The Mechanism of GABA in Attenuating Neuroinflammation in Alzheimer’s Disease: CP/CEBPα/miR-34a-Mediated Suppression of HDAC2/3 in Astrocytes

Abstract

As a widely available dietary supplement, γ-Aminobutyric acid (GABA) exhibits potential for early intervention against Alzheimer’s disease (AD). This study demonstrates that GABA alleviates AD neuroinflammation, and its suppression of astrocytic pro-inflammatory cytokine expression through histone deacetylase (HDAC2/3) inhibition contributes to this effect. Here, in both the cerebral cortex of AD mice and Aβ-exposed U251 cells, pro-inflammatory cytokines and HDAC2/3 expression levels were elevated, whereas the levels of creatine phosphate (CP), CCAAT/enhancer-binding protein α (CEBPα) and microRNA34a (miR-34a) were decreased. GABA treatment counteracted these alterations. Silencing HDAC2 or HDAC3 suppressed pro-inflammatory cytokines. Transfection with miR-34a mimics suppressed pro-inflammatory cytokines and HDAC2/3 expression in U251 cells, while miR-34a inhibitors had the opposite effect. A luciferase reporter assay confirmed HDAC2 as a direct miR-34a target via 3′UTR binding. Knockdown of CEBPα suppressed miR-34a expression, thereby elevating HDAC2/3 and pro-inflammatory cytokine expression in U251 cells. In CP-treated U251 cells, CEBPα and miR-34a expression was elevated, while pro-inflammatory cytokine and HDAC2/3 expression was down-regulated. In conclusion, GABA alleviates neuroinflammation in AD model mice. This effect may be partially attributed to its suppression of astrocyte-derived pro-inflammatory cytokine expression via HDAC2/3 inhibition. The CP/CEBPα/miR-34a pathway mediates the inhibitory effect of GABA on HDAC2/3 expression.

1. Introduction

γ-Aminobutyric acid (GABA), a naturally occurring bioactive amino acid ubiquitous across plant and animal species, has been developed as a functional food ingredient and is widely used globally [1]. Its favorable safety profile has led to its approval as a food additive by regulatory bodies such as the U.S. FDA, with a recommended supplemental dose ranging from 15 to 2250 mg/day for humans. This approval has catalyzed its widespread incorporation into various consumer products, including functional beverages, dairy alternatives, baked goods, and sleep-aid formulations. GABA is naturally present in certain foods, notably in fermented products like kimchi and yogurt, as well as in germinated grains such as brown rice. However, its concentration in natural sources is generally low, making direct extraction inefficient and costly. Currently, microbial fermentation is the predominant method for producing food-grade GABA, which is widely used as a food additive or functional ingredient. GABA exhibits a broad spectrum of physiological benefits, including the promotion of sleep, reduction in anxiety, and modulation of blood pressure and glucose metabolism [1].

As the primary inhibitory neurotransmitter in the central nervous system, GABA is essential for maintaining brain health and function [2]. Given its fundamental role in maintaining brain homeostasis, GABA has emerged as a promising dietary agent not only for modulating pathological processes but also for potentially preventing or delaying the onset of neurodegenerative disorders, particularly Alzheimer’s disease (AD). AD is an age-associated neurodegenerative condition defined by progressive cognitive and behavioral deterioration. The core neuropathological hallmarks of AD are characterized by extracellular amyloid-β (Aβ) deposits and intracellular neurofibrillary tangles (NFTs) of hyperphosphorylated tau [3]. The intricate pathogenesis of AD complicates therapeutic development. Current pharmacotherapies offer only symptomatic relief without halting disease progression and are often limited by adverse effects [4]. Thus, identifying safe and effective early interventions remains imperative. Our previous studies demonstrate that oral GABA supplementation (2 mg/mL in drinking water for mice, equivalent to a human dose of approximately 1751 mg/day) ameliorates cognitive deficits and pathological features in AD mouse models [5], suggesting its potential as an early intervention. However, the precise molecular mechanisms underlying GABA-mediated attenuation of AD progression require further elucidation.

Although the blood–brain barrier (BBB) inherently restricts GABA passage, GABA transporter 2 (GAT-2) on cerebral microvascular endothelial cells enables bidirectional transport [6,7]. However, physiological conditions favor a net GABA efflux from the brain to the periphery due to high cerebral GABA concentrations [6]. Notably, cerebral GABA levels decline significantly with aging and are further reduced in AD patients relative to controls [8,9,10]. Astrocytes, which highly express the GABA transporter GAT-3, not only recycle synaptically released GABA but also preferentially utilize peripherally derived GABA [11,12]. This implies GABA may serve important functions in astrocytes independent of classical neurotransmitter transmission.

Neuroinflammation is a central driver of AD pathogenesis, in which microglia-mediated chronic inflammation accelerates core pathologies like Aβ deposition [13,14]. Increasing evidence highlights the critical role of aberrantly activated astrocytes in AD neuroinflammation [15,16]. Specifically, these cells release pro-inflammatory cytokines such as interleukin-1β (IL-1β), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNFα) [15,16]. This exacerbates microglial inflammatory cascades, creating a self-sustaining vicious cycle of neuroinflammation that fuels AD progression [15,16]. Evidence indicates that GABA suppresses the release of IL-6 from astrocytes following LPS stimulation [17], suggesting its potential to mitigate AD-associated neuroinflammation by modulating astrocytic activation. Nevertheless, the molecular mechanisms remain unclear.

Histone deacetylases (HDACs) are key enzymes regulating histone acetylation. Studies indicate that pan-HDAC inhibitors reduce the levels of IL-1β, IL-6 and TNFα [18,19]. Moreover, substantial evidence points to the contribution of HDACs to AD pathogenesis, with HDAC2 and HDAC3 exhibiting the strongest pathological links and thus representing promising therapeutic targets [20,21]. Our previous research reveals that GABA inhibits HDAC2 and HDAC3 expression [22]. Thus, GABA may attenuate AD progression by suppressing HDAC2/3-mediated inflammatory cytokine release from astrocytes. Nevertheless, how GABA regulates HDAC2/3 expression in astrocytes remains unclear.

Bioinformatics analyses predict microRNA34a (miR-34a) targets HDAC2/3. The miR-34a promoter contains a CCAAT box, potentially binding CCAAT/enhancer-binding protein (CEBP) transcription factors. Of the six CEBP family members highly expressed in the brain, we have previously identified CEBPα as being down-regulated in AD model mice [5]. Furthermore, studies in leukemia indicate that CEBPα regulates miR-34a expression [23]. Creatine phosphate (CP), a high-energy phosphate compound critical for cellular energy transfer, inhibits CEBP homologous protein [24,25,26], suggesting potential suppression of CEBPα. While GABA modulates brain energy metabolism, it remains unexplored whether GABA modulates CP levels to regulate CEBPα expression, thereby up-regulating miR-34a and ultimately suppressing HDAC2/3 expression.

This study demonstrates that GABA, a prominent functional food component, attenuates neuroinflammation in AD. Mechanistically, GABA suppresses HDAC2/3 expression in astrocytes through the CP/CEBPα/miR-34a signaling axis, leading to reduced pro-inflammatory cytokine expression, which may contribute to its anti-inflammatory effects in AD. Our work identifies a novel mechanism for GABA-mediated attenuation of AD progression, advancing the scientific rationale for using GABA-enriched foods or supplements as part of nutritional interventions against AD.

2. Materials and Methods

2.1. Reagents

GABA was purchased from Shanghai Sangon Biotech Co., Ltd. (Shanghai, China). It was food-grade GABA produced through microbial fermentation. The toxic Aβ_25–35 fragment (American Peptide Company, 131602-53-4, Sunnyvale, CA, USA) was solubilized in sterile water and aggregated at 37 °C for 7 days prior to use. Commercial sources of reagents and kits were as follows: the siRNA transfection reagent from Guangzhou RiboBio Co., Ltd. (Guangzhou, China); rabbit anti-CEBPα polyclonal antibody from Wuhan Huamei Laboratory Equipment Co., Ltd. (Wuhan, China); a series of rabbit polyclonal antibodies (anti-β-ACTIN; anti-IL-1β, IL-6, TNFα, and HDAC2; anti-HDAC3); and all secondary antibodies (goat anti-rabbit IgG, FITC-donkey anti-rabbit IgG, and CY3-donkey anti-rat IgG) from Shanghai Sangon Biotech Co., Ltd. (Shanghai, China). Cell culture medium and dual-luciferase reporter gene assay kits were sourced from Hyclone Inc. (Logan, UT, USA) and Shanghai Hanheng Biotechnology (Shanghai, China), respectively.

2.2. In Vivo Study

2.2.1. Animals and Treatment

As previously described [5], six-month-old APP/PS1 and wild-type (WT) littermate mice (Beijing Huafukang Biotechnology Co., Ltd., Beijing, China) were randomly divided into four groups (n = 10, 5 males and 5 females): AD, AD + GABA, WT, and WT + GABA. The GABA-treated groups received GABA-supplemented drinking water (2 mg/mL) for six months, while controls received plain water. Under standard housing conditions, followed by a 12 h fast, the mice were euthanized. For molecular analyses, brain tissues were harvested and stored at −80 °C. The Animal Care and Use Committee of China Medical University approved all experimental procedures involving animals.

2.2.2. qRT-PCR Assay of Cortex Samples

Total mRNA and microRNAs were extracted and reverse-transcribed according to our previously described methods [5]. Following cDNA synthesis, quantitative real-time PCR was performed using established methods [5]. All primers were obtained from Sangon Biotech (Dalian, China), with sequences detailed in

Table 1. List of qRT-PCR primer sequences for cortex samples.

Category	Primer Sequence (5′–3′)	Length
IL-1β	F: CACTACAGGCTCCGAGATGAACAAC R: TGTCGTTGCTTGGTTCTCCTTGTAC	145 bp
IL-6	F: CTTCTTGGGACTGATGCTGGTGAC R: TCTGTTGGGAGTGGTATCCTCTGTG	91 bp
TNFα	F: CCCTCACACTCAGATCATCTTCT R: GCTACGACGTGGGCTACAG	123 bp
CEBPα	F: TCGGTGGACAAGAACAGCAACG R: CGGTCATTGTCACTGGTCAACTCC	140 bp
HDAC2	F: AGTGGAGATGAGGATGGAGAAGACC R: GCAACATTCCTACGACCTCCTTCAC	128 bp
HDAC3	F: ATCCGCCAGACAATCTTTGA R: CTCGGGACCTCTCTCTTCAG	132 bp
β-actin	F: CATCCGTAAAGACCTCTATGCCAAC R: ATGGAGCCACCGATCCACA	171 bp
miR-34a	F: CTGGCAGTGTCTTAGCTGGTTGT R: CGCTTCACGAATTTGCGTGTCAT	_
U6	F: GCTTCGGCAGCACATATACTAAAAT R: CGCTTCACGAATTTGCGTGTCAT	_

. Gene expression levels were normalized to β-actin mRNA or U6 (n = 10). We adopted the comparative CT (2^−△△CT) method for data analysis.

2.2.3. Western Blot Analysis of Cortex Samples

Western blot analysis was conducted following a previously established protocol [5]. In brief, cerebral cortex samples were thawed and homogenized in RIPA buffer containing 0.1% protease inhibitor (Amerso, Solon, OH, USA). Supernatant protein concentrations were quantified using the Bradford assay with a bovine serum albumin standard curve. Equal amounts of soluble protein (40 μg) were used for Western blot test, using rabbit anti-IL-1β (1:1000), anti-IL-6 (1:1000), anti-TNFα (1:1000), anti-CEBPα (1:1000), anti-HDAC2 (1:1000), anti-HDAC3 (1:1000), or anti-β-actin (1:1000) antibody. We expressed the Western blot results as fold changes relative to the WT group after normalization to β-actin.

2.2.4. ELISA Analysis of Cortex Samples

CP levels in the cerebral cortex were measured using an ELISA kit following the manufacturer’s instructions. Standard wells, sample wells, and blank wells were set up accordingly. The optical density (OD) of each well was measured at 450 nm. A standard curve was generated, and CP concentrations in cortical homogenates were calculated.

2.3. In Vitro Study

2.3.1. Cell Culture and Treatment

Human U251 cells (KCB2006107YJ, Chinese Academy of Sciences Cell Bank) were grown in complete DMEM/F12 medium (10% FBS, 1% penicillin-streptomycin) at 37 °C in a 5% CO₂ atmosphere. Following this culture, the cells were utilized for experiments, all of which were carried out in duplicate and independently replicated thrice.

For GABA and Aβ treatments, cells plated in six-well plates were incubated for 3 h in antibiotic-free medium containing 10 nM GABA (GABA and GABA + Aβ groups) or vehicle (control and Aβ groups). Subsequently, the Aβ and GABA + Aβ groups were subjected to a 24 h treatment with 10 μM Aβ_25–35. The concentrations of GABA (10 nM) and Aβ (10 μM) were selected based on the literature [5,22] and preliminary cell viability assays, which showed no significant impact on cell viability or morphology.

For CP treatment, cells were divided into control and CP treatment groups. The CP treatment group was incubated with medium containing a final concentration of 5 mM CP for 24 h. This concentration was selected based on cell viability assays confirming no significant effects on cell viability or morphology.

2.3.2. qRT-PCR, Western Blot and ELISA Analyses of Cells

qRT-PCR, Western blot, and ELISA analyses of cell samples were performed as described for cortex samples in Section 2.2.2, Section 2.2.3 and Section 2.2.4. Briefly, qRT-PCR was performed using primers listed in

Table 2. List of qRT-PCR primer sequences for U251 cells.

Category	Primer Sequence (5′–3′)	Length
IL-1β	F: GCCAGTGAAATGATGGCTTATT R: AGGAGCACTTCATCTGTTTAGG	85 bp
IL-6	F: CACTGGTCTTTTGGAGTTTGAG R: GGACTTTTGTACTCATCTGCAC	101 bp
TNFα	F: TGGCGTGGAGCTGAGAGATAACC R: CGATGCGGCTGATGGTGTGG	134 bp
CEBPα	F: GACAAGAACAGCAACGAGTAC R: TCATTGTCACTGGTCAGCTC	131 bp
HDAC2	F: AGGTTGAAGCCATTCTCCTG R: ATCCCAGCACTTTGGAAGG	179 bp
HDAC3	F: GAGGGATGAACGGGTAGACA R: CAGGTGTTAGGGAGCCAGAG	137 bp
β-actin	F: CATCCGTAAAGACCTCTATGCCAAC R: ATGGAGCCACCGATCCACA	171 bp
miR-34a	F: CTGGCAGTGTCTTAGCTGGTTGT R: CGCTTCACGAATTTGCGTGTCAT	_
U6	F: GCTTCGGCAGCACATATACTAAAAT R: CGCTTCACGAATTTGCGTGTCAT	_

(Sangon, Dalian, China). For Western blot, the primary antibodies employed (all at 1:1000 except where noted) targeted IL-1β, IL-6, TNFα, CEBPα, HDAC2, HDAC3, and β-actin (loading control). CP levels were measured by ELISA following the manufacturer’s instructions. Results from three independent experiments with duplicate measurements are presented as protein expression folds relative to the control.

2.3.3. Cell Transfection with miRNA Mimics and Inhibitors

Commercially synthesized miR-34a mimics, inhibitors, and their negative controls (Guangzhou RiboBio Co., Ltd., Guangzhou, China) were utilized for transfection. Briefly, after seeding in six-well plates, cells were transfected with 200 nM of each RNA oligonucleotide and incubated for 24 h at 37 °C in antibiotic-free medium, employing the riboFECT™ CP Transfection Kit (RiboBio) as per the protocol. Subsequently, the cells were harvested for the simultaneous extraction of total mRNA, miRNA, and protein. In addition, the expression of HDAC2, HDAC3, IL-1β, IL-6 and TNFα was analyzed at the mRNA level by qRT-PCR and at the protein level by Western blot, as described above. All Western blot data were quantified by densitometry and expressed as fold change over the respective control. The experiment included three biological replicates with technical duplicates.

2.3.4. Small Interfering RNA (siRNA)

To achieve knockdown HDAC2, HDAC3 and CEBPα, we performed transfection with gene-specific siRNAs. All siRNA oligonucleotides were purchased from Guangzhou RiboBio Co., Ltd. The sequences of the oligonucleotides are listed below: HDAC2: 5′-TCCGTAATGTTGCTCGATG-3′; HDAC3: 5′-GCATTGATGACCAGAGTTA-3′; CEBPα: 5′-ACGAGACGTCCATCGACAT-3′. siRNA transfection was performed using the method previously described in detail [5]. Knockdown efficiency was assessed by qRT-PCR and Western blot analyses. The mRNA and protein expression levels of HDAC2, HDAC3, IL-1β, IL-6 and TNFα, as well as the expression level of miR-34a, were measured by the above methods. All Western blot data were quantified by densitometry and expressed as fold change over the respective control. The experiment included three biological replicates with technical duplicates.

2.3.5. Dual-Luciferase Reporter Gene Assay

To create luciferase reporter constructs, we cloned the wild-type and mutant 3′UTR regions of HDAC2, which contain the miR-34a-3p binding sites, into the psiCHECK-2 vector (Hanheng Biotechnology, Shanghai, China). After seeding in 96-well plates at 70% confluence, 293T cells were co-transfected with a mixture containing 2 μg of either the wild-type or mutant reporter vector and 2 μg of miR-34a-3p mimic or negative control (NC) miRNA. Luciferase activity was measured after 48 h using a dual-luciferase assay system. Results are presented as the ratio of Renilla to Firefly luciferase activity. All transfections and assays were performed in duplicate and repeated in three independent experiments.

2.3.6. MTS Assay

Cell viability was measured with an MTS Cell Proliferation Assay Kit (CellTiter 96, Madison, WI, USA) according to our previously described methods [22]. In brief, U251 cells were seeded in 96-well culture microplates and allowed to attach overnight in the presence of FBS. The cells were then separately exposed to 0, 2.5, 5, 10, 20 or 40 mM CP for 24 h. Then, 20 µL of MTS was added to each well, and the plates were incubated for 4 h. Absorbance was measured at 492 nm with a Microplate Reader (BioTek Instruments, Inc., Winooski, VT, USA). The experiment was performed in duplicate and repeated three times.

2.4. Statistical Analyses

Statistical analysis was performed with SPSS 20.0 (IBM Corp., Armonk, NY, USA), and data are presented as means ± standard deviation (SD). Inter-group differences were evaluated using Student’s t-test for two groups or one-way ANOVA with Fisher’s LSD post hoc test for multiple groups. A p-value below 0.05 was deemed statistically significant.

3. Results

3.1. GABA Suppresses the Increase in Pro-Inflammatory Cytokines in AD Models

To investigate whether GABA ameliorates AD by modulating neuroinflammation, we assessed the mRNA and protein expression levels of pro-inflammatory cytokines (IL-1β, IL-6 and TNFα) in vivo (Figure 1A–C) and in vitro (Figure 1D,E). In AD model mice, a marked increase was observed in the expression levels of IL-1β, IL-6, and TNF-α compared to WT mice (p < 0.01). GABA treatment significantly reduced the expression levels of IL-1β, IL-6 and TNFα in AD mice relative to untreated AD controls. Consistently, in vitro experiments in Aβ-exposed cells showed a marked increase in IL-1β, IL-6, and TNFα expression compared to controls (p < 0.01). This upregulation was significantly attenuated by GABA pretreatment (p < 0.01). GABA pretreatment alone significantly reduced the expression levels of IL-1β, IL-6 and TNFα in cells compared to controls (p < 0.05; p < 0.01).

3.2. GABA Suppresses the Upregulation of HDAC2/3 Expression in AD Models

To investigate the underlying mechanism through which GABA modulates the expression of pro-inflammatory cytokines in AD, we analyzed HDAC2 and HDAC3 expression both in vivo (Figure 2A–C) and in vitro (Figure 2D,E). A significant upregulation of HDAC2/3 mRNA and protein expression was observed in the cerebral cortex of AD model mice relative to WT controls (p < 0.01). GABA treatment markedly suppressed this increase in AD mice (p < 0.01). Consistently, in vitro experiments revealed that Aβ exposure significantly elevated HDAC2/3 expression in cells relative to controls (p < 0.01). This elevation was attenuated by GABA pretreatment (p < 0.01). Furthermore, GABA pretreatment significantly reduced the mRNA and protein expression levels of HDAC2/3 compared to the control group (p < 0.05; p < 0.01).

3.3. Silencing HDAC2 or HDAC3 Caused the Downregulation of Pro-Inflammatory Cytokine Expression in U251 Cells

To further determine whether HDAC2/3 regulates the expression of pro-inflammatory cytokines, endogenous HDAC2 and HDAC3 were knocked down in U251 cells using specific siRNA duplexes (Figure 3). Silencing of either HDAC2 or HDAC3 led to a significant reduction in IL-1β, IL-6 and TNFα expression levels compared to scrambled siRNA controls (p < 0.01).

3.4. GABA Suppresses the Downregulation of miR-34a, CEBPα and CP Expression Levels in AD Models

To explore the upstream mechanisms through which GABA modulates HDAC2/3, we assessed miR-34a, CEBPα and CP levels both in vivo (Figure 4A,B) and in vitro (Figure 4C–F) across experimental groups. AD model mice showed notably lower levels of miR-34a and CP in the cerebral cortex compared to WT controls (p < 0.01). GABA treatment notably increased both miR-34a expression and CP levels in AD mice (p < 0.01). Aβ exposure led to a marked reduction in miR-34a, CEBPα and CP levels relative to untreated controls (p < 0.05; p < 0.01). Conversely, GABA pretreatment in the absence of Aβ significantly elevated these measures relative to controls (p < 0.05; p < 0.01). Importantly, in Aβ-exposed cells, co-treatment with GABA significantly rescued the suppression of miR-34a, CEBPα, and CP levels observed with Aβ treatment alone (p < 0.05; p < 0.01).

3.5. miR-34a Regulates HDAC2/3 Expression in U251 Cells

We performed transfection of U251 cells with a miR-34a mimic or inhibitor to further investigate its regulation of HDAC2/3 (Figure 5). Transfection with the miR-34a mimic significantly reduced the expression of HDAC2/3 and the pro-inflammatory cytokines (IL-1β, IL-6, TNF-α) at both mRNA and protein levels (p < 0.01), whereas the miR-34a inhibitor conversely increased their expression (p < 0.05; p < 0.01). To validate direct targeting, dual luciferase reporter assays were performed. Wild-type (HDAC2-WT) and mutant (HDAC2-Mut) luciferase reporter vectors were constructed. Using site-directed mutagenesis, 20 base pairs within the putative miR-34a binding sites of the HDAC2 3′UTR were mutated in the mutant vector (Figure 5G,H). We co-transfected HDAC2-WT or HDAC2-Mut into 293T cells with miR-34a mimics, and then the luciferase activity was examined. The result showed that the luciferase activity of wild-type vectors was significantly reduced (p < 0.01) compared with the control (Figure 5I).

3.6. Silencing CEBPα Caused the Downregulation of miR-34a Expression in U251 Cells

To determine the regulatory role of CEBPα in miR-34a expression, we knocked down endogenous CEBPα mRNA using a specific siRNA duplex. As shown in Figure 6, in CEBPα-silenced cells, miR-34a expression was strongly decreased (p < 0.01), and the expression of HDAC2/3, IL-1β, IL-6 and TNFα was obviously increased (p < 0.01), compared with the cells exposed to the scrambled siRNA construct (scrambled cells).

3.7. Upregulation of CEBPα by CP in U251 Cells

We determined cell viability by the MTS assay. Treatment with CP at concentrations ranging from 2.5 to 40 mM induced a progressive reduction in cell viability in a dose-dependent manner. A significant decrease in viability was observed at CP concentrations of 20 mM and 40 mM compared to the control group (p < 0.05, p < 0.01; Figure 7A). Exposure to 10 mM CP significantly upregulated the expression of CEBPα and miR-34a, while conversely downregulating HDAC2/3, IL-1β, IL-6 and TNFα compared to the control (p < 0.01; Figure 7B–E).

4. Discussion

Accumulating evidence implies that neuroinflammation is a crucial driver in the pathogenesis of AD [13,14]. Our findings confirm the significant upregulation of pro-inflammatory cytokines IL-1β, IL-6 and TNF-α in the AD brain, consistent with numerous studies [27,28,29]. As a widely available dietary supplement, GABA has been shown to possess anti-inflammatory properties [30,31]. Under physiological conditions, GAT-2 primarily mediates net GABA efflux due to high brain GABA concentrations [6]. Correspondingly, we found that GABA administration did not alter the expression of pro-inflammatory cytokines in wild-type mice. The present study also found that GABA suppressed IL-1β, IL-6 and TNFα expression in the cerebral cortex of AD mice, confirming its anti-inflammatory role in AD. Targeting inflammation is considered a promising strategy to improve AD symptoms [13,14]. Our previous study indicated that drinking water containing GABA ameliorated cognitive deficits in AD mice [5]. These findings suggested that GABA alleviates AD progression at least in part through mitigating neuroinflammation.

Astrocytes are key regulators of inflammatory cascades in AD. These cells are activated in the early stage of AD and release pro-inflammatory cytokines that propagate neuronal damage and amplify local inflammation through microglial crosstalk [15,16]. Astrocytes have been reported to preferentially take up peripherally derived GABA via their high GAT-3 expression, in addition to recycling neurotransmitter from synapses [11,12]. Considering that cerebral GABA levels are significantly decreased in AD [8,9,10], it is speculated that peripherally derived GABA may be actively taken up by astrocytes via transporter GAT-3 to compensate for neuronal GABA deficiency. This study demonstrated that GABA significantly inhibited the ability of Aβ to upregulate the levels of IL-1β, IL-6 and TNFα in U251 cells. The above indicates that GABA likely attenuates neuroinflammation in AD in part through inhibiting pro-inflammatory cytokine expression levels in astrocytes.

Of note, the in vivo data from cortical tissue homogenates reflect an integrated response across multiple cell types. While our in vitro data indicate that astrocytes contribute to these effects, the specific contributions of other cell types such as neurons and microglia remain to be determined. Additionally, given the important role of microglia in AD neuroinflammation, future studies should examine microglial responses and their potential interaction with astrocytes to better understand how GABA exerts its anti-inflammatory effects. Based on the available evidence [6,7,9,11], it has been suggested that the GABA-deficient state in AD may enhance central uptake of circulating GABA via transporters such as GAT-2 and GAT-3. However, the precise brain bioavailability, optimal dosing, and pharmacokinetics of oral GABA supplementation have not been established. These aspects warrant further investigation to better understand the physiological relevance of GABA as a dietary intervention for AD.

GABA exerts its physiological function mainly by binding to its receptors. However, Lee and Kuhn et al. demonstrated that GABA receptor agonists only partially recapitulate the anti-inflammatory effects of GABA [17,32]. It is suggested that GABA inhibits pro-inflammatory cytokines through additional receptor-independent pathways. Our previous study demonstrated that GABA inhibits HDACs in a receptor-independent manner [22]. Notably, HDAC2 and HDAC3 emerge as critical molecular determinants in AD pathogenesis, with established roles in driving AD-related neuropathology [20,21]. In this study, GABA significantly inhibited the upregulation of HDAC2/3 in both in vivo and in vitro models of AD. Further investigation revealed that silencing HDAC2 or HDAC3 significantly elevated IL-1β, IL-6 and TNFα expression levels in U251 cells. Consequently, GABA attenuates pro-inflammatory cytokine expression through a pathway involving the downregulation of HDAC2 and HDAC3. It should be noted that the in vitro experiments were performed using the U251 astrocytoma cell line, which may not fully replicate the physiological characteristics of primary astrocytes. Future validation using primary astrocytes is needed to further strengthen the conclusions.

Bioinformatic analysis predicted miR-34a as a putative simultaneous regulator of both HDAC2 and HDAC3, which was experimentally validated by using the miR-34a mimic and inhibitor. The luciferase reporter assay was employed to validate the specific binding of miR-34a to HDAC2, with no significant interaction detected with HDAC3, suggesting distinct miR-34a regulatory mechanisms for HDAC2 and HDAC3. Meanwhile, the present study demonstrated a marked decrease in miR-34a expression in the cerebral cortex of AD model mice and Aβ-treated U251 cells, which is in accordance with some studies [33,34]. These findings suggest that the downregulation of miR-34a expression is, at least partly, attributed to the upregulation of HDAC2 and HDAC3 in AD. In addition, this study showed that GABA alleviated the downregulation of miR-34a expression both in vivo and in vitro. It is indicated that miR-34a acts as a key mediator by which GABA suppresses HDAC2/3 expression, ultimately leading to the downregulation of pro-inflammatory cytokines in astrocytes.

CEBPα is a subfamily of basic region-leucine zipper transcription factors. Replicating our previous finding in the AD mouse cortex [5], this study demonstrated that CEBPα downregulation was also evident in U251 cells exposed to Aβ. Further silencing experiments identified CEBPα as a transcriptional activator of miR-34a expression. Accordingly, HDAC2/3 and pro-inflammatory cytokine expression levels were decreased in CEBPα-silenced U251 cells. Notably, GABA treatment reversed CEBPα downregulation in both in vivo and in vitro models of AD. These findings indicate that CEBPα upregulates miR-34a expression, a mechanism through which GABA inhibits HDAC2/3 expression.

The expression of CEBP homologous protein is known to be regulated by CP [24]. Here, we found that CP can also increase CEBPα expression, which accordingly upregulated miR-34a and decreased expression of HDAC2/3 and pro-inflammatory cytokines. However, reported CP levels in AD brains are inconsistent across studies [35,36,37,38]. In line with the majority of reports [35,36,37], we observed reduced CP levels in both in vivo and in vitro models of AD. These discrepancies may be attributed to variations in the examined brain regions and the methodological approaches employed. In this study, GABA was found to increase the CP levels in U251 cells. Furthermore, our results demonstrated that GABA rescued the CP reduction in both in vivo and in vitro models of AD. These findings suggest that GABA upregulates CEBPα expression through elevating CP levels in astrocytes.

The mechanism by which GABA increases CP levels remains to be fully elucidated. GABA is related to energy metabolism, which can be catalyzed by GABA aminotransferase to enter the tricarboxylic acid cycle and promote ATP production. The ATP generated through this pathway can then be utilized for the synthesis of CP from creatine in the mitochondrial membrane gap. This may account for the observed increase in CP following GABA treatment. However, it is necessary to validate this mechanism in future studies. Additionally, we cannot exclude the possibility of feedback loops, and further investigation is needed to determine whether downstream signals might in turn influence CP metabolism.

5. Conclusions

Collectively, our results indicate that GABA, a food-grade bioactive compound, alleviates neuroinflammation in AD model mice. This effect may be partially attributed to its suppression of astrocyte-derived pro-inflammatory cytokine expression through the inhibition of HDAC2 and HDAC3. Furthermore, we identified the CP/CEBPα/miR-34a signaling axis as the pathway through which GABA suppresses HDAC2/3 expression. However, it is prudent to be able to extrapolate the conclusions from experimental AD models to humans. Collectively, this study elucidates a novel GABA-driven mechanism underlying AD neuroprotection. Given its safety profile and regulatory approval as a food additive, this study provide a theoretical basis for considering GABA as a dietary supplement for the prevention of AD.