DNA damage and neurodegenerative phenotypes in aged Ciz1 null mice

Abstract

Cell-cycle dysfunction and faulty DNA repair are closely intertwined pathobiological processes that may contribute to several neurodegenerative disorders. CDKN1A interacting zinc finger protein 1 (CIZ1) plays a critical role in DNA replication and cell-cycle progression at the G1/S checkpoint. Germline or somatic variants in CIZ1 have been linked to several neural and extra-neural diseases. Recently, we showed that germline knock-out (KO) of Ciz1 is associated with motor and hematological abnormalities in young adult mice. However, the effects of CIZ1 deficiency in much older mice may be more relevant to understanding age-related declines in cognitive and motor functioning and age-related neurological disorders such as isolated dystonia and Alzheimer disease. Mouse embryonic fibroblasts from Ciz1^−/− mice showed abnormal sensitivity to the effects of γ-irradiation with persistent DNA breaks, aberrant cell-cycle progression, and apoptosis. Aged (18-mo-old) Ciz1^−/− mice exhibited marked deficits in motor and cognitive functioning, and, in brain tissues, overt DNA damage, NF-κB upregulation, oxidative stress, vascular dysfunction, inflammation, and cell death. These findings indicate that the deleterious effects of CIZ1 deficiency become more pronounced with aging and suggest that defects of cell-cycle control and associated DNA repair pathways in post-mitotic neurons could contribute to global neurological decline in elderly human populations. Accordingly, the G1/S cell-cycle checkpoint and associated DNA repair pathways may be targets for the prevention and treatment of age-related neurodegenerative processes.

1. Introduction

Preservation of genomic integrity by the DNA repair response is fundamental to the function and survival of cells. The DNA damage response in post-mitotic neurons includes cell-cycle arrest coupled with DNA repair (Lavin et al., 2005, Friedberg, 2003). Accumulation of DNA breaks can result in apoptosis, senescence or aberrant differentiation. In this regard, several studies have suggested that defects in cell-cycle and DNA repair pathways lead to uncontrolled cell death and play an important role in the pathogenesis of several age-related neurodegenerative disorders (Adamec et al., 1999, Best, 2009, Chen et al., 2010, Gensler and Bernstein, 1981, Katsel et al., 2013, Ioannidou et al., 2016). Identification of genes and their encoded proteins critically involved in DNA checkpoints and repair within the central nervous system (CNS) establishes mechanistic precedents for unraveling the pathobiology of age-related deterioration in global neurological functioning (McKay et al., 2017, Seidler et al., 2010) and specific age-related neurological disorders including dementia (Silva et al., 2014, Saykin et al., 2015) and dystonia (Ledoux et al., 2013).

CDKN1A interacting zinc finger protein 1(CIZ1), a cell-cycle protein widely expressed in neural and extra-neural tissues, plays an important role in DNA synthesis at the G1/S cell-cycle checkpoint (Coverley et al., 2005, Liu et al., 2016, Mitsui et al., 1999) and localization of Xist RNA to the inactive X chromosome (Ridings-Figueroa et al., 2017). Depletion of CIZ1 transcripts restrains cell proliferation by inhibiting entry to S phase. CIZ1 mutations are associated with isolated cervical dystonia, possibly by precipitating G1/S cell-cycle dysregulation in the mature human brain (Xiao et al., 2012). Young adult Ciz1^−/− mice exhibit motoric and hematological abnormalities including elevated monocyte and neutrophil counts in peripheral blood (Xiao et al., 2016). Gene expression studies showed that transcription factor 7-like 2 (TCF7L2), a member of the Wnt/β-catenin signaling pathway, was a major hub for down-regulated genes, whereas NF-κB was a major hub for up-regulated genes in Ciz1^−/− mouse brain, suggesting that CIZ1 deficiency is associated with synaptic plasticity and neurodegeneration. Besides its potential roles in dystonia and the maintenance and/or differentiation of post-mitotic neurons within the CNS, CIZ1 may also be an important contributor to several other neural and extra neural diseases (Higgins et al., 2012, Liu et al., 2015, Yin et al., 2013, Judex et al., 2003, Bageghni et al., 2017, Dahmcke et al., 2008, Warder and Keherly, 2003). For instance, dysfunction of CIZ1 caused by aberrant splicing of exon 8 may contribute to the pathogenesis of Alzheimer disease (AD) by inhibiting proliferation and differentiation of neural progenitor cells in the hippocampus (Dahmcke et al., 2008). In other studies, a CIZ1 variant was found to be a circulating biomarker for early-stage lung cancer (Higgins et al., 2012) and overexpression of CIZ1 in the heart reduced the impact of myocardial injury (Bageghni et al., 2017). CIZ1 has been shown to co-regulate estrogen receptor α (ERα) by enhancing estrogen transactivation activity by promoting the recruitment of the estrogen complex to target gene chromatin (den Hollander et al., 2006). Estrogen binds to ERα and is neuroprotective in cellular and mouse models of neurodegeneration (Fitzpatrick et al., 2002, Rosario et al., 2010, Hwang et al., 2015).

The relationship of CIZ1 variants to age-related neurological disorders (AD and adult-onset isolated dystonia) prompted us to use our Ciz1 null mouse model to examine to role of defective cell-cycle control and associated DNA repair pathways in neurodegeneration. A growing literature has pointed to a potential role for defective DNA repair in human neurodegenerative disorders (Coppede and Migliore, 2009, Chow and Herrup, 2015, Ross and Truant, 2017) including dystonia (Ledoux et al., 2013, Prudente et al., 2013). Moreover, a significant percentage of human populations with dementia do not have AD-type pathology on postmortem brain examination (Grammas, 2011, Graff-Radford et al., 2016). Alternative mechanisms for dementia include the cumulative effects of neuronal loss due to defects in DNA repair pathways and microvascular dysfunction (Chui and Ramirez Gomez, 2017). Therefore, model systems to study DNA repair and neurovascular risk factors are essential for understanding global neurological decline.

Herein, we found that CIZ1 deficiency leads to cell-cycle dysfunction, the accumulation of DNA breaks, and increased apoptosis in mouse embryonic fibroblasts (MEFs). Aged Ciz1^−/− mice showed persistent DNA breaks, motor and cognitive deficits, increased oxidative and inflammatory markers in brain, neurovascular dysfunction, and neuronal apoptosis. Overall, our data indicate that the deleterious effects of CIZ1 deficiency become more pronounced with aging and suggest that defects in cell-cycle checkpoints and associated DNA repair pathways may contribute to the pathobiology of neurodegenerative disorders and age-related global decline in neurological functioning.

2. Materials and methods

2.1. Cell Culture and irradiation

MEFs were isolated from embryos at 12 to 14 d gestation and grown and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 1% L-glutamine and 1% penicillin-streptomycin. Cells were plated on gelatin-coated T25 tissue culture flasks in an incubator at 37°C with 5% CO ₂. After the fourth passage, MEFs were exposed to γ-radiation (0, 5, 10 and 20 Gy) using a ¹³⁷Cs source. In pilot experiments, we performed dose-response effects (5, 10 and 20 Gy) of γ-irradiation to establish the optimal dose of γ-irradiation (20 Gy) for comparative analysis of DNA damage. Immediately after irradiation, cells were returned to the incubator for recovery.

2.2. Analyses of DNA damage, the cell-cycle, and apoptosis

For assessment of DNA damage, cells were plated on poly-l-lysine-coated coverslips and fixed with 4% paraformaldehyde. After fixation, cells were washed twice with 1X phosphate-buffered saline (PBS) and permeabilized with 0.3% Triton X-100 for 15 min followed by blocking with 1% bovine serum albumin (BSA) in PBS. Cells were incubated with rabbit anti-phospho-H2A.X (Ser139) (#2577, Cell Signaling) or rabbit anti-53BP1 (ab21083, Abcam) for 2 h at room temperature, rinsed and incubated with a fluorescently-tagged secondary antibody (Jackson ImmunoResearch Laboratories) for 1 h and then washed 3X with PBS. Coverslips were mounted onto slides with medium containing DAPI (H-1200, Vector).

For cell-cycle analysis, cells were fixed with ice-chilled 70% ethanol for 1 h. Fixed cells were then washed with PBS 3X followed by incubation with PI staining buffer (50 μg/mL) for 30 min. After incubation, cells were analyzed by flow cytometry (BD Biosciences LSR II Flow Cytometer). FITC Annexin V/Dead Cell Apoptosis Kit (V13242, Thermofisher) was used to detect apoptosis in MEFs after irradiation according to the manufacturer’s instructions. Briefly, cells were harvested, washed in cold PBS and then centrifuged to discard the supernatant. Pellets were suspended with 1X annexin-binding buffer. Cells were incubated with 5 μL of FITC annexin V and 1 μl of 100 μg/ml propidium iodide (PI) working solution for 15 min and then analyzed with a BD Biosciences LSR II Flow Cytometer.

2.3. Alkaline comet assay

To characterize DNA damage, the alkaline comet assay was performed on cultured MEFs after 4 passages using the protocol provided with Trevigen's Comet Assay kit. Comet assays were also performed on cells isolated from the brains of Ciz1^−/− mice and their sex-matched littermates. The comet assay or single cell gel electrophoresis is a sensitive technique for quantifying DNA damage in individual cells. The resulting image resembles a comet with a head and tail. The tail contains broken DNA. Percent DNA in the tail or % tail DNA is a normalized measure of the percentage of total DNA found in the tail. Briefly, tissues were homogenized in a glass teflon dounce homogenizer by 10 strokes and cells were collected through a 100 μm cell strainer (ThermoFisher Scientific). In brief, 1 × 10⁵ cells were mixed with 500 μl of pre-warmed low melting temperature agarose (1:10, v/v) and 50 μl of the mixture was plated onto slides (CometSlide™, Trevigen). After agarose solidified and attached to slides, slides were immersed in pre-chilled lysis solution for 1 h on ice, then in alkaline unwinding solution (200 mM NaOH, 1 mM EDTA) for 1 h at room temperature. Electrophoresis was performed in a pre-chilled alkaline electrophoresis solution (200 mM NaOH, 1 mM EDTA) at 4°C for 30 min at 21 V. Slides were then washed for 5 min with distilled water 2X, followed by a 5 min-incubation with 70% ethanol. Slides were air-dried at 37°C for 15 min in the dark and agarose gels were stained with SYBR® Gold for 5 min at 4°C. After SYBR Gold was removed and slides were air-dried, images were acquired with fluorescence microscopy. The DNA-bound SYBR® Gold emits green light when excited at 425 – 500 nm. Assays were carried out in triplicate and mean % tail DNA for at least 50 cells/slide was determined by two investigators blinded to genotype. Means were used for statistical comparisons.

2.4. Maintenance and breeding of mice

Ciz1^−/− mice and Ciz1^+/+ littermates (18-mo-old) were used in the experiments described herein. Ciz1^−/− founders were backcrossed to C57BL/6J mice for at least 6 generations and genotyped as previously described (Xiao et al., 2016). All experiments used Ciz1^+/+ littermates on the same genetic background. All mouse experiments were performed in accordance with the National Institutes of Health’s Guidelines for the Care and Use of Laboratory Animals and approved by our Institutional Animal Care and Use Committee.

2.5. Behavioral assessments

Ciz1^−/− mice and sex-matched Ciz1^+/+ littermates at 18-mo of age were subjected to a battery of sensorimotor and cognitive tests including open-field activity, rotarod, vertical rope climbing, raised-beam task, grip strength, dominance tube, cross maze and Morris water maze as previously described (Xiao et al., 2016, Xiao et al., 2017, Caccamo et al., 2014). Tests were performed by investigators blinded to genotype. An expanded version of these methods is available in the supporting information.

2.6. Immunohistochemistry

Ciz1^−/− mice and WT littermates were overdosed with pentobarbital (> 100 mg/kg) and transcardially perfused, first with ice cold saline and then 4% paraformaldehyde in 0.1M phosphate buffer (PB, pH 7.4). Brains were removed, weighed and post-fixed with 4% paraformaldehyde and cryoprotected with 30% sucrose in 0.1M PBS. Brains were sectioned on a cryostat at 20 μm. Endogenous peroxidases were quenched with 1% H₂O₂ in PBS and sections were rinsed with PBS followed by blocking with 2% non-fat dry milk and 0.3% Triton X-100. Sections were incubated overnight with rat anti-mouse Ly-6B.2 monoclonal antibody (Bio-Rad, MCA771GT) followed by biotinylated goat anti-rat IgG (Vector Laboratories, BA-9400). After rinsing with PBS X 3, sections were developed using the Vectastain ABC Kit from Vector Laboratories and DAB (3,3’-diaminobenzidine). Sections were counterstained with hematoxylin. For immunofluorescence, we used primary antibodies to 53BP1 (ab21083, Abcam), ionized calcium binding adaptor molecule 1(Iba-1) [1:500, polyclonal rabbit, Wako- #019-19741], NeuN (1:100, mouse monoclonal, EMG Millipore - MAB377) and glial fibrillary acidic protein (GFAP) [1:500, polyclonal rabbit, Abcam - #7260]. One of the following fluorescent secondary antibodies was used: Alexa Flour 568 anti-rabbit, Alexa Flour 488 anti-rabbit. These sections were then washed and mounted with Vectashield® mounting medium containing DAPI (H-1200, Vector Labs). The numbers of GFAP and Iba-1 positive cells were counted at 400X magnification (4 or 5 fields/region and genotype) by an investigator blinded to genotype.

2.7. Western blot analyses

Dissected brain tissues and MEFs were lysed with ice-cold RIPA lysis buffer (ThermoFisher Scientific) containing Halt™ protease and phosphatase inhibitor cocktail (ThermoFisher Scientific). Lysed issue and cells samples were microcentrifuged for 15 min at 14,000 rpm and the supernatants were collected. Proteins were separated by SDS-PAGE (4–12% Bis-Tris gels) and transferred to PVDF membranes using a BioRad wet transfer system. Subsequently, membranes were blocked for 2 h in 5% non-fat dry milk and incubated overnight with rabbit anti-phospho-H2A.X (Ser139) (#2577, Cell Signaling) in tris-buffered saline with Tween 20 (TBST) or rabbit anti-4-hydroxynonenal (HNE) primary antibody (ab46545; Abcam) in phosphate-buffered saline with Tween 20 (PBST) containing 5% non-fat dry milk. Membranes were washed 3X with 1XTBST and 1X PBST respectively for 15 min and then incubated with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories) for 1 h with constant rocking at room temperature. Signal was detected using enhanced chemiluminescence (Amersham) and quantified with NIH image J software.

2.8. Myeloperoxidase (MPO) activity assay

Brain tissues were sonicated in 50 mM potassium phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide. After centrifugation, the supernatant was diluted in reaction solution containing o-dianisidine hydrochloride and H₂O₂. The rate of change in optical density (OD) for 1 min was measured at 460 nm to calculate MPO activity as previously described (Khan et al., 2017).

2.9. ELISAs

Mice were anesthetized with pentobarbital (> 100 mg/kg) and transcardially perfused with ice-cold saline and decapitated. Brains were collected and homogenized in tissue lysis buffer (50 mM Tris HCl, pH 8.0, 5 mM NaCl, and 1% Triton X-100) containing Halt™ protease and phosphatase inhibitor cocktail. Nuclear fractions were prepared with NE-PER™ nuclear and cytoplasmic extraction kit (ThermoFisher Scientific, #78833) according to the manufacturer’s instructions. Briefly, lysed tissues were microcentrifuged for 5 min at 14,000 rpm and supernatants were collected for interleukin-6 (IL-6) and reduced glutathione (GSH) assays. Insoluble pellets were resuspended in ice-cold nuclear extraction reagent. Pellets were placed on ice and vortexed for 15 s every 10 min, for a total of 40 min. After incubation, pellets were microcentrifuged for 10 min at 14,000 rpm and nuclear fractions were collected for the analysis of NF-κB. Protein concentrations were determined using a bicinchoninic acid assay (BCA) kit (Thermo Fisher Scientific, #23227).

An IL-6 ELISA kit (R & D System, DY406-05) was used for determination of IL-6 levels in accordance with the manufacturer’s instructions. Briefly, the capture antibody was diluted to the working concentration in PBS and used to load a 96-well microplate with 100 μL per well. The plate was sealed and incubated overnight at room temperature.. After washing, standards and samples were pipetted into the wells and incubated for 2 h. After washing away unbound substances, an enzyme linked polyclonal antibody specific for IL-6 was added to the wells. Following a wash to remove any unbound antibody–enzyme reagent, a substrate solution was added to the wells and color developed in proportion to the amount of bound IL-6. The reaction was terminated by the addition of 2N sulfuric acid. The absorbance was measured at 450 nm with a microtiter plate reader (SpectraMax M2e, Molecular Devices).

GSH was measured with a commercial ELISA kit (Cayman Chemical, #703002) in accordance with the manufacturer’s instructions. Briefly, 50 μl of standards and samples were added to 150 μl of a reaction mixture containing nicotinamide adenine dinucleotide phosphate (NADP⁺), glutathione reductase, glucose-6-phosphate, and 5,5′-dithiobis-2-nitrobenzoic acid (DTNB). The reaction was carried out at 37 °C for 10 min, and then GSH levels were determined by absorbance at 412 nm with a microtiter plate reader (SpectraMax M2e).

The activity of NF-κB was measured in nuclear extracts with a commercial NF-κB (p65) Transcription Factor Assay kit (Cayman Chemical, #10007889) in accordance with the manufacturer's instructions. Briefly, a specific double-stranded DNA sequence containing the NF-κB response element was immobilized to wells of a 96-well plate, nuclear fractions were added to wells, and NF-κB (p65) was detected by addition of a specific primary antibody. A secondary antibody conjugated to HRP was added to provide a sensitive colorimetric readout at 450 nm. Optical density was normalized using the amount of protein present in the nuclear fraction.

2.10. Relative quantitative real-time reverse-transcriptase PCR (QRT-PCR) for expression of vascular adhesion factors

Relative levels of mouse intercellular adhesion molecule 1 (ICAM-1) and vascular adhesion molecule (VCAM) mRNA were determined in hippocampus harvested from 3 mice (18 mo) of each genotype (Ciz1^{−/ −} and Ciz1^+/+). TaqMan (Roche)-based QRT-PCR was performed using two primer pairs and probes (ICAM: forward, tggaagctgtttgagctgag; reverse, tgccacagttctcaaagcac; probe 56; VCAM, forward, gagaatgaacactcttacctgtgc; reverse, cacgtggatacttcgttccag; probe 21) with a LightCycler® 480 System (Roche). Mouse β-actin was used as the endogenous control.

2.11. Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL)

TUNEL staining was performed using a commercially available in situ Cell Death Detection Kit (Roche Diagnostics). The assay was performed according to the manufacturer’s instructions. Data were analyzed by determining the mean number of TUNEL-positive cells per 200X magnification field. A rater blinded to genotype examined eight microscopic fields per slide.

2.12. Statistical analysis

ANOVA with post-hoc tests was used to determine the effects of genotype and sex on parametric behavioral measures. ANOVA with post-hoc tests was used to determine the effects of genotype and radiation on % tail DNA. The Mann-Whitney test was used to determine the effects of genotype within sex for a non-parametric behavioral measure (slips on the raised beam task). Two-tailed t-tests were used to determine the effects of genotype on gene expression, TUNEL labeling, biochemical measures, and hematological parameters. Fisher’s exact test was used to determine the effects of genotype on the results of dominance tube testing. An alpha (α) of 0.05 was chosen for statistical significance.

3. Results

3.1. CIZ1 deficiency causes persistent DNA damage, cell-cycle defects and apoptosis in MEFs

To determine the contributions of CIZ1 to the DNA damage response, we subjected MEFs isolated from Ciz1^+/+ (wild-type, WT) and Ciz1^{−/ −} (knock-out or null) mice to γ-irradiation. MEFs isolated from Ciz1^{−/ −} mice showed increased sensitivity to irradiation. At 24 hr after IR (20 Gy), we assessed the persistence of DNA break markers γ-H2A.X (ser139) and 53BP1 and degree of DNA damage with comet assays. MEFs isolated from null mice displayed increased and sustained γ-H2A.X(ser139)- (Fig. 1 A and B) and 53BP1- (Fig. S1) immunoreactivity. There were significant effects of irradiation (F_1,8 = 271,8, P < 0.0001) and genotype (F_1,8 = 13.8, P = 0.0059) on % tail DNA (Fig. 1 C and D). Using flow cytometry, MEFs from null mice showed cell-cycle defects (Fig. 1E) and increased apoptosis (two-tailed t(8) = 3.14, P = 0.014; Fig. 1F) in comparsion with MEFs from WT littermates. In aggregate, these results indicate that CIZ1-deficient MEFs are more sensitive to irradiation than WT MEFs, as evidenced by increased DNA damage, cell-cycle disruption and cell death.

Fig. 1.

CIZ1 deficiency is associated with increased and persistent DSBs, DNA tails, cell-cycle defects and apoptosis in irradiated MEFs. MEFs isolated from Ciz1^−/− and Ciz1^+/+ embryos were subjected to 20 Gy γ-IR. At 24 h after γ-IR, DNA damage was assessed by γ-H2A.X (ser139) immunohistochemistry (A), Western blotting (B), comet assays (C and D) and flow cytometry (E and F). Untreated MEFs from WT and Ciz1^−/− mice showed negligible expression of γ-H2A.X (ser139) and largely absent DNA tails. In contrast, Ciz1^−/− MEFs exposed to γ-IR showed increased γ-H2A.X (ser139) expression along with increased % tail DNA. (E) Representative images of flow cytometry analysis of the cell cycle in Ciz1^+/+ and Ciz1^−/− MEFs. (F) In comparison with Ciz1^+/+ MEFs, a higher percentage of Ciz1^−/− MEFs were apoptotic. Value are expressed as means ± SEM. *P < 0.05. Scale bar for A, 50 μm.

3.2. CIZ1 deficiency is associated with DNA damage in the brains of aged mice

First, we confirmed that CIZ1 is widely expressed in neural tissue and present within nuclear foci in mouse neurons, astrocytes and microglia (Fig. S2). CIZ1 is known to interact with cyclin A1 (Greaves et al., 2012), a cell-cycle protein implicated in the repair of DNA double-strand breaks (DSBs). Therefore, we sought to determine if CIZ1 deficiency contributes to impaired DNA repair in brain. For this purpose, we investigated DNA damage in the cerebellum and hippocampus of aged WT and null mice with comet assays (Fig. 2A–D) and immunohistochemistry (Fig. 2F–I). We found increased % tail DNA (Fig. 2E) in the cerebellum (two-tailed t(10) = 12.14, P <0.0001) and hippocampus (two-tailed t(10) = 14.07, P <0.0001) of aged Ciz1^{−/ −} mice as compared with WT littermates. Similarly, 53BP1-immunoreactive cells (Fig. 2J) were also more prevalent in the cerebellum (two-tailed t(4) = 4.90, P = 0.008) and hippocampus (two-tailed t(5) = 4.12, P = 0.0091) of aged Ciz1^{−/ −} mice. Taken together, these results implicate CIZ1 as a participant in the DNA damage response.

Fig. 2.

CIZ1 deficiency is associated with increased % tail DNA (A – E) and persistent DSBs (F –J ) in the brains of aged Ciz1^−/− mice (A–J). Shown are representative images of comet assays [(A and B; cerebellum) and (C and D; hippocampus)] and their quantification (E). Right-sided panels (F – I) show 53BP1 immunohistochemistry and its quantification (J) in WT and Ciz1^−/− mice. Value are expressed as means ± SEM. * P < 0.05. Scale bar for F – I, 100 μm.

3.3. Oxidative stress is seen in the brains of aged Ciz1^{−/ −} mice

Oxidative stress generally refers to increased levels of reactive oxygen species (ROS) which can damage DNA, RNA, proteins and lipids. Oxidative stress, most commonly linked to impaired mitochondrial function, plays an important role in the pathogenesis of aging and neurodegeneration (Gibson et al., 2008, Butterfield et al., 2001). ROS are a major cause of DSBs and other deleterious changes to DNA. Analysis of lipid peroxidation (LPO) and GSH levels in Ciz1^{−/ −} mice was motivated by the known functional interplay between the cell-cycle, DNA damage response and oxidative stress (Yan et al., 2014). LPO, a classic marker of oxidative stress, is seen in the earliest stages of neurodegeneration (Bradley-Whitman and Lovell, 2015). Levels of the antioxidant glutathione, which is essential for detoxification of ROS, are commonly reduced in neurodegenerative diseases (Mazzetti et al., 2015). In comparison to WT littermates, Ciz1^{−/ −} mice exhibited increased LPO (two-tailed t(4) = 4.14, P = 0.014) and decreased GSH (two-tailed t(7) = 2.56, P = 0.037) levels (Fig. 3 A and B).

Fig. 3.

CIZ1 deficiency contributes to oxidative stress as measured by lipid peroxidation (4-HNE) and GSH levels in the brains of Ciz1^−/− mice and WT littermates. Protein levels of 4-HNE (A) and GSH (B) in brain homogenates were determined by Western blotting and ELISA, respectively. Values are expressed as means ± SEM. *P < 0.05.

3.4. Aged Ciz1^{−/ −} mice show increased glial activation in the CNS

Activation of microglia and astrocytes, hallmarks of neuroinflammation, has been reported in several neurodegenerative disorders (Glass et al., 2010, Khan et al., 2015). In the present study, we analyzed astrocytic (GFAP) and microglial (Iba1) activation in the brains (cerebellum and hippocampus) of aged Ciz1^{−/ −} mice and WT littermates. GFAP immunostaining exposed astroglial activation and increased numbers of activated astrocytes in the cerebellae (two-tailed t(4) = 3.67, P = 0.021) of Ciz1^{−/ −} mice as compared with WT littermates (Fig. 4). Differences did not reach statistical significance in the hippocampi (two-tailed t(4) = 2.17, P = 0.096). There were significantly increased numbers and amoeboid features of Iba1-positive activated microglia in Ciz1^{−/ −} mice compared with WT littermates in both the cerebellum (two-tailed t(3) = 8.92, P = 0.0030) and hippocampus (two-tailed t(4) = 3.93, P = 0.017) (Fig. 4).

Fig. 4.

Glial activation in the brains of 18-mo-old Ciz1^−/− mice and Ciz1^+/+ littermates. The expression of microglial (Iba1) and astrocytic (GFAP) markers was determined by immunofluorescence. Marked expression of GFAP (green) and Iba1 (red) positive cells in brain was observed in Ciz1^−/− mice as compared with Ciz1^+/+ littermates. Scale bar, 20 μm.

3.5. NF-κB is up-regulated in Ciz1^{−/ −} mice

NF-κB signaling is at the center of the inflammatory network in aging and plays an important role in myriad physiological and pathophysiological conditions (Mattson and Camandola, 2001, Bauernfeind et al., 2009). In previous work using gene-expression microarrays, we found that NF-κB was a major hub for up-regulated genes in Ciz1^−/− mice (Xiao et al., 2016). Consistent with our gene expression data, we observed increased protein levels of NF-κB (p65) in the nuclear extracts from the cerebellae (two-tailed t(9) = 5.18, P = 0.0006) and hippocampi (two-tailed t(7) = 5.02, P < 0.0015) of aged Ciz1^{−/ −} mice as compared with WT littermates (Fig. S3).

3.6. CIZ1 deficiency triggers inflammation

We then tested the hypothesis that increased DNA damage and NF-κB upregulation are associated with increased neuroinflammation. For this purpose, we analyzed MPO activity and proinflammatory cytokine IL-6 levels in supernatant fractions of brain (Fig. S4). MPO, an abundant pro-inflammatory enzyme stored in neutrophils and monocytes, is also expressed by perivascular macrophages and amoeboid microglia in the CNS (Kaindlstorfer et al., 2015). MPO-immunoreactive cells are prominent in brain regions undergoing neurodegeneration in Parkinson and Alzheimer diseases (Gellhaar et al., 2017). The pro-inflammatory cytokine IL-6 is elevated in both classic neurodegenerative disorders like Parkinson and AD and primary neuroinflammatory conditions such as multiple sclerosis (Alam et al., 2016). IL-6 signaling promotes DNA repair (Chen et al., 2015). Neutrophils were more abundant in the brains of Ciz1^{−/ −} mice compared with WT littermates (Fig. S4 A), and Ciz1^{−/ −} mice demonstrated significantly increased MPO activity (two-tailed t(10) = 2.83, P = 0.018; Fig. S4 B) and IL-6 levels (two-tailed t(13) = 2.26, P = 0.042; Fig. S4 C). We also observed a significantly increased percentage of neutrophils in the blood of aged Ciz1^{−/ −} female mice compared with WT female littermates (two-tailed t(10) = 5.25, P = 0.0004; Table S1). These findings demonstrate that CIZ1 deficiency triggers inflammatory responses.

3.7. Aged Ciz1 null mice show evidence of neurovascular dysfunction

To determine if CIZ1 deficiency causes endothelial dysfunction, we measured the expression of ICAM-1 and VCAM with QRT-PCR. Induction of ICAM-1 and VCAM expression are well-established markers of endothelial cell activation and vascular dysfunction (Engelhardt et al., 1994, Seo et al., 2015). VCAM levels are elevated in patients with both Alzheimer and vascular dementia (Zuliani et al., 2008). Aged Ciz1 null mice showed higher expression levels of ICAM (two-tailed t(7) = 4.47, P = 0.0029) and VCAM (two-tailed t(7) = 2.72, P = 0.030) compared with WT littermates (Fig. S5).

3.8. Deficiency of CIZ1 is associated with cell death in the brains of aged mice

Next, we tested the hypothesis that the accumulation of DNA breaks contributes to cell death in Ciz1^{−/ −} mice. TUNEL labeling showed that markers of DNA damage, inflammation, and neurovascular dysfunction were accompanied by increased apoptosis in sensorimotor (cerebellum; two-tailed t(9) = 5.18, P = 0.0006) and cognitive (hippocampus; two-tailed t(7) = 5.02, P = 0.0015) regions of the CNS in Ciz1 null mice in comparison with WT littermates (Fig. 5).

Fig. 5.

CIZ1 deficiency contributes to apoptosis. Brain tissue was harvested from 18-mo-old Ciz1^−/− mice and WT littermates and processed for TUNEL staining. The left panel shows representative photomicrographs stained for apoptotic cells in cerebellum and hippocampus. The right panel shows quantification of TUNEL-positive cells. In comparison with WT littermates, Ciz1^−/− mice had a significant increase in the number of apoptotic events. Values are expressed as means ± SEM. *P < 0.005. Scale bar, 100 μm.

3.9. CIZ1 deficiency contributes to sensorimotor, behavioral, and cognitive deficits in aged mice

Our earlier work showed that CIZ1 deficiency causes mild motor abnormalities in 3-mo-old mice (Xiao et al., 2016). Now, we provide multiple lines of evidence that aged (18-mo-old) Ciz1 mice exhibit marked relative deficits in sensorimotor and behavioral functioning in comparison with WT littermates. There were no overall effects of genotype on body or brain weights (Table S2, Fig. S6). In comparison with WT littermates, both male (F_1,31 = 11.19, P = 0.0022) and female (F_1,31 = 12.63, P = 0.0024) Ciz1^{−/ −} mice performed poorly on the rotarod (Fig. 6A) and raised-beam tasks (Fig. 6 B and C). On the 12-mm square beam, there were overall effects of genotype (F_1,41 = 20.10, P < 0.0001) and sex (F_1,41 = 10.54, P < 0.0023) on traversal times but no genotype*sex interaction. On the 12-mm round beam, there were overall effects of genotype (F_1,40 = 37.14, P < 0.0001) and sex (F_1,40 = 19.29, P < 0.0001) on traversal times and a weak genotype*sex interaction (F_1,40 = 4.12, P < 0.049) with relatively poorer performances by female than male Ciz1^{−/ −} mice. Analysis of slips with the Mann-Whitney U statistic exposed similar effects of genotype on performance. On the 12-mm square beam, both male (Mann-Whitney U = 16, P = 0.0002) and female (Mann-Whitney U = 1, P < 0.0001) Ciz1^{−/ −} mice exhibited more slips than their WT littermates (Fig. 6C). Likewise, on the 12-mm round beam, both male (Mann-Whitney U = 10, P < 0.0001) and female (Mann-Whitney U = 14, P < 0.018) Ciz1^{−/ −} mice exhibited more slips than WT littermates (Fig. 6C). There was no overall effect of genotype on grip strength or normalized grip strength. However, normalized grip strength showed a significant genotype*gender interaction (F_1,48 = 8.38, P = 0.0057) with female Ciz1^{−/ −} mice showing lower values for grip strength/body weight in comparision with WT littermates (P = 0.011). A similar genotype*gender interaction (F_1,30 = 11.88, P = 0.0017) was seen for normalized rope climbing times with female Ciz1^{−/ −} mice showing larger values for rope climbing time/body weight in comparision with WT littermates (P = 0.003).

Fig. 6.

Aged Ciz1^−/− mice and sex-matched WT littermates were analyzed with a battery of behavioral tests (N = 10 – 16/group). (A) Rotarod- latency to fall from an accelarating rotating drum was measured. Both male and female Ciz1^−/− mice had a shorter latency to fall in comparison to WT littermates. (B and C) Raised beam tasks assessed the ability of mice to traverse narrow beams to reach a dark box. Overall, Ciz1^−/− mice moved slower and slipped more frequently than WT littermates (B and C, respectively). (D) Ciz1^−/− mice showed reduced percentages of correct choices in the cross maze task, a measure of working memory. The effect of genotype was promiment in both female and male mice. (E and F) Morris water maze- aged Ciz1^−/− mice exhibited decreased escape latencies (E) and spent significantly less time in the target quadrant (F) when compared to WT littermates. Value are expressed as means ± SEM. *significant effect of genotype within sex (P < 0.05).

Overall, male mice were more active than female mice, particularly in the vertical plane, and Ciz1^−/- mice were less active than their WT littermates (Table S2). Ciz1^{−/ −} mice traveled shorter distances (F_1,48 = 16.99, P < 0.0001), had lower ambulatory counts (F_1,48 = 16.45, P = 0.0002) and fewer ambulatory episodes (F_1,48 = 11.69, P = 0.0013) than their WT littermates in the open field and there were no differential effects of sex for any of these three measures. Likewise, there were overall effects of genotype on stereotypic (F_1,48 = 17.68, P < 0.0001), vertical (F_1,48 = 5.58, P = 0.022) and jump (F_1,48 = 10.49, P = 0.0022) counts. The genotype*sex interaction (F_1,48 = 6.63, P = 0.013) reached statistical significance for jump counts with more pronounced relative reductions in male than female Ciz1^{−/ −} mice. There was no overall effect of genotype on average velocity but post-hoc analyses showed the male Ciz1^{−/ −} mice were slower than their WT male littermates (P = 0.028).

Behaviorally, male and female Ciz1^{−/ −} mice were more aggressive than their WT littermates in the dominance tube (P < 0.0001, for both; Table S2). Similar to motor and behavioral findings, we also observed cognitive dysfunction in Ciz1^{−/ −} mice as assessed by cross maze and Morris water maze testing (Fig. 6D–F). The was an overall effect of genotype (F_1,31 = 24.61, P < 0.0001) but no effect of sex or the genotype*sex interaction on cross maze scores. There were modest effects of the null genotype on escape latencies in both male (F_1,14 = 4.38, P = 0.054) and female (F_1,14 = 4.65, P = 0.049) mice. There was a strong overall effect of the null genotype on the probe trial (Fig. 6F; F_1,26 = 14.17, P = 0.0009) but no effect of the genotype*sex interaction. These data indicate that the deleterious effects of CIZ1 deficiency become more pronounced with aging and CIZ1 deficiency contributes to the development of both cognitive and motor dysfunction.

4. Discussion

Several studies have demonstrated the roles of ectopic cell-cycle events and DNA damage in neuropathological conditions and altered expression of cell-cycle proteins has been signified as a contributor to and indicator of cell death (Suberbielle et al., 2015, Katsel et al., 2013, Shen et al., 2016). CDKN1A-interacting zinc finger protein-1 (CIZ1), a nuclear protein that plays an important role in DNA replication and cell-cycle progression (Mitsui et al., 1999, Coverley et al., 2005), has been linked to cervical dystonia through G1/S cell-cycle dysregulation in humans (Xiao et al., 2012). Recently, we reported that germline knock-out of Ciz1 causes motor and hematological abnormalities in young adult mice (Xiao et al., 2016). However, our previous work did not explore the significance of CIZ1 deficiency in the aged nervous system or contributions of CIZ1 to DNA repair pathways. In the present study, we learned that CIZ1 deficiency leads to impaired DNA repair capacity, cell-cycle defects, and apoptosis in MEFs. Concordant results were identified in the brains of aged Ciz1 null mice with increased DNA breaks accompanied by neuroinflammation, vascular dysfunction, and cell death. Consistent with the effects of CIZ1 deficiency at the cellular level, aged Ciz1 null mice had overt deficits in motor and cognitive functioning. Our findings expand upon previous work linking cell-cycle proteins to the DNA damage response in post-mitotic neurons (Suberbielle et al., 2015, Jordan-Sciutto et al., 2003, Kim and Tsai, 2009, Katsel et al., 2013, Shen et al., 2016) by tying a monogenic defect to the complete constellation of motor, cognitive, neurovascular, and neuroinflammatory phenotypes seen in age-related global neurological deterioration and neurodegenerative disorders. It should be noted, however, that young controls were not included in our experimental design and the influence of age and the age*genotype interaction on phenotypes was not directly tested by assessing Ciz1^+/+ and Ciz1^{−/ −} mice at multiple development time points.

DNA is constantly subjected to damage from numerous sources and defects in DNA repair processes can lead to senescence and uncontrolled cell death. However, neurons and other cell types within the CNS are equipped with a complex network of DNA repair mechanisms to preserve genomic integrity (Friedberg, 2003, Chow and Herrup, 2015). To understand the role of the cell-cycle protein CIZ1 in DNA damage response pathways, we first isolated MEFs from WT and Ciz1^−/− mice. We found that MEFs isolated from Ciz1^−/− mice were abnormally sensitive to γ-irradiation. Following irradiation, MEFs from Ciz1^−/− mice displayed increased DNA damage, cell cycle defects, and increased apoptosis. In particular, we found persistent 53BP1 and γ-H2A.X-immunoreactivity at the sites of DSBs along with increased % tail DNA with comet assays (Rogakou et al., 1998, Lips and Kaina, 2001).

Mutations in genes that encode proteins involved in the cell-cycle and DNA repair have been linked to several age-related neurodegenerative and/or neurodevelopmental disorders including AD, ataxia telangiectasia (ATM), Cockayne syndrome (ERCC8), and xeroderma pigmentosum (ERCC4) (Best, 2009, Jeppesen et al., 2011, Katsel et al., 2013). Moreover, genome-wide association studies have consistently identified variants in cell-cycle genes as susceptibility factors for mild cognitive impairment and dementia (Saykin et al., 2015) and two polymorphisms in CDKN1A are associated with increased risk and earlier age-of-onset of AD (Yates et al., 2015). ATM shows interesting parallels to CIZ1. Both genes encode G1/S cell-cycle proteins and sequence variants in both genes have been linked to isolated dystonia and neoplasias. Loss of ATM in rats leads to accumulation of unrepaired DNA, increased cytokine production and activation of microglia (Quek et al., 2017a). Innate recognition of cytosolic self-DNA may trigger neuroinflammation in ATM and CIZ1 deficient rodents (Quek et al., 2017b, Li et al., 2017).

Among the types of DNA damage induced by ionizing radiation, DSBs may be the most lethal. DSBs are particularly deleterious in neurons because of their reduced DNA repair capacity when compared with proliferating cells (Kanungo, 2013). The accumulation of DNA damage is believed to result in a slow build-up of DNA adducts which may trigger an immune system response and inflammation and contribute to cell death (Ioannidou et al., 2016, Gasser and Raulet, 2006, Paludan, 2015). Deficiency of the cell-cycle protein ATM is associated with genomic instability and contributes neuronal death in mouse and human AD brain (Shen et al., 2016). Similarly, reduced neuronal expression of BRCA1 occurs in AD brain, causes cognitive dysfunction in mice and is characterized by sustained DSBs and abnormal chromatin remodeling (Suberbielle et al., 2015). Consistent with studies of the G1/S cell-cycle proteins ATM and BRCA1, our aged Ciz1^−/− mice displayed increased DNA damage in the brain as demonstrated with persistent 53BP1-positive cells and increased percentages of DNA tails in neurons. The presence of extensive DNA damage and neuroinflammation in the aged brain may be manifestations of defective DNA repair pathways and accumulative oxidative damage to DNA (Silva et al., 2014, Chen et al., 2010).

Brain lipids, proteins and DNA are sensitive to oxidative stress (Radak et al., 2011, Butterfield et al., 2001, Markesbery and Lovell, 2007). In our study, we found a significant association between CIZ1 deficiency and oxidative stress in the brains of aged mice. Ciz1^−/− mice exhibited increased LPO and decreased GSH. Oxidative damage to lipids can lead to structural and functional disruption of cell membranes and inactivation of enzymes, which ultimately leads to cell death. A reduction in GSH may impair H₂O₂ clearance and increases the free radical load, which promotes oxidative stress and consequently disrupts homeostasis. Moreover, altered cellular redox status can activate redox-sensitive transcription factors such as NF-κB (Janssen-Heininger et al., 2000). In view of our aggregate findings, it is worth considering the possibility that the increased oxidative stress identified in Ciz1^−/− mice works as an amplification loop to further accelerate DNA damage and neurodegeneration.

The importance of glia and pro-inflammatory mediators in neurodegenerative diseases has been supported by the analysis of diseased postmortem brains (Ouchi et al., 2009, Glass et al., 2010). Importantly, causal relationships between the accumulation of damaged DNA, glial activation, and age-related neurodegeneration was provided by work with Ercc1 mutant mice (Ercc1Δ/−) (Borgesius et al., 2011). Consistent with study of ERCC1 which is involved in multiple DNA repair pathways, we observed that aged Ciz1^−/− mice harbored more activated astrocytes and microglia than their WT littermates.

NF-κB, a master regulator of inflammation, plays an important role in myriad physiological and pathological conditions (Mattson and Camandola, 2001, Balistreri et al., 2013). NF-κB is upregulated in response to accumulation of damaged DNA and pharmacological inhibition of NF-κB activation leads to attenuation of age-related pathologies (Tilstra et al., 2012). Moreover, several studies have suggested that activated microglia also upregulate NF-κB nuclear translocation and inhibition of NF-κB suppresses pro-inflammatory cytokines and neuronal loss in mouse models of neurodegenerative diseases (Frakes et al., 2014, Khan et al., 2015, Ghosh et al., 2007). Consistent with our gene expression data in younger Ciz1^−/− mice, we observed increased NF-κB (p65) and cytokine IL-6 in the brains of aged Ciz1^−/− mice as compared with WT littermates. Furthermore, we provided evidence that upregulation of NF-κB was associated with damaged DNA and neuronal death in Ciz1^−/− mice. Neutrophil influx into the CNS has been positively correlated with neuroinflammation and neuronal death (Khan et al., 2012, Soehnlein et al., 2017, Zenaro et al., 2015). In our study, we found that aged Ciz1^−/− mice display increased MPO activity and neutrophil recruitment into the CNS as compared with their WT littermates. Although not assessed in our study, aged female Ciz1^−/− mice have been reported to exhibit pathological evidence of a diffuse lymphoproliferative disorder resembling non-Hodgkin follicular-type lymphoma with enlarged organs, particularly the spleen (Ridings-Figueroa et al., 2017).

Recent studies have provided evidence that the accumulation of DNA breaks may be a direct contributor to neurovascular damage in several pathological conditions (Uryga et al., 2016, Bautista-Nino et al., 2016). In aged Ciz1^−/− mice, we observed increased expression of ICAM1 and VCAM when compared with WT littermates. Increased expression of VACM-1 and ICAM-1 has been associated with increased risk of metabolic disorders and cognitive impairment (Hall et al., 2013, Zuliani et al., 2008).

Behavioral findings are often closely correlated with overall measures of CNS pathology in both humans and mice. Aged Ciz1^−/− mice exhibited motor deficits on the rotarod, grip strength, rope climbing, open field activity, and a raised-beam task. Ciz1^−/− mice also displayed cognitive deficits as measured by cross and Morris water mazes when compared with WT littermates. Our findings correlate well with our analyses of young adult Ciz1^−/− mice and work with other mouse models in which motor and cognitive deficits were tied to defective cell-cycle proteins and DNA damage repair pathways (Xiao et al., 2016, Suberbielle et al., 2015).

Our findings that lack of CIZ1 is associated with the accumulation of damaged DNA and neurodegeneration in aged mice was not unexpected given CIZ1’s established roles in DNA replication and cell-cycle progression at the G1/S checkpoint in mitotic cells. However, its role in aged post-mitotic neurons had not been investigated. Our study exposes an important role for the cell-cycle protein CIZ1 in the DNA damage repair pathways of post-mitotic neurons and potential contributions of CIZ1 and associated cell-cycle proteins to dementia. Overall, our results provide, for the first time, a rationale model system for studying the aggregate and combinatorial effects of DNA damage and neuroinflammation in global neurological decline in elderly human populations. Additional work will be required to determine the exact mechanisms through which CIZ1 contributes to DNA repair in the adult brain and delineate pathways linking DNA damage, neuroinflammation and vascular dysfunction.

Highlights.

• Embryonic fibroblasts from Ciz1 null mice are abnormally sensitive to irradiation

• Aged Ciz1 null mice exhibit marked deficits in motor and cognitive functioning

• Aged Ciz1 null mice show increased DNA damage and NF-κB upregulation in brain

• Defects in DNA repair may contribute to global neurological decline in elderly humans

Appendix A. Supplementary information

Supplementary information associated with this article can be found, in the online version.