chang S (2006) Pot1 deficiency initiates DNA damage checkpoint activation and aberrant homologous recombination at telomeres. Cell 126(1):49–62. doi: 10.1016/j.cell.2006.05.037 CrossRefPubMedGoogle 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.