Elevated Expression of Two Pore Potassium Channel THIK-1 in Alzheimer’s Disease: An Inflammatory Mechanism

Abstract

Introduction:

Tandem pore domain halothane-inhibited K+ channel 1 (THIK-1, coded by KCNK13) provides an upstream regulation of the activation of the NOD-like receptor pyrin domain containing 3 (NLRP3) inflammasome, which has been suggested as one of the key mechanisms of the pathological process in neurodegeneration mainly from in vitro and in vivo model systems studies. However, unequivocal evidence from neurodegenerative disorders has been lacking.

Objective:

To investigate the involvement of the THIK-1/NLRP3 pathway in the pathological process of Alzheimer’s disease (AD) and Parkinson’s disease (PD).

Methods:

This study investigated gene expression of markers in the THIK-1/NLRP3 pathway in an animal model representing AD as well as in human postmortem brains of AD and PD by quantitative real-time PCR. THIK-1 protein expression was determined using automated capillary electrophoresis immunoblotting. Furthermore, DNA methylation of KCNK13 was analysed in AD cohort by pyrosequencing.

Results:

A substantial upregulation of KCNK13, glial activation markers, NLRP3 inflammasome components, and IL1B was observed in the animal study. Increased expression of KCNK13 support an inflammatory glial cell activation in both advanced AD and PD. The increase in KCNK13 expression was also supported by downregulation in DNA methylation of KCNK13 in AD.

Conclusions:

The association between THIK-1 K+ channels expression and pathology changes indicates a THIK-1-induced activation of this glial subtype in AD and PD. Therefore, specific blocks of the microglial THIK-1 K+ channels at the early stage of AD and PD may be beneficial for the patients.

Keywords

Alzheimer’s disease DNA methylation neuroinflammation NLRP3 inflammasome Parkinson’s disease THIK-1 potassium channel

INTRODUCTION

Neurodegenerative diseases are the leading cause of morbidity and disability for the aging population worldwide. Among the neurodegenerative diseases, Alzheimer’s disease (AD) and Parkinson’s disease (PD) represent the most common two proteinopathies. Specific proteins, amyloid-β (Aβ) and tau in AD and alpha-synuclein (α-Syn) in PD, along with their misfolding, oligomerization and aggregation, underlie the pathogenic processes of these proteinopathies that eventually leads to neurodegeneration [1]. However, the mechanisms leading to neurodegeneration are multifactorial, caused by endogenous (genetic and aging-related) and exogenous (environmental) factors and their interplay, and remain poorly defined [2]. Several hypotheses have been postulated to explain the pathogenesis of AD and PD, one of which is the involvement of neuroinflammation emerging early in the neurodegenerative process and contributing to its progression [3].

In recent years, the mechanisms of neuroinflammation in the pathological changes of neurodegenerative diseases have been extensively studied. For example, by using animal and cell models, researchers have shown that exposure of glial cells to Aβ or α-Syn results in the elevated secretion of pro-inflammatory cytokines including tumor necrosis factor alpha, interleukin 6 (IL-6), IL-18, and IL-1β, which are detrimental to neuronal function [4 –6]. Data from clinical research also links neuroinflammation to the pathogenesis of neurodegenerative diseases. The dysregulation of cytokines and chemokines has been observed in plasma, cerebrospinal fluid (CSF) and brain of AD and PD patients [7 –9]. Indeed, accumulating experimental and clinical evidence support the view that immunological mechanisms are closely involved in the onset and progression of neurodegenerative diseases, although whether they act primarily as the cause of pathology or may solely be a consequence of, for example, Aβ deposition is not fully understood.

One potentially important component of the neuroinflammatory process is activation of the NOD-like receptor pyrin domain containing 3 (NLRP3) inflammasome pathway. The NLRP3 inflammasome is composed of three proteins: a cytosolic sensor molecule NLRP3, an apoptosis-associated speck-like protein containing a caspase-activating recruitment domain (ASC, coded by PYCARD gene), and an effector molecule, cysteine protease pro-caspase-1 (CASP1) [10]. This mediator of inflammation in the brain controls microglial activation and IL-1β secretion, with evidence indicating its central role in neurodegenerative disease processes [11]. This mechanism of modulation of innate immunity is in turn controlled by two-pore potassium (K2P) channels; potassium efflux mediated by K2P in glial cells is an upstream signal activating the NLRP3 inflammasome [12, 13]. Tandem pore domain halothane-inhibited K+ channel (THIK-1, coded by KCNK13 gene) is a key K2P channel in this respect, owing to its close association with glial cells [14] and the main potassium channel expressed in microglia [15, 16].

THIK-1 exhibits a distinct function of regulating microglial ramification, surveillance and IL-1β release [14]. Moreover, inhibiting the function of THIK-1 through pharmacological action or gene knockout can depolarize microglia, jeopardize the monitoring function and reduce IL-1β secretion as a result of affecting NLRP3 inflammasome activation [13]. We recently reported that pharmacological inhibition of THIK-1 inhibited caspase-1 activation and IL-1β release from mouse macrophages, mixed glia, and microglia in response to NLRP3 agonists. Similarly, macrophages and microglia from THIK-1 knockout mice had reduced NLRP3-dependent IL-1β release [17].

This evidence suggests that manipulation of THIK-1 may serve as a novel therapeutic approach in neurodegeneration, despite the current absence of direct evidence linking THIK-1 to AD and PD. Therefore, we aimed to determine whether the evidence for a role for THIK-1 activation in cell and animal models also extends to equivalent effects in the human brain. To achieve this, we first determined gene expression of THIK-1 potassium channels and the NLRP3 inflammasome pathway in rats after intracerebral administration of a presumed pathogenic factor in AD, amyloid-β oligomers (Aβo). We subsequently determined expression of KCNK13 and THIK-1 protein in postmortem brain tissue of AD and PD patients, as well as control subjects. Given the role of epigenetic mechanisms in the regulation of inflammasome expression [18], a further investigation into DNA methylation of KCNK13 was undertaken to investigate a potential epigenetic mechanism for the observed upregulation of this potassium channel in AD.

METHODS

Animal husbandry and research ethics committee approval

20 adult female Lister hooded rats, aged 8 weeks, weighing 222±19 g (mean±SD), were obtained from Charles River and housed in groups of five in individually ventilated double decker GR1800 cages (Tecniplast, Italy) on a 12-h light-dark cycle switching at 07 : 00 and 19 : 00 respectively. Environmental conditions were controlled at a temperature of 21±2°C and humidity at 55±5%. Animals had continuous ad libitum access to food and water (Special Diet Services, UK). Animal welfare was assessed daily. Animals were weighed weekly until the day of surgery, after which they were weighed at day 1, 3, and 7.

All surgical procedures and behavioral tests were conducted in strict accordance with the UK Animals (Scientific Procedures) Act of 1986 and the Animal Welfare and Ethical Review Board (AWERB) of the University of Manchester. The protocol was reviewed and approved by the University of Manchester Research Ethical Committee.

Surgical procedure

Rats were randomly assigned to either vehicle or Aβo group (n = 10 per group). Aβo were prepared as described by Rushworth et al [19]. Briefly, synthetic biotin-labelled-Aβ_1 - 42 (Anaspec) was dissolved in Hexafluoroisopropanol (HFIP) to disaggregate for 1 h, then aliquoted and HFIP was evaporated off using vacuum centrifugation. Peptide films were dissolved in dimethyl sulfoxide (DMSO) to bring to a 1 mM concentration and then diluted in Ham’s F-12 (Lonza) to a final monomer concentration of 100μM. It was then allowed 16-18 h to aggregate at room temperature. The aggregates tubes were centrifuged at 14,000×g for 20 min to pellet out any fibrillary material and the supernatant was retained as the Aβo. Aβo in the F-12 media was characterized using dot blot before being administered to the animals.

The surgical procedure was carried out as described in earlier publications [20, 21]. Briefly, rats were anaesthetized with 4% (v/v) isoflurane (Piramal) in oxygen (maintained at 2-3% isoflurane) and settled firmly into a stereotaxic frame. Animals then received a sub-cutaneous administration of 0.1 mg/kg of analgesic (Vetergesic^® containing 0.3 mg/mL buprenorphine (Orion Pharma)), followed by the bilateral intrahippocampal injections (AP - 3.5 mm, ML 2 mm, DV - 4 mm) [22] of 10μL of either vehicle (DMSO and Ham’s F-12) or 10 nmol Aβo on each side on Day 0. Seven days after surgery (Day 7), animals underwent novel object recognition (NOR) behavioral testing.

Source of human brain tissue and research ethics committee approval

Neuropathologically confirmed postmortem AD brain tissue samples (Table 1) were obtained from the Manchester Brain Bank at Salford Royal NHS Foundation Trust and PD brain tissue samples (Table 2) were obtained from the Multiple Sclerosis and Parkinson’s Tissue Bank at Imperial College London. Matched control samples were also supplied by the two brain banks for each cohort respectively. The Manchester Brain Bank has been approved by the Newcastle & North Tyneside 1 Research Ethics Committee (REC reference 09/H0906/52 + 5). The MS and Parkinson’s Tissue Bank has been approved as a Research Tissue Bank by the Wales Research Ethics Committee (REC reference 08/MRE09/31 + 5).

Table 1

Demographic parameters for the postmortem AD human brain samples

	Control	Moderate AD	Advanced AD
	(Braak Stage 0-II, N = 20 per sex)	(Braak Stage III-IV, N = 20 per sex)	(Braak Stage V-VI, N = 20 per sex)
Age of death (y)
All (N = 120)	82.6 ± 11.5	85.5 ± 6.3	78.1 ± 8.8
Male (N=60)	84.5 ± 9.4	84.0 ± 6.5	73.3 ± 4.7
Female (N=60)	80.7 ± 13.3	87.1 ± 6.0	83.0 ± 9.3
Postmortem Interval (h)
All (N=120)	81.2 ± 36.9	79.4 ± 38.4	86.2 ± 43.0
Male (N=60)	85.7 ± 41.4	81.2 ± 37.0	82.7 ± 48.9
Female (N=60)	76.5 ± 31.8	77.5 ± 40.6	89.6 ± 37.1

Table 2

Demographic parameters for the postmortem PD human brain samples

	Control	PD
	(Male N = 8,	(Male N = 6,
	Female N = 7)	Female N = 9)
Age of death (y)
All (N = 30)	82.6±9.8	80.3±8.2
Male (N = 14)	80.1±8.4	78.0±8.6
Female (N = 16)	85.4±11.1	81.8±8.0
Postmortem interval (h)
All (N = 30)	25.9±8.0	18.1±7.8
Male (N = 14)	27.8±9.8	17.5±7.6
Female (N = 16)	23.7±7.8	18.6±8.4

The brain regions that have been examined in this project were temporal cortex for the AD cohort, substantia nigra (SN) and frontal cortex (FC) for the PD cohort, as these brain regions were indicated as relatively more severely affected by underlying molecular pathologies in the two diseases [23, 24].

Standard procedures were carried out including postmortem brain sample storage, handling and examination according to UK research ethics. Unfixed samples were stored at –80°C immediately after dissection until tissue homogenization. Research was conducted in compliance with the principles of the Declaration of Helsinki.

Clinical data was provided by the brain banks. For the AD cohorts, only demographic parameters and clinical diagnosis were available, while extra clinical data was available for the PD cohort, including age of disease onset, duration of disease, family medical history, clinical history, complications, autonomic and psychiatric symptoms, medication, and clinical cause of death.

Quantitative real-time PCR

RNA expression was quantified using reverse transcriptase semi-quantitative real-time PCR (RT-qPCR) as previously described [25]. All the procedures were performed according to the MIQE guidelines [26].

Briefly, total RNA from postmortem brain tissue of human and rat was extracted using TRIzol™ Reagent (Thermo Fisher, UK) followed by RNA clean up using Monarch^® RNA Cleanup Kit (New England Biolabs, UK). RNA purity and concentrations were assessed using a NanoDrop 2100 (Thermo Fisher, UK). Single-stranded cDNA was synthesized from 1 ng of total RNA using M-MLV Reverse Transcriptase (ThermoFisher, UK). cDNA concentrations were measured on a NanoDrop 2100 and diluted to 500 ng/μL in RNase-free water.

Reference genes (RGs) were selected based on previous studies. The stability of RGs were ranked by RefFinder [27], and the top two stable RGs were chosen for delta Ct analysis. Primer pairs were designed by using Primer-BLAST (NCBI, http://www.ncbi.nlm.nih.gov/tools/primer-blast/, accessed on 11 February 2019), exon junctions were included wherever possible to avoid amplification of genomic DNA (Supplementary Table 1). Primers were synthesized by ThermoFisher. Primer specificity was analyzed by dissociation curve. RT-qPCR reactions were carried out using Power SYBR™ Green PCR Master Mix (ThermoFisher, UK) on a 7900HT Real-Time PCR System with 384-well format (Applied Biosystems, UK). The final volume for each reaction was 10μL with 1μL (0.25μM) of corresponding primers, 1μL (500 ng/μL) of total cDNA, 3μL RNase-free water, and 5μL SYBR green. All samples were run in triplicate. The thermal cycler parameters were as follows: UDG activation at 50°C for 2 min, and DNA polymerase activation at 95°C for 10 min, followed by amplification of cDNA for 40 cycles with denaturation at 95°C for 15 s and annealing/extension at respective annealing temperature for each gene for 1 min. Dissociation curve analyses were carried out at the end of each run for PCR product verification. The intra-assay variation was controlled as standard deviation of triplicates < 0.2 and was analyzed by using raw Ct value of RGs and there were no significant variations between plates.

Automated capillary electrophoresis immunoblotting

Protein expression of THIK-1 was quantified using automated capillary electrophoresis immunoblotting as previously described [25]. Briefly, 40–60 mg human postmortem brain tissue was homogenized by 10-fold volume (mg:μL) of lysis buffer (pH 7.4, 10 mM Trizma base, 2 mM EDTA, 320μM sucrose (Sigma-Aldrich, UK)). cOmplete™ Protease Inhibitor Cocktail (Sigma-Aldrich, UK), phenylmethylsulfonyl fluoride (PMSF), and sodium orthovanadate were added to the lysis buffer immediately before use at 4%, 1%, 1% (v/v) concentrations respectively. The differential centrifugation method was used for protein purification [28]. In brief, the tissue homogenate was centrifuged for 15 min (800 x g, 4°C). The pellet containing nuclei was discarded and the supernatant was centrifuged for 20 min (12,000 x g, 4 °C). The supernatant containing soluble protein was stored at -20°C for protein detection. All the procedures were performed on ice wherever possible.

Automated capillary electrophoresis immuno-quantification was conducted on a Wes^TM instrument for target protein analyses, according to the manufacturer’s protocol (ProteinSimple) [29]. In brief, soluble protein concentration was firstly assessed using Bradford assay, then optimization of sample concentration and primary antibodies was carried out. Sample was diluted to the optimized concentration at 0.4 mg/mL. After addition of a master mix containing dithiothreitol and fluorescent molecular weight marker, samples were heated at 95°C for 5 min for denaturation. Then, prepared samples, blocking reagent (Wes^TM antibody diluent), target protein primary antibodies, horseradish peroxidase (HRP)-conjugated secondary antibodies, and chemiluminescent substrate were loaded into the allocated wells of a Wes plate, pre-loaded with sample and stacking matrices. Instrument default settings were used. The size-based separation electrophoresis, immobilization, and immunodetection were automatically processed in the capillary array system and the chemiluminescence was detected at multiple exposure times and quantified by the ProteinSimple v3.1 software. Primary antibody used was rabbit anti-THIK-1 (1 : 200, NBP2-41132, Novus Biologicals). The total protein value detected by Total Protein assay (ProteinSimple) of each sample was used for normalization.

DNA extraction, bisulphite conversion, and pyrosequencing

Genomic DNA from temporal cortex of human postmortem tissue of AD and matched controls was purified using QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. EpiTec Fast DNA Bisulfite Kit (Qiagen) was used to convert unmethylated cytosine to uracil in the CpG sites with a calculated mean conversion of 99%. For analysis, a sequence containing 5 CpGs (TGCGCCCGTTCCCCGGGACCTTCAGCGTTAAGGAGAACTCCGAGG) in the promoter region of KCNK13 (coding gene for THIK-1) (GRCh38.p13, Chromosome 14, 90,061,889 –90,061,933. GenBank accession number BC012779.2, positions 125–169) that contained potential transcription factor binding sites was identified and amplified by PCR using primers, including a biotinylated reverse primer, as follows: 5^′-AGGGAGTAGGGGTGGAAATTG-3^′ (forward) and 5^′-[btn]CTCCACTCCCCCTACTCT-3^′ (reverse) (Eurofins MWG Operon). Then, methylation status of the sequence was determined with a PyroMark Q48 Autoprep (Qiagen UK) using 10μl PCR product and employing a sequencing primer, 5^′- AAGTAGAAGTGAATTT-3^′ (Eurofins MWG Operon). PCR reactions, amplification conditions and the methylation profile were carried out according to our previous studies [30, 31].

Statistical analysis

For the clinical sample, demographic and clinical data were analyzed using descriptive statistics. mRNA analysis was carried out using the 2(-Delta Delta Ct) Method [32]. Raw Ct values were normalized using geometric mean of two RGs as an endogenous internal standard. Data are displayed as mean±SEM. Gaussian distribution was evaluated using the Shapiro-Wilk normality test. Homogeneity of variance was evaluated using Levene’s test. Statistical analyses were conducted in IBM SPSS Statistics 23 and GraphPad Prism 7 software using non-paired parametric Student’s t-test, Welch’s t-test for unequal variances, one-way ANOVA followed with a post hoc test, and Pearson correlation or Spearman rank correlation. Statistical significance was set at p-value≤0.05 (*≤0.05, **≤0.01, ***≤0.001).

RESULTS

Glial cells and NLRP3 inflammasome were activated 7 days after Aβo administration

Fig. 1

Gene expression of markers of NLRP3 inflammasome and glial cell activation. mRNA expression of (A) NLRP3 inflammasome and (B) glial cell markers were significantly higher in Aβo group on Day 7. Student t-tests were conducted. Data are presented as the means±SEM. (*p≤0.05, **p≤0.01, n = 10 per group).

Relative gene expression of inflammatory markers of the NLRP3 pathway determined by qPCR in the left dorsal hippocampus were compared between vehicle and Aβo injected animals 7 days after surgery, when the inflammation is considered to be chronic [33]. As illustrated in Fig. 1 above, gene expression of PYCARD, CASP1, downstream cytokine IL1B as well as the glial markers were significantly elevated in Aβo group, with the highest expression of GFAP (glial fibrillary acidic protein) (Fig. 1).

Gene expression of THIK-1 potassium channels were elevated 7 days after Aβo administration

Similar to NLRP3 activation markers, on Day 7, the mRNA expression of KCNK13 was significantly higher in rats following Aβo administration (Fig. 2).

Fig. 2

Gene expression of THIK-1 potassium channel. mRNA expression of THIK-1 potassium channels were significantly higher in Aβo group on Day 7. Student t-tests were conducted. Data are presented as the means±SEM. (*p≤0.05, **p≤0.01, n = 10 per group).

Further analysis of the correlation KCNK13 and NLRP3 inflammasome activation, found gene expression of KCNK13 was positively correlated with gene expression of NLRP3, PYCARD, CASP1, and IL1B as well as GFAP (Table 3).

Table 3

Correlation between gene expression of THIK-1 and functional indicators of NLRP3 inflammasome

dCT value of gene expression		THIK-1
NLRP3	Rho	0.814**
	p	0.000
PYCARD	Rho	0.644**
	p	0.000
CASP1	Rho	0.770**
	p	0.000
IL1B	Rho	0.645**
	p	0.000
AIF1	Rho	–0.177
	p	0.275
GFAP	Rho	0.731**
	p	0.000

^**Spearman’s correlation was conducted. n = 20.

RNA expression of THIK-1 potassium channel was increased in advanced Alzheimer’s disease

Age, postmortem interval (PMI), or sex did not correlate with any of the quantification value of mRNA or protein expression assessed in this study. Relative gene expression of THIK-1 potassium channel in the temporal cortex of the AD cohort was determined by qPCR. A significant association was seen between KCNK13 expression and AD, with a higher expression in AD (t (112)=2.163, p = 0.033). Further analysis demonstrated an association with the severity of tau pathology of AD showing that KCNK13 was significantly elevated in advanced AD (p = 0.008, 95% CI=(-2.19, -0.29)) compared with controls as well as compared with moderate AD (p = 0.048) but not significantly so in moderate AD (p = 0.581, 95% CI=(-1.34, 0.58)) (Fig. 3).

Fig. 3

RNA expression of THIK-1. Significant elevation of THIK-1 (KCNK13) expression was observed in advanced AD, but not in moderate AD. Control, n = 37; moderate AD, n = 39; advanced AD, n = 39. Data are presented as the mean±SEM and analyzed by the one-way ANOVA test followed by the Dunnett post hoc test. *p≤0.05, **p≤0.01.

Protein expression of THIK-1 was not significantly changed in Alzheimer’s disease

Protein expression of THIK-1 was studied by automated capillary electrophoresis immunoblotting. Three peaks/bands (equivalent to molecular weight (MW): 27 kDa, 49 kDa, and 58 kDa) were detected in the protein assay of THIK-1 (Fig. 4A). The theoretical MW of THIK-1 is 45 kDa (GeneCards, https://www.genecards.org/cgi-bin/carddisp.pl?gene=KCNK13. accessed on October 13, 2021); however, the observed MW by western blot indicated for the anti-THIK-1 antibody is 62 kDa according to product literature (NBP2-41132, Novus Biologicals). Due to the antigen being unavailable, we could not confirm the MW of THIK-1 for our samples. Therefore, expression data for 49 kDa and 58 kDa bands were analyzed on the assumption that they corresponded to THIK-1 protein. No significant differences of 49 kDa protein (one-way ANOVA, F (2. 108)=1.852, p = 0.162) or 58 kDa protein (F (2. 108)=0.977, p = 0.380) expression were observed between groups (Fig. 4B).

Fig. 4

Automated capillary electrophoresis immuno-quantification analysis of THIK-1 protein expression. A) Results are shown as gel-like image view in left panel and electropherograms in the right panel. Three peaks were detected by the antibody. B) Two potential peaks of THIK-1 were analyzed, both of which were unchanged between groups. Control, n = 37; moderate AD, n = 38; advanced AD, n = 36. Data are presented as the mean±SD and analyzed by the one-way ANOVA test followed by the Dunnett post hoc test.

DNA methylation of KCNK13 was lower in Alzheimer’s disease

DNA methylation of KCNK13 was studied to follow up the result observed showing an increase of KCNK13 expression in AD. DNA methylation (mean±SEM) on CpG4 in the target sequence of THIK-1 in AD was lower (4.87±2.55) compared with control (6.28±3.85) (t (93)=2.127, p = 0.036) (Fig. 5). However, when compared between AD subgroups, the changes did not reach the significance (one-way ANOVA, F(2. 92)=2.248, p = 0.111). Methylation of either the single CpG site or the mean value of the 5 CpGs was not significantly related to age, PMI or sex in the sample.

Fig. 5

DNA methylation of THIK-1 in AD and controls. A) CpG sites (numbered 1–5) are shown in bold in the promoter region of DNA sequence of THIK-1 that underwent pyrosequencing. B) AD group presents significant lower level of CpG4. Control, n = 32; AD, n = 63. Data are presented as the mean±SEM. *p≤0.05.

Interestingly, although the methylation status of any CpG site studied, including CpG4, was not significantly correlated with KCNK13 expression, mean percentage methylation, as well as the mean methylation level of each of the five CpGs, were negatively correlated with immunoreactivity of THIK-1 protein (Table 4). This effect was strongest for CpG3-5.

Table 4

Correlation between % methylation at five CpGs and gene and protein expression of THIK-1 in AD cohort

		CpG1	CpG2	CpG3	CpG4	CpG5	Mean of five CpGs
dCT value of THIK-1	Rho	–0.057	0.141	–0.147	–0.164	0.106	–0.008
	p	0.590	0.177	0.161	0.112	0.311	0.939
Protein level of THIK-1 (peak49)	Rho	–0.229*	–0.217*	–0.468**	–0.336**	–0.272**	–0.349**
	p	0.030	0.037	0.000	0.001	0.009	0.001
Protein level of THIK-1 (peak58)	Rho	–0.204	–0.063	–0.427**	–0.313**	–0.309**	–0.297**
	p	0.055	0.550	0.000	0.002	0.003	0.004

Analyzed by the Spearman’s correlation test. *p<0.05, **p<0.01.

RNA expression of THIK-1 potassium channel was increased in Parkinson’s disease

Relative gene expression of KCNK13 was determined in two brain regions of the PD cohort. As shown in Fig. 6, in SN KCNK13 expression was significantly higher in PD compared with control (t (28)=-2.74, p = 0.011). In FC, an elevation of KCNK13 was also seen in the PD group (t (28)=-4.054, p < 0.001).

Fig. 6

RNA expression of THIK-1 in PD cohort. There was a significant elevation in PD group of THIK-1 expression in both brain regions. Control, n = 15; PD, n = 15. Data are presented as the mean±SEM. *p≤0.05, ***p≤0.001.

In addition to the pathology, the brain bank has also provided clinical information including cognitive function in the PD cohort. Given that cognition is more closely associated with FC than SN, and KCNK13 expression was substantially increased in AD in the present study, further analyzing the KCNK13 expression in the symptom-based subgroups of PD (PD with dementia and PD without dementia) was carried out. In the PD with dementia subgroup, the increase of KCNK13 was only found in the FC (p = 0.001, 95% CI=(-2.62, -0.72)) but not in SN (p = 0.248, 95% CI=(-1.90, 0.41)) (Fig. 7). Re-analysis of the FC data after removing one outlier made essentially no difference to the results, with higher expression of KCNK13 in the FC of PD with dementia (one-way ANOVA, F (2, 24)=8.690, p = 0.001). In PD without dementia subgroup, KCNK13 expression was significantly higher in SN (p = 0.05, 95% CI=(-2.02, 0.01)) and FC (p = 0.014, 95% CI=(-1.87, -0.20)) (Fig. 7).

Fig. 7

Comparison of RNA expression of THIK-1 between PD subgroups and controls in different brain regions. In PD without dementia subgroup, THIK-1 expression was significantly higher in both brain regions. However, in the PD with dementia subgroup, the elevation of THIK-1 expression was only shown in the frontal cortex. Control, n = 13; PD without dementia, n = 9; PD with dementia, n = 6. Data are presented as the mean±SEM. *p≤0.05, **p≤0.01, ***p≤0.001.

DISCUSSION

We have shown here the involvement of the THIK-1 potassium channel in the activated NLRP3 pathway in an acute animal model of AD. Data at Day 7, representing a sub-chronic inflammatory response to Aβo injection in the brain [33], are consistent with the hypothesis that Aβo can promote NLRP3 activation, potentially via the activation of THIK-1, resulting in an increase in the secretion of pro-inflammatory cytokines. In addition, an elevated expression of KCNK13 gene occurs in human postmortem brains of subjects with AD and with PD with dementia. This highlights the potential role of THIK-1 potassium channel dysfunction in the process of NLRP3 inflammasome activation in dementia-related pathologies. Based on these findings, we further demonstrate a reduction in DNA methylation of the KCNK13 gene promoter sequence in the AD postmortem brains.

The observed increase in expression of CASP1 which mediates the production of IL1B, would be expected to be reflected by an increase of this pro-inflammatory cytokine. The increase in ASC (coded by PYCARD) gene expression, in the absence of a significant elevation of NLRP3, could still indicate an increase in inflammasome-like activity [34]. It was reported that overexpressed ASC can rapidly polymerize and aggregates in the absence of the sensor domain of the inflammasome, such as NLRP3. The ASC speck is functional in regulating caspase-1 activity via recruiting pro-caspase-1 or nucleating CARD domain of caspase-1 filaments [35]. Interestingly, the strong indications of an increase in inflammasome activity leading to a functional pro-inflammatory state in this study occur in the context of a decrease in the expression of AIF-1. While this reduction appears to be inconsistent with NLRP3 activation; however, other such inconsistent results of AIF-1 expression in AD have been reported in previous studies [36]. The results with this microglial marker could reflect a change in cellular form and function which is associated with an inflammatory response. Meanwhile, a substantial elevation in GFAP, indicative of astrocyte activation and proliferation, is also apparent following Aβo injection, which is in accordance with the results of the postmortem study we previously reported [25].

Even though potassium efflux has been known to be an essential upstream event in the process of NLRP3 activation [14 , 37], the main potassium channel responsible for AD-related NLRP3 activation is undefined. In the present animal model study, THIK-1 potassium channels showed an increased expression profile in the Aβo group. Importantly, KCNK13 gene expression was positively correlated with NLRP3, PYCARD, CASP1, and IL1B as well as GFAP. The involvement of THIK-1 in AD was further supported by the human postmortem study, with an increased expression of KCNK13 in both advanced AD and PD. THIK-1 function has been little studied in the pathologies of neurodegenerative diseases, the increased expression of the KCNK13 is consistent with another postmortem study [38]. The increase in PD is greatest in the cortex of those patients with dementia compared with other subgroups, although this is not a significant effect, perhaps reflecting the small sample sizes in the PD subgroups. The association between KCNK13 expression and pathological changes, as well as its role in regulating microglial function highlighted in other research discussed above, indicate a THIK-1-induced activation of this glial subtype in AD and PD. Therefore, targeting of microglial THIK-1 channels at the early stage of AD and PD may be beneficial for the patients.

Of note, data of the THIK-1 protein expression was not associated with AD. The THIK-1 protein level in postmortem brain might be low and it can be easily affected by the postmortem process of the brain sample, such as PMI and brain PH [39]. Moreover, agonal factors, including fever, infection, and unconsciousness can contribute to brain gene and protein expression changes [40]. Therefore, our findings for the protein expression of THIK-1 need to be further verified.

DNA methylation, through its effects in modulating gene expression, is found to be involved in cell differentiation, synapse formation, synaptic plasticity, learning and memory behavior in the healthy brain [41]. Thus, it is conceivable that alteration of DNA methylation could have a drastic impact on the ethology and pathology of brain diseases, such as neurological and psychiatric disorders. One potential underlying mechanism is DNA methylation-induced alteration of glial functions in these diseases. It is reported that DNA methylation promotes heightened microglial activation in aged mouse brain [42]. Changes in DNA methylation have also been observed in glial cells in the postmortem AD brain [43]. As THIK-1 is strongly associated with microglia function and NLRP3 inflammasome activation, the reduction in methylation of DNA from a KCNK13 promoter sequence in AD might be expected to enhance transcription and so is consistent with the gene expression findings. The reduction in DNA methylation in the AD group is found to be significant at CpG4, which is a binding site for multiple transcription factors (TFs) (Fig. 8). CpG methylation can negatively, or occasionally positively, influence the binding of TFs to DNA and vice versa [44]. The interactions between TFs and methylated DNA can further affect gene expression, splicing regulation, and chromatin remodeling [45]. Of the TFs with a binding site at CpG4, the forkhead box protein 3 (FOXP3) is particularly involved in immune response.

Fig. 8

Predicted transcription factor and their binding sites in targeted KCNK13 sequences. The transcription factors and the corresponding binding sites were identified by using online laboratory PROMO (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3, accessed on 13/10/2021). In the KCNK13 sequences of interest, there are 5 CpG sites and each site is within the transcription factor binding sites.

FOXP3 is mainly expressed in a subset of T cells that play a suppressive role in the immune system [46, 47]. FOXP3 acts also as a TF that suppresses or activates transcription of many genes including cytotoxic T-Lymphocyte antigen4, CD25 and glucocorticoid-induced TNF receptor family gene [46]. Abnormal methylation at this site would suggest a dysregulation of the transcriptional effects of this response, although whether reduced methylation might be a cause or consequence of increased FOXP3 function is unclear. Interestingly, this TF is reportedly increased both in AD brain and in an animal model of AD [48].

The methylation data at CpG4 does not significantly correlate with the expression of mRNA. A mismatch between DNA methylation and transcription is not a rare observation, potentially due to differences in time course as well as other regulatory mechanisms that also contribute to gene expression [49]. Nevertheless, KCNK13 methylation at CpG4 does show a highly significant negative correlation with THIK-1 protein, as does methylation at most of the 5 CpG sites studied (Table 4). This suggests that methylation in the promoter sequence studied might have a functional effect in regulating the amount, and perhaps the activity, of the resultant THIK-1 channels, although the findings across gene and protein expression are not fully consistent in this study.

In summary, the inflammatory outcome in the animal model of relevance to AD is apparent and supports the hypothesis of an involvement of THIK-1/NLRP3 mediated neuroinflammation as a consequence of one pathological feature of AD. The findings suggest a chronic inflammatory change in response to Aβo injection which is consistent with the hypothesis and with human studies. However, in the absence of cognitive behavior changes, assessed by the NOR test (data shown in the Supplementary Material), it is difficult to draw a conclusion whether the inflammatory response at this stage is neuroprotective or neurotoxic. Animal studies do not, of course, suffer from the many limitations of postmortem human investigations, particularly the problem that tissue taken after death is primarily limited to subjects at the very end stage of the disease process. Thus in an animal model, the progression of the pathological process can be monitored at different time-points. In the context of the current study, it would be possible, for example, to determine the temporal relationships between amyloid deposition and the appearance of inflammatory markers. Future studies with a longer time course with different timepoints to assess the THIK-1/NLRP3 induced inflammatory effect on behavioral changes are essential. In addition, THIK-1 knockout models can provide more direct information on its role in regulating NLRP3 activation, which need to be further studied.

While study of human postmortem samples has proved enormously useful in understanding neurobiological disease, it has limitations, as discussed in our previous study [25]. For example, the effects of agonal state can severely reduce the integrity of RNA; while this can to some extent be controlled for with housekeeping genes, it introduces another limitation in interpreting results. Nevertheless, the use of samples from well-matched non-AD subjects provides some control for many of these confounding effects. In addition, even accurate determination of mRNA is only representative of gene expression at a single time point. Moreover, our results in AD are limited to temporal cortex. It would be important in future studies to determine whether elevated KCNK13 is found in other brain regions affected by AD, particularly the hippocampus. While we have a relatively substantial sample size in comparison to many such human brain studies, the sample remains underpowered to identify anything other than relative large effects. Furthermore, changes in the cellular expression of markers may not always be reflected by results in tissue due to recruitment or depletion of the relevant cell [50]. Future research with higher cell specificity and closer examination of protein function in this pathway in human cohorts with clinical information, especially cognitive data, is urgently needed.

Footnotes

ACKNOWLEDGMENTS

We are indebted to patients and their families, who donated these precious samples that made this study possible. We thank contributors who collected samples used in this study. We are grateful for the collaboration with Professor Nigel M. Hooper’s group for the preparation of amyloid-β oligomers.

FUNDING

Opening Foundation of Clinical Medical Centre of Neurology of First People’s Hospital of Yunnan Province. Grant Number: 2021LCZXXF-SJ03, 2022LCZXKF-SJ02.

Hao Tang was funded by China Scholarship Council - The University of Manchester Joint Scholarship Programme, which supported her PhD study. The work presented in this manuscript was conducted as part of her PhD work and reported in her thesis.

Tissue samples were supplied by The Manchester Brain Bank, which is part of the Brains for Dementia Research program, jointly funded by Alzheimer’s Research UK and Alzheimer’s Society.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

All data generated or analyzed during this study are included in this published article and its supplementary information files. Raw CT value of qPCR study are available from the corresponding author on reasonable request.

The supplementary material is available in the electronic version of this article: .

References

Soto

, Pritzkow

(2018) Protein misfolding, aggregation, and conformational strains in neurodegenerative diseases. Nat Neurosci 21, 1332–1340.

Madore

, Yin

, Leibowitz

, Butovsky

(2020) Microglia, lifestyle stress, and neurodegeneration. Immunity 52, 222–240.

Leng

, Edison

(2021) Neuroinflammation and microglial activation in Alzheimer disease: Where do we go from here? Nat Rev Neurol 17, 157–172.

Patel

, Paris

, Mathura

, Quadros

, Crawford

, Mullan

(2005) Inflammatory cytokine levels correlate with amyloid load in transgenic mouse models of Alzheimer’s disease. J Neuroinflammation 2, 1–10.

Wong

, Krainc

(2017) α-synuclein toxicity in neurodegeneration: Mechanism and therapeutic strategies. Nat Med 23, 1–13.

Cornell

, Salinas

, Huang

, Zhou

(2022) Microglia regulation of synaptic plasticity and learning and memory. Neural Regen Re 17, 705–716.

Ransohoff

(2016) How neuroinflammation contributes to neurodegeneration. Science 353, 777–783.

Zuena

, Casolini

, Lattanzi

, Maftei

(2019) Chemokines in Alzheimer’s disease: New insights into prokineticins, chemokine-like proteins. Front Pharmacol 10, 622.

, Yang

, Zhao

, Luo

, Niu

, Kang

, Xie

, Lu

, Chen

, Liu

(2022) Parkinson’s disease peripheral immune biomarker profile: A multicentre, cross-sectional and longitudinal study. J Neuroinflammation 19, 116.

10.

Rathinam

, Fitzgerald

(2016) Inflammasome complexes: Emerging mechanisms and effector functions. Cell 165, 792–800.

11.

Heneka

, Kummer

, Stutz

, Delekate

, Schwartz

, Vieira-Saecker

, Griep

, Axt

, Remus

, Tzeng

, Gelpi

, Halle

, Korte

, Latz

, Golenbock

(2013) NLRP3 is activated in Alzheimer’s disease and contributes to pathology in APP/PS1 mice. Nature 493, 674–678.

12.

Muñoz-Planillo

, Kuffa

, Martínez-Colón

, Smith

, Rajendiran

, Núñez

(2013) K+ efflux is the commontrigger of NLRP3 inflammasome activation by bacterial toxins andparticulate matter. Immunity 38, 1142–1153.

13.

, Xiong

, Ye

, Malireddi

RKS

, Kometani

, Zhong

, Mittal

, Hong

, Kanneganti

, Rehman

, Malik

(2018) The TWIK2 potassium efflux channel in macrophages mediates NLRP3 inflammasome-induced inflammation. Immunity 49, 56–65.

14.

Madry

, Kyrargyri

, Arancibia-Carcamo

, Jolivet

, Kohsaka

, Bryan

, Attwell

(2018) Microglial ramification, surveillance, and interleukin-1β release are regulated by the two-pore domain K+channel THIK-1. Neuron 97, 299–312.

15.

Zhang

, Chen

, Sloan

, Bennett

, Scholze

, O’Keeffe

, Phatnani

, Guarnieri

, Caneda

, Ruderisch

, Deng

, Liddelow

, Zhang

, Daneman

, Maniatis

, Barres

, Wu

(2014) An RNA-sequencing transcriptome and splicing database of glia, neurons, and vascular cells of the cerebral cortex. J Neurosci 34, 11929–11947.

16.

Zhang

, Sloan

, Clarke

, Caneda

, Plaza

, Blumenthal

, Vogel

, Steinberg

, Edwards

MSB

, Li

, Duncan

3rd, Cheshier

, Shuer

, Chang

, Grant

, Gephart

MGH

, Barres

(2016) Purification and characterization of progenitor and mature human astrocytes reveals transcriptional and functional differences with mouse. Neuron 89, 37–53.

17.

Drinkall

, Lawrence

, Ossola

, Russell

, Bender

, Brice

, Dawson

, Harte

, Brough

(2022) The two pore potassium channel THIK-1 regulates NLRP3 inflammasome activation. Glia 70, 1301–1316.

18.

Poli

, Fabi

, Bellet

, Costantini

, Nunziangeli

, Romani

, Brancorsini

(2020) Epigenetic mechanisms of inflammasome regulation. Int J Mol Sci 21, 5758.

19.

Rushworth

, Griffiths

, Watt

, Hooper

(2013) Prion protein-mediated toxicity of amyloid-β oligomers requires lipid rafts and the transmembrane LRP1. J Biol Chem 288, 8935–8951.

20.

Daniels

MJD

, Rivers-Auty

, Schilling

, Spencer

, Watremez

, Fasolino

, Booth

, White

, Baldwin

, Freeman

, Wong

, Latta

, Yu

, Jackson

, Fischer

, Koziel

, Pillot

, Bagnall

, Allan

, Paszek

, Galea

, Harte

, Eder

, Lawrence

, Brough

(2016) Fenamate NSAIDs inhibit the NLRP3 inflammasome and protect against Alzheimer’s disease in rodent models. Nat Commun 7, 12504.

21.

Watremez

, Jackson

, Almari

, McLean

, Grayson

, Neill

, Fischer

, Allouche

, Koziel

, Pillot

, Harte

(2018) Stabilized low-n amyloid-β oligomers induce robust novel object recognition deficits associated with inflammatory, synaptic, and GABAergic dysfunction in the rat. J Alzheimers Dis 62, 213–226.

22.

Paxinos

, Franklin

KBJ

(2019) Paxinos and Franklin’s the Mouse Brain in Stereotaxic Coordinates. Academic Press, 357 p.

23.

Braak

, Ghebremedhin

, Rüb

, Bratzke

, Del Tredici

(2004) Stages in the development of Parkinson’s disease-related pathology. Cell Tissue Res 318, 121–134.

24.

Braak

, Thal

, Ghebremedhin

, Del Tredici

(2011) Stages of the pathologic process in Alzheimer disease: Age categories from 1 to 100 years. J Neuropathol Exp Neurol 70, 960–969.

25.

Tang

, Harte

(2021) Investigating markers of the NLRP3 inflammasome pathway in Alzheimer’s disease: A human post-mortem study. Genes 12, 1753.

26.

Bustin

, Benes

, Garson

, Hellemans

, Huggett

, Kubista

, Mueller

, Nolan

, Pfaffl

, Shipley

, Vandesompele

, Wittwer

(2009) The MIQE guidelines: Minimum information for publication of quantitative real-time PCR experiments. Clin Chem 55, 611–622.

27.

Xie

, Xiao

, Chen

, Xu

, Zhang

(2012) miRDeepFinder: A miRNA analysis tool for deep sequencing of plant small RNAs. Plant Mol Biol 80, 75–84.

28.

Lodish

, Berk

, Zipursky

, Matsudaira

, Baltimore

, Darnell

(2000) Section 3.5, Purifying, detecting, and characterizing proteins. Molecular Cell Biology, 4th edition. W.H. Freeman, New York.

29.

Harris

(2015) Protein detection by simple Western^TM analysis. In Western Blotting: Methods and Protocols, Kurien BT, Scofield RH, eds. Springer, New York. pp 465–468.

30.

Tang

, Dalton

, Srisawat

, Zhang

, Reynolds

(2014) Methylation at a transcription factor-binding site on the 5-HT1A receptor gene correlates with negative symptom treatment response in first episode schizophrenia. Int J Neuropsychopharmacol 17, 645–649.

31.

Fachim

, Loureiro

, Corsi-Zuelli

, Shuhama

, Louzada-Junior

, Menezes

, Dalton

, Del-Ben

, Reynolds

(2019) GRIN2B promoter methylation deficits in early-onset schizophrenia and its association with cognitive function. Epigenomics 11, 401–410.

32.

Livak

, Schmittgen

(2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 25, 402–408.

33.

Moore

, Wu

, Shaftel

, Graham

, O’Banion

(2009) Sustained expression of interleukin-1beta in mouse hippocampus impairs spatial memory. Neuroscience 164, 1484–1495.

34.

Franklin

, Latz

, Schmidt

(2018) The intra- and extracellular functions of ASC specks. Immunol Rev 281, 74–87.

35.

, Magupalli

, Ruan

, Yin

, Atianand

, Vos

, Schröder

, Fitzgerald

, Wu

, Egelman

(2014) Unified polymerization mechanism for the assembly of ASC-dependent inflammasomes. Cell 156, 1193–1206.

36.

Hopperton

, Mohammad

, Trepanier

, Giuliano

, Bazinet

(2018) Markers of microglia in post-mortem brain samples from patients with Alzheimer’s disease: A systematic review. Mol Psychiatry 23, 177–198.

37.

Izquierdo

, Attwell

, Madry

(2019) Ion channels and receptors as determinants of microglial function. Trends Neurosci 42, 278–292.

38.

Ossola

, Rifat

, Rowland

, Hunter

, Drinkall

, Bender

, Hamlischer

, Teall

, Burley

, Barker

, Cadwalladr

, Dickson

, Lawrence

JMK

, Harvey

JRM

, Lizio

, Xu

, Kavanagh

, Cheung

, Sheardown

, Lawrence

, Harte

, Brough

, Madry

, Matthews

, Doyle

, Page

, Powell

, Brice

, Bürli

, Carlton

, Dawson

(2023) Characterisation of C101248: A novel selective THIK-1 channel inhibitor for the modulation of microglial NLRP3-inflammasome. Neuropharmacology 224, 109330.

39.

McKeel

(2004) The Role of Agonal Factors in Human Postmortem CNS Research. Washington University School of Medicine, St. Louis. https://files.alz.washington.edu/presentations/2004/summer/McKeel.pdf

40.

Dai

, Chen

, Dai

, Jiang

, Tian

, Liu

, Xu

, Li

, Zhou

, Liu

, Chen

(2021) Agonal factors distort gene-expression patterns in human postmortem brains. Front Neurosci 15, 614142.

41.

Moore

, Le

, Fan

(2013) DNA methylation and its basic function. Neuropsychopharmacolog 38, 23–38.

42.

Matt

, Lawson

, Johnson

(2016) Aging and peripheral lipopolysaccharide can modulate epigenetic regulators and decrease IL-1β promoter DNA methylation in microglia. Neurobiol Aging 47, 1–9.

43.

Gasparoni

, Bultmann

, Lutsik

, Kraus

TFJ

, Sordon

, Vlcek

, Dietinger

, Steinmaurer

, Haider

, Mulholland

, Arzberger

, Roeber

, Riemenschneider

, Kretzschmar

, Giese

, Leonhardt

, Walter

(2018) DNA methylation analysis on purified neurons and glia dissects age and Alzheimer’s disease-specific changes in the human cortex. Epigenetics Chromatin 11, 41.

44.

Yin

, Morgunova

, Jolma

, Kaasinen

, Sahu

, Khund-Sayeed

, Das

, Kivioja

, Dave

, Zhong

, Nitta

, Taipale

, Popov

, Ginno

, Domcke

, Yan

, Schübeler

, Vinson

, Taipale

(2017) Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science 356, eaaj2239.

45.

Zhu

, Wang

, Qian

(2016) Transcription factors as readers and effectors of DNA methylation. Nat Rev Genet 17, 551–565.

46.

Kim

(2009) FOXP3 and its role in the immune system. Adv Exp Med Biol 665, 17–29.

47.

Pereira

LMS

, Gomes

STM

, Ishak

, Vallinoto

ACR

(2017) Regulatory T Cell and Forkhead Box Protein 3 as Modulators of Immune Homeostasis. Front Immunol 8, 605.

48.

Di Benedetto

, Burgaletto

, Carta

, Saccone

, Lempereur

, Mulas

, Loreto

, Bernardini

, Cantarella

(2019) Beneficial effects of curtailing immune susceptibility in an Alzheimer’s disease model. J Neuroinflammation 16, 166.

49.

Dhar

, Saha

, Mitra

, Nag Chaudhuri

(2021) DNA methylation and regulation of gene expression: Guardian of our health. Nucl Int J Cytol Allied Top 64, 259–270.

50.

, Nehorai

, Dougherty

(2013) Cell type specific analysis of human brain transcriptome data to predict alterations in cellular composition. Syst Biomed Austin Tex 1, 151–160.