Introduction
Epigenetic regulations are functional chromatin alterations that are not caused by changes in nucleotide sequences [
1]. As one of the key epigenetic modifications, mammalian genomic DNA methylation is nearly universal in eukaryotic genomes. Hypomethylation and hypermethylation result in increased or decreased expression of genes, respectively. In mammals, epigenetic modifications of DNA predominantly involve the addition of a methyl group to carbon five of the cytosine base (C), subsequently generating 5-methylcytosine (5mC), which is catalyzed by DNA methyltransferases (DNMTs) [
2]. DNA methylation has been considered irreversible until the discovery of the structure and function of the demethylase TET (ten-eleven translocation) protein which renders DNA methylation a dynamic, reversible, and actively regulated process. TET oxidizes 5mC to 5-hydroxymethylcytosine (5hmC) [
3]. Sequentially, TET enzymes hydroxylate 5hmC to produce 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). Thymine DNA Glycosylase (TDG) recognizes the intermediary bases 5fC and 5caC and disrupts the glycosidic bond, leading to an apyrimidinic site. After that, base-excision repair (BER) enzymes subsequently repair the apyrimidinic site and T:G mismatches to generate cytosine [
4]. TET belongs to the family of Fe (II)—and α-ketoglutarate (α-KG)-dependent dioxygenases. Different TET isoforms are expressed in various cells and organs. TET1 expression is present in embryonic stem cells, and TET3 is mainly enriched in oocytes, whereas TET2 is widely distributed temporally and spatially [
5].
TET activity, as measured by the catalytic product of the TET enzyme, 5hmC, is significantly downregulated across different conditions, such as during the development of stem cells [
6], deprivation of oxygen in hypoxic cells [
7], 2-hydroxyglutarate (2-HG) treatment which acts as an α-KG antagonist by binding to the same space in the catalytic site and competitively inhibits the activity of TET proteins [
8]. All above indicate a potentially fundamental role of 5hmC and cytosine epigenetic modification in various physiological and pathological conditions.
Meisel reported that part of the intestinal barrier of Tet2-KO mice was damaged as evidenced by decreased expression of a tight junction protein—zonula occludens 1 (ZO-1) by jejunal epithelial cells and increased permeability of the intestinal barrier, which results in lactobacillus residing in the small intestine entering the blood and the spleen. Ectopic growth of lactobacillus induces the expression of interleukin-6 (IL-6) which stimulates myeloid cell proliferation and subsequent preleukemic myeloproliferation [
9].
The blood–brain barrier (BBB) is another critical barrier of the human body, whose structure is very similar to that of the intestine, such as the tight junction proteins of claudins and ZO-1 [
10]. These tight junction proteins, along with the immune barrier composed of phagocytes and other inflammatory cells, form a potential physical barrier confining the transgression of pathogens, diffusion of solutes in the blood, and the entry of macromolecules or hydrophilic molecules. This highly selective semi-permeable barrier maintains the homeostasis of the internal environment [
11]. Based on the similarity between these two barriers, we speculated that inhibition of TET2 activity of cerebrovascular endothelial cells might decrease the expression of ZO-1 in the brain, leading to the structural damage of BBB.
The accumulation of oxidative stress participants in aging induces BBB damage [
12]. Oxidative stress has been reported to influence TET hydroxylase activity [
13]. It is, therefore, reasonable to hypothesize that oxidative stress during aging induces the decrease in TET2 activities of cerebrovascular endothelial cells, leading to decreased expression of tight junction proteins, BBB destruction, and the potential occurrence of AD. The present study aimed to investigate the developmental changes of 5hmC and ZO-1 expression of endothelial cells in aged human brains and to explore evidence of direct regulation of ZO-1 expression by TET2 in Tet2-knockout mice and cultured endothelial cells.
Methods
Antibodies and reagents
Antibodies against ZO-1 (Abcam, ab190085; dilution 1:1,000), 5hmC (Active Motif, #39769, dilution 1:1,000), GFAP (Abcam, ab7260, dilution 1:1000), NeuN (Abcam, ab177487, dilution 1:1000), CD68 (Abcam, ab237988, 1:1000) and Claudin-5 (Abcam, ab131259, dilution 1:500) were used for immunohistochemistry staining. Antibody against 5hmC (Active Motif, #39769, dilution 1:5,000) was used for dot-blotting. Antibody against ZO-1 (Proteintech, 22601-1-AP, dilution 1:1000) and TET2 (Proteintech, 21207-1-AP, dilution 1:1000) were used for western-blotting.
Hydrogen peroxide solution (H2O2) (Sigma, 323381), N-acetyl cysteine (NAC) (Sigma, A7250), and DAPI (Sigma, D9542) were commercially obtained.
Fluorescein isothiocyanate (FITC)-conjugated dextran (40 kDa) and FITC-conjugated dextran 10 kDa were purchased from Thermo Fisher Scientific Inc. (Waltham, MA USA).
Cell culture and treatment
Human umbilical vein endothelial cells (HUVECs), human brain microvascular endothelial cells (HBMECs), and immortalised mouse brain endothelial bEnd.3 cells were obtained from the American Type Culture Collection (ATCC). HUVECs, HBMECs, and bEnd.3 cells were grown in the endothelial cell medium (ECM) (Cat. No. 1001, ScienCell, USA) containing 5% FBS (Cat. No. 0025, ScienCell, USA), 1% endothelial cell growth supplement (ECGS, ScienCell, US), and 1% antibiotic solution (P/S, ScienCell, USA). Following 12-h serum deprivation synchronization, HUVECs and bEnd.3 cells were treated with H2O2 and NAC supplement. Cells were plated in complete culture media with 10 μM H2O2 for 6 h, then exchanged the medium supplemented with 1 mM NAC for another 6 h.
For TET2 siRNA transfection, TET2 siRNAs were commercially acquired from Hanbio Ltd, Shanghai, China (sequences were list in Table
2). HUVECs and HBMECs were seeded in 6-well plates and transfected with control siRNA or siRNA for TET2 using riboFECT™ CP Transfection Kit (Hanbio Ltd, Shanghai, China) according to the manufacturer’s instructions after 70–80% cell confluence was reached. The final concentration of siRNA solution was 20 μM. MISSION
® siRNA Universal Negative Control (Sigma-Aldrich) was used as control.
Animals
Tet2-floxed mice and EIIa-Cre mice were kindly provided by Dr. Ye Dan (Fudan University), which were described previously [
14]. Quantitative PCR confirmed that the Tet2 mRNA level was reduced by 90% or more in the tail tissue isolated from Tet2-KO mice. Littermate control mice were used as controls. These mice were housed in the laboratory animal centre of Fudan University. All experiments were approved by the Laboratory Animal Management and Ethics Committee of Fudan University by complying with the Chinese Guidelines on the Care and Use of Laboratory Animals (20180302-051).
Human brain tissue
Post-mortem human archival brain tissue was collected by the Department of Pathology, School of Basic Medical Sciences, Fudan University. All procedures associated with the acquisition and usage of the human brain samples were approved by the Ethics Committee of Fudan University (2019-Y004). These procedures were conducted by complying with the Helsinki Declaration (1964, amended most recently in 2013) [
15]. Clinical histories and autopsy records were reviewed, and samples from patients with cardiovascular disorders were excluded. Brain samples with known hypoxic damage were excluded. Brain samples were evaluated by two neuropathologists as no microscopic abnormality, except brains of old people with mild age-related changes. The average post-mortem interval was 40.13 ± 15.59 h. Formalin-fixed and paraffin-embedded tissue was obtained from the temporal cortex (Brodmann area 21/22). Eight cases were in the adult group with their age less than 60, and 8 cases in the aged group. Case details were provided in Table
1.
Table 1
Demographics of Human brain donors
Aged | 66 | Male | Lobular pneumonia | 40 | None |
98 | Female | Lobular pneumonia | 36 | None |
66 | Male | Pulmonary thromboembolism | 72 | None |
90 | Female | Multiple organ dysfunction syndrome | 36 | None |
83 | Male | Lobular pneumonia | 48 | None |
88 | Female | Cachexia | 24 | None |
74 | Female | Lobular pneumonia | 32 | Mild edema |
89 | Female | Cachexia | 36 | none |
Adult | 29 | Female | Septic shock | 56 | Mild edema |
28 | Male | Acute complete intestinal obstruction | 70 | Mild edema |
40 | Female | pulmonary thromboembolism | 50 | None |
52 | Male | Lobular pneumonia | 38 | None |
54 | Female | Respiratory tract obstruction | 20 | None |
| 52 | Male | Heart failure | 36 | None |
| 54 | Female | Asphyxia | 24 | None |
| 58 | Female | Cachexia | 24 | None |
5hmC dot-blotting
The procedure for the dot-blotting assay was modified from the procedure described previously [
13]. Briefly, genomic DNA was spotted on a nitrocellulose membrane (Whatman), which was placed under an ultraviolet lamp for 20 min to crosslink the DNA. Subsequently, the membrane was blocked with 5% milk in TBST for 1 h, followed by the incubation with the primary anti-5hmC antibody at 4 °C overnight. After incubation with horseradish peroxidase-conjugated anti-rabbit IgG (GenScript) for 1 h at room temperature, the membrane was washed with TBST three times and then scanned with a Typhoon scanner (GE Healthcare).
Immunofluorescence and microscopy
Cells were plated at 1 × 106 per well on glass coverslips in 6-well culture plates and stimulated as described above. Cells were incubated overnight at 4 °C with a rabbit anti-ZO-1 antibody (1:1000; Chemicon). The primary antibody was detected with an isotype-specific anti-rabbit secondary antibody (1:1000) conjugated with Alexa 488 (Invitrogen). Negative control samples were performed without the primary antibody. 4′,6-Diamidino-2-phenylindole (DAPI, Invitrogen) was used for counterstaining.
Fluorescent images were captured under an LSM 510META confocal microscope (Carl Zeiss).
Western blot analysis
For western blot analysis, proteins were extracted by the radioimmunoprecipitation assay buffer (RIPA) containing complete protease and phosphatase inhibitor cocktails (Selleck, Houston, TX, USA). Protein concentrations were determined using the BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA). The blots were incubated with primary antibodies followed by the incubation with HRP-conjugated secondary antibodies (Abcam, Cambridge, UK). The protein–antibody complexes were visualised using an enhanced chemiluminescence kit (Thermo Fisher Scientific, Waltham, MA, USA).
hMeDIP-qPCR analyses
Genomic DNA from cells was prepared using a DNA isolation kit (QIAGEN). The hMeDIP assay was performed as previously described [
16]. Briefly, genomic DNA was sonicated to 300–500 bp fragments, denatured, then immunoprecipitated with an anti-5hmC antibody (Active Motif) or IgG control antibody (Active Motif) and protein A magnetic Dynabeads (Millipore). After washing three times, beads were treated with proteinase K for 3 h. Chip’d DNA was purified with a QIAquick PCR purification kit (Qiagen) by following the instructions from the manufacturer. 2 μg of each genomic DNA was used to run quantitative Real-time PCR (qPCR). qPCR amplification was performed using the SYBR Green system (Takara, Dalian, China) in an Applied Biosystems thermal cycler with StepOnePlus™ RT‐PCR System (ThermoFisher, MA, USA). Relative quantification of target genes was conducted using the 2 − ΔΔCT method. Isotype-specific IgG was used as a control. All primer sequences for qRT-PCR were listed in Table
2.
Table 2
primers of Tet 2 siRNA and hMeDIP-qPCR
Tet2 siRNA-1 | F: 5′-CCCAGAGUCCUAAUCCAUCUATT-3′ R: 5′-UAGAUGGAUUAGGACUCUGGGTT-3′ |
Tet2 siRNA-2 | F: 5′-AAGCUAGCGUCUGGUGAAGAATT-3′ R: 5′-UUCUUCACCAGACGCCUAGCUUTT-3′ |
Tet2 siRNA-3 | F: 5′-CAAGGCAGUGCUAAUGCGUAATT-3′ R: 5′-UUAGGCAUUAGCACUGCCUUGTT-3′ |
SiRNA NC | F: 5′-UUCUCCGAACGUGUCACGUTT-3′ R: 5′-ACGUGACACGUUCGGAGAATT-3′ |
Homo Tjp1-CpG1- Homo Tjp1-CpG- | F: 5′-ATAAGTAATACAGCATGAAA-3′ R: 5′-TGCTGGCCGGGAGCTGCTGT-3′ |
Homo Tjp1-CpG2 | F: 5′-AGCTATGCACCTGCCCAGTA-3′ R: 5′-TCCCTCGCTGGCGGCTGTCT-3′ |
Homo Tjp1-CpG3 | F: 5′-CAAATAAACATCTCCCGAGA-3′ R: 5′-AGTGGGTGGCCGCGCTCGGA-3′ |
Homo Tjp1-CpG4 | F: 5′-CCTAGAGACCTGTGACATCT-3′ R: 5′-AACAAGAATTTCAGCAATCTT-3′ |
Immunohistochemistry staining (IHC)
For IHC analysis, the labelled streptavidin–biotin (LSAB) method was applied using commercial kits (Dako Corporation, Santa Barbara, CA, USA). Briefly, paraffin embedded sections were deparaffinised and rehydrated following the standard protocol. The sections were then transferred to the antigen retrieval solution (Tris–EDTA pH: 8.0) for incubation for 10 min, followed by incubation with 3% H2O2 to eliminate the endogenous peroxidase activity, and blocked with 5% normal goat serum. After treatment with 2 N HCl for 15 min at room temperature, sections were neutralised with 100 mM Tris–HCl (pH 8.5) for 10 min and then washed three times with PBS, followed by incubation with primary anti-5hmC antibody at 37 °C for 1 h. Horseradish peroxidase (HRP)-conjugated secondary antibody (Dako Corporation) was then applied and incubated at 37 °C for 1 h. For ZO-1, CD68, NeuN, and GFAP IHC staining, the LSAB method was used as described above, omitting the treatment with 2 N HCl and the ensuing Tris–HCl neutralization.
To semi-quantify the IHC staining of ZO-1, GFAP, CD68, and NeuN, five fields (173 µm2 each) from each sample were randomly selected and microscopically examined. Images were captured using a charge-coupled device (CCD) camera and analyzed using the IMT i-Solution software (IMT i-Solution, Inc., Burnaby, BC, Canada) to get the staining intensity. All IHC image analysis values were calculated as mean ± SEM.
To quantify the endothelial 5hmC intensity, immunoreaction scores (IRS) of 5hmC on endothelial cells were calculated based on the staining intensity (SP) and the positive staining percentage (SI) of the cells. SI was scored as follows: 0: 0–1%; 1: 2–25%; 2: 26–50%; 3: 51–75%; and 4: 75–100%. SP was subjectively scored as follows: 0, no staining; 1, weak but definite staining; 2, moderate staining; and 3, intense staining. The IRS was calculated as IRS = SP × SI. The total score was 12, and the range was from 0 to 12.
Measurement of the permeability of the endothelial monolayer
To evaluate the paracellular permeability, HUVECs and HBMECs cells were seeded in the upper chamber of 12-well Transwell inserts (3460, CORNING, New York, NY, USA) to form a confluent monolayer. After washing these cells with the blank medium, FITC-Dextran (1 mg/ml, Sigma, St. Louis, MO, USA) was added to the upper chamber. After incubating at 37 °C for 6 h, the fluorescence intensity of FITC-Dextran transgressed to the lower chamber was measured by using a multimode microplate reader (Synergy H1, Biotek, Inc., Suite F Layton, UT, USA) at the excitation wavelength of 485 nm and emission at 530 nm. The permeability coefficient was calculated using the following equation: Pdextran = (RFUlower/RFUupper) (V) (1/time) (1/area) and normalized to control untreated cultures.
Dextran of variable sizes coupled with FITC was used to study BBB permeability. Mice received 20 mg/ml of FITC-dextran (10 kDa or 40 kDa) dissolved in isotonic saline at a dose of 0.4 ml/kg. Five minutes after injection, mice were anesthetized with isoflurane and sacrificed. The brains were isolated and fixed with 4% PFA for 6 h, then incubated in 10% glycerol for 18–24 h at 4 °C. After being embedded in OCT and snap frozen, brains were cut into 40 μm thick coronal sections. Fluorescent images were captured under an LSM 510 META confocal microscope (Carl Zeiss, Thornwood, NY, USA).
Statistical analysis
GraphPad Prism 8 was used to assess the significance of all data. One-way ANOVA followed by the post-hoc test (e.g., Dunnett’s Test) was used to assess parameters between multiple groups, and an unpaired two-tailed Student’s t-test for comparisons between two groups. Statistical significance was set at p < 0.05.
Discussion
The present study examined the impact of Tet2 activity on the expression of ZO-1, a BBB tight junction protein, and the underlying mechanism. We found that the level of 5hmC decreased in brain microvascular endothelial cells of aged mice and it was even weaker in Tet2 knockout mice. Similar decrease was observed in the expression of ZO-1. In cultured endothelial cells, Tet2 knock-down significantly downregulated the expression of ZO-1 in endothelial cells accompanied by decreased levels of 5hmC in the ZO-1 promoter region, suggesting the causal relationship between Tet2 activity and ZO-1 expression. Findings from post-mortem human brains support the findings from cultured endothelial cells and Tet2 knockout mice.
Meisel et al. reported that the intestinal barrier of the jejunum of Tet2-KO mice was damaged, and they found no changes in ZO-1 in epithelial cells of the colon [
9]. There are three isotypes of TET in mammalian cells. The distribution of three TET enzymes is related to the development stage and cell types in the mammalian body. TET1 is mainly enriched in embryonic stem cells, TET3 in oocytes, and TET2 is widely distributed spatially and temporally [
5]. Here, we found that TET2 activity modified the expression of ZO-1 in cerebral endothelial cells. ZO-1 serves as an important linker molecule of tight junctions which comprise the main component of mechanical barriers in the human body. Whether TET2 dependent ZO-1 expression modulates the permeability of other mechanical barriers such as blood-testis-barrier still needs further verification.
TET activity is suppressed under a variety of pathophysiological conditions. Various metabolites, such as 2-hydroxyglutarate (2-HG), α- KG antagonist, competitively inhibit the activity of TET enzymes in IDH mutant glioma cells [
8], and itaconic acid (ITA) inhibits Tet activity in macrophages after LPS treatment [
20]. TET enzymes are also inactivated under hypoxia [
7] or by mutations in about 15% of myeloid cancers [
21]. The decrease of TET activity will decrease the level of 5hmC, and this epigenetic modification of cytosine has a potential role in various physiological and pathological conditions. Previous studies on TET2 in the nervous system often focused on the relationship between the 5hmC level of neurons and their function. Using antibody-based 5hmc immunoprecipitation combined with deep sequencing (hMeDIP SEQ), Gontier identified 345 differential 5-hydroxymethylated regions (DHMRS) in the hippocampus of adult and elderly mice. All DHMRS showed that 5hmC in the hippocampus was partially lost with age and confirmed that TET2 activity was related to the methylation modification of the brain during aging [
22]. However, another team found that TET1, TET2, and 5hmC levels in the hippocampus of aged mice were lower than those of adult mice, and there was a positive correlation trend between TET2 and 5hmC, but this decrease was not significant [
23]. In our study, we also found that no changes of 5hmC were observed in aged human brain parenchymal cells. Even in Tet2-KO mice, there were no changes of 5hmC in neurons. Since TET proteins have three isoforms and neurons remain relatively static, that may explain the insignificant change in 5hmC of neurons to certain extents.
Tight junction proteins between brain endothelial cells greatly restrict paracellular transport. Therefore, the expression and function of the tight junction proteins are used as a marker of BBB integrity [
24]. ZO-1, one of the tight-junction proteins, interact with other tight junction proteins to anchor and tether them to the actin cytoskeleton [
25]. Alterations and breakdown of the BBB can occur naturally with aging even in the absence of concomitant conditions that cause cognitive decline and dementia [
26]. During healthy aging, there is an age-dependent loss of BBB integrity. BBB integrity can become compromised at the beginning of the loss of tight-junction proteins of the endothelial cell disrupted by excessive oxidative stress accumulated over time. Here, we found that the expression of tight junction protein ZO-1 in BBB endothelial cells significantly decreased in Tet2-KO mice. We also confirmed that the decreased expression of ZO-1 was positively correlated with the decreased level of 5hmC of endothelial cells in aged human brains. In cultured endothelial cells, we verified that oxidative stress significantly attenuated TET2 function, increased the methylation level of CpG islands in the promoter region of the ZO-1 gene, and inhibited ZO-1 expression. Due to the accumulation of oxidative metabolites in the process of aging, we hypothesize that inhibition of TET2 demethylase activity of endothelial cells leads to the destruction of BBB integrity, which plays a role in aging. Sasaki et al. found that isocitrate dehydrogenase 1 (IDH1) mutation produced a large amount of metabolites, such as 2-HG, which competitively inhibited α- kG dependent TET2 activity, leading to altered regulation of type IV collagen maturation and incomplete basement membrane of cerebral microvessels [
27]. These lead to perinatal intracerebral hemorrhage and death in mice. Since cerebral microvascular endothelial cells are one of the critical components of BBB, our findings strengthen the correlation between the destruction of BBB and the inhibition of TET2 activity.
In a human study, BBB impairment in aged brains was also correlated with increased inflammatory markers in the cerebrospinal fluid [
22]. BBB impairment might occur after the decreased expression of ZO-1, which is related to aging-induced cytokine changes [
28]. It recognized that plasma proteins leak into the brain parenchyma after BBB damage. Plasma proteins, such as thrombin, plasmin, albumin, and other proteins like immunoglobin, not only cause direct damage to neurons but also activate microglia and astrocytes to aggravate inflammatory neurotoxicity [
29]. However, Goodall et al. reported that the human brain staining showed BBB damage in aged brains and associated reactive astrogliosis, but no age-associated changes in microglia pathology, microvascular density, and pericyte coverage [
28]. Whether compromise of the integrity of BBB is involved in decreasing the number of neurons in aged brains and the role of astrocyte and oligodendrocyte activation in variable neurological diseases still need to be further studied.
Disruption of the BBB coincides with healthy aging, but the breakdown can become further exacerbated in neurodegenerative disorders. BBB disruption, supported by the reduction of tight junction proteins, is a hallmark of a number of neurodegenerative diseases, such as AD, Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), multiple sclerosis (MS), and Huntington’s disease (HD) [
29,
30]. The disruption of the BBB in neurodegenerative disorders makes it a clear target for developing new therapeutics. If the BBB disruption could be restored, at least to some extent, it may be possible to slow down the progression of certain conditions. Emerging research on the gut-brain axis and the protection or disruption of gut microbial-derived metabolites on BBB integrity is only at the very beginning stage. To better understand the role of epigenetic modulation of the BBB integrity, especially the participation of TET enzymes whose activity could be regulated by various factors and pathological conditions, it will be interesting to incorporate and utilize these interactions to develop therapeutics aiming to either restore or protect against BBB breakdown. A good example is the administration of vitamin C, which is a potent enhancer of TET activity. It can be taken as a supplement to prevent damage to the BBB and associated CNS conditions [
30‐
32].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.