ZFP36-mediated ZBP1 degradation inhibits microglia pro-inflammatory and NLRP3 inflammasome activation in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is a heterogeneous disease with limited treatment efficacy. Identifying novel molecular targets and mechanisms is therefore crucial for developing therapeutic strategies. Zinc finger protein 36 (ZFP36) has not been reported in AD. This study found that the hippocampus of APP/PS1 mice showed ZFP36 upregulation. Using recombinant adeno-associated virus to overexpress ZFP36 improved the cognitive function of APP/PS1 mice, as assessed by Morris maze and Y maze tests. Furthermore, ZFP36 overexpression reduced Aβ deposition, expression of pro-inflammatory markers, and inhibited NLRP3 inflammasome activation in the hippocampus. These inhibitory effects of ZFP36 overexpression on the aforementioned proteins were also observed in Aβ₁₋₄₂-treated BV-2 cells. mRNA sequencing identified Z-DNA Binding Protein 1 (ZBP1) as a target of ZFP36. After ZFP36 overexpression, ZBP1 was downregulated in the hippocampus and Aβ_1-42-treated BV-2 cells. The interaction between ZFP36 and ZBP1 RNA was verified by RIP-PCR, and ZFP36 was shown to promote the degradation of ZBP1 mRNA. The inhibitory effects of ZFP36 on the NLRP3 inflammasome activation and microglial pro-inflammatory activation was reversed by ZBP1 overexpression. In summary, ZFP36 inhibits microglia pro-inflammatory and NLRP3 inflammasome activation through promoting the degradation of ZBP1 mRNA, thereby ameliorating cognitive deficits of APP/PS1 mice.

Graphical Abstract

Schematic diagram of the process of exploring the effect of ZFP36 on microglia activation and NLRP3 inflammasome activation in APP/PS1 transgenic mice and Aβ_1-42-treated BV-2 cells.

Supplementary information

The online version contains supplementary material available at 10.1007/s10565-026-10139-6.

Introduction

Alzheimer's disease (AD) often characterized by progressively severe cognitive impairment (including memory disorders, learning disabilities, attention disorders, spatial cognitive disorders, and problem-solving disorders) (Lane et al. 2018). AD mostly affects people over 60 years old, but in recent years, younger patients have also gradually increased. It is generally believed that AD is a complex and heterogeneous disease, and multiple factors may be involved in the disease, such as genetic factors, neurotransmitters, immune factors, and environmental factors (Tiwari et al. 2019). There is no specific drug for AD, and current treatments are less effective. Therefore, further exploration of the pathogenesis of AD is very important for the future treatment of AD.

Activation of the NLRP3 inflammasome in microglia, triggered by AD-related pathological agents such as Aβ oligomers and phosphorylated tau, constitutes a sustained driver of neuroinflammation and neuronal damage (Ising et al. 2019). This activation promotes the cleavage of pro-IL-1β and pro-IL-18 into their active forms and can induce pyroptotic cell death, thereby amplifying the inflammatory cascade. Pharmacological or genetic inhibition of NLRP3 has been demonstrated to effectively suppress microglial pro-inflammatory activation and ameliorate pathological hallmarks in AD models (Heneka et al. 2013; Wang et al. 2023). Therefore, targeting the NLRP3 pathway represents a promising strategy to modulate neuroinflammation. ZFP36 (Zinc finger protein 36, also known as Tristetraprolin) is an RNA-binding protein that promotes deadenylation and degradation of target mRNA by binding to the adenosine-uridine element rich in the 3’-UTR of target mRNA (Tiedje et al. 2016). Previous study showed that ZFP36 alleviated neuronal apoptosis in cerebral ischemia re-injury (Guo et al. 2022). ZFP36 can significantly inhibit the activation of NLRP3 inflammasome (Haneklaus et al. 2017). In addition, the expression of ZFP36 is significantly upregulated in the cerebral tissue of Aβ_1–42-induced AD rats (GSE129055, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE129055) and human AD patients (GSE5281, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE5281), but the role of ZFP36 in AD has not yet been elucidated. We speculate that the compensatory increase in ZFP36 may be involved in alleviating the development of AD by inhibiting microglial activation and NLRP3 inflammasome activation. This project used mRNA-seq to screen out the differential genes that might be regulated by ZFP36 in microglia.

Z-DNA Binding Protein 1 (ZBP1) participates in promoting the development process of AD (Guo et al. 2023). ZBP1 promotes NLRP3 inflammasome activation in influenza virus-infected cells (Kuriakose et al. 2016). ZBP1 is involved in regulating inflammatory responses in microglia (Kaurani et al. 2024). The mRNA-seq results showed that compared with the Aβ_1–42 + Vector group, the expression of ZBP1 was significantly down-regulated in the Aβ_1–42 + ZFP36 group. ENCORI database predicted that the binding position of ZFP36 to ZBP1 mRNA is in the 3’-UTR. Therefore, ZBP1 may serve as a downstream factor of ZFP36 and participate in regulating the microglia polarization.

Methods

Animal experiments

Seven-month-old APP/PS1 transgenic mice on a C57BL/6 J background, along with wild-type (WT) littermate controls of the same genetic background were used in this work. Both groups consisted of equal numbers of males and females. APP/PS1 transgenic mice were injected with recombinant adeno-associated virus serotype 2 (rAAV2) overexpressing ZFP36 ((1.5 × 10⁸ vg/mL, 5 µL on each side) in the bilaterally hippocampus (anterior–posterior position −2.0 mm, medial–lateral position ± 1.5 mm, 1.5 mm dorsoventral from the bregma) (Rivera-Escalera et al. 2014). rAAV2-vector was used as a control viral vector. After 4 weeks of injection, the following behavioral experiments were performed on the mice according to reported literatures (Gao et al. 2022; Kraeuter et al. 2019). Finally, Mice were euthanized with excessive carbon dioxide, and hippocampal tissue was collected. Mice were transcardially perfused with 0.9% sodium chloride injection, immediately followed by perfusion with 4% paraformaldehyde to rapidly and uniformly preserve the tissue architecture for subsequent analysis.

Behavioral testing

The Y-maze behavioral test was employed to assess short-term working memory (spontaneous alternation test) and long-term spatial reference memory (novel arm recognition test) in mice. All tests were conducted in a quiet, uniformly lit laboratory room. Mice were acclimated to the testing room for at least 1 h prior to each session, and all procedures were performed by an experimenter blinded to the group assignments. For the spontaneous alternation test, a mouse was placed at the distal end of one of the three labeled arms (A, B, or C), facing the center, and allowed to explore freely for 8 min. A valid arm entry was defined as all four paws crossing into an arm. The total number of arm entries and the number of sequential entries into three different arms (alternations) were recorded using Xeye Aba software. The percentage of spontaneous alternation was calculated using the formula: Alternation % = (Number of alternations)/(Total arm entries—2) × 100. Mice with fewer than 8 total arm entries were excluded from analysis to ensure valid exploratory behavior. Twenty-four hours after the spontaneous alternation test, the novel arm recognition test was conducted. During the training phase, one arm was blocked (novel arm), and the mouse was placed in an open arm for 15 min of free exploration. After a 1-h interval, the test phase began: the blockade was removed, and the mouse was returned to the same starting arm for a 5-min free exploration session. The total distance traveled and the exploration preference for each arm were recorded. The maze was thoroughly cleaned with 70% ethanol after each mouse to eliminate olfactory cues.

Morris water maze test was used to evaluate mice spatial learning and memory. On the first day of the experiment, a visible platform (the platform is above the water) was used to record the time it took the mice to find the platform and their swimming speed, and the results were provided to screen mice that meet the requirements in terms of vision, motor ability, and desire to find the platform. From the second day of the experiment, a concealed platform method was used to conduct positioning and navigation training for 5 days, and the latency time was recorded. On the 7th day of the experiment, the platform was removed and a 60 s space exploration experiment was conducted. The number of times the mice crossed the platform and the time spent exploring in the target quadrant were recorded.

Vector construction

Recombinant adeno-associated viruses (rAAVs) were constructed and prepared using AAV-293 cells (purchased from iCell Biotechnology Co., Ltd., iCell-h404); the cells were cultured in a humidified incubator at 37 °C with 5% CO₂, and transfection was carried out when the cell confluence reached 60%. Viral packaging was performed with a three-plasmid co-transfection system, where the shuttle vector was pAAV-CMV (Takara, No. 6230) with the ZFP36 gene inserted, at a dosage of 7 μg per 10 cm culture dish, and the helper vectors were pHelper and pRC2-mi342 serotype plasmid (Takara, No. 6230) with dosages of 14 μg and 7 μg per 10 cm culture dish, respectively. The transfection reagent used was Lipofectamine 3000 (Invitrogen, L3000015). The viruses were harvested at 72 h post-transfection; the cells were lysed by repeated freeze–thaw cycles, the impurities were removed by filtration through a 0.45 μm filter, and the high-titer recombinant adeno-associated viruses were finally obtained via ultracentrifugation for concentration and purification. The overexpression plasmids were purchased from Youbao Biotechnology (Changsha, China), and siZFP36 was purchased from Jiangsu Genecefe Biotechnology Co., Ltd (Jiangsu, China).

Cell culture and treatment

BV-2 cells were purchased from iCell Bioscience. BV-2 cells were cultured in DMEM containing 10% fetal bovine serum in an incubator at 37 °C and 5% CO₂. An in vitro model of Aβ-induced inflammatory activation was established by treating BV-2 cells with 5 μM Aβ_1–42 for 24 h. followed by relevant assays. Prior to this treatment, the cells were transfected with ZFP36-overexpressing plasmids or siZFP36 for 48 h. For mRNA stability analysis, following Aβ_1–42 treatment, cells were incubated with 2 µg/mL actinomycin D for 0, 2, 4, and 6 h. The mRNA levels of ZBP1 were subsequently quantified by quantitative real-time PCR.

Thioflavin-S staining

Hippocampal tissue was dehydrated through graded alcohol, embedded in paraffin, and cut into coronal sections of 5 μm. After dewaxing to water, tissue slides were incubated with Thioflavin-S solution (5 g of Thioflavin-S was diluted in 100 ml 50% ethanol) for 8 min. Subsequently, the sections were washed three times with 50% ethanol and another three times with distilled water. The staining was observed under a BX53 microscope.

Immunohistochemistry

The hippocampal tissue sections were put into antigen retrieval solution and continuously heated for 10 min under low heat. Incubation with 3% H₂O₂ for 15 min was used to eliminate endogenous peroxidase activity. After blocked with 1% BSA, the sections were incubated with Iba1 antibody (RRID: AB_3711469, A19776, ABclonal, China) at 4℃ overnight and HRP-labeled goat anti-rabbit IgG (RRID: AB_228341, 31,460, ThermoFisher, USA) at 37℃ for 1 h. DAB and hematoxylin were used to color development. The sections were dehydrated in ethanol, cleared in xylene, and mounted with neutral gum. The quantification and analysis of the IHC staining were performed using Image-Pro Plus version 6.0 software.

Immunofluorescence (IF)

IF was performed on the 5-μm-thick hippocampal sections and fixed BV-2 cells. For hippocampal sections, incubation with antigen retrieval solution was performed for 10 min under low heat. For BV-2 cells, incubation with 0.1% Triton X-100 was conducted at room temperature for 30 min. Sections of hippocampal tissues or BV-2 cells were washed with PBS three times, and then incubated with 1% BSA for 15 min to reduce non-specific antibody binding. Primary antibody was used to incubate the section at 4℃ overnight, followed by three PBS washes, and then incubation with secondary antibody. Nuclei were stained with DAPI. Antibody information is indicated in Table 1. The quantification and analysis of the IF staining were performed using Image-Pro Plus version 6.0 software. To eliminate background interference, we performed background subtraction for each fluorescence channel. In every field of view, a cell-free region was selected to measure its average fluorescence intensity and standard deviation. This value was then used as the local background and subtracted from the entire image. Following this, a cell was only considered positive for a given channel if its mean fluorescence intensity exceeded the local background value by more than three standard deviations.

Table 1.

The antibody information

	Dilution	RRID	Cat#	Manufacturer
ZFP36	1:500	AB_10988939	sc-374305	Santa Cruz
NLRP3	1:1000	AB_2923634	68,102–1-lg	Proteintech
ASC	1:10,000	AB_2174862	10,500–1-AP	Proteintech
Cleaved caspase-1	1:1000	AB_2845464	AF4022	Affinity
IL-1β	1:1000	AB_2769945	A16288	ABclonal
Iba1	1:100	AB_3711469	A19776	ABclonal
Aβ	1:100	AB_2881451	60,342–1-Ig	Proteintech
iNOS	1:50	AB_627810	Sc-7271	Santa Cruz
β-Actin	1:10,000	AB_626632	sc-47778	Santa Cruz
HRP-labeled goat anti-rabbit IgG	1:100	AB_228341	31,460	ThermoFisher
Goat anti-rabbit IgG	1:200	AB_955021	ab6939	Abcam
Goat anti-mouse IgG	1:200	AB_955241	ab6785	Abcam
Goat anti-mouse IgG	1:200	AB_10680176	ab97035	Abcam
Goat anti-rabbit IgG-HRP	1:3000	AB_2797593	SE134	Solarbio
Goat anti-mouse IgG-HRP	1:3000	AB_2797595	SE131	Solarbio

ELISA

Mouse TNF-α ELISA Kit, Mouse IL-6 ELISA Kit, and Mouse IL-1β ELISA Kit were purchased from the Multi Sciences Biotech (Hangzhou, China). The protein concentration of samples was detected using the BCA kit (Beyotime, Shanghai, China). All steps were carried out according to the instructions of the kit. The OD values at 450 nm and 570 nm were detected by an ELX-800 microplate reader (Biotek, VT, USA).

Western blot

Total protein was extracted using RIPA Lysis Buffer, and its concentration was qualified using BCA kit. The polyacrylamide gel consisted of 5% stacking gel and 8%, 12%, and 14% separating gel. SDS-PAGE of 20 μl samples (15 μg protein) was conducted to separate the target protein. A PVDF film was used to transfer the protein and then blocked with the blocking solution (SW3010, Solarbio, China) for 1 h. Following primary antibody incubation at 4℃ overnight, the PVDF film was incubated with secondary antibody for 1 h at 37℃. Finally, the protein bands were visualized with ECL chemi-luminescence substrate. Antibody information is indicated in Table 1.

Real time PCR

The first strand of cDNA was synthesized from 1 μg RNA using All-in-One First-Strand SuperMix (Magen Biotech, Guangzhou, China). Real time PCR was performed using the ExicyclerTM 96 fluorescence quantifier (Bioneer, Daejeon, Korea) with 96-well plates. The reaction system was established based on TaqMan real-time PCR system and SYBGreen real-time PCR system (Solarbio, Beijing, China). The PCR was done according to the following steps: 1) 95℃ for 5 min, 2) 40 cycles of 95℃ for 10 s, 60℃ for 10 s and 72℃ for 15 s, 3) 72℃ for 90 s, 4) 40℃ for 1 min. The delta-delta- Ct method was employed to determine the relative gene expression level of gene of interest normalized to the β-actin. The Primers were listed in Table 2.

Table 2.

The sequence of primer

Target	Forward	reverse
ZFP36	5’-TCCCTGTCCTCTTGTTCCTT-3’	5’-CGCTGCTGGCGTAGTCAT-3’
ZBP1	5’- CCCTGCGATTATTTGTC-3’	5’-GAGGCTCATCTCGTTGT-3’
β-actin	5’-CATCCGTAAAGACCTCTATGCC-3’	5’-ATGGAGCCACCGATCCACA-3’

RIP-PCR

Cells were collected at 80% confluency and lysed in RIP Lysis Buffer. Protein A/G magnetic beads were pre-bound with 5 µg of specific antibody in RIP Wash Buffer for 30 min at room temperature. The antibody-bead complexes were then incubated with the pre-cleared cell lysate overnight at 4 °C, followed by six washes with cold RIP Wash Buffer. RNA was recovered by adding Proteinase K Buffer to the beads and incubating at 55 °C for 30 min. Following de-crosslinking, RNA was extracted using the phenol–chloroform method. For cDNA synthesis, 1 µg of RNA was treated with dsDNase and reverse transcribed using All-in-One First-Strand cDNA Synthesis SuperMix. PCR amplification was performed using gene-specific primers and Taq PCR Master Mix. The products were analyzed by 2% agarose gel electrophoresis and visualized under a gel documentation system.

mRNA-seq and bioinformatic analysis

Total RNA was extracted from cells using the TRIzol method. RNA integrity and concentration were assessed with an Agilent 2100 Bioanalyzer to ensure they met the requirements for library construction. mRNA was enriched using Oligo(dT) magnetic beads. The enriched RNA was fragmented at 94 °C, followed by first-strand cDNA synthesis primed with random hexamers. Second-strand cDNA was then synthesized using DNA Polymerase I/RNase H. The double-stranded cDNA underwent end repair, A-tailing, and ligation of Illumina adapters. Fragments of 300–500 bp were selected using AMPure XP beads and enriched by PCR. The resulting library was purified again and quantified by qPCR. Qualified libraries were sequenced on the NovaSeq 6000 platform. Raw sequencing data were processed with fastp to remove adapters and low-quality reads. Clean reads were aligned to the Ensembl GRCm39 genome using HISAT2. Gene expression levels were quantified with HTSeq and normalized as FPKM. Differential expression analysis was performed using DESeq2, with significant differences defined as p < 0.05 and |log2FC|> 1. PCA, Volcano plot, and Heatmap were generated by R package (ggplot2 v3.5.0 and pheatmap v1.0.12). GO and KEGG enrichment analyses were conducted using clusterProfiler.

Statistical analysis

Data are presented as mean ± SD. Statistical analysis was performed using GraphPad Prism. After confirming assumptions of normality and homogeneity of variance, comparisons between two groups were made using unpaired two-tailed Student’s t-tests. For comparisons among three or more groups, one-way ANOVA was employed, followed by Tukey’s post hoc test for multiple comparisons. For the analysis of escape latency in the Morris water maze task, a repeated-measures two-way ANOVA was applied. Cell experiments represent three independent replicates; animal experiments involve six mice per group; and histopathological quantifications are based on three randomly selected animals per group. A p-value of less than 0.05 was considered statistically significant.

Results

ZFP36 is upregulated in the hippocampus tissue of APP/PS1 mice

The hippocampus tissues of wild-type mice and APP/PS1 transgenic mice were isolated, and real-time PCR and western blot found that the expression of ZFP36 in the hippocampus tissues of transgenic mice was significantly increased (Fig. 1A -B). Immunofluorescence staining of ZFP36 and Iba1 in the hippocampal tissue showed that a marked increase of ZFP36 expression and microglia activation in the APP/PS1 transgenic mice compared with wild-type mice (Fig. 1C).

Fig. 1.

ZFP36 is upregulated in the hippocampus tissue of APP/PS1 mice. A Relative mRNA level of ZFP36 in the hippocampus tissue of APP/PS1 transgenic mice. B Protein level of ZFP36 in the hippocampus tissue. C Immunofluorescence staining of ZFP36 and Iba1 in mouse hippocampus tissue and statistical analysis results of positive cells, bar = 500 μm (40 ×), 50 μm (400 ×), white arrow indicates double-positive cells. Quantitative analysis was performed on images of the CA1 region at 400 × magnification. N = 3–6 replicates. Wt: wild type mice (C57BL/6), Tg: APP/PS1 transgenic mice with a C57BL/6 background. ***: p < 0.001, **: p < 0.01

ZFP36 overexpression improves the cognitive ability of APP/PS1 mice

To understand the role of ZFP36 in cognitive ability, we injected adeno-associated virus overexpressing ZFP36 into the hippocampus of APP/PS1 mice. In the Y-maze test, transgenic mice exhibited significant reductions in both spontaneous alternations and total distance traveled, whereas ZFP36-overexpressing mice showed marked increases in these parameters (Fig. 2A-B, Supplementary Fig. 1A). Furthermore, ZFP36 overexpression significantly increased the frequency of entries into the novel arm and prolonged the time spent exploring it (Fig. 2C-D, Supplementary Fig. 1B). These data collectively indicate that ZFP36 enhances spatial working memory and short-term memory performance and promotes explorative behavior toward novel spatial contexts. Morris water maze test was used to evaluate the effects of ZFP36 overexpression on the spatial learning and memory in APP/PS1 mice. Mice with ZFP36 overexpression had a significantly shorter escape latency on training day 5 and 6 compared to APP/PS1 transgenic mice (Fig. 2E), and showed better memory retention. The space exploration experiment found that overexpression of ZFP36 increased the time of transgenic mice in the target quadrant (Fig. 2F, Supplementary Fig. 1C) and the number of times they crossed the platform (Fig. 2G). The overexpression efficiency of ZFP36 was determined by real-time PCR and Western blot (Fig. 2H-I). Thioflavin S staining of hippocampal tissue was used to assess Aβ plaque formation. As shown in Fig. 2J, Aβ plaques were increased in the transgenic mice and decreased in mice with ZFP36 overexpression. Immunofluorescence staining of Aβ showed less positive staining in the ZFP36 overexpression group compared with the Ev group (Fig. 2K).

Fig. 2.

ZFP36 overexpression improves the cognitive ability of APP/PS1 mice. A Quantification of spontaneous alteration rate in the Y-maze experiment among the APP/PS1 transgenic mice and wild type mice. B Total moving distance of mice in 5 min in the Y maze test. C Number of entries in each arm of the Y-maze test, S: starting arm, O: old arm, N: new arm. D Time spent in the three arms of Y-maze test. E Escape latency of Morris water maze during learning time. F Time spent in target quadrant in the Morris water maze. G The number of times each group of mice crossed the platform in the water maze test. H The RNA levels of ZFP36 in the hippocampus of mice. I The blot of ZFP36 and quantitative analysis of blot density. J Thioflavin-S staining of amyloid plaques in the hippocampus, bar = 500 μm. K Double-immunofluorescence staining to detect localization of Aβ and Iba-1 in the hippocampus, bar = 50 μm. N = 3–6 replicates. ***: p < 0.001, **: p < 0.01, *: p < 0.05

ZFP36 prevents the microglia pro-inflammatory activation in the hippocampal tissue

Immunohistochemical staining of Iba1 in hippocampal tissue showed that the mean integrated optical density in the ZFP36 overexpression group was significantly decreased compared to the vector group (Fig. 3A). To investigate the effect of ZFP36 overexpression on microglia polarization, immunofluorescence of iNOS and Iba1 found that double positive cells was increased in transgenic mice and decreased in mice with ZFP36 overexpression (Fig. 3B-C). In the hippocampus, the levels of pro-inflammatory factors including TNF-α, IL-1β and IL-6 were elevated in the vector group and decreased in the ZFP36 overexpression group (Fig. 3D).

Fig. 3.

ZFP36 prevents the microglia pro-inflammatory activation in the hippocampal tissue. A Immunohistochemical detection of Iba-1 in the hippocampus, bar = 500 μm (40 ×), 50 μm (400 ×), Right panel: quantification of mean integrated optic density (IOD) of Iba1. B & C Double-immunofluorescence staining to detect co-localization of iNOS and Iba-1 in the hippocampus, bar = 500 μm (40 ×), 50 μm (400 ×). White arrow indicates double-positive cells. C Quantification of double positive cells in the immunofluorescence staining of iNOS and Iba-1. Quantitative analysis was performed on images of the CA1 region at 400 × magnification. D ELISA for the content of TNF-α, IL-1β and IL-6 in the hippocampus. N = 3–6 replicates. ***: p < 0.001, **: p < 0.01, *: p < 0.05

ZFP36 inhibits NLRP3 inflammasome activation in the hippocampal tissue

Existing research has shown that inhibiting the NLRP3 inflammasome regulated the activation of microglia, thereby affecting the progression of AD (Heneka et al. 2013; Wang et al. 2023). In the hippocampus with ZFP36 upregulation, the protein levels of cleaved caspase-1, ASC, NLRP3, and cleaved IL-1β were reduced (Fig. 4A-D). The protein levels of pro-IL-1β had no significant changes following ZFP36 overexpression (Fig. 4D). Immunofluorescence of NLRP3 showed reduced positive staining in the ZFP36 overexpression group, at the same time, double positive staining of NLRP3 and Iba1 was reduced (Fig. 4E).

Fig. 4.

ZFP36 inhibits NLRP3 inflammasome activation in the hippocampal tissue. A Protein level of cleaved caspase-1 in the hippocampus of mice. B Protein level of ASC in the hippocampus. C Protein level of NLRP3 in the hippocampus. D Protein level of pro-IL-1β and cleaved IL-1β in the hippocampus. E Double-immunofluorescence staining to detect co-localization of NLRP3 and Iba-1 in the hippocampus, bar = 50 μm (400 ×). Quantitative analysis was performed on images of the CA1 region at 400 × magnification. N = 3–6 replicates. ***: p < 0.001, **: p < 0.01

ZFP36 inhibits the pro-inflammatory activation of BV-2 cells and NLRP3 inflammasome activation induced by Aβ_1–42 treatment

Real time PCR and western blot were used to detect the efficiency of ZFP36 overexpression in the BV-2 cells (Fig. 5A-C). BV-2 cells with ZFP36 overexpression were treated with 5 μM Aβ_1–42 for 24 h. After Aβ_1–42 treatment, the positive staining of iNOS and CD68 was increased. In the BV-2 cells with ZFP36 overexpression, the promoting effect of Aβ_1–42 on iNOS and CD68 expression was inhibited (Fig. 5D). Aβ_1–42 treatment promoted the expression of TNF-α, IL-1β and IL-6. Compared with Aβ_1–42 + empty vector group, ZFP36 overexpression decreased the expression of TNF-α, IL-1β and IL-6 (Fig. 5E). These results revealed that ZFP36 inhibited the pro-inflammatory activation of BV-2 cells. Aβ_1–42 treatment increased the protein level of cleaved caspase-1, ASC, NLRP3, and cleaved IL-1β, and the increasing trend of these four proteins was attenuated by ZFP36 overexpression (Fig. 6A-E). To further validate the role of ZFP36, we knocked down its expression using siRNA (Fig. 7A) and performed parallel assays. The results showed that in BV-2 cells, knockdown of ZFP36 followed by Aβ1–42 treatment increased the level of TNF-α and IL-1β (Fig. 7B-C). Meanwhile, the positive staining of iNOS and CD68 were significantly higher than those in the Aβ_1–42 + siNC group (Fig. 7D). In addition, knockdown of ZFP36 further increased the protein levels of cleaved caspase-1, NLRP3, and cleaved IL-1β (Fig. 7E-F). These results, which are opposite to those of the overexpression experiments, indicate that knockdown of ZFP36 exacerbates Aβ_1–42-induced pro-inflammatory activation, inflammatory response, and NLRP3 inflammasome activation in BV-2 cells, thereby further supporting the inhibitory role of ZFP36 in microglial inflammatory regulation.

Fig. 5.

ZFP36 overexpression inhibits the pro-inflammatory activation of BV-2 cells induced by Aβ_1–42 treatment. A Relative RNA level of ZFP36 in the BV-2 cells with ZFP36 overexpression and Aβ_1–42 treatment. B & C Protein level of ZFP36 in the BV-2 cells. D Immunofluorescence staining of iNOS and CD68, bar = 50 μm. E. TNF-α, IL-1β, and IL-6 levels were assessed by ELISA in cell supernatants. N = 3 replicates. ***: p < 0.001, **: p < 0.01

Fig. 6.

ZFP36 overexpression inhibits NLRP3 inflammasome activation in A_1–42 treated BV-2 cells. A Representative blots of cleaved caspase-1, ASC, NLRP3, and IL-1β in BV-2 cells with Aβ_1–42 treatment and ZFP36 overexpression. B Quantitative analysis of cleaved caspase-1 protein level. C Quantitative analysis of ASC protein level. D Quantitative analysis of NLRP3 protein level. E Quantitative analysis of cleaved IL-1β protein level. N = 3 replicates. ***: p < 0.001, **: p < 0.01, *: p < 0.05

Fig. 7.

ZFP36 knockdown enhances the pro-inflammatory and NLRP3 inflammasome activation of BV-2 cells induced by Aβ_1–42 treatment. A Protein level of ZFP36 in the BV-2 cells. B & C TNF-α and IL-1β levels were assessed by ELISA in cell supernatants. D Immunofluorescence staining of iNOS and CD68, bar = 50 μm. E Representative blots of cleaved caspase-1, NLRP3, and IL-1β in BV-2 cells with Aβ_1–42 treatment and ZFP36 overexpression. F Quantitative analysis of cleaved caspase-1, NLRP3, and cleaved IL-1β level. N = 3 replicates. ***: p < 0.001, **: p < 0.01

Exploration of downstream targets of ZFP36

mRNA isolated from the BV-2 cells with Aβ_1–42 treatment and ZFP36 overexpression was sent to mRNA-seq. PCA analysis of mRNA-seq data was presented in Fig. 8A, which showed significant differences between the two groups. Under the screening criteria of absolute value of log2FC > 1, and P value < 0.05, 470 up-regulated genes and 388 down-regulated genes were identified (Fig. 8B). ZFP36 was reported to promote the target degradation, therefore, we mainly focus on the down-regulated genes. Heat map (Fig. 8C), KEGG and GO enrichment analysis (Fig. 8D) was performed on the down-regulated genes. To determine the target of ZFP36 in the direction of NLPR3 inflammasome, we searched for ZFP36 targets on the ENCORI website and searched for NLRP3 inflammasome-related genes on GeneCards. These genes and down-regulated genes were then subjected to Venn plot analysis (Fig. 8E). A total of 13 genes was obtained. Among these genes, we excluded some cytokines (6 genes), factors with opposite effects to ZFP36 (4 genes), and factors with clear functions (2 genes), and finally ZBP1 remained (Fig. 8F). The expression of ZBP1 was presented in Fig. 8G.

Fig. 8.

Exploration of downstream targets of ZFP36. The RNA samples of BV-2 cells with Aβ_1–42 treatment and/or ZFP36 overexpression were sent to mRNA-seq. A PCA analysis of mRNA-seq data. B Volcano plot of differentially expressed genes. C Heat map of differentially downregulated genes. D GO and KEGG enrichment analysis of differentially down-regulated genes. E Venn diagram of downregulated genes from mRNA-seq, potential targets of ZFP36 (mouse) in ENCORI database and genes related to NLRP3 Inflammasome in Genecards database. F Intersection analysis identified 13 genes. G The mRNA level of ZBP1. N = 3–6 replicates. *: p < 0.05

ZFP36 inhibits NLRP3 inflammasome activation and microglia pro-inflammatory activation through ZBP1

ZBP1 was downregulated in the hippocampal tissue with ZFP36 overexpression (Fig. 9A) and cells with ZFP36 overexpression and Aβ_1–42 treatment (Fig. 9B-C). RIP-PCR verified the binding of ZFP36 to ZBP1 mRNA in cells treated with Aβ_1–42 (Fig. 9D). Cells overexpressing ZFP36 and treated with Aβ_1–42 were treated with 2 µg/ml actinomycin D for 0, 2, 4 and 6 h. It found that the degradation of ZBP1 RNA was enhanced by ZFP36 overexpression (Fig. 9E). To clarify whether ZFP36 functions through ZBP1, we performed rescue experiments at the cellular level through ZBP1 overexpression. The efficiency of ZBP1 overexpression was verified in the cells (Fig. 9F). The protein level of cleaved caspase-1 was decreased in the Aβ_1–42 + ZFP36 group and increased in the Aβ_1–42 + ZFP36 + ZBP1 group (Fig. 9G). Immunofluorescence of iNOS showed reduced positive staining after ZBP1 overexpression (Fig. 9H). ELISA found that the levels of TNF-α, IL-1β and IL-6 was decreased in the cells overexpressing ZFP36 alone and increased in the cells with ZFP36 and ZBP1 overexpression (Fig. 9I). These results demonstrated that ZBP1 overexpression reversed the effect of ZFP36 on the NLRP3 inflammasome activation and microglial activation.

Fig. 9.

ZFP36 regulates the activation of NLRP3 inflammasome and microglia through ZBP1. A RNA level of ZBP1 in the hippocampus tissue of APP/PS1 transgenic mice. B RNA level of ZBP1 in the cells with Aβ_1–42 treatment and/or ZFP36 overexpression. C Protein level of ZBP1 in the cells with Aβ_1–42 treatment and/or ZFP36 overexpression. D The binding of ZFP36 and ZBP1 mRNA was verified by RIP-PCR. E The mRNA level of ZBP1 in cells after treatment with actinomycin D for 0, 2, 4 and 6 h. F RNA level of ZBP1 in the cells with ZBP1 overexpression. G Protein level of cleaved caspase-1 in cells with Aβ_1–42 treatment and/or ZFP36 overexpression and ZBP1 overexpression. H Immunofluorescence staining of iNOS, bar = 50 μm. I. TNF-α, IL-1β, and IL-6 levels were assessed by ELISA in cell supernatants. N = 3–6 replicates. ***: p < 0.001, **: p < 0.01, *: p < 0.05

Discussion

The core pathological features of AD include abnormal deposition of Aβ and persistent neuroinflammation, in both of which microglia play a critical role (Arranz & De Strooper 2019; Lynch et al. 2010; Scheltens et al. 2016). A major cause of the excessive accumulation of neurotoxic Aβ in the brain is the imbalance between the production and clearance of Aβ (Deng et al. 2021). Notably, microglial phagocytosis is one of the multiple strategies for Aβ clearance (Su et al. 2021). This study found that in APP/PS1 mice and Aβ-treated BV-2 microglial cells, the expression of the RNA-binding protein ZFP36 was compensatory up-regulated; however, such endogenous elevation was insufficient to halt disease progression. This observation aligns with the phenomenon seen in certain pathological conditions where endogenous compensatory responses are activated but remain below the optimal therapeutic window (Guo et al. 2022).

Our study further reveals that hippocampal overexpression of ZFP36 significantly improves cognitive function in APP/PS1 mice, concurrently reducing Aβ deposition and suppressing the pro-inflammatory activation of microglia. In contrast to previous strategies primarily focused on directly inhibiting downstream effector molecules (Venegas et al. 2017), ZFP36 functions as an upstream regulator, exerting a hub role through broader post-transcriptional regulation (Ferri et al. 2025; Qin et al. 2025). Notably, ZFP36 overexpression significantly lowered the levels of NLRP3, ASC, cleaved caspase-1, and cleaved IL-1β in the hippocampus and Aβ_1–42 treated BV-2 microglia. This suggests that ZFP36 may exert a suppression of NLRP3 inflammasome activation. Also, in a neurodegenerative diseases Parkinson's disease, ZFP36 is reported to protects against dopaminergic oxidative injury (Sun et al. 2020). Mechanistically, we have identified ZBP1 as a key target of ZFP36 in the context of AD-related neuroinflammation. ZBP1 is an intracellular nucleic acid sensor known to recognize aberrant nucleic acid structures and activate downstream signaling pathways in viral defense and sterile inflammation (Chen et al. 2026). Previous study suggested that ZBP1 drives neuroinflammation in AD (Guo et al. 2023; Song et al. 2025). Aβ-induced oxidative stress leads to mtDNA fragmentation and the formation of a Z-DNA conformation, which is recognized by ZBP1 in both microglia and neurons. This subsequently recruits and activates RIPK1 and IRF3, respectively(Chen et al. 2026). One axis amplifies IFN/TLR/NLRP3 inflammatory transcription and drives hyperactivation of microglia, while the other axis triggers neuronal pyroptosis. Genetic or pharmacological blockade of ZBP1 simultaneously reduces Aβ burden, neuroinflammation, and cognitive deficits, thus establishing ZBP1 as a potential therapeutic target. Our experimental data confirm that ZFP36 binds to and degrades ZBP1 mRNA, thereby inhibiting its downstream pro-inflammatory signaling. Therefore, the ZFP36-ZBP1 axis may be an important upstream pathway regulating microglia pro-inflammatory and NLRP3 inflammasome activation. This newly discovered regulatory mechanism not only links the function of ZBP1 to AD neuroinflammation but also provides a fresh molecular perspective for understanding abnormal NLRP3 activation.

A critical question is whether the observed decline in Aβ plaques is the cause or the consequence of diminished neuroinflammation. We propose a synergistic model rather than a unidirectional relationship. The primary and direct effect of ZFP36 overexpression is the potent inhibition of the microglial pro-inflammatory state, mainly through degrading ZBP1 mRNA. This resolution of the chronic inflammatory milieu likely contributes to reduce Aβ levels via two interconnected mechanisms(Jung et al. 2025; Sastre et al. 2006; Zimmer et al. 2014). First, a dampened inflammatory environment is known to decrease neuronal amyloidogenic processing, thereby reducing de novo Aβ production. Second, alleviating the dysfunctional inflammatory state may restore the efficiency of intrinsic microglial homeostatic clearance pathways. Therefore, the reduction in Aβ plaques is best interpreted as a beneficial outcome of the overall cerebral environmental improvement orchestrated by ZFP36, where attenuated neuroinflammation and optimized microglial function jointly contribute to the effect. This model also provides a framework for interpreting our subsequent cellular findings.

It is noteworthy that while ZFP36 overexpression produces a potent anti-inflammatory effect, it is accompanied by the downregulation of the phagocytosis-related marker CD68. This finding contrasts with the prevailing view that simply enhancing microglial phagocytic function directly facilitates Aβ clearance (Etani et al. 2025; Pilat et al. 2025). We posit that this may stem from the core function of ZFP36, which is to reverse the pathological pro-inflammatory state of microglia. Given that a chronic inflammatory environment itself can impair Aβ clearance and promote its generation (Tejera et al. 2019; Xie et al. 2021), the ZFP36-mediated attenuation of inflammation may indirectly reduce Aβ generation. Simultaneously, by fostering a less inflammatory microenvironment, it may restore a more effective homeostatic clearance function in microglia, rather than merely increasing their phagosome count. This suggests that the phenotypic regulation of microglia is highly complex, and functional improvement cannot be simplistically equated with the upregulation of specific markers.

It should be acknowledged that the non-specific expression mediated by the CMV promoter in vivo complicates the precise attribution of effects solely to microglia. Nevertheless, this limitation is substantively addressed by our complementary in vitro experiments. The robust and reproducible functional changes, elicited by direct ZFP36 manipulation in pure microglial cultures confirm its potent and intrinsic regulatory role within these cells. This study primarily utilized the BV-2 microglial cell line for mechanistic exploration. While this model provides convenience for genetic manipulation and high-throughput screening, we acknowledge the inherent differences in transcriptional profiles and functional responses between immortalized cell lines and primary cells in vivo. Therefore, conclusions regarding specific phenotypes such as CD68 downregulation and their functional implications require further validation and elucidation in future studies employing primary cells or more complex in vivo models.

In summary, this study not only reveals for the first time the protective role of ZFP36 in AD models but elucidates a novel functional axis. ZFP36 alleviates AD pathology and improves cognitive function by targeting and degrading ZBP1 mRNA, thereby inhibiting excessive microglial inflammatory responses, including NLRP3 inflammasome activation, at the post-transcriptional level. This discovery provides a new upstream potential target for the immunomodulatory treatment of AD. Future research should further validate this mechanism in models closer to the physiological and pathological context and explore the therapeutic potential of targeting the ZFP36-ZBP1 axis.

Supplementary information

Below is the link to the electronic supplementary material.

Author contribution

Ting Liu: Investigation, Writing–review & editing, Methodology. Dan Chen: Investigation, Writing–original draft. Fengjie Liu: Formal analysis, Writing–review & editing. Yun Sun: Conceptualization, Supervision.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval

All animal experiments were approved by the Ethical Committee of Dalian Medical University (AEE22067).

Clinical trial number

Not applicable.

Conflicts of interest

The authors declare no competing interests.