Abstract
Abnormal modification of 5-hydroxymethylcytosine (5hmC) is closely related to the occurrence of Alzheimer’s disease (AD). However, the role of 5hmC and its writers, ten-eleven translocation (Tet) proteins, in regulating the pathogenesis of AD remains largely unknown. We detected a significant decrease in 5hmC and Tet2 levels in the hippocampus of aged APPswe/PSEN1 double-transgenic (2×Tg-AD) mice that coincides with abundant amyloid-β (Aβ) plaque accumulation. On this basis, we examined the reduction of Tet2 expression in the hippocampus at early disease stages, which caused a decline of 5hmC levels and led young 2×Tg-AD mice to present with advanced stages of AD-related pathological hallmarks, including Aβ accumulation, GFAP-positive astrogliosis and Iba1-positive microglia overgrowth as well as the overproduction of pro-inflammatory factors. Additionally, the loss of Tet2 in the 2×Tg-AD mice at 5 months of age accelerated hippocampal-dependent learning and memory impairments compared to age-matched control 2×Tg-AD mice. In contrast, restoring Tet2 expression in adult neural stem cells isolated from aged 2×Tg-AD mice hippocampi increased 5hmC levels and increased their regenerative capacity, suggesting that Tet2 might be an exciting target for rejuvenating the brain during aging and AD. Further, hippocampal RNA sequencing data revealed that the expression of altered genes identified in both Tet2 knockdown and control 2×Tg-AD mice was significantly associated with inflammation response. Finally, we demonstrated that Tet2-mediated 5hmC epigenetic modifications regulate AD pathology by interacting with HDAC1. These results suggest a combined approach for the regulation and treatment of AD-related memory impairment and cognitive symptoms by increasing Tet2 via HDAC1 suppression.
Introduction
Alzheimer’s disease (AD) is the most common degenerative disorder of the central nervous system, which is characterized by progressive memory loss and cognitive dysfunction with a gradual and consistent course of development process that is strongly associated with aging (1,2). The classical pathological hallmarks include deposition of amyloid-β (Aβ) plaques, accumulation of hyperphosphorylated tau protein and formation of neurofibrillary tangle (3), but it is also accompanied by other key pathological changes such as neuroinflammation and neuronal and synaptic loss (4,5). The AD patients suffer both mental and physical agony as dementia worsens (6). Unfortunately, the etiology of AD remains obscure, and no current interventions can modify the underlying disease mechanisms. Several treatment strategies aimed at eliminating Aβ burden have failed to prevent and/or reverse the pathological process (7), suggesting that the amyloid-cascade hypothesis cannot comprehensively explain the complex symptoms of AD (8,9). Therefore, identification of novel molecular targets may provide for a more effective preventive and therapeutic strategy to treat AD.
Recent studies on the pathological mechanisms of AD have begun to explore the regulation of epigenetic modifications (10,11). Current evidence suggests that the cytosine modifications including 5-methylcytosine and 5-hydroxymethylcytosine (5hmC) are potentially involved in the pathogenesis of AD (12,13). Interestingly, 5hmC has been reported to be significantly higher in the nervous system and in particular located at genes known to have neuronal and synaptic functions (14–16), suggesting 5hmC is a crucial epigenetic mark in the central nervous system and may have a role in the initiation and/or progression of neurodegenerative diseases including AD. The overall levels of 5hmC are significantly decreased in the hippocampus of aged APPswe/PSEN1 double-transgenic mice (2×Tg-AD) (17,18). The genes associated with differential hydroxymethylation not only modulate Tau-induced neurotoxicity but also perturb neuronal function (19). Therefore, 5hmC-mediated epigenetic regulation exhibits a positive correlation with the pathogenesis of AD. Mechanistically however, the involvement of Tet2 in regulating the pathological process of AD has yet to be investigated.
The enzymes that catalyze the production of 5hmC are the ten-eleven translocation (Tet) proteins, which consist of Tet1, Tet2 and Tet3, and are recognized as important epigenetic mediators (20–22). Tet family members have distinct expression patterns and biological functions (23). Previous studies have shown that Tet-mediated epigenetic regulation is involved in neurological diseases. For instance, Tet1 is critical for neuronal activity-regulated gene expression and, mice lacking Tet1 exhibit impaired hippocampal neurogenesis accompanied by poor learning and memory (24,25). In addition, Tet1 and Tet2 could regulate chronic restraint stress-induced expression-like behavior in mice (26). Additionally, Tet3 is critical for neural progenitor cell maintenance and terminal differentiation of neurons (27). Emerging evidence also suggests that Tet2 epigenetically regulates critical genes, controlling proliferation and differentiation of adult neural stem cells (aNSCs) during adult neurogenesis (28). Furthermore, Tet1 and Tet3 have been found to be involved in memory extinction although not in long-term memory acquisition, whereas Tet2 has a unique role in cognitive function distinct from the other Tet family members (29,30). Tet2-mediated 5hmC modification could affect age-related regenerative decline and influence cognitive function in the aged mouse brain (31). However, the 2×Tg-AD mice not only faced general problems of aging but also suffered from the complex pathological events of AD. However, it is not yet known whether Tet2-mediated 5hmC modification can affect the microenvironment of the 2×Tg-AD mice brain and whether its alteration may ameliorate or reverse the pathological process. Moreover, how Tet2-mediated 5hmC modifications may affect the memory loss and other neuropsychiatric symptoms of AD remains largely unexplored.
Here we demonstrate that Tet2 depletion in the hippocampus exacerbates AD pathology and cognitive dysfunction at early disease stages. Initially, we detect a significant decrease in 5hmC levels which are predominantly catalyzed by Tet2 associated with an increased accumulation of Aβ plaques in the hippocampus at later AD stages. This is followed by the progression of classical AD symptoms such as Aβ deposition, glial fibrillary acidic protein (GFAP)-positive astrocytes and Iba1 positive microglia overgrowth as well as pro-inflammatory cytokine accumulation. Therefore, the loss of Tet2 exacerbates cognitive impairments in young 2×Tg-AD mice. Conversely, overexpression of Tet2 (Overexpression Tet2, OE-Tet2) in aNSCs isolated from aged 2×Tg-AD mice hippocampi can reestablish regenerative capacity paralleled with a global increase of 5hmC levels. Moreover, the differentially expressed genes (DEGs) identified in the hippocampus of both Tet2 knockdown and control young 2×Tg-AD mice are strongly involved in inflammation reaction. Interestingly, we also observed that loss of Tet2 alters expression of the histone deacetylase 1 (HDAC1). Moreover, we demonstrated the physical interaction between endogenous HDAC1 and Tet2 in the hippocampus. Together, these results indicate that Tet2-mediated 5hmC epigenetic modifications can regulate the AD pathological process, possibly via an interaction with HDAC1.
Results
5hmC levels significantly decline in the aged hippocampus of 2×Tg-AD mice and are associated with increased Aβ deposition
To determine whether 5hmC-mediated epigenetic modification plays a crucial role in AD pathogenesis, we first investigated the changes of 5hmC levels in the hippocampi of wild-type (WT) and 2×Tg-AD mice by immunohistochemical staining and dot blot analysis and then evaluated the relationship between 5hmC expression and Aβ plaque accumulation. About 20 μm coronal brain sections of the hippocampi of WT and 2×Tg-AD mice were stained with anti-5hmC and anti-Aβ (6E10) antibodies (Fig. 1A). DNA extracted from the hippocampi was analyzed for 5hmC levels by dot blot (Fig. 1B and E). The result showed that no significant differences were found in 5hmC expression between WT and 2×Tg-AD mice at 24 weeks old (Fig. 1B and C). In this young stage, only a few Aβ plaques were found in the hippocampus region of 2×Tg-AD mice (Fig. 1A). However, at the aged stage (67 weeks old), we observed a precipitous decrease in 5hmC expression associated with an abundant accumulation of Aβ plaques in the hippocampus of 2×Tg-AD mice compared with age-matched WT mice (Fig. 1A, E and F). Furthermore, the 5hmC levels in the cortex of 2×Tg-AD mice at 67 weeks of age declined significantly compared to age-matched WT mice (Supplementary Material, Fig. S1A–D). Altogether, these results show a correlation between Aβ plaque buildup and decreased 5hmC levels in advanced stages of AD.
5hmC levels significantly decline in the hippocampus of aged 2×Tg AD mice associated with the accumulation of abundant Aβ plaques. (A) Immunofluorescence staining of 5hmC and Aβ in the hippocampus of 24- and 67-week-old WT and 2×Tg-AD mice. 5hmC fluorescence intensity did not show an observable difference in the hippocampus of 24-week-old WT and 2×Tg-AD mice, but it was decreased in the hippocampus of aged (67 weeks old) 2×Tg-AD mice than the age-matched WT mice. Aβ plaques began to appear at 24 weeks of age in 2×Tg-AD mice and steadily increased with aging. Scale bar: 200 μm. (B) Representative 5hmC dot blot assay of 24-week-old WT and 2×Tg-AD mice hippocampal lysates (n = 3 per group). (C) Quantitative analysis indicates that 5hmC levels were not significantly different between WT and 2×Tg-AD mice at 24 weeks of age. (D) Tet1, Tet2 and Tet3 mRNA transcripts were detected in the hippocampus of 67-week-old WT and 2×Tg-AD mice by real-time PCR (n = 3 per group). 18s mRNA levels were used as an internal control. Tet2 mRNA was elevated compared to Tet1 and Tet3 mRNA in aged WT mice, while it was significantly decreased in aged 2×Tg-AD mice. (E) Representative 5hmC dot blot assay of hippocampal lysates from 67-week-old WT and 2×Tg-AD mice (n = 3 per group). (F) Quantitative analysis showing that 5hmC levels was significantly decreased in hippocampal lysates of 2×Tg-AD mice compared to WT mice at 67 weeks of age (n = 3 per group). (G) Quantitative reverse-transcription PCR of Tet2 mRNA from 24- and 67-week-old WT and 2×Tg-AD mice hippocampus (n = 3 per group). 18s mRNA levels were used as an internal control. Data are presented as mean ± SEM; unpaired t-test; *P < 0.05; **P < 0.01; ***P < 0.001.
We next determined the expression of the Tet enzymes using fluorescence-quantitative reverse transcription-polymerase chain reaction (qRT-PCR). At the 67-week time point, there were no significant changes in Tet1 and Tet3 mRNA levels between WT and 2×Tg-AD mice hippocampi (Fig. 1D). However, the Tet2 mRNA level in the 2×Tg-AD mice hippocampus was significantly lower compared to that of age-matched WT mice (Fig. 1D). From 24 to 67 weeks, there was a sharp decline in Tet2 mRNA level in the hippocampi of both aged WT and 2×Tg-AD mice (Fig. 1G). Notably, Tet2 mRNA level in the aged AD mice was decreased by a much larger extent than in the WT mice (Fig. 1G). Taken together, these results indicate that Tet2 and 5hmC levels decline in the aging hippocampus and are potentially related to the progression of AD.
Decreased Tet2 in the dentate gyrus leads to an increased Aβ burden and decreased synaptic proteins in the early stage of 2×Tg-AD mice
To explore the role of Tet2-mediated 5hmC modification in the pathological process of AD, we knocked down Tet2 expression in the hippocampi of young WT and 2×Tg-AD mice (5-month-old) utilizing an in vivo adeno-associated virus (AAV)-mediated shRNA approach. Microinfusions (0.6 μl/side) of AAV-sh-Scramble or AAV-shTet2 were injected into the dentate gyrus (DG) regions of the hippocampi using a stereotaxic apparatus. The mice were sacrificed for biochemical analysis 4 weeks after the viral infusions (Fig. 2A). The infection efficiency of the AAV was determined by enhanced green fluorescent protein (eGFP) using fluorescence microscopy (Fig. 2B). Western blot and real-time PCR results confirmed that the expression of Tet2 in AAV-shTet2 groups significantly decreased compared with the AAV-sh-Scramble groups, indicating a high degree of Tet2 knockdown (Fig. 2C–E). To verify that this Tet2 knockdown could alter 5hmC levels, we then carried out a 5hmC dot blot assay with hippocampal DNA from 2×Tg-AD mice (Fig. 2F). Quantitative analysis of the dot blot result demonstrated that the levels of 5hmC were significantly decreased in AAV-shTet2 mice compared to AAV-sh-Scramble mice (Fig. 2G). Moreover, we examined 5hmC levels in the young 2×Tg-AD and WT mice infected with adenovirus by immunohistochemical analysis. Loss of Tet2 in the hippocampus resulted in an obvious decrease in the fluorescence signal intensity of 5hmC in the 2×Tg-AD mice compared to the Scramble control 2×Tg-AD mice (Supplementary Material, Fig. S2A), while the eGFP fluorescence signal intensity indicated similar infection levels in each group. This data verified that abrogation of Tet2 could significantly decrease 5hmC levels in the hippocampi of AD mice.
Loss of Tet2 in the hippocampus results in decreased 5hmC levels. (A) 20-week-old WT and 2×Tg-AD mice were given bilateral stereotactic injections of AAV encoding either shRNA targeting Tet2 or Scramble control sequences in tandem with an eGFP reporter into the bilateral DG of the ventral hippocampus. The mice were killed for biochemical analysis 4 weeks after the viral infusions. (B) Successful microinjection of the AAV-sh-Scramble and AAV-shTet2 vectors was evaluated by the expression of eGFP (green), which was observed by fluorescence microscopy. Scale bar = 200 μm. (C) Representative immunoblots of Tet2 protein detected by western blotting in hippocampal lysates from 2×Tg-AD mice. (D) Quantitative analysis of Tet2 protein expression in the hippocampus normalized to β-actin (n = 3 per group). (E) Tet2 mRNA transcript levels were detected in the hippocampus of 2×Tg-AD mice infected with virus by real-time PCR (n = 3 per group). 18s mRNA levels were used as an internal control. Tet2 mRNA was significantly downregulated in AAV-shTet2 compared to AAV-sh -Scramble. (F) Representative 5hmC dot blot analysis of hippocampal lysates from 2×Tg-AD mice infected with AAV-sh-Scramble and AAV-shTet2 (n = 3 per group). (G) Quantitative analysis showing that the level of 5hmC was significantly decreased by Tet2 knockdown in 2×Tg-AD mice hippocampus (n = 3 per group). Data are presented as mean ± SEM, unpaired t-test and ***P < 0.001.
Since deposition of Aβ plaques is one of the earliest neuropathological hallmarks in AD-transgenic mice that coincides with AD-related cognitive deficits (32). We next evaluated whether reducing Tet2 would accelerate progression of Aβ pathology in the hippocampus at the early stages of AD. Immunohistochemical staining showed that the reduction of Tet2 expression in the young AD hippocampus led to a significant increase in the number of Aβ-positive cells compared with the Scramble control 2×Tg-AD mice hippocampus (Fig. 3A). This result suggested that Tet2 deficiency may also accelerate additional changes in 2×Tg-AD mice that coincide with memory deficits, such as the initial synaptic loss that is closely related to cognitive deficits (33). The postsynaptic density 95 (PSD95) and synaptophysin (SYP) proteins are crucial for synaptic plasticity and memory function (34). Therefore, we measured the levels of PSD95 and SYP in hippocampus lysates of young WT and 2×Tg-AD mice which had been infected with AAV-sh-Scramble or AAV-shTet2 (Fig. 3B). Notably, western blot analysis showed that the loss of Tet2 led to a precipitous decrease in PSD95 and SYP protein expression in 2×Tg-AD mice compared with the control 2×Tg-AD mice (Fig. 3C and D). Collectively, these findings indicate that Tet2 reduction not only results in the deposition of aggregated Aβ proteins but also leads to the damage of hippocampal synaptic plasticity in the early stage of AD.
Knockdown of Tet2 significantly increases Aβ plaque levels while decreasing the levels of synaptic and NeuN protein in 2×Tg-AD mice. (A) Representative immunofluorescence staining images of GFP+ cells and 6E10+ Aβ plaques in the hippocampus of 24-week-old WT and 2×Tg-AD mice infected with AAV-sh-Scramble or AAV-shTet2 vectors. Tet2 deficiency significantly accelerated the accumulation of Aβ plaques in 2×Tg-AD mice. Scale bar: 250 μm. (B) Representative western blotting images showing the protein levels of PSD95 and SYP from the hippocampus (n = 4 per group). GAPDH was used as a loading control. (C) Quantitative analysis showing the protein levels of PSD95 significantly decreased in 2×Tg-AD mice compared to WT mice and in 2×Tg-AD mice injected with AAV-shTet2 compared to 2×Tg-AD mice injected with AAV-sh-Scramble (n = 4 per group). (D) Quantification analysis showing that the SYP protein level in the shTet2 groups was significantly decreased compared with that in the sh-Scramble control groups (n = 4 per group). (E) Representative immunoblots of NeuN protein in hippocampal lysates of WT and 2×Tg-AD mice normalized to GAPDH. (F) Quantitation analysis showing that loss of Tet2 significantly decreased hippocampal NeuN protein expression levels (n = 4 per group). Data are presented as mean ± SEM; unpaired t-test; and ***P < 0.001.
Tet2 depletion in the DG leads to nerve cells alterations in the early stages of AD progression in 2×Tg-AD mice
Characteristics of pathological injury in aged 2×Tg-AD mice include neuronal necrosis, apoptosis and proliferation of glial cells (35). Additionally, glial cell activation and the resulting inflammatory reaction are major causes of AD-related deterioration of neurological function (36). To examine whether decreasing Tet2 levels in the DG would affect nerve cells in young 2×Tg-AD mice, we evaluated the levels of the marker proteins Neuronal Nuclei (NeuN), GFAP and Iba1 for mature neurons, astrocytes and microglia, respectively. Immunostaining with anti-NeuN showed that the fluorescence signal intensity of NeuN-positive cells in 2×Tg-AD mice treated with AAV-shTet2 was weaker than 2×Tg-AD mice treated with AAV-sh-Scramble (Supplementary Material, Fig. S2B). Western blotting was performed to further confirm the findings from immunofluorescence staining. The immunoblotting results also showed decreased expression of NeuN protein in 24-week-old 2×Tg-AD mice treated with AAV-shTet2 compared to both WT and 2×Tg-AD mice treated with AAV-sh-Scramble (Fig. 3E and F), demonstrating that Tet2 suppression promoted the loss of mature neurons in young 2×Tg-AD mice.
Immunostaining analysis revealed that GFAP-positive (GFAP+) astrocytes in the hippocampus of 2×Tg-AD mice infected with AAV-shTet2 proliferated much more rapidly compared to the corresponding groups of WT and 2×Tg-AD mice infected with AAV-sh-Scramble at 6 months of age (Fig. 4A). Notably, immunoblotting showed that ablation of Tet2 significantly upregulated the expression of GFAP protein levels in the hippocampal lysates of 2× Tg-AD mice compared with control 2×Tg-AD mice (Fig. 4B and C). Furthermore, Iba1-positive (Iba1+) microglial cells were also evaluated utilizing immunohistochemistry, which showed that the reduction of Tet2 led to a significant increase of Iba1+microglias in the hippocampi of 2×Tg-AD mice (Fig. 5A and B and Supplementary Material, Fig. S3A and B). Taken together, these results suggest that the absence of Tet2 in the hippocampus altered nerve cell growth, accompanied by an elevated inflammatory response in the early stage of 2×Tg-AD mice.
Tet2 deficiency increases astrocyte numbers in the hippocampus of 2×Tg-AD mice. (A) Representative immunofluorescence staining images of GFP-positive and GFAP-positive cells in the hippocampus of 24-week-old WT and 2×Tg-AD mice after infection with AAV-sh-Scramble or AAV-shTet2 vectors. Scale bar: 200 μm. (B) Representative immunoblots of GFAP protein in hippocampal lysates of WT and 2×Tg-AD mice normalized to β-actin. (C) Quantitative analysis showing that GFAP protein levels was higher in 2×Tg-AD mice compared to WT and that loss of Tet2 significantly increased hippocampal GFAP protein levels in both WT and 2×Tg-AD mice (n = 4 per group). Data are presented as mean ± SEM; unpaired t-test; **P < 0.01; and ***P < 0.001.
Tet2 depletion results in an increase of the microglial marker Iba1 and increased pro-inflammatory cytokine accumulation in the hippocampus of 2×Tg-AD mice. (A) Representative images showing Iba1+ cells and GFP+ cells in the hippocampi of WT and 2×Tg-AD mice 1 month after the AAV-sh-Scramble and AAV-shTet2 injection. Scale bar: 250 μm. (B) Quantitative analysis demonstrating that Tet2 depletion significantly increases Iba1 levels in hippocampi of 2×Tg-AD mice (n = 10 per group). (C–E) The levels of IL-6 (C), IL-β (D) and TNF-α (E) were evaluated by ELISA in the infected hippocampi extracted from WT and 2×Tg-AD mice (n = 6 per group). Data are presented as mean ± SEM; unpaired t-test; *P < 0.05; **P < 0.01; and ***P < 0.001.
Reducing Tet2 in the DG exacerbates the accumulation of pro-inflammatory cytokines in the early stage of AD in 2×Tg-AD mice
Numerous studies have implicated neuroinflammation as a crucial event leading to neuronal damage during the pathogenesis of AD (37,38). Additionally, glial cell proliferation may further exacerbate the inflammatory damage associated with AD (39). Moreover, increased levels of pro-inflammatory cytokines including interleukin-6 (IL-6), interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) might cause neuroinflammation in mice. We therefore investigated whether Tet2 reduction in the hippocampus would exacerbate the accumulation of pro-inflammatory factors in the early stage of AD in 2×Tg-AD mice. As predicted, enzyme-linked immunosorbent assay (ELISA) analysis of IL-6, IL-1β and TNF-α all showed a significantly higher baseline levels in the hippocampal lysates of 2×Tg-AD mice than in the hippocampus of WT mice in the control AAV-sh-Scramble groups (P < 0.05, Fig. 5C–E). Interestingly, eliminating Tet2 significantly increased the production of IL-6 (P < 0.001, Fig. 5C), but not IL-1β or TNF-α (ns, Fig. 5D and E), in the hippocampal lysates of 2×Tg-AD mice compared with WT mice. However, the levels of IL-6 (P < 0.01), IL-1β (P < 0.05) and TNF-α (P < 0.001) were all significantly increased in 2×Tg-AD mice treated with AAV-shTet2 compared with 2×Tg-AD mice treated with control AAV-sh-Scramble (Fig. 5C–E). Similarly, the levels of IL-6 (P < 0.01), IL-1β (P < 0.001) and TNF-α (P < 0.001) were all significantly higher in WT mice after Tet2 knockdown than in WT mice treatment with control AAV-sh- Scramble (Fig. 5C–E). These results indicate that Tet2 depletion potentiates neuroinflammation characteristic of AD development even at an early stage in 2×Tg-AD mice.
OE-Tet2 in 2×Tg-AD aNSCs promotes neural regeneration in vitro
Neural stem cells have the potential to differentiate into neurons, astrocytes and oligodendrocytes as well as the ability to self-renew even in the adult brain (40). To investigate whether restoring Tet2 would be sufficient to ameliorate the regenerative decline of aNSCs when nerve cells are damaged, we analyzed neural stem cells from the aged 2×Tg-AD mice hippocampus utilizing an in vitro acute overexpression assay. WT and 2×Tg-AD mice aged 9 months were sacrificed, and the hippocampi were dissected to isolate the neural stem cells. Isolated neural stem cells in the proliferation stage were then transfected with control PCDNA or OE-Tet2 plasmids by electroporation. We then evaluated the differentiation capacity of aNSCs after treatment using immunocytochemistry for cell identification and quantification (Fig. 6A and B). This analysis showed that the number of β III-tubulin (Tuj1)-positive mature neurons was significantly decreased in 2×Tg-AD aNSCs compared to WT aNSCs from the PCDNA control groups (Fig. 6C), indicating that the decreased differentiation of aNSCs occurred within 9 months in 2×Tg-AD mice. Importantly, increasing Tet2 expression resulted in a significant increase in the number of Tuj1-positive neurons in 2×Tg-AD aNSCs compared with the control transfected 2×Tg-AD aNSCs, leading the number of Tuj1-positive neurons in 2×Tg-AD mice close to that of control WT mice (Fig. 6C). Moreover, overexpression of Tet2 also enhanced the differentiation ability of WT aNSCs compared with the control PCDNA-treated WT aNSCs (Fig. 6C). Quantitative qRT-PCR confirmed that overexpression plasmid could significantly increase Tet2 transcripts in both WT and 2×Tg-AD aNSCs (Fig. 6D). Furthermore, Tet2 overexpression increased global 5hmC levels in both WT and 2×Tg-AD aNSCs at the differentiated state (Fig. 6E and F). Together, these data strongly support the notion that restoring Tet2 expression not only increases 5hmC levels but also is sufficient to rescue the regenerative decline and promote neuronal rejuvenation in aNSCs from 2×Tg-AD mice.
Increasing Tet2 expression rescues neural regeneration in aNSCs from the hippocampus of aged 2×Tg-AD mice. (A) Differentiated aNSCs isolated from the hippocampi of 9-month-old WT and 2×Tg-AD mice were stained with anti-Tuj1 and anti-GFAP antibodies. DNA (nuclei) was stained with DAPI. Scale bar: 50 μm. (B) Representation IF staining in aNSCs from the hippocampi of aged WT and 2×Tg-AD mice which were transfected Tet2 overexpression plasmids by electroporation. Differentiated neurons were stained with the marker Tuj1 and astrocytes, the marker of GFAP. Scale bar: 50 μm. (C) Quantification of Tuj1+ cells indicating that overexpression of Tet2 results in a significant increase of Tuj1+ neurons in 2×Tg-AD aNSCs compared to 2×Tg-AD aNSCs transfected with PCDNA control plasmids (n = 10 per group). (D) Tet2 mRNA transcripts were detected in differentiated aNSCs. 18s mRNA levels were used as an internal control. Tet2 mRNA was very significantly increased in OE-Tet2 groups compared to PCDNA control groups (n = 3 per group). (E) Representative dot blot images using anti-5hmC antibody showing a decrease of 5hmC levels in aNSCs from aged 2×Tg-AD mice, whereas Tet2 overexpression results in a significant increase of 5hmC levels in aNSCs compared to PCDNA control groups (n = 3 per group). (F) Quantification of 5hmC dot blot intensities showing a significant decrease of 5hmC in aNSCs from 2×Tg-AD mice that is rescued by Tet2 overexpression (n = 3 per group). Data are presented as mean ± SEM; unpaired t-test; *P < 0.05; **P < 0.01; and ***P < 0.001.
Reduction of Tet2 in the DG severely exacerbates cognitive impairment in the early stage of AD in 2×Tg-AD mice
We next explored whether decreasing Tet2 levels in the DG caused the cognitive dysfunction in young 2×Tg-AD mice to progress more rapidly (Fig. 7A). The mice locomotor activity was first assessed by the open-field test (OFT). The results show that although there was no significant difference in baseline motor function between the WT and 2×Tg-AD mice at this age, the loss of Tet2 does lead to a significant decline in motor function that is significantly more pronounced in the 2×Tg-AD mice compared to WT mice (Supplementary Material, Fig. S4A and B). We then evaluated the impact of Tet2 loss on memory in the mice at 5 months of age utilizing the novel object recognition (NOR) tests. The discrimination index during the retention session was significantly lower in AAV-shTet2 groups than in the AAV-sh-Scramble groups, although there was no significant difference between WT and 2×Tg-AD mice. These results support the notion that reducing Tet2 is sufficient to cause impairments of recognition in mice (Supplementary Material, Fig. S4C and D).
Tet2 depletion rapidly impairs hippocampal-dependent spatial learning and memory of young 2×Tg-AD mice. (A) Diagram illustrating the experimental scheme to investigate the behavior of 20-week-old WT and 2×Tg-AD mice given bilateral stereotactic injections of AAV virus encoding shRNA targeting Tet2 (shTet2) Scramble or a Scramble control sequence (sh-Scramble) into the DG. The mice were then tested with the OFT, NOR test, MWM and radial arm maze (RAM) 4 weeks after virus administration as indicated. (B) Representative images showing the swim path of the MWM test of 24-week-old WT and 2×Tg-AD mice given bilateral stereotactic injection of AAV-sh-Scramble or shTet2 into the DG. WT, n = 9 per group; 2×Tg-AD, n = 9 per group. (C) During 4 days of navigation training, no significant difference was observed in the escape latency between AD and WT mice targeted with AAV-sh-Scramble. On the other hand, loss of Tet2 significantly increased the escape latency of both AD and WT mice. Furthermore, the 2×Tg-AD mice targeted with AAV-shTet2 had a significantly longer escape latency compared with that of the WT mice targeted with AAV-sh Tet2. (D) The swim path length of WT and 2×Tg-AD mice injected with either AAV-sh-Scramble or AAV-shTet2 in the orientation navigation experiment. (E) Quantitative comparison of escape latency in the fourth day of training session. (F) The number of platform crossings among every group in the spatial probe test. (G) The percentage of time each group spent in the target quadrant (NW) and the other quadrants. (H) The reference memory error ratios for each day were assessed for WT and 2×Tg-AD mice treated with AAV-sh-Scramble or AAV-shTet2. Quantification of the ratio of entry errors showed that the 2×Tg-AD mice targeted with AAV-shTet2 had higher overall error ratios than the 2×Tg-AD mice targeted with AAV-sh-Scramble. Inset: schematic illustration of the eight-arm maze test (above). WT, n = 6 per group; 2×Tg-AD, n = 6 per group. Data are presented as mean ± SEM.; unpaired t-test; *P < 0.05; **P < 0.01; ***P < 0.001.
The spatial learning and memory ability of the treated mice were further examined by the Morris water maze (MWM) test. At day 35 postsurgery, mice performed the training trial and probe trial. Figure 7B shows representative images of the swimming routes of WT and 2×Tg-AD mice that were injected with AAV-sh-Scramble or AAV-shTet2. In training trials for 4 days, WT and 2×Tg-AD mice treated with AAV-sh-Scramble did not show a significant difference in escape latency (Fig. 7C). However, both WT and 2×Tg-AD mice treated with AAV-shTet2 spent a significantly longer time to find the platform when compared to the WT and 2×Tg-AD mice treated with the AAV-sh-Scramble control (Fig. 7C). Moreover, both WT and 2×Tg-AD mice treated with the AAV-sh-Scramble displayed similar swimming paths throughout the 4-day training trials, while both the WT and 2×Tg-AD mice treated with AAV-shTet2 displayed a significantly longer swimming path than that of the Scramble control groups (Fig. 7D). Furthermore, on the last day of the training session, further analysis indicated that the 2×Tg-AD mice in the AAV-shTet2 group took a significantly longer time to reach the platform when compared with the 2×Tg-AD mice in the control group (Fig. 7E). On the fifth day, mice were allowed to swim for 90 min before removing the platform in the probe trial. Notably, the number of platform crossings was very significantly decreased in AAV-shTet2-treated WT and 2×Tg-AD mice compared to the AAV-sh Scramble-treated WT and 2×Tg-AD mice, whereas there was no difference in platform crossings between the WT and 2×Tg-AD mice in the Scramble control group (Fig. 7F). Moreover, both the WT and 2×Tg-AD mice in the AAV-sh-Scramble group displayed a clear preference searching in the target quadrant compared with both WT and 2×Tg-AD mice in the AAV-shTet2 group (Fig. 7G). Although the number of platform crossings appears to be lower in the 2×Tg-AD mice compared to WT mice in the Tet2 knockdown group, this difference was not statistically significant (Fig. 7F). Altogether these results suggest that Tet2 knockdown impairs learning and memory and exacerbates the spatial memory deficits in young 2×Tg-AD mice.
In addition, we further tested long-term hippocampal-dependent spatial learning and memory using the eight-arm radial maze (also known as radial maze, RAM). During the first 4 days of training, all mice displayed similar learning capacities. By contrast, 2×Tg-AD mice with reduced Tet2 displayed a significantly higher reference error ratio compared to Scramble control 2×Tg-AD mice during the short term (day 5). Meanwhile, from day 6 to day 10, the 2×Tg-AD mice with Tet2 knockdown showed significantly higher errors in the total entries to the target arms than the other groups (Fig. 7H). These results demonstrate that decreased Tet2 in the DG not only impairs short-term hippocampal-dependent spatial reference learning and memory but also damaged long-term cognitive competence in the early stage of AD.
Reduction of Tet2 in the DG alters gene expression in young 2×Tg-AD mice
To determine the effect of Tet2 on gene expression associated with AD-related pathology and cognitive dysfunction at early AD stages, we carried out RNA sequencing (RNA-seq) using RNA isolated from the hippocampus of 24-week-old WT and 2×Tg-AD mice treated with AAV-sh-Scramble or AAV-shTet2. Information on sequencing quality is shown in Supplementary Material, Table S1. A total of 1822 genes with differential changes were identified between Tet2 knockdown and control 2×Tg-AD mice, of which 1360 genes were upregulated and 462 genes were downregulated (Supplementary Material, Tables S2 and S3). Further analyses revealed that 534 unique genes were identified in the group of Tet2 knockdown and control 2×Tg-AD mice compared to the other three groups (Fig. 8A, Supplementary Material, Table S4). Heat map showed the expression profiles of homogeneity and difference genes in each sample (Fig. 8B). Hierarchical clustering analysis revealed distinct expression landscapes in the 2×Tg-AD mice treated with AAV-shTet2 compared to the other groups, indicating that Tet2-modulated genes played a critical role in the early AD development. MA plot was used to assess the expression abundance and changes of each gene compared from Tet2 knockdown and control 2×Tg-AD mice (Fig. 8C). Gene ontology (GO) analysis of DEGs may provide clues about the exacerbation of AD. We next performed the GO analysis of 1822 DEGs (termed as Tet2 modulated genes in the young 2×Tg-AD mice, Supplementary Material, Table S5) which were highly enriched in critical pathways, including immune effector process, cytokine production, cell activation and inflammatory response, among others (Fig. 8D), which are mainly involved in inflammation response. Further, 534 unique genes (termed as Tet2 modulated unique genes in the young 2×Tg-AD mice) were significantly enriched in pathways related to inflammation response (Fig. 8E). The above results of ELISA (Fig. 5C–E) also confirm this notion. To further explore the function of DEGs at the protein level, we compared the 534 DEGs with the STRING database and the analysis showing the interaction network of proteins related to the inflammation response (Fig. 8F). Together, these results suggest that Tet2 knockdown triggers neuroinflammation in the early stage of AD brain and then exacerbates the progression of AD.
Identification and characterization of altered gene expression caused by the loss of Tet2 in the hippocampus of young 2×Tg-AD mice. (A) Venn diagram illustrating the numbers of DEGs identified in the hippocampus of young WT and 2×Tg-AD mice treated with AAV-sh-Scramble or AAV-shTet2. 1822 DEGs were discovered between Tet2 knockdown and control 2×Tg-AD mice, of which 534 were unique genes. The FDR < 0.01 and |log2(FoldChange)| ≥ 2 were set as the threshold for significantly differential expression. WT, n = 3 per group; 2×Tg-AD, n = 3 per group (pooled tissues from three mice per sample). (B) Heat map showed hierarchical clustering of all DEGs in each sample. (C) MA map showed the DEGs between Tet2 knockdown and control 2×Tg-AD mice. In the figure, the green dots (462 genes) represented significantly downregulated genes, the red dots (1360 genes) represented significantly upregulated genes, and the black dots represented genes with no significant expression differences. (D) Gene ontology (GO) analysis showing altered expression of 1822 genes identified in the group of Tet2 knockdown and control 2×Tg-AD mice was enriched in pathways related to inflammatory response. (E) Further analysis showed 534 unique genes were also significantly involved in inflammation response pathways. (F) The interaction network showing 25 DEGs based on 534 unique genes related to inflammatory response, aging and cell overgrowth, in which 1 gene was downregulated (green) and 24 genes were upregulated (red).
Reduction of Tet2 in the DG increases HDAC1 levels in young 2×Tg-AD mice
Although the results from this study show a clear effect of Tet2 on the AD-related pathological hallmarks and accelerated cognitive dysfunction at early AD stages in 2×Tg-AD mice, it is not clear what mechanism is responsible for this phenomenon. Interestingly, among the proteins from hippocampal tissue homogenates that were measured, we observed that HDAC1 levels were significantly increased in both WT and 2×Tg-AD mice with Tet2 knockdown compared to control mice at both the protein and mRNA level (Fig. 9A–C). However, no significant difference in HDAC1 expression was observed in 2×Tg-AD mice compared to WT mice in both the AAV-sh-Scramble control and Tet2 knockdown groups (Fig. 9A–C), indicating that Tet2 suppresses HDAC1 expression in both groups of WT and 2×Tg-AD mice. To confirm HDAC1 directly interacts with Tet2 protein, we performed co-immunoprecipitation (co-IP) experiments followed by immunoblotting assays using hippocampal extracts with anti-HDAC1 and anti-Tet2 antibodies. We found that endogenous HDAC1 could be readily precipitated by Tet2 and vice versa (Fig. 9D), suggesting that HDAC1 directly interacts with Tet2 to regulate AD pathology.
The coordinative role of Tet2-HDAC1 was critical in modulating AD pathology. (A) Representative western blot images showing HDAC1 levels in the hippocampus of Tet2 knockdown and control WT and 2×Tg-AD mice. (B) Quantitative analysis indicates that Tet2 depletion increased HDAC1 protein levels in the hippocampus of both WT and 2×Tg-AD mice. n = 4 per group. (C) Reverse-transcription qPCR of HDAC1 mRNA from the hippocampus of WT and 2×Tg-AD mice treated with AAV-sh-Scramble or shTet2. 18s mRNA levels were used as an internal control. n = 3 per group. (D) Endogenous HDAC1 and Tet2 co-IP experiments followed by immunoblotting analyses were performed using hippocampal lysates of the 24-week-old WT mice. Endogenous HDAC1 interacted with Tet2 in hippocampus. (E) Schematic illustration of Tet2-mediated control of HDAC1 expression in the mouse hippocampus. The enzyme Tet2 catalyzes the production of 5hmC in the HDAC1 promoter region, which results to silencing of the expression of HDAC1. Unpaired t-test; *P < 0.05; ***P < 0.001.
Discussion
In this study we explored how the 5hmC epigenetic modification catalyzed by Tet2 altered the pathogenesis of AD. We found no significant changes in levels of 5hmC in the hippocampi of 6-month-old 2×Tg-AD mice, which is when small amounts of Aβ plaques begin to appear in the brain but is prior to the emergence of cognitive decline. This result is consistent with previous reports (17,33). On the other hand, the aged 2×Tg-AD mice (67-week-old) displayed an abundance of Aβ plaques and showed a significant decrease in 5hmC levels in the hippocampus. These results suggest that the plaques are deposited at an earlier AD stage prior to changes in 5hmC levels and that the gradual accumulation of Aβ plaques may accelerate the 5hmC decline in later stages. We then hypothesized that the converse may also be true that a reduction of 5hmC could also accelerate the accumulation of amyloid plaques. If this were true, then 5hmC profiling may also be a relevant biomarker for the extent of cognitive decline in 2×Tg-AD mice.
Previous studies have indicated that altered DNA demethylation occurs at specific regulatory regions involved in adult hippocampal neurogenesis during the pathological course of AD (41), and our prior published results showed that the overall level of 5hmC decreases in the hippocampus during the pathogenesis of AD (17). Moreover, we find that 5hmC status in the hippocampus of 2×Tg-AD mice is mainly controlled by Tet2. Tet2 can regulate the adult neurogenesis that occurs in specific regions of the adult mammalian brain (28,42). Aberrant adult neurogenesis has been linked to intellectual disabilities and the development of neurodegenerative disorders (43). Furthermore, adult hippocampal neurogenesis is impaired prior to the onset of AD pathology (41,44). Therefore, the current study primarily focused on adult hippocampal neurogenesis occurring in 2×Tg-AD mice and sought to establish a causal link between alterations of Tet2-mediated 5hmC levels and the pathophysiological symptoms of AD. To do so, we modulated 5hmC levels via suppression or induction of Tet2 and assayed the biochemical as well as neuropsychiatric and memory symptoms of AD at early disease stages.
Previous research has shown that Tet2 abrogation in the adult hippocampus results in cognitive impairments in the normal healthy brain (31). However, whether Tet2 plays a role in maintaining hippocampal organization and function under the pathological conditions of AD was unknown. To mimic the loss of Tet2 in the aged AD brain, young adult (5-month-old) 2×Tg-AD mice were given bilateral stereotaxic injections of AAV-shTet2 into the DG. A month after surgery, we evaluated whether elimination of Tet2 could promote the transition from early AD stages to advanced AD stages. First, we demonstrated that depletion of Tet2 can decrease global 5hmC levels. Interestingly, we also found that Tet2 depletion in the early AD hippocampus could exacerbate Alzheimer’s-type pathology consistent with advanced symptoms, including Aβ burden, neuron loss, overgrowth of GFAP+ astrocytes and Iba1+ microglial cells and decreased levels of synaptic proteins. Importantly, our behavioral assessments found that 6-month-old 2×Tg-AD mice did not show significantly different behavioral features or abnormal cognitive functions compared with age-matched WT groups. On the other hand, reduction of Tet2 in the hippocampi of 5-month-old 2×Tg-AD mice led to a striking decrease in cognitive function and reduced learning and memory capacity. Altogether, our data shows that Tet2 is a crucial regulator of cognition, learning and memory in both healthy and AD mice and its loss exacerbates the progression of AD.
Neuroinflammation acting as a trigger is one of the prime contributors to the initiation and progression of neurodegeneration in AD (36). Epigenetic mechanisms can regulate astrocyte function and their response to neuroinflammatory signals (36,45). Previous studies have shown that Tet2-deficient mice display a more severe inflammatory phenotype and have increased IL-6 production (46). To further investigate the physiological role of Tet2 in control of the inflammatory response to AD pathology in vivo, we used Tet2-knockdown 2×Tg-AD mice to assess the accumulation of pro-inflammatory markers, including TNF-α, IL-1β and IL-6. Our data demonstrated higher production of TNF-α, IL-1β and IL-6 in the Tet2- knockdown AD hippocampus than that of control AD hippocampus or Tet2- knockdown WT hippocampus, suggesting that 2×Tg-AD mice are more susceptible to the Tet2-induced inflammatory reaction. Hippocampal RNA-seq data analysis also revealed that Tet2-modulated genes in the early stage of 2×Tg-AD mice were highly enriched in pathways related to inflammatory response. These data indicate that Tet2 depletion could exacerbate inflammation in the presence of an unhealthy, early symptomatic AD brain. In addition, Tet2 depletion leads to a deteriorated microenvironment and more severe brain damage in the young 2×Tg-AD mice.
Recent mounting evidence indicates that a promising approach to broadly rejuvenate regenerative capacity during aging is to improve adult stem cell function (47,48). Epigenetic modifiers play a critical role in establishing and maintaining stem cell function (49,50). Moreover, emerging evidence indicates that reducing Tet2 in aNSCs impairs the neuronal differentiation process, whereas restoring Tet2 in the aging brain could rescue rejuvenation capacity (28,31). Given these findings, we investigated the possibility of nerve regeneration in the AD condition by increasing Tet2 expression in the aNSCs isolated from the hippocampus of aging (9-month-old) 2×Tg-AD mice. Excitedly, we found that boosting Tet2 expression was sufficient to rescue mature neurons (Tuj1 positive cells) that were lost in the aged 2×Tg-AD mice, paralleled by significantly enhanced 5hmC levels. Furthermore, it has been demonstrated that intervention factors which facilitate neurogenic rejuvenation concomitantly promote rejuvenation in peripheral tissues (51,52). Together, our data show that increasing Tet2-mediated 5hmC levels could restore nerve regeneration in the aged AD condition and may be therapeutically useful for reversing AD-related pathology and cognitive dysfunction in humans.
Importantly, understanding the mechanisms that drive nerve cells damage in AD and how to counteract them is a critical step for promoting tissue repair and maintenance. Prior studies have shown that Tet2 can directly interact with other epigenetic modifiers to positively regulate gene transcription (53,54). It has been found that HDAC1 interacts with Tet2, and knocking down HDAC1 increases endogenous Tet2 acetylation. Moreover, HDAC1 is a primary Tet2 deacetylase, while acetylation enhances Tet2 enzymatic activity (55). However, whether the epigenetic modifications of both DNA demethylation and histone deacetylation by Tet2 and HDAC1 coordinate together in the regulation of AD pathology has not been clearly determined. To address this, we performed western blotting and real-time PCR to evaluate the relationship between changes of HDAC1 and Tet2 levels in the AD hippocampus. Our results showed a significant increase of HDAC1 in both WT and 2×Tg-AD mice hippocampus upon Tet2 knockdown. However, we observed no major difference in HDAC1 expression between control and AD hippocampi. In addition, our present data revealed endogenous HDAC1 directly interacts with Tet2 using co-IP experiments followed by western blot analyses. Studies have shown that targeting Tet2 acetylation is protective against abnormal DNA methylation during oxidative stress (55). Inhibition of class I HDACs was shown to be a promising avenue for treating the cognitive deficits associated with early stage AD (56). Tet2 mediates active repression of IL-6 through HDACs to erase histone acetylation during inflammation resolution (46). Alterations in HDAC2 are potentially useful to assess the hallmarks of AD in 5×FAD mice (57). Moreover, HDAC inhibitors may be therapeutically useful for targeting neuroinflammation in brain injuries and neurodegenerative disease (58). Consistent with these studies, our results indicate a coordinative role of Tet2-HDAC1 in modulating 5hmC levels of the downstream genes, and ultimately regulating gene expression, which plays a pivotal role in AD-related pathological hallmarks and cognitive deficits (Fig. 9E).
In summary, our current work demonstrates Tet2-mediated modulation of the pathological process of AD. We demonstrate that targeting the Tet2-mediated 5hmC epigenetic modification can have a significant influence on AD progression. Moreover, increasing Tet2 accompanied by inhibition of HDAC1 expression in the early stages of AD pathogenesis may protect against neuronal cell death, decrease amyloid deposition, and ameliorate the inflammatory response. The combinatorial effect of DNA demethylation and histone acetylation may provide a novel therapeutic strategy for AD.
Materials and Methods
Animals
B6C3-Tg (APPswe/PSEN1dE9) double-transgenic mice (2×Tg-AD; B6C3F1 background) co-overexpress mutant forms of human amyloid precursor protein (APP) and presenilin 1 (PS1) under control of the mouse prion protein promoter. 2×Tg-AD mice were purchased from the Model Animal Research Center of Nanjing University (Certificate No. 201501556; license No. SCXK (Su) 2015-001) and were housed in the animal facility of Ningbo University Medical School (Ningbo, China). All animals were maintained in groups of six per cage in standard environmental conditions, including a 12-h light (07:00–19:00 h)/12-h dark (19:00–07:00 h) cycle with ad libitum access to water and food pellets at 22 ± 3°C and 60 ± 5% relative humidity. All animal experiments were performed according to the guidelines of the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of the Medical School of Ningbo University (license No.SYXK (Zhe) 2013-0191).
The littermates of WT and age-matched 2×Tg-AD male mice at the age of 5 months (young) and 9 months (old) were used in this study. Five-month-old mice were injected with AAV-sh-Scramble or AAV-shTet2. The mice were approximately 6 months old at the time of behavioral testing and biochemical and pathological analyses. All behavioral experiments were conducted during the light period (between 9:00 A.M. and 6:00 P.M.). WT and 2×Tg-AD untreated mice were sacrificed at 9 months of age, and their bilateral hippocampi were removed to isolate neural stem cells.
Construction and stereotactic injection of AAV-sh-Scramble and AAV-shTet2
AAV vectors were purchased from Obio Technology Corp., Ltd. (Shanghai, China). AAV vectors encoding small hairpin RNA (shRNA) targeting Tet2 or Scramble control sequences in tandem with an eGFP reporter were designed using the following procedures. The shRNA targeting mouse Tet2 and the Scramble shRNA were constructed in the original plasmid pShuttle-CMV (Cat#240007, Agilent, China). 5’-GGTTATCAGGCTTTTTACAAC-3′ was used for Tet2-shRNA, and 5’-CGCTGAGTACTTCGAAATGGC-3′ was used for Scrambled-shRNA. The Tet2 shRNA sequence and scramble shRNA sequence were integrated into the pAKD-CMV-bGlobin-eGFP-H1-shRNA vector for adenovirus packaging, and the AAV2/8 serotype was selected. These recombinant plasmids were further co-transfected into HEK293 cells with adenovirus packaging plasmids (such as pAAV helper and pAAV-Rep/Cap) for 3 days of infection. The medium and transfected cells were then harvested and centrifuged, followed by concentration and purification of the viral particles by CsCl density gradient centrifugation and ultrafiltration. Quantitative PCR was applied to detect the viral titer, which was above 1012 vg/ml.
Mice were anaesthetized with an intraperitoneal (i.p.) administration of 5% pentobarbital sodium (50 mg/kg) before mounting on a stereotaxic apparatus (RWD Life Science, China). AAV vectors containing an eGFP reporter (0.6 μl/side) were implanted into bilateral DG regions of the hippocampi using the following coordinates: −2.3 mm AP from bregma, ±1.50 mm ML from midline and −2.10 mm DV from pia mater. The injection speed was set at 0.2 μl/min, and the postinjection needle residence time was an additional 5 min. The mice were allowed to recover for 4 weeks prior to initiating the experiments.
aNSC culture and assays
Hippocampus-derived neural stem cells were isolated from the DG of WT and AD littermate aged mice (9 months old) as described previously (59). The proliferating aNSCs were maintained in DMEM/F-12 medium supplemented with 2% B27(without vitamin A), 2 mm L-glutamine, 2 ng/ml EGF, 2 ng/ml FGF and 100 U penicillin-streptomycin and grown in an incubator at 37°C with 5% CO2. For the differentiation assay, the proliferating aNSCs growing on coverslips were treated with DMEM/F-12 medium containing 1 mm retinoic acid (Sigma) and 5 mm forskolin (Sigma) for 48 h, which induced aNSCs to differentiate into neurons and astrocytes. After 48 h growth, the coverslips were washed in PBS three times followed by fixing with 4% paraformaldehyde (PFA) for 30 min. Then, the coverslips were washed with PBS for 15 min and blocked with blocking buffer (PBS containing 5% normal goat serum and 0.1% Triton X-100) for 1 h at room temperature (RT). And then the coverslips were incubated with primary antibodies at 4°C overnight, followed by immunofluorescence staining with the primary antibodies: mouse anti-Tuj1 (1:2000, Promega, Cat#G7121) and rabbit anti-GFAP (1:1000, DAKO, Cat#Z0334). On the following day, the secondary antibodies were used: goat anti-rabbit Alexa Fluor 488 (1:500, Invitrogen, Cat#A-11008) and goat anti-mouse Alexa Fluor 568 (1:500, Invitrogen, Cat# A-11031), and the coverslips were incubated for 1 h RT. DNA (nuclei) was stained with 4′6-diamidino-2-phenylindole (DAPI, 1:5000, Sigma, Cat#D9542) for 1 h RT. After final washes, coverslips were mounted onto glass slides with mounting medium (Vector Laboratories, CA). All immunostaining experiments were repeated at least three times, and representative images were photographed with a laser scanning confocal microscope (Olympus IX81-FV1000). The numbers of Tuj1 and GFAP positive cells were quantified using Image J software.
Electroporation
Cultured WT and AD aNSCs were electroporated with the mouse neural stem cell Nucleofector solution kit (Lonza, cat#S-05279) by an Amaxa electroporator (Lonza). In brief, the cells were pelleted at 2000 g for 2 min, and cell pellets were resuspended in 100 ml nucleofector solution containing 4ug of Tet2 overexpression or empty vector control PCDNA plasmids. After electroporation, cells were plated onto coverslips coated with poly-ornithine and laminin and cultured for 48 h.
Immunofluorescence staining
The mice were deeply anesthetized with pentobarbital sodium (50 mg/kg, i.p.) and transcardially perfused with 60 ml of saline, followed by perfusion with 40 ml of 4% PFA. Brains were gently removed and fixed in 4% PFA at 4°C for 3 days and then kept in a 30% sucrose solution for dehydration at 4°C until they sank. The coronal brain sections were cut with a thickness of 20 μm by a microtome cryostat (Leica). Brain sections were washed in PBS for 15 min and then were blocked with blocking buffer (PBS containing 5% normal goat serum and 0.1% Triton X-100) for 1 h RT. For 5hmC immunostaining, the sections were pretreated with 1 M HCl and incubated at 37°C for 30 min. The sections then were incubated with primary antibodies overnight at 4°C. The following primary antibodies were used: rabbit anti-5hmC (Active motif, 36 769, 1:500), mouse anti-Aβ (Millipore, MABN10, 1:1000), mouse anti-NeuN (Millipore, MAB377, 1:500), rabbit anti-GFAP (Dako, z0334, 1:500), goat anti-Iba1(Abcam, ab5076, 1:250) and mouse anti-Tuj1 (Promega, G7121, 1:300). On the next day, the brain sections were washed in PBS and incubated with secondary antibodies for 1 h at RT. Fluorophore-conjugated secondary antibodies were used: goat anti-rabbit Alexa Fluor 488 (Invitrogen; A11008, 1:500), goat anti-mouse Alexa Fluor 568 (Invitrogen, A11031, 1:500), goat anti-rabbit Alexa Fluor 568 (Invitrogen, A11036, 1:500), donkey anti-goat Alexa Fluor 568 (Invitrogen, A11057, 1:500) and goat anti-mouse Alexa Fluor 405 (Invitrogen, A31553, 1:500). DNA (nuclei) was stained with DAPI (Sigma-Aldrich, #B2261, 1:2000). The brain sections were mounted onto glass slides and coverslips with mounting medium (Vector Laboratories, CA). Immunostained sections were imaged using an Olympus confocal microscope (Olympus IX81-FV1000) and analyzed with Image J software.
Western blot
Hippocampus tissues were dissolved in ice-cold RIPA lysis buffer (Solarbio, cat#R0020) containing protease and phosphatase inhibitors (Sigma). Subsequently, hippocampal extracts were centrifuged at 20 817g for 20 min at 4°C, and then the supernatants were harvested and transferred into pre-cooled microfuge tubes. Protein concentrations of the samples were quantified using the BCA Protein Assay Kit (Beyotime, Beijing, China) and estimated by a BioPhotometer (Eppendorf). About 30 μg of protein from each sample was electrophoresed in 10% SDS-PAGE gels, followed by transferring to PVDF membranes (0.22 μm, Cat#v1620177, Bio-Rad, USA). After washing in TBST buffer (Tris-buffer saline and 0.1% Tween-20), the membranes were blocked with 5% non-fat milk in TBST for 1 h at RT and then incubated with primary antibodies in the blocking solution at 4°C overnight. The following primary antibodies were used: rabbit anti-Tet2 (GeneTex, GTX131099, 1:500), rabbit anti-PSD95 (Abcam, ab76115, 1:1000), rabbit anti-SYP (Abcam, ab32127, 1:250), mouse anti-NeuN (Millipore, MAB377, 1:1000), mouse anti-GFAP (Cell Signaling Technology, 3670S, 1:1000), rabbit anti-HDAC1 (Abcam, ab213701, 1:200), rabbit anti-beta Actin (Abcam, ab8227, 1:10000) and mouse anti-GAPDH (Ambion, AM4300, 1:0000). Then, after washing three times with TBST the following day, the membranes were incubated with secondary antibody (anti-rabbit/mouse IgG conjugated, 1:0000) for 1 h at RT. The specific bands were coated with Pierce ECL Western Blotting Substrate (Cat#190115-32, Advansta, USA) and then visualized with a Molecular Imager Imaging System (Tanon, China). The intensity of images was quantified with Image Studio Software (LI-COR Biosciences, Lincoln, NE).
Dot blot
Briefly, the hippocampus samples were dissected and homogenized in lysis buffer (100 mm Tris–HCl, 200 mm NaCl, 5 mm EDTA, 0.2% SDS, pH 8.5) containing proteinase K and then digested overnight at 56°C. Residual RNA contamination was completely removed by RNase treatment at 37°C for 1 h. Afterward, genomic DNA was extracted with an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1, P-3803, Sigma) as well as isopropanol precipitation. Subsequently, DNA samples were spotted onto the Amersham Hybond-N+ membrane (GE Healthcare) using a Bio-Dot Apparatus (Bio-Rad), followed by drying at 80°C for 45 min. The membrane was blocked in TBS (50 mm Tris, 150 mm NaCl) with 5% milk and 0.1% Triton for 1 h at RT and incubated with rabbit polyclonal 5hmC antibody (Active Motif, #39769) at a dilution ratio of 1:5000 at 4°C overnight. Then, after sufficient washing three times the following day, the blot was incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at RT with gentle shaking. Chemiluminescent detection imaging was then performed using a Molecular Imager Imaging System (Tanon, China). The density of each blot signal was analyzed by Image J software.
Quantitative real-time PCR
Total RNA was isolated from fresh mouse hippocampus with the TRIzol Reagent (Invitrogen, cat#15596026). Hippocampus tissues were dissected and homogenized, added into 200 μl chloroform, mixed completely, and centrifuged at 20 817g for 10 min at 4°C. Supernatants were transferred into pre-cooled microfuge tubes, and then 600 μl isopropanol was added to precipitate RNA. RNA pellets were washed with 70% ethanol and then dissolved in RNase-free water. Total RNA was treated with DNase I to ensure that the residual genomic DNA was degraded. The concentration of RNA was measured with a NanoDrop 2000 Spectrophotometer. 1 μg of total RNA was used as a template for reverse transcription according to the manufacturer’s instructions (Invitrogen). About 1 ml of cDNA was generated using the Power SYBR Green PCR Master Mix (Invitrogen). 18s rRNA was used as an endogenous control for all samples. All real-time PCR reactions were run on an Mx3005P Instrument from Agilent Technologies (Santa Clara, CA) in triplicate, and the data were analyzed using the ΔΔCt method. All primer sequences were listed in Supplementary Material, Table S6.
Enzyme-linked immunosorbent assay
Fresh hippocampus tissues were homogenized in ice-cold lysis buffer (150 mm sodium chloride; 1.0% Triton X-100; 0.5% sodium deoxycholate; 0.1% SDS; 50 mm Tris; pH 8.0) containing 1X protease inhibitor cocktail (Sigma). After sonication, hippocampal extracts were centrifuged at 2000g at 4°C for 20 min. Supernatants were collected, and protein concentrations were subsequently measured by a spectrophotometer (Thermo Scientific) at 450 nm with the BCA protein assay kit (Cat# CW0014S, Beyotime, China). The concentrations of IL-6, IL-1β and TNF-α were measured by ELISA kits (Cat# EK206/3, Cat# EK201B/3, Cat# EK282/3, MultiSciences Biotech, China) according to the manufacturer’s protocol.
Briefly, 300 μl washing buffer (PBS with 1% Tween-20) was added per well and allowed to sit in the wells for about 30 seconds before aspiration. Then, 100 μl twofold diluted standards in duplicate and 100 μl diluted standard to blank wells in duplicate were added. About 90 μl of assay buffer (PBS with 0.5% Tween-20 and 5% BSA) and 10 μl sample were added to the sample wells, followed by adding 50 μl of diluted detection antibody to each well. After incubating at RT for 2 h on a microplate shaker set at 446 g, each well was aspirated and washed six times. Then, 100 μl diluted streptavidin-HRP was added to each well and incubated at RT for 45 min. After washing, 100 μl substrate solution was added to each well and incubated for 30 min at RT. Finally, after adding 100 μl of stop solution into each well, the absorbance at 450 nm was measured by a spectrophotometer within 30 min. The concentrations of IL-6, IL-1β and TNF-α were calculated according to the standard curve and presented as pg/mg.
Open-field test
The experimental room for the OFT was a white plexiglass box with dimensions of 50 × 50 × 39 cm. The bottom surface was divided into four identical squares (each of which was 25 cm × 25 cm) with black lines. The mice were individually placed at the center of the field and allowed to explore for 5 min. The video system was used to record changes in the behavior of the mice for 5 min. After the experiments were completed, the number of line crossings (three or more claws entering into the neighboring square) and rearings (two front claws vacate or climb the walls) were counted blindly by an independent researcher. The data were used to evaluate the exploratory behavior and locomotor activity, respectively. The box was cleaned with 10% alcohol to remove residual odor interference between each test.
NOR tests
The experiment was performed in an open-field arena (30 × 30 × 30 cm), which was divided into four quadrants of equal area. On the first day, the animals were allowed to move freely to familiarize themselves with the environment in the open box for 5 min. On the first day of the training session, two identical objects (black square, 5 × 5 × 5 cm) were located in diagonal positions of the box. Then the experimental mice were gently placed at the center of the box and allowed to explore them for 5 min. On the following day for the recognition session, one of the familiar objects was replaced by a novel object with a completely different color and shape (yellow triangle, 5 × 5 × 7 cm), and the mice were placed back to re-explore the familiar and novel objects for 5 min. The animal’s exploration learning time for objects were defined as the time when the animal’s nose sniffed and/or the forepaws touched the object at a distance of less than 2 cm. Sitting on or turning around the objects was not counted as exploratory behavior. The exploration time was calculated as the identification index. The apparatus was decontaminated with 10% ethanol after each test.
MWM test
The MWM was performed in a round tank painted black (outer diameter 120 cm, depth 50 cm) which was filled with water consistently maintained at 22°C. Titanium dioxide was added to the water to turn it milky white. The platform with a diameter of 10 cm was positioned in the target quadrant (such as NW), which was hidden at 1 cm below the water surface. The wall around the pool was affixed with different shapes of color cards as a reference mark for the mice to identify its position. Swimming paths were captured by a video camera linked to a computer-based image system. Spatial memory was assessed by orientation navigation tests and space exploration tests. On the day before the training, the mice were allowed to swim freely in a pool without hidden platforms to adapt to the environment and water temperature. During the training, a pool was divided into four quadrants of equal area (such as NE, SE, SW, NW), and the mice were randomly placed into the four quadrants each day. If the mice reached the hidden platform within 90 s and then stood on the platform for 3 s, then the time spent on finding the platform was defined as escape latency. If the mice did not reach the platform within 90 s, they were then manually guided to it, and the escape latency was counted as 90 s. In addition, the platform was removed for the space exploration tests on the last day. The trained mice were allowed to swim for 90 s. The number of platform crossings and the time spent in each of the four quadrants were calculated. The water was drained, and the pool was cleaned with 10% ethanol to avoid scent clues between each test. All trials were video-recorded and evaluated with MazeScan software (Actimetrics, China).
Eight-arm RAM test
The equipment was radially distributed about a central area. The different shapes and colors of cards were affixed in the test room as a reference. A video tracking system (Med Associates Inc) was used to simultaneously monitor the mice’s trajectory. The mice were fed less to reduce their normal body weight by 20%. On the first day, an artificial bait (2–3 g chocolate pellets) was scattered in the end of all arms, and then each mouse was placed on the center platform and allowed to explore freely to eat the food rewards for 10 min. The training trial was started on the following day. During the training period, each mouse conducted two trials daily for 4 days. The same four arms were baited; thus mice learned to ignore the empty arm that never contained a reward. Mice were allowed to explore freely until all chocolate pellets were received or until 10 min had elapsed. Entries into non-baited arms with all four paws were defined as reference memory errors, and the total number of arms entered was counted. After each mouse experiment, the maze was cleaned with 10% alcohol to minimize olfactory cues.
RNA sequencing
Total RNAs were extracted from AAV-treated WT and 2×Tg-AD mice hippocampal tissues with the TRIzol Reagent (Invitrogen, cat#15596026) for RNA-seq. The RAN concentration was quantified using the NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE). The integrity of the RAN samples was assessed by the RNA Nano 6000 Asssay Kit of the Agilent Bioanalyzer 2100 System (Agilent Technologies, CA, USA). Only samples with the RNA integrity value (RIN) of ≥8 and OD260:280 ratio of ≥1.9 were deemed qualified. Firstly, the Ribo-Zero™ removal kit (Epicentre, Madison, WI, USA) was used to remove rRNA from a total amount of 1.5 μg RNA per sample. Briefly, the extracted mRNA was fragmented into small pieces. cDNA library construction was generated using NEBNextR Ultra™ Directional RNA Library Prep Kit for IlluminaR (cat#RS-100-0801, NEB, USA) as previously described (26). And afterward, these cDNA fragments were ligated with NEBNext Adaptor with hairpin loop structure. The final cDNA libraries were enriched with PCR. Subsequently, PCR products were sequenced on the Illumina HiSeq2000 machines (Illumina, San Diego, USA). Raw sequencing output was filtered, and the clean reads were mapped to the Mus musculus reference genome sequence using HISAT2 soft. The read counts for each sequenced library were adjusted by edgeR program package through one scaling normalized factor. The resulting false discovery rate (FDR) were adjusted using the posterior probability of being DE (PPDE). The FDR < 0.01 and |log2(FoldChange)| ≥ 2 were set as the threshold for significantly differential expression.
Co-immunoprecipitation
Hippocampus samples were dissected and lysed in ice-cold RIPA lysis buffer (Solarbio, cat#R0020) with protease inhibitor cocktail (Roche). After sonication, hippocampal extracts were centrifuged at 20 817g for 25 min at 4°C. 40 ul of the supernatants were used as input, and the remaining supernatants were then incubated either anti-Tet2 (Proteintech, cat#21207-1-AP, 10ug) or anti-HDAC1 (Abcam, ab7028, 3ug) overnight at 4°C. The anti-IgG (ProteinFind, HS201-01, 3ug) was used as control IgG. Then, on the following day, after sufficiently prewashing 100 ul protein A/G magnetic beads (Bimake, cat#B23202) three times, protein complex was incubated with beads for another 2 h at 4°C. And then protein complex-containing beads were washed with RIPA lysis buffer four times. Afterward, the immunoprecipitates were eluted and boiled in 1% SDS loading buffers at 98°C for 5 min and subjected to SDS-PAGE analysis. Immunoblotting assays were carried out to detect the target proteins as indicated with primary antibodies: rabbit anti-HDAC1 (1:200 for western, Abcam, ab213701) and rabbit anti-Tet2 (1:500 for western, GeneTex, GTX131099). These target proteins that appeared in HDAC1-group or Tet2-group but not in IgG-group were considered.
Statistical analysis
All data were presented as mean ± s.e.m. and were analyzed with GraphPad Prism software (GraphPad Software, La Jolla, CA). Comparisons of means between two groups were conducted with a two-tailed unpaired Student’s t-test. Means for multiple group comparisons were analyzed with two-way ANOVA, followed by Tukey’s post hoc test. P value <0.05 was considered statistically: *P < 0.05; **P < 0.01; ***P < 0.001. Sample sizes and quantity collection were provided in each figure legend.
Conflict of Interest statement
The authors declare no competing financial interests.
Funding
National Natural Science Foundation of China (U1503223 to Q.W.); National Key Research and Development Project (2016YFC1305900 to Q.W.); Natural Science Foundation of Zhejiang Province (LQ19H090005 to L.P.); Natural Science Foundation of Ningbo (2018A610305 to L.P.); Scientific research fund project of Ningbo University (XYL20030 to L.P.); Student Research, Innovation Program (SRIP) of Ningbo University and the K.C. Wong Magna Fund in Ningbo University.
Acknowledgements
We would like to thank Xuekun Li for assistance with figure generation and providing useful comments. We would like to thank Xiaofeng Jin for technical assistance and Wei Cui for revising the manuscript and statistical analyses.
