Human microglial HMC3 cells (ATCC, Manassas, VA, USA) were maintained under standard conditions, authenticated by STR profiling, and confirmed as mycoplasma-free. Pyroptosis was induced with 25 µM A$\beta$25–35 for 24 h. DHA was applied at concentrations of 5, 10, and 25 µM. Cells were transfected with HOXA9 plasmids and treated with 1 µM MCC950 (Sigma-Aldrich, St. Louis, MO, USA) to inhibit NLRP3, as detailed in the Supplementary Methods [20, 21].

Total RNA was extracted from cultured HMC3 cells and mouse brain tissue using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). cDNA synthesis was performed using a commercial reverse transcription kit (Takara, Japan). RT-qPCR was conducted with SYBR Green chemistry, and gene expression levels were normalized. Primer sequences, thermal cycling conditions, and reagent details are provided in the Supplementary Methods [22, 23].

Total proteins from cells and brain tissues were isolated using radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors. Equal protein samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to polyvinylidene fluoride membranes. Following blocking, membranes were probed with primary antibodies targeting HOXA9 (PA5-102516; Invitrogen, Carlsbad, CA, USA), NLRP3 (MA5-32255, Invitrogen), apoptosis-associated speck-like protein containing a caspase-recruitment domain (PA5-50915, Invitrogen), p30-gasdermin D (ab215203; 1:1000; Abcam, Cambridge, UK), cleaved-caspase-1 (PA5-38099, Invitrogen), interleukin (IL)-1$\beta$ (MA5-23691, Invitrogen), IL-18 (PA5-79479, Invitrogen), and GAPDH (MA5-15738; 1:1000; Invitrogen), and then incubated with horseradish peroxidase (HRP)‑linked secondary antibodies. Immunoreactive bands were visualized by enhanced chemiluminescence and analyzed using ImageJ (National Institutes of Health, Bethesda, MD, USA). Full antibody details and protocols are available in the Supplementary Methods.

Pro-inflammatory cytokines (TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6) in cell culture supernatants and brain tissue homogenates were quantified using ELISA kits (BioLegend, San Diego, CA, USA).

Cell viability was evaluated using the CCK-8 assay. HMC3 cells were seeded into 96-well plates at 5 $\times$ 10³ cells per well and treated with varying concentrations of DHA. After 24 h, 10 µL of CCK-8 reagent (Dojindo, Kumamoto, Japan) was added to each well, and absorbance was measured at 450 nm after 2 h of incubation. Viability was expressed as a percentage relative to the control.

ChIP assays were performed using the EZ-ChIP™ Kit (Millipore, Burlington, MA, USA) to investigate HOXA9 binding to the NLRP3 promoter. Cells were cross-linked with formaldehyde, lysed, and sonicated to form shear chromatin. The sheared chromatin was immunoprecipitated using anti-HOXA9 antibodies, followed by DNA purification and qPCR analysis to detect enrichment at the NLRP3 promoter site.

HEK-293T cells were co-transfected with NLRP3 promoter constructs (wild-type or mutant) and either HOXA9 overexpression or control plasmids by Lipofectamine 3000 (Thermo Fisher Scientific). Cells were harvested 48 h post-transfection, and luciferase activities were measured using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA). Relative luciferase activity was calculated by normalizing firefly luciferase signals to Renilla luciferase signals.

Immunofluorescence analysis was performed to detect p30-GSDMD levels in both mouse brain tissue sections and cultured HMC3 microglial cells. Following fixation, permeabilization, and blocking, samples were incubated with primary antibodies and subsequently with fluorescence-conjugated secondary antibodies. Nuclear staining was performed using DAPI, and images were obtained by confocal microscopy (Leica, Wetzlar, Germany).

Male amyloid precursor protein / Presenilin‑1 double-transgenic (APP/PS1) transgenic mice were randomly assigned to control and treatment groups. Mice received daily oral DHA or vehicle for 8 weeks, with some groups additionally receiving intracerebroventricular HOXA9-overexpressing adenovirus or the NLRP3 inhibitor MCC950 [24]. Mice were anesthetized with isoflurane (induction at 4% for 3 minutes, then maintenance at 2% in oxygen for approximately 30 minutes) until loss of pedal reflex, with body temperature maintained on a heating pad. At the experimental endpoint, animals were euthanized by intraperitoneal injection of sodium pentobarbital (200 mg/kg). After deep anesthesia was confirmed, transcardial perfusion with phosphate-buffered saline and 4% paraformaldehyde was performed for histological analysis, or PBS alone for molecular assays. Death was verified by cessation of heartbeat and respiration. Cognitive function was assessed using the Morris water maze test, and brain tissues were analyzed by RT-qPCR, Western blotting, and immunofluorescence.

Data were presented as mean $\pm$ standard deviation and analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for multiple comparisons. For behavioral data, repeated-measures ANOVA was used. A p-value of less than 0.05 was considered statistically significant.

To assess the influence of DHA on A$\beta$25–35-induced pyroptosis in HMC3 microglial cells, we first assessed the cell viability. CCK-8 data revealed that DHA concentrations from 0 to 25 µM had no significant toxic effects on HMC3 cells (Fig. 1A). Accordingly, we used 0–25 µM DHA for subsequent experiments. Compared to the control group, A$\beta$25–35 treatment markedly reduced cell viability, while DHA treatment (0–25 µM) dose-dependently increased cell viability (Fig. 1B). Immunofluorescence staining revealed that A$\beta$25–35 significantly increased p30-GSDMD fluorescence intensity compared to control, while DHA (0–25 µM) dose-dependently decreased p30-GSDMD fluorescence intensity (Fig. 1C). Western blot data suggested that A$\beta$25–35 exposure upregulated pyroptosis-associated proteins (p30-GSDMD, cleaved-caspase-1, IL-1$\beta$, and IL-18) and inflammasome factors (NLRP3 and ASC), whereas DHA administration led to a gradual reduction in their expression (Fig. 1D). ELISA results demonstrated that A$\beta$25–35 significantly increased TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6 levels, while DHA treatment (0–25 µM) dose-dependently decreased these inflammatory factors (Fig. 1E). These results indicate that DHA effectively suppresses NLRP3 activation and decreases A$\beta$25–35-induced microglial pyroptosis.

Fig. 1.

Docosahexaenoic acid (DHA) inhibits NLRP3 inflammasome activation and reduces A$\beta$-induced microglial pyroptosis. (A) Cell viability of HMC3 cells treated with various DHA concentrations (0–25 µM) for 24 h, measured by CCK-8 assay. (B) Cell viability of HMC3 cells exposed to A$\beta$ with or without DHA treatment (0–25 µM) for 24 h, measured by CCK-8 assay. (C) Representative immunofluorescence images and quantification of p30-GSDMD in HMC3 cells treated with A$\beta$ and various DHA concentrations (0–25 µM). Scale bar: 25 µm. (D) Western blot analysis of pyroptosis-related proteins (p30-GSDMD, cleaved-caspase-1, IL-1$\beta$, IL-18) and inflammasome components (NLRP3, ASC) in treated HMC3 cells. (E) ELISA measurements of pro-inflammatory cytokines (TNF-$\alpha$, IL-1$\beta$, IL-18, IL-6) in culture supernatants of treated HMC3 cells. *p 0.05, **p 0.01, ***p 0.001. Three independent experiments with three technical replicates were performed for each experiment.

We explored the involvement of HOXA9 in the protective effects of DHA against A$\beta$25–35-induced pyroptosis. Treatment with A$\beta$25–35 resulted in a pronounced enhancement in HOXA9 mRNA and protein expression, while DHA supplementation suppressed this upregulation. Overexpression of HOXA9 (oe-HOXA9) in DHA-treated cells further elevated HOXA9 expression compared with the negative control (oe-NC) (Fig. 2A,B). Cell viability assays revealed that A$\beta$25–35 significantly reduced cell viability, which was rescued by DHA treatment. However, HOXA9 overexpression partially reversed the protective effects of DHA (Fig. 2C). Immunofluorescence staining showed that A$\beta$25–35 increased p30-GSDMD fluorescence intensity, indicating enhanced pyroptosis. DHA treatment reduced this effect; however, HOXA9 overexpression counteracted DHA’s protective action (Fig. 2D). Western blotting indicated that A$\beta$25–35 dramatically enhanced the expression of pyroptosis-related proteins and inflammasome components. DHA treatment decreased the levels of these proteins; however, HOXA9 overexpression partially reversed this effect (Fig. 2E). ELISA measurements showed that A$\beta$25–35 increased TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6 expression but attenuated using DHA. Consistently, HOXA9 overexpression partially negated the inflammation‑suppressive abilities of DHA (Fig. 2F). These results suggest that DHA inhibits A$\beta$25–35-induced microglial pyroptosis, at least in part, through suppression of HOXA9.

Fig. 2.

Docosahexaenoic acid (DHA) inhibits A$\beta$-induced microglial pyroptosis via HOXA9 suppression. (A) qRT-PCR analysis of HOXA9 mRNA expression in HMC3 cells under different treatments. (B) Western blot analysis of HOXA9 protein levels. (C) Cell viability assessed using CCK-8 assay. (D) Representative immunofluorescence images and quantification of p30-GSDMD. Scale bar = 25 µm. (E) Western blot analysis of pyroptosis-related proteins (p30-GSDMD, cleaved-caspase-1, IL-1$\beta$, IL-18) and inflammasome components (NLRP3, ASC). (F) ELISA measurements of pro-inflammatory cytokines (TNF-$\alpha$, IL-1$\beta$, IL-18, IL-6) in the culture supernatants. *p 0.05, **p 0.01, ***p 0.001. Three independent experiments with three technical replicates were performed for each experiment.

To elucidate the cellular pathways regarding DHA’s involvement in NLRP3, we used bioinformatics analysis, which predicted a HOXA9 binding site within the NLRP3 promoter region (Fig. 3A). Dual-luciferase assays indicated that knockdown of HOXA9 (sh-HOXA9) significantly decreased the transcriptional activity of the wild‑type NLRP3 promoter without affecting the mutant construct the mutant promoter (Fig. 3B). This finding suggests that HOXA9 directly regulates NLRP3 transcription via its predicted binding sites. ChIP assays further indicated enrichment of anti-HOXA9 at the NLRP3 promoter sequence in HEK-293T cells (Fig. 3C), providing strong evidence for the direct binding of HOXA9 to the NLRP3 promoter. To evaluate the functional impact of this interaction, we examined NLRP3 expression levels under various conditions. RT-qPCR data revealed that DHA treatment dramatically suppressed NLRP3 mRNA levels, while HOXA9 overexpression promoted NLRP3 mRNA expression (Fig. 3D). Western blot analysis was consistent with these results, demonstrating reduced NLRP3 protein expression following DHA treatment and increased expression with HOXA9 overexpression (Fig. 3E). Taken together, these data demonstrate that DHA suppresses NLRP3 activation by inhibiting HOXA9, which directly regulates NLRP3 transcription.

Fig. 3.

Docosahexaenoic acid (DHA) suppresses NLRP3 activation through HOXA9 inhibition. (A) JASPAR prediction of HOXA9 binding site in the NLRP3 promoter region. (B) Dual-luciferase reporter assay results showing the effect of HOXA9 knockdown on NLRP3 promoter activity in HEK-293T cells. (C) ChIP assay results demonstrating HOXA9 enrichment at the NLRP3 promoter in HEK-293T cells. (D) qRT-PCR analysis of NLRP3 mRNA levels under DHA treatment and HOXA9 overexpression conditions. (E) Western blot analysis of NLRP3 protein levels under DHA treatment and HOXA9 overexpression conditions. **p 0.01, ***p 0.001. Three independent experiments with three technical replicates were performed for each experiment.

Further experiments were conducted to clarify how DHA influences NLRP3 transcriptional activation via HOXA9 in microglial cells. Three experimental groups were compared: A$\beta$25–35+DHA+oe-NC, A$\beta$25–35+DHA+oe-HOXA9, and A$\beta$25–35+DHA+oe-HOXA9+NLRP3 inhibitors (MCC950). Cell viability was significantly reduced by HOXA9 overexpression but partially restored by MCC950 treatment (Fig. 4A). Immunofluorescence analysis of p30-GSDMD, a key pyroptosis marker, suggested a dramatic enhancement in fluorescence intensity in oe-HOXA9 group. However, this effect was markedly attenuated by MCC950 treatment (Fig. 4B). Western blot analysis indicated that HOXA9 overexpression increased pyroptosis-related proteins and inflammasome components, whereas MCC950 reduced their expression in HOXA9-overexpressing cells (Fig. 4C). ELISA assays revealed that HOXA9 overexpression elevated TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6, but MCC950 treatment decreased these cytokine levels (Fig. 4D). Collectively, these results demonstrate that DHA inhibits cellular pyroptosis by regulating NLRP3 transcriptional activation through HOXA9. The ability of the NLRP3 inhibitor MCC950 to reverse the effects of HOXA9 overexpression further confirms the pivotal involvement of the HOXA9-NLRP3 axis.

Fig. 4.

Docosahexaenoic acid (DHA) inhibits cellular pyroptosis by regulating NLRP3 transcriptional activation via HOXA9. (A) Cell viability of HMC3 cells under different treatments, assessed using the CCK-8 assay. (B) Representative immunofluorescence images and quantification of p30-GSDMD. Scale bar = 25 µm. (C) Western blot analysis of pyroptosis-related proteins (p30-GSDMD, cleaved-caspase-1, IL-1$\beta$, IL-18) and inflammasome components (NLRP3, ASC). (D) ELISA measurements of pro-inflammatory cytokines (TNF-$\alpha$, IL-1$\beta$, IL-18, IL-6) in culture supernatants. *p 0.05, **p 0.01, ***p 0.001. Three independent experiments with three technical replicates were performed for each experiment.

Learning and memory functions of APP/PS1 transgenic mice were assessed using the Morris water maze. The AD model group exhibited a significantly prolonged escape latency, fewer platform crossings, and reduced time spent in the target quadrant. DHA improved these cognitive impairments, as evidenced by a shortened escape latency, increased platform crossings, and prolonged target quadrant time. However, overexpression of HOXA9 (DHA+oe-HOXA9 group) partially reversed the cognitive benefits conferred by DHA, resulting in increased escape latency, reduced platform crossings, and decreased target quadrant time compared with the DHA+oe-NC group. Notably, cotreatment with the NLRP3 inhibitor MCC950 (DHA+oe-HOXA9+MCC950 group) mitigated the adverse effects of HOXA9 overexpression and restored cognitive performance to levels (Fig. 5A).

Fig. 5.

Docosahexaenoic acid (DHA) improves learning and memory abilities and reduces pyroptosis levels in Alzheimer’s disease mice via the HOXA9-NLRP3 Axis (n = 10 mice in each group). (A) Morris water maze performance of APP/PS1 mice across different treatment groups. Escape latency, platform crossings, and time spent in the target quadrant were measured. (B,C) qRT-PCR and Western blot analyses of HOXA9 and NLRP3 expression levels in the brain tissues of the different groups. (D) ELISA results showing the levels of inflammatory cytokines (TNF-$\alpha$, IL-1$\beta$, IL-18, IL-6) in brain tissues. (E) Immunofluorescence staining of p30-GSDMD in brain tissues, indicating pyroptosis levels. Scale bar = 25 µm. *p 0.05, **p 0.01, ***p 0.001.

The AD model group showed significant upregulation of HOXA9 and NLRP3 expression. Correspondingly, the levels of pyroptosis-related proteins were markedly increased in AD models. DHA treatment dramatically reduced the expression of these proteins, indicating its protective effect against neuroinflammation and pyroptosis. In contrast, overexpression of HOXA9 in DHA-treated mice (DHA+oe-HOXA9 group) resulted in a dramatically increase in HOXA9 and NLRP3 levels, along with higher levels of pyroptosis markers, compared with the DHA+oe-NC group. Co-administration of MCC950 effectively counteracted these effects, resulting in reduced expression of HOXA9, NLRP3, and associated pyroptosis markers in the DHA+oe-HOXA9+MCC950 group (Fig. 5B,C).

The levels of these cytokines were markedly higher in the AD model group, indicating a pronounced inflammatory response. DHA treatment markedly reduced TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6 levels, suggesting its anti-inflammatory potential. Conversely, HOXA9 overexpression in the DHA-treated group (DHA+oe-HOXA9) enhanced the expression of these cytokines. Importantly, the addition of MCC950 in the DHA+oe-HOXA9+MCC950 group effectively reduced the TNF-$\alpha$, IL-1$\beta$, IL-18, and IL-6 levels, further supporting the role of the HOXA9-NLRP3 axis in mediating inflammation in AD (Fig. 5D).

Immunofluorescence staining of p30-GSDMD was performed to visualize and quantify pyroptosis in brain tissues. The results were consistent with those of the biochemical analyses; the AD model group showed elevated p30-GSDMD levels, indicating increased pyroptosis. DHA treatment reduced p30-GSDMD staining, whereas HOXA9 overexpression in DHA-treated mice (DHA+oe-HOXA9) increased p30-GSDMD levels. Co-treatment with MCC950 in the DHA+oe-HOXA9+MCC950 group significantly diminished p30-GSDMD staining, corroborating the involvement of the HOXA9-NLRP3 axis in pyroptosis regulation in AD (Fig. 5E). These results emphasize the importance of the HOXA9-NLRP3 axis in mediating the neuroprotective effects of DHA against AD-associated cognitive decline and neuroinflammation.