Dl-3-n-Butylphthalide Inhibits NLRP3 Inflammasome and Mitigates Alzheimer's-Like Pathology via Nrf2-TXNIP-TrX Axis

Dl-NBP treatment mitigates neuronal apoptosis in the hippocampus of the APP/PS1 mouse brain

It is considered and reviewed that neuronal cells are vulnerable to apoptosis in the AD brain (42, 57). Caspases are critical mediators of apoptosis and inflammation (24). Caspase-1 activation induces the maturation of IL-1β and subsequently potentiates inflammatory responses (35). Chronic activation of caspase-3 is involved in mediating the degenerative process in the aging brain (60). Increases of caspase-1 and caspase-3 immunoreactivity have been observed in the postmortem brains with AD (21, 63). To determine the level of genotype-related neurodegeneration and whether Dl-NBP could be neuroprotective in the APP/PS1 mouse brain, apoptotic markers were measured. As shown in Figure 1A and Supplementary Figure S1, Western blot analysis for cleaved caspase-1 (CC1) in the hippocampus revealed significant effects of genotype [F(1,20) = 14.293, p < 0.01] and Dl-NBP treatment [F(1,20) = 10.857, p < 0.01] in the absence of genotype × Dl-NBP interaction [F(1,20) = 0.305, p > 0.05]. Dramatic increases in the levels of CC1 were observed in APP/PS1 mice with respect to WT mice (p < 0.01). The effects of genotype [F(1,20) = 20.376, p < 0.01] and Dl-NBP [F(1,20) = 13.814, p < 0.01] or genotype × Dl-NBP interaction [F(1,20) = 0.409, p > 0.05] on the protein expression of procaspase-1 (Pro-Casp-1) were observed as well. We then assess the ratios of CC1 to Pro-Casp-1. Two-way analysis of variance (ANOVA) for CC1/Pro-Casp-1 ratios showed evident effects of genotype [F(1,20) = 86.171, p < 0.01] and Dl-NBP [F(1,20) = 124.117, p < 0.01], but not genotype × Dl-NBP interactions [F(1,20) = 1.237, p > 0.05]. CC1/Pro-Casp-1 ratios were higher in the hippocampus of the Tg groups in comparison with the age-matched WT groups (p < 0.01, Fig. 1A), which indicated increases in the CC-1 from Pro-Casp-1 in the APP/PS1 mouse brain. Western blot analysis for cleaved caspase-3 (CC3) and procaspase-3 (Pro-Casp-3) revealed significant main effects of genotype (p < 0.01) and Dl-NBP (p < 0.01), but not a genotype × Dl-NBP interaction (p > 0.05). The CC3/Pro-Casp-1 ratios were also increased in Tg mouse brains compared with those in the age-matched WT controls (p < 0.01, Fig. 1A). The caspase-1 activities were higher in the Tg group than those in the WT group (p < 0.01, Fig. 1B). Dl-NBP treatment reduced the ratios of CC1 to Pro-Casp-1 in both WT and Tg mice compared with those in the vehicle-treated controls (p < 0.01, Fig. 1A). The ratio of CC3 to Pro-Casp-3 in Dl-NBP-treated group was comparable with respect to that in the vehicle-treated controls in the hippocampus of WT and Tg mice (p > 0.05, Fig. 1A), although a trend toward a decrease. To confirm the level of Dl-NBP-induced mitigation of neuronal apoptosis in the hippocampus, we performed immunostaining analysis using TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) which reveals DNA fragmentation associated with apoptotic signaling cascades. As shown in Figure 1C and D, the number of TUNEL-positive cells was markedly higher in the hippocampus of the brain from the Tg mice than in the WT mice. Two-way ANOVA for TUNEL-positive cells revealed evident main effects of genotype [F(1,20) = 18.443, p < 0.01] and Dl-NBP [F(1,20) = 16.539, p < 0.01], but not a genotype × Dl-NBP interaction [F(1,20) = 2.062, p > 0.05, Fig. 1C, D]. The number of TUNEL staining cells was significantly lower in the hippocampus of Dl-NBP-treated group in both mouse genotypes than in the vehicle-treated controls (p < 0.01, Fig. 1C, D). Dl-NBP administration did not significantly affect the body weight and motility of the mice (data not shown).

OPEN IN VIEWER

FIG. 1.

Effects of Dl-NBP administration on alleviating neuronal degeneration in the APP/PS1 mouse brain. APP/PS1 Tg and age-matched WT C57BL/6 controls were given Dl-NBP at a dose of 20 mg/kg body weight/day or vehicle (vegetable oil) by oral gavage from 4 months of age for 5 months. n = 6 in each group. (A) Western blot assays show the expression levels of Procaspase-1 (Pro-Casp-1), CC1, Procaspase-3 (Pro-Casp-3) and CC3 in Tg and WT mouse brain. The total amount of GAPDH served as the internal loading control. The ratio of cleaved caspase to procaspase was calculated. (B) The activity of caspase-1 in the brains of both mouse genotypes was determined with a caspase-1 activity assay kit. (C) Neuronal apoptosis in the mouse brain was detected using TUNEL staining. (D) Representative images showing the apoptotic neuronal GCs in the hippocampus of mouse brain. Nuclear chromatin was labeled with DAPI (left panels, low magnification). The middle panels show the results of TUNEL signal in low magnification, and the right panels show the confocal microscopy images of apoptotic cells in the square boxes of left and middle panels in high magnification. Statistical significance: **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with two-way ANOVA followed by Student's t-test. Scale bar = 30 μm. ANOVA, analysis of variance; CC1, cleaved caspase-1; CC3, cleaved caspase-3; DAPI, 4′,6-diamidino-2-phenylindole; Dl-NBP, Dl-3-n-butylphthalide; GCs, granulosa cells; Tg, transgenic; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; WT, wild type.

Dl-NBP treatment inhibits the activation of NLRP3 inflammasomes in the APP/PS1 mouse brain

It was observed that caspase-1-induced release of mature IL-1β can be mediated by the activation of NLRP3 inflammasome (21). To determine whether Dl-NBP-mediated inhibition of caspase-1 processing in the APP/PS1 mouse brain was related to the inhibition of NLRP3 inflammasome formation, we detected the inflammasome components, including NLRP3, adaptor molecule apoptosis-associated speck-like protein containing a CARD (ASC), and IL-1β in the APP/PS1 mouse brain. Two-way ANOVA for the protein expression of NLRP3 showed a significant genotype effect [F(1,20) = 8.137, p < 0.01] together with a significant main effect of Dl-NBP [F(1,20) = 49.853, p < 0.01, Fig. 2A and Supplementary Fig. S1] in the absence of genotype × Dl-NBP interaction [F(1,20) = 0.173, p > 0.05]. We observed that the protein expression of NLRP3 was significantly higher in the brains from Tg mice than in the age-matched WT group (p < 0.01, Fig. 2A and Supplementary Fig. S1). The upregulations of inflammasome-related components indicate the activation of the NLRP3 inflammasome (65). Once the NLRP3 is activated, it recruits ASC and procaspase-1. The assembled NLRP3 inflammasome induces the transformation of procaspase-1 to CC1, cleaving IL-1β into mature IL-1β (33). In the present study, Dl-NBP administration led to a decrease in NLRP3 protein expression (p < 0.01, Fig. 2A) and NLRP3 mRNA levels (p < 0.01, Fig. 2B) in both genotypes compared with the vehicle-treated controls, and the effects of Dl-NBP on ASC were similar (p < 0.01, Fig. 2A, B). As demonstrated by Western blot analysis for mature, 17 kDa IL-1β, evident main effects of genotype [F(1,20) = 55.691, p < 0.01, Fig. 2A] and Dl-NBP treatment [F(1,20) = 28.027, p < 0.01, Fig. 2A, B] were observed, but not significant genotype × Dl-NBP interactions [F(1,20) = 1.279, p > 0.05, Fig. 2A]. Dl-NBP treatment repressed the increase of 17 kDa IL-1β in the Tg and WT mouse brains (p < 0.01, Fig. 2A, B). The results of 17 kDa IL-1β immunofluorescence staining confirmed the effects of Dl-NBP on the expression of the inflammatory molecule (Fig. 2C): the number of IL-1β⁺ cells in the hippocampus of APP/PS1 mice was markedly higher than that of the WT mice (p < 0.01). Under Dl-NBP administration, the number of IL-1β immune-stained cells was significantly reduced in both mouse genotypes compared with controls (p < 0.01, Fig. 2C). Furthermore, coimmunoprecipitation (Co-IP) analyses showed that significant increases in ASC-associated NLRP3 protein were detected in the APP/PS1 mouse brain relative to those in the age-matched WT group. Two-way ANOVA for ASC-associated NLRP3 revealed significant main effects of genotype [F(1,20) = 26.331, p < 0.01] and Dl-NBP [F(1,20) = 13.281, p < 0.01, Fig. 2A and Supplementary Fig. S1], but not a genotype × Dl-NBP interaction [F(1,20) = 0.161, p > 0.05] (Fig. 2D and Supplementary Fig. S2). NLRP3 protein was discovered to be complexed together with ASC, whereas the process of NLRP3 inflammasome formation was interrupted by Dl-NBP administration. Dl-NBP treatment triggered a decrease in ASC-associated NLRP3 protein expression in the brains of WT and Tg mice (p < 0.01, Fig. 2D). We also evaluated the inflammatory response by detecting the distribution and expression of GFAP⁺ (an astrocytic marker) and Iba1⁺ (a microglial marker) cells. We have previously reported that Dl-NBP treatment reduced the levels of soluble Aβ and Aβ oligomer in the APP/PS1 mouse brain, but did not alter the fibrillar Aβ levels (67). In the present study, double immunofluorescence staining with anti-Aβ and glial fibrillary acidic protein (GFAP) antibodies showed that there was more severe astrocytosis in Tg mouse brain (Fig. 3). The diffuse plaques were surrounded by astrocyte clusters, which displayed dystrophic characteristics with large somas, hypertrophic primary processes, and intense GFAP immunoreactivity (56). The intensity of Aβ-associated GFAP⁺ cells in Dl-NBP-treated APP/PS1 mice brain was less than that in the vehicle-treated APP/PS1 mice (p < 0.01), although there was no statistical difference in the fluorescence intensity of Aβ plaque between these two groups (p > 0.05). Two-way ANOVA for debris score of astrocytes revealed significant main effects of genotype [F(1,20) = 142.962, p < 0.01] and Dl-NBP [F(1,20) = 94.308, p < 0.01] in the absence of genotype × Dl-NBP interaction [F(1,20) = 1.724, p > 0.05]. As the GFAP-immunoreactive astrocytes, Iba1⁺ microglia formed significant inflammatory foci surrounding the Aβ plaque in the Tg mice (Fig. 4). The cell bodies of ionized calcium binding adaptor molecule 1 (Iba1)-immunoreactive microglial cells were swollen in the Tg mouse brain. The fluorescence intensity of Aβ plaque-associated Iba1⁺ cells was reduced in the Dl-NBP-treated Tg mice than in the vehicle-treated Tg controls (p < 0.01). Two-way ANOVA for the fragmented microglia revealed significant main effects of genotype [F(1,20) = 89.206, p < 0.01] and Dl-NBP [F(1,20) = 51.847, p < 0.01], but not genotype × Dl-NBP interactions [F(1,20) = 1.422, p > 0.05]. Dl-NBP treatment reduced microglial debris in both WT and Tg mice with respect to those in their vehicle controls (p < 0.01). Interestingly, the microglial cells in the brain of Dl-NBP-administered Tg mice were closely tracked to the boundaries of the compact Aβ plaques and formed a barrier-like structure with respect to vehicle-treated Tg mice (p < 0.01).

OPEN IN VIEWER

FIG. 2.

Effects of Dl-NBP management on inhibiting NLRP3 inflammasome activation in the APP/PS1 mouse brain. (A) Western blot analysis revealed the protein expression of NLRP3, ASC, IL-1β, and mature 17 kDa IL-1β in the brains of APP/PS1 Tg and age-matched WT mice with Dl-NBP (20 mg/kg body/day) or vehicle (vegetable oil) oral administration. n = 6 in each group. (B) The mRNA levels of NLRP3 and ASC in the hippocampal tissues of the Tg and WT mice are shown. (C) Representative immunofluorescence images showing the labeling of mature IL-1β within the CA1 area of the hippocampus from the brain slices of Tg and WT mice. The relative number of immunoreactive cells was calculated. (D) Co-IP assays were performed to analyze the NLRP3-ASC protein interactions in Dl-NBP-treated Tg and WT mouse brain. Brain tissue protein extracts were immunoblotted with an antibody against NLRP3. The immune complexes were fractionated by SDS-PAGE, and the blots were then probed with an antibody to ASC. Statistical significance: **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with two-way ANOVA followed by Student's t-test. Scale bar = 50 μm. ASC, adaptor molecule apoptosis-associated speck-like protein containing a CARD; Co-IP, coimmunoprecipitation; IL-1β, interleukin-1β; NLRP3, nucleotide binding oligomerization domain-like receptor family, pyrin domain containing 3; SDS-PAGE, sodium dodecyl sulfate/polyacrylamide gel electrophoresis.

OPEN IN VIEWER

FIG. 3.

Dl-NBP treatment mitigates Aβ plaque-associated astrocytosis and reduces dystrophic astrocytes in the APP/PS1 mouse brain. Aβ plaques and astrocytes in the brain sections of APP/PS1 Tg and age-matched WT mice with Dl-NBP (20 mg/kg body/day) or vehicle (vegetable oil) oral administration were visualized. An anti-Aβ antibody was used to label Aβ plaques (green). Extensive GFAP-immunoreactive astrocytes (red) were shown in the Tg mice. Right panels show the representative images in high magnification. Arrows indicate the fragmented astrocytes losing discernable cell shape. Arrow head shows the reactive astrocytes. The fluorescence intensity of Aβ plaques and Aβ plaque-associated GFAP⁺ cells was quantified. Dystrophic astrocytes were determined by calculating the debris score. Statistical significance: **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with Student's t-test or two-way ANOVA followed by Student's t-test. Scale bars: 60 μm. Scale bar is 20 μm in the high magnification of right panels. Aβ, β-amyloid; GFAP, glial fibrillary acidic protein.

OPEN IN VIEWER

FIG. 4.

Effects of Dl-NBP administration on impairing Aβ plaque-associated microglial dystrophy in the APP/PS1 mouse brain. Double immunofluorescence labeling of Aβ (green) and microglia (Iba1, red) in brain sections of APP/PS1 Tg and age-matched WT mice with Dl-NBP (20 mg/kg body/day) or vehicle (vegetable oil) oral administration. Right panels show the dystrophic microglial cells or Aβ-related microglia in high magnification. Arrows show the fragmentations of microglial cells without discriminable cell morphology. Arrow head indicates the reactive microglia. The fluorescence intensity of Aβ plaque-associated Iba1⁺ cells was determined. Fragmented microglia was assessed by calculating the debris score. Polarization of microglia near Aβ plaques was quantified. Statistical significance: **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with Student's t-test or two-way ANOVA followed by Student's t-test. Scale bar: 30 μm. Scale bar is 20 μm in high magnification of right panels. Iba1, ionized calcium binding adaptor molecule 1.

Dl-NBP treatment reduces TXNIP expression and suppresses TXNIP-NLRP3 interaction

TXNIP is the endogenous negative regulator of TrX, which is a major cellular antioxidant and antiapoptotic protein (47). Studies have demonstrated the potential role of TXNIP to serve as a binding partner for NLRP3 and for activating the NLRP3-caspase-1 inflammasome in retinal capillary endothelial cells (51) and in retinal Muller glia (15). In the present study, we assessed the expression of TXNIP in human postmortem brains of AD patients and normal controls. As shown in Figure 5A, the immunoreactivity of TXNIP was markedly induced in the cerebral cortex of patients with AD. The cortex of AD samples demonstrated significant increases in the number of cells with high levels of TXNIP expression (> 40% of total cell) (p < 0.01). The intensity of TXNIP staining was increased in the AD group relative to those in the control subjects (p < 0.01). The high-power micrographs of bottom panels in Figure 5A illustrate that the immunoreactive TXNIP is tightly associated with the neuron and glia morphology. To investigate whether TXNIP is an effective inducer for the activation of the NLRP3 inflammasome in AD-like pathology and to explore whether treatment with Dl-NBP exerts anti-inflammatory effects, the distribution and expression of TXNIP in the APPP/PS1 and age-matched WT mouse brains were analyzed. Western blot assays for TXNIP protein expression revealed the significant main effects of genotype [F(1,20) = 63.112, p < 0.01, Fig. 5B] and Dl-NBP treatment [F(1,20) = 43.699, p < 0.01, Fig. 5B], but not significant genotype × Dl-NBP interactions [F(1,20) = 3.325, p > 0.05, Fig. 5B]. TXNIP expression was significantly increased in the brains of APP/PS1 mice relative to that in the age-matched WT group (p < 0.01, Fig. 5B and Supplementary Fig. S2). Quantitative real-time polymerase chain reaction (PCR) assays showed that the TXNIP mRNA levels were significantly elevated in the Tg mouse brain compared with those in the WT group (p < 0.01, Fig. 5C). In contrast, the protein expression and mRNA levels of TXNIP were markedly lower under Dl-NBP administration than those with the vehicle treatment in both mouse genotypes (p < 0.01, Fig. 5B, C). Co-IP analysis demonstrated the evident binding of TXNIP to NLRP3 in the APP/PS1 mouse brain. Two-way ANOVA for TXNIP-associated NLRP3 showed significant main effects of genotype [F(1,20) = 26.331, p < 0.01] and Dl-NBP [F(1,20) = 13.281, p < 0.01], but not genotype × Dl-NBP interaction [F(1,20) = 0.616, p > 0.05] (Fig. 5D and Supplementary Fig. S2). On the contrary, Dl-NBP-mediated inhibition of TXNIP was accompanied by the prevention of TXNIP-NLRP3 interaction in the APP/PS1 mouse brain (p < 0.01, Fig. 5D). Double immunofluorescence labeling showed the morphological distribution and expression of TXNIP. TXNIP staining in the neuronal cells and neuroglia in the brains of mice was observed (Fig. 5E–H). TXNIP-positive astrocytes and microglia were significantly identifiable in the Tg mice, as exemplified by the cortical and hippocampal section shown, whereas with the decreases of immunoreactive astrocytes and microglia in the brains of Dl-NBP-treated Tg mice (as described in Figs. 3 and 4), the weaker expression of TXNIP immunoproducts was observed in the cytoplasm of granulosa neurons, in the GFAP⁺ cells of the hippocampus, and in the cortex of Dl-NBP-administered Tg mouse brain (Fig. 5E–H).

OPEN IN VIEWER

FIG. 5.

Effects of Dl-NBP treatment on regulating TXNIP expression and TXNIP-NLRP3 interaction. (A) TXNIP immunohistochemical staining in the cerebral cortex of patients with AD and normal human control (control), Scale bar = 60 μm; the right panels show the representative images of TXNIP-positive cells in high magnification. Scale bar = 30 μm. TXNIP-immunoreactive cells show the neuron-like (arrowheads) and glia-like morphology (arrows). The number of cells with high levels of TXNIP expression and the intensity of TXNIP staining were quantified. **p < 0.01 versus the controls. Data were analyzed using Student's t-test. n = 3 in each group. (B) Representative images of Western blot assays are shown for the TXNIP expression in the APP/PS1 Tg and age-matched WT mouse brain under treatment with Dl-NBP (20 mg/kg) or vehicle (vegetable oil) control. (C) mRNA levels of TXNIP in the Tg and WT brain were determined by quantitative real-time PCR. (D) TXNIP-precipitated proteins from Tg and WT mouse brains were immunoblotted with an antibody to NLRP3. Immunofluorescence double-labeled images showing the distribution and expression of TXNIP (red) within the CA1 (E) and CA3 (F) regions of hippocampal subareas, and in the cortex of WT and Tg mouse brain (G). An antibody against GFAP was used to detect the astrocytes (green). (H) TXNIP distribution in microglia of the cortex of WT and Tg mouse brain was determined by coimmunofluorescence staining with anti-TXNIP antibody (red) and a microglial marker, Iba1 (green). High-magnification images are depicted in the right panels of (G, H). Statistical significance (B) to (E): **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with two-way ANOVA followed by Student's t-test. n = 6 in each group. Scale bars: F = 60 μm, E = 30 μm, G = 50 μm, H = 20 μm. AD, Alzheimer's disease; PCR, polymerase chain reaction; TXNIP, thioredoxin-interacting protein.

Dl-NBP-mediated upregulation of Nrf2 inhibits TXNIP transcription and enhances TrX activity

We previously reported that Dl-NBP treatment upregulated Nrf2 signaling in the APP/PS1 mouse brain (67). It has been reported that Nrf2 boosted TrX activity by suppressing TXNIP in the streptozotocin-induced diabetes mouse model (20). To explore whether the Dl-NBP-mediated inhibition of TXNIP is involved in the upregulation of Nrf2 signaling, the protein expression, activity, and mRNA level of TrX were detected. As shown in Figure 6A and Supplementary Figure S3, Western blot analysis showed that the protein expression of TrX in the APP/PS1 mouse brain was comparable with that in the WT group: neither significant genotype effect [F(1,20) = 1.213, p > 0.05] nor Dl-NBP effect [F(1,20) = 0.298, p > 0.05] or genotype × Dl-NBP interaction [F(1,20) = 0.016, p > 0.05] was observed, and similar results were obtained for the mRNA levels (p > 0.05, Fig. 6B). Attenuation of TrX activity was observed in the APP/PS1 mouse brain compared with that in the WT mouse brain (p < 0.01, Fig. 6C). Under Dl-NBP treatment, the activity of TrX was significantly higher in both mouse genotypes than that in the vehicle controls (p < 0.01, Fig. 6C). The binding of TXNIP to reduced-TrX inhibits TrX activity and impairs TrX-apoptosis signal-regulating kinase 1 (ASK1) interaction, leading to the release of ASK1, which mediates stress-induced cellular apoptosis (70). Considering the negative regulation of TrX to ASK1 (61), we evaluated the protein expression of ASK1 and in the mouse brain. As shown in Figure 6A, two-way ANOVA for ASK1 protein expression indicated that neither significant genotype effect [F(1,20) = 0.030, p > 0.05] nor Dl-NBP effect [F(1,20) = 2.949, p > 0.05] or genotype × Dl-NBP interaction [F(1,20) = 0.107, p > 0.05] was observed. Phosphorylates of ASK1 at Ser83 indicate the inhibition of ASK1 activity, promoting cell survival (32). In the present study, the levels of phosphorylated Ser 83 of ASK1 (p-ASK1) were markedly lower in the Tg mouse brain compared with those in the WT mouse brain (p < 0.01, Fig. 6A). Dl-NBP treatment led to an increase of p-ASK1 level in both mouse genotypes compared with those in the controls (p < 0.01, Fig. 6A). Furthermore, co-IP assays for TrX-associated ASK1 showed significant main effects of genotype [F(1,20) = 28.743, p < 0.01] and Dl-NBP [F(1,20) = 22.264, p < 0.01], but not significant genotype × Dl-NBP interactions [F(1,20) = 2.420, p > 0.05] (Fig. 6D and Supplementary Fig. S3). The TrX-ASK1 interaction in the Tg mouse brain was less than of the WT group (p < 0.01, Fig. 6D). Thus, according to previous work reported by Zhou and colleagues (75) and our results, the dissociation of TXNIP from oxidized TrX likely facilitates TXNIP activation and interactions with the NLRP3 inflammasome in AD. We then performed chromatin immunoprecipitation (ChIP) to analyze the presence of Nrf2 at ARE sites on the promoters of TXNIP genes in the brains of Tg and WT mice. As shown in Figure 6E, the binding of Nrf2 to the TXNIP ARE was lower in the APP/PS1 mouse brain than WT mice (p < 0.01). Dl-NBP treatment significantly increased the binding of Nrf2 to the TXNIP ARE in the brains of both mouse genotypes compared with the vehicle controls (Fig. 6E, p < 0.01). MondoA, a transcription factor of the basic helix-loop-helix leucine zipper family, is required for TXNIP induction (43, 59). The binding of MondoA to carbohydrate response element (ChoRE) is involved in the positive regulation for TXNIP transcription (20, 71). It is reported that Nrf2 may suppress MondoA-ChoRE interaction to keep TXNIP expression at a low level, preventing diabetes (20). To investigate whether the effects of Nrf2 on TXNIP were related to Nrf2-mediated blocks on MondoA-ChoRE interaction, we assessed the binding of MondoA to TXNIP ChoRE-a by ChIP. The increases of MondoA-ChoRE were observed in the APP/PS1 mouse brain relative to that in the age-matched WT mice (Fig. 6F, p < 0.01), whereas Dl-NBP management significantly reduced the recruitment of MondoA to ChoRE in comparison with those of controls in both mouse genotypes (Fig. 6F, p < 0.01). We previously reported that the nuclear levels of Nrf2 were reduced in the APP/PS1 mouse brain compared with those of the age-matched WT mouse brain (67). In the present study, we detected the location and expression of Nrf2 in the CA1 region of hippocampus and the frontal cortex from WT and Tg mouse brain. Immunofluorescence labeling showed that Nrf2 reactivity was observed in the cytoplasm or in addition within the nucleus of neuronal granulosa cells (Fig. 6G) and in the neurons of the cortex (Fig. 6H). The immunofluorescence staining for nuclear Nrf2 in the hippocampus showed significant main effects of genotype [F(1,20) = 17.640, p < 0.01] and Dl-NBP [F(1,20) = 15.089, p < 0.01], but not genotype × Dl-NBP interaction [F(1,20) = 2.073, p > 0.05] (Fig. 6G and Supplementary Fig. S4). The quantifications of cytosolic Nrf2 staining in the hippocampus were observed as well. Similar results of Nrf2 staining in the frontal cortex were observed (Fig. 6H and Supplementary Fig. S4). The intensity of nuclear and cytosolic Nrf2 staining was weaker in the Tg group than in the WT mice (p < 0.01). In contrast, Dl-NBP management caused significant increases in the intensity of nuclear and cytosolic Nrf2 in both mouse genotypes relative to their vehicle controls (p < 0.01).

OPEN IN VIEWER

FIG. 6.

Dl-NBP-mediated Nrf2 upregulation increases Nrf2 recruitment to TXNIP ARE and enhances TrX activity. (A) Quantification of the protein expression of TrX, ASK1, and p-ASK1 is shown in Tg APP/PS1 and age-matched WT C57BL/6 mice treated with DL-NBP (20 mg/Kg) or vehicle (vegetable oil). (B) The relative mRNA levels of TrX normalized to that of β-tubulin are shown. (C) The activity of TrX was analyzed. (D) The binding of TrX-ASK1 is evaluated by Co-IP analysis. (E, F) The effects of Dl-NBP treatment on regulating TXNIP transcription in the Tg and WT mouse brain are indicated. ChIP analyses were conducted to explore the binding of Nrf2 to TXNIP. The chromatin was first immunoprecipitated with an antibody against Nrf2, or anti-MondoA. The binding of Nrf2 to the TXNIP ARE promoter or the binding of MondoA to ChoRE-a was detected by quantitative real-time PCR. The immunoprecipitation data are represented as the percentage of input DNA (internal control), which is relative to the amount of nonimmune rabbit IgG used. The results of the recruitment of MondoA to the TXNIP promoter are shown. (G) Immunofluorescence images showing the Nrf2 reactivity (red) in the CA1 area of hippocampus. The nuclei were labeled with DAPI (blue). Scale bars in G = 50 μm. High-magnification image of Nrf2-positive cells is depicted in the right panels, and scale bar = 20 μm. (H) Representative images of Nrf2 staining in frontal cortex. Brain sections are labeled for Nrf2 (red) plus the neuronal marker NeuN (green). DAPI (blue) is a nuclear stain. Scale bars in H = 30 μm. High-magnification image (right panels) shows the localization of Nrf2 and NeuN, and scale bar = 10 μm. Statistical significance: **p < 0.01 versus the vehicle group; ^## p < 0.01 versus the WT group. Data were analyzed with two-way ANOVA followed by Student's t-test. n = 6 in each group. ASK1, apoptosis signal-regulating kinase 1; ChIP, chromatin immunoprecipitation; DAPI, 4′,6-diamidino-2-phenylindole; Nrf2, nuclear factor erythroid 2-related factor 2; p-ASK1, phosphate apoptosis signal-regulating kinase 1.

Dl-NBP treatment inhibits IL-1β secretion and caspase-1 activation, mitigating amyloidosis in neuronal cells

Considering the hypothesis that Nrf2 is the “master regulator” of the antioxidant response and modulates immunological signaling (25), we verified the effect of Dl-NBP on inflammatory mediator-triggered cell damage in human-derived glial A172 cells in vitro. As a major component of all gram-negative bacterial cell walls, lipopolysaccharides (LPS) are considered to be an inducer to trigger proinflammatory responses. It has been reported that suppression of the LPS-triggered inflammatory system was Nrf2 pathway relevant in macrophages in vitro (44). In the present study, LPS was used to induce neuroinflammation in A172 cells. Western blot analysis revealed that LPS (1 μg/mL) stimulated caspase-1 activation, as confirmed by the appearance of processed caspase-1 (p < 0.01, Fig. 7A and Supplementary Fig. S5). Elevated secretion of mature IL-1β was also observed in the culture medium of LPS-treated A172 cells relative to vehicle controls (p < 0.01, Fig. 7B). However, treatment with Dl-NBP reduced LPS-mediated protein expression of processed caspase-1 (p < 0.01, Fig. 7A) in A172 cells. A significant decrease in IL-1β levels was observed in the culture medium supernatants of Dl-NBP-treated A172 cells compared with those in the LPS-treated group (p < 0.01, Fig. 7B). In addition to caspase-1, we also assessed cell apoptosis by analyzing annexin V-FITC/propidium iodide (PI) staining. As shown in Figure 7C, flow cytometry assays showed that the apoptosis rate was elevated following LPS stimulation compared with controls (p < 0.01), whereas treatment with Dl-NBP significantly reduced the percentage of apoptotic cells (p < 0.01; Fig. 7C).

OPEN IN VIEWER

FIG. 7.

Dl-NBP treatment abrogates the effects in which the gliacyte-secreted IL-1β induces increase of Aβ in neuron cells. Human glioblastoma A172 cells were first treated with LPS (1 μg/mL) for 24 h. After the culture medium was replaced with serum-free medium, the cells were incubated with Dl-NBP at a final concentration of 10 μM for 12 h. Cells treated with vehicle were used as controls. (A) The protein expression of procaspase-1 and the processed caspase-1 in whole-cell lysates and precipitated supernatants were determined by Western blot analyses. (B) IL-1β secretion was measured in culture supernatants by ELISA. (C) Annexin V-FITC/PI staining and flow cytometry were performed to assess the extent of apoptosis. The ratio of annexin-V-positive to PI-negative cells indicates cell apoptosis. (D) SH-SY5Y cells overexpressing human APPswe or empty vector (neo) were cultured with medium that was collected from A172 cells in the presence or absence of LPS (A172 Condi or Control, respectively) with or without the Dl-NBP treatment. Aβ1–40 and Aβ1–42 levels in cell supernatants were evaluated by sandwich ELISA. The ratio of Aβ1–40 to Aβ1–42 was quantified. (E) Western blot assays show the protein expression of BACE-1 and p-APP668 in neo cells and APPswe cells under indicated cultures. (F) Specific peptides conjugated to the reporter molecules EDANS and DABCYL were used to assess the β- and γ-secretase activity of neo and APPswe cells. Data are representative of at least three independent studies. Statistical significance: **p < 0.01 versus the control group; ^## p < 0.01 versus the LPS-treated group (A–C); **p < 0.01 versus the group cultured with A172 vehicle medium (A172 vehicle); and ^## p < 0.01 versus the group treated by LPS-primed A172 conditioned medium and ^$$ p < 0.01 versus the neo cells. (D–F). Data were analyzed with one-way ANOVA with post-hoc Fisher's PLSD tests. APPswe, Swedish mutant form of APP; ELISA, enzyme-linked immunosorbent assay; FITC, fluorescein isothiocyanate; LPS, lipopolysaccharides; PI, propidium iodide.

It has been reported that elevated IL-1β levels contribute to an increased risk of AD (45). We previously observed that Dl-NBP treatment led to decreases in soluble Aβ1–40, soluble Aβ1–42, and Aβ oligomers in the APP/PS1 mouse brain (67). It is speculated that sustained activation or impairment of glia results in a chronic inflammatory process, giving rise to neuronal dysfunction in AD-like pathology (7). To investigate whether the increased secretion of IL-1β in A172 cells facilitates the production of Aβ in human neuronal cells and to confirm the effects of Dl-NBP-mediated inhibition of neuroinflammation on Aβ accumulation, experiments were performed to treat SH-SY5Y cells overexpressing the APPswe or empty vector (neo) with conditioned medium from A172 cells with or without Dl-NBP treatment. IL-1β antibody (1 μg/mL) was added to confirm that IL-1β secretion from LPS-primed A172 cells was essential for inducing amyloidosis in APPswe or neo cells. We discovered that Dl-NBP (p < 0.01, Fig. 7D) or IL-1β antibody (data not shown) incubation abrogated the effects of LPS-treated conditioned medium from A172 cells, and the conditioned medium facilitated the increases of Aβ1–40 and Aβ1–42 secretion in both APPswe and neo cells (p < 0.01, Fig. 7D). Compared with the neo cells, the protein expression of β-site of APP-cleaving enzyme (BACE-1) and phospho-APP (p-APP668) was significantly higher in the APPswe cells cultured with the conditioned medium from LPS-primed A172 cells (p < 0.01; Fig. 7E and Supplementary Fig. S5), and the alterations of the activity in the β- and γ-secretase were as well (p < 0.01, Fig. 7F), however, Dl-NBP treatment partly inhibited the protein expression of BACE-1 and p-APP668 (p < 0.01, Fig. 7E), and repressed the upregulation of β- and γ-secretase activity (p < 0.01, Fig. 7F). Our results highlight an inflammatory-mediated reciprocal regulation between glial and neuronal cells that may contribute AD-like pathology, and suggest that Dl-NBP-mediated inhibition of IL-1β secretion facilitates the non-amyloidogenic processing in neuronal cells.

Dl-NBP-mediated upregulation of Nrf2 induces TXNIP inhibition and reduces TXNIP-NLRP3 binding

Previous study has demonstrated that TXNIP is inhibited by Nrf2 in the diabetic mouse heart (20). Considering our present findings that Dl-NBP treatment upregulates Nrf2 and inhibits TXNIP transcription in the APP/PS1 mouse brain, we examined the effects of Dl-NBP on the inflammation-induced signaling cascade of TXNIP in A172 cells in vitro. Our data revealed that LPS chronic treatment reduced nuclear protein expression (p < 0.01, Fig. 8A and Supplementary Fig. S5) and DNA binding activity of Nrf2, when compared with A172 cells on vehicle incubation (p < 0.01, Fig. 8B). However, the protein expression of cytosolic Nrf2 was not significantly altered (p > 0.05, Fig. 8A). Marked increases in protein expression (p < 0.01, Fig. 8C and Supplementary Fig. S6) and mRNA levels (p < 0.01, Fig. 8D) of TXNIP were observed in LPS-primed A172 cells compared with those of vehicle controls. There were no significant differences in the protein expression (p > 0.05, Fig. 8C) and mRNA levels (p > 0.05, Fig. 8D) of TrX among groups under indicated treatment. The activity of TrX in LPS-primed A172 cells was lower than in the controls (p < 0.01, Fig. 8E). Interestingly, the incubation of A172 cells with Dl-NBP (10 μM) significantly increased the levels of nuclear Nrf2 (p < 0.01, Fig. 8A) and Nrf2 DNA binding activity (p < 0.01, Fig. 8B). Dl-NBP treatment also inhibited the LPS-induced upregulation of TXNIP (p < 0.01, Fig. 8C). Accordingly, in the LPS-primed A172 cells, the TrX activity was higher under Dl-NBP administration than that without Dl-NBP addition (p < 0.01, Fig. 8D). IP assays showed a differential increase of TXNIP-associated NLRP3 protein expression in the LPS-primed A172 cells, when compared with A172 cells on vehicle incubation (p < 0.01, Fig. 8F and Supplementary Fig. S6). In contrast, Dl-NBP treatment caused a significant reduction in the interaction between TXNIP and NLRP3 relative to those of controls or LPS-primed A172 cells (p < 0.01, Fig. 8F). Interestingly, under treatment with the A172 condition medium, the binding of TXNIP to NLRP3 in the APPswe cells was higher than that in the neo cells (p < 0.01, Fig. 8G and Supplementary Fig. S6). Importantly, the TXNIP-associated NLRP3 protein was reduced under Dl-NBP treatment in both neo cells and APPswe cells, when compared with those treated by conditioned medium without Dl-NBP addition (p < 0.01, Fig. 8G). To verify the specificity of the pharmacological effects of Dl-NBP, A172 cells were transfected with small interfering RNAs (siRNAs) targeting Nrf2, treated with adenoviral overexpression of TXNIP (Ad-TXNIP) or incubated with 10 μM trigonelline (C7H7NO2, Nrf2 inhibitor) (1) before being incubated with Dl-NBP (10 μM). Co-IP analysis showed that Nrf2 siRNA (Fig. 8H and Supplementary Fig. S7), trigonelline treatment (Fig. 8I and Supplementary Fig. S7), or Ad-TXNIP (Fig. 8J and Supplementary Fig. S7) partly abrogated Dl-NBP-mediated inhibition of NLRP3-TXNIP interaction (p < 0.01) in A172 cells. These data further indicate that Dl-NBP-induced upregulation of Nrf2 inhibits TXNIP transcription and expression, blocks the TXNIP-NLRP3 interaction, and enhances TrX-mediated redox homeostasis.

OPEN IN VIEWER

FIG. 8.

Dl-NBP inhibits NLRP3 inflammasome activation via the NRF2-TXNIP-TrX axis. (A) A172 cells were first treated with LPS (1 μg/mL) for 24 h followed by administration with Dl-NBP (10 μM) for 12 h. Cells treated with vehicle served as controls. Western blot analyses were performed to determine the protein expression of nuclear and cytosolic Nrf2. (B) Results of ELISA show the Nrf2 DNA binding activity. (C) Representative images showing the protein expression of TXNIP, TrX, and GAPDH in A172 cells with or without LPS treatment. (D) Quantitative real-time PCR results show the mRNA level of TXNIP and TrX. (E) The activity of TrX was detected. TXNIP-associated NLRP3 protein alterations were determined by Co-IP assays (F). (G) SH-SY5Y cells overexpressing human APPswe or empty vector (neo) were treated with conditioned medium obtained from A172 cells in the presence (A172 Condi) or absence (Control) of LPS with or without the Dl-NBP addition. The binding of TXNIP to NLRP3 in the neo and APPswe cells was probed by Co-IP assays. (H) A172 cells were transfected with 60 nM of Nrf2 siRNA before treating the cells with Dl-NBP (10 μM) for 12 h. Scrambled siRNA served as control. The interactions of TXNIP-NLRP3 were detected. Co-IP NLRP3 is shown. (I) Trigonelline (Nrf2 inhibitor) was administered to identify the role of Dl-NBP on regulating Nrf2 in A172 cells. The cells were subsequently treated with Dl-NBP (10 μm, 12 h) followed by trigonelline (10 μM, 12 h). The levels of TXNIP-associated NLRP3 protein were determined. (J) Overexpression of TXNIP was performed by gene transfer using an adenovirus vector encoding human TXNIP (Ad-TXNIP) in A172 cells before treating with Dl-NBP (10 μM, 12 h). Co-IP analyses showed that the decreases induced by Dl-NBP on TXNIP-associated NLRP3 protein are partly repressed under the treatment of Ad-TXNIP. Data are representative of at least three independent studies. Statistical significance: **p < 0.01 versus the vehicle control; ^## p < 0.01 versus the LPS-treated group (A–G); **p < 0.01 versus controls; ^## p < 0.01 versus the Dl-NBP treatment group (H–J). Data were analyzed with one-way ANOVA with post-hoc Fisher's PLSD tests. Ad-TXNIP, adenoviral overexpressed TXNIP; siRNA, small interfering RNA; TrX, thioredoxin.

Animals and drug treatments

Male APP/PS1 (B6C3-Tg [APPswe, PSEN1dE9] 85Dbo/Mmjax) double Tg mice were originally generated by and purchased from the Jackson Laboratory (West Grove, PA). The APP/PS1 mouse expresses a chimeric mouse/human Swedish mutation for amyloid precursor protein and a mutant form of human presenilin 1 driven by a central nervous system-specific neuronal promoter. Age-matched C57BL/6 mice (Jackson Laboratory) were used as WT controls. The experimental procedures were performed in compliance with the Chinese regulations involving animal protection and were approved by the Ethics Committee for Animal Use of China Medical University.

The procedures for Dl-NBP administration were performed exactly as previously described (67). Briefly, 4-month-old APP/PS1 mice and age-matched WT mice were randomly divided into two groups: one was treated with Dl-NBP at a dosage of 20 mg/kg body weight by oral gavage once daily for 5 months and the other was administered vehicle (vegetable oil, 50 μL). Dl-NBP (purity, 99.6%; Shijiazhuang Pharma Group NBP Pharmaceutical Co., Ltd, Shijiazhuang, Hebei, China) was dissolved in vegetable oil at 10 mg/mL. The general behavior of all mice was monitored daily, and the body weight of each mouse was recorded weekly.

Tissue preparation

Twenty-four hours after the last oral administration with Dl-NBP or vehicle, mice were intraperitoneally anesthetized with sodium pentobarbital at a dose of 50 mg/kg. After transcardial perfusion with ice-cold saline, the mice were sacrificed by decapitation. The brains were sagittally dissected into two halves on an ice-cold surface. The right hemisphere was prepared for morphological assessment. The left hemisphere was snap-frozen in liquid nitrogen and stored at −80°C for biochemical analyses.

Quantification of apoptosis

For the analysis of TUNEL, a commercial kit was used (In Situ Cell Death Detection Kit, Roche Diagnostics, Indianapolis, IN) according to the manufacturer's instructions. Briefly, coronal frozen mouse brain sections (6 μm) underwent 4% paraformaldehyde fixation, and antigen retrieval was performed by boiling in Tris, EDTA, and glycine buffer (pH 8.0) for 4 min in a microwave oven. After washes with phosphate-buffered saline (pH 7.4), the slides were treated with 0.1% sodium citrate plus 0.1% Triton X-100 for 8 min to permeate the nucleus. After rinsing, the slides were blocked in blocking buffer for 30 min and then were incubated with TUNEL reaction mixture (60 min, 37°C). After several washes, the slides were analyzed under a confocal laser scanning microscope (Model SP2; Leica, Wetzlar, Germany). The images of apoptosis-positive cells were captured using an excitation filter of 488 nm.

After the indicated treatment periods in vitro, apoptotic cells were quantified by flow cytometry using the annexin V-FITC/PI apoptosis detection kit (Biovision, Milpitas, CA). Briefly, after trypsinization, the cultured cells were treated with serum-containing media. The cells were rinsed using ice-cold rinsing buffer (pH 7.4) containing 10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl₂. The cells were then incubated using 500 μL of binding buffer with 5 μL of annexin V-FITC and 5 μL of PI. The samples were then incubated for 15 min at room temperature in the dark. Cell apoptosis was quantified using the FITC (FL1) and PI (FL2) signal detectors.

Determination of TrX activity

TrX activity was measured by colorimetry with the Thioredoxin Reductase Assay Kit (Sigma-Aldrich, St. Louis, MO). Briefly, brain tissues or treated cells were homogenized and lysed in lysis buffer containing 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% sodium dodecyl sulfate (SDS), and 50 mM Tris (pH 8.0) plus protease inhibitor cocktail (Sigma-Aldrich). The lysate was centrifuged at 10,000 g for 10 min at 4°C. The total protein concentration was determined using a UV 1700 PharmaSpec ultraviolet spectrophotometer (Shimadzu, Tokyo, Japan). The substrate in the kit is 5,5-dithiobis (2-nitrobenzoic) acid (DTNB). The TrX reductase reduction reactions occur as follows: two moles of 5-thio-2-nitrobenzoic acid (TNB) are formed for every 1 mol of NADPH oxidized. The reduction of DTNB with NADPH to TNB produces a maximum absorption peak at 412 nm. The levels of total DTNB reduction by the sample and the amount of DTNB reduction by the sample in the presence of the reductase inhibitor solution were determined. The difference between these two results indicates the DTNB reduction, which is attributed to the level of TrX reductase activity.

Western blot analysis

Mouse brain tissues and cultured cells were collected and homogenized by sonication in ice-cold RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, and 50 mM Tris, pH 8.0) with a protease inhibitor cocktail addition (Sigma-Aldrich). After centrifugation at 15,000 g at 4°C for 30 min, the supernatants were collected, and protein concentrations were measured using the UV 1700 PharmaSpec ultraviolet spectrophotometer. Equal amounts of protein (20 μg) were separated by 10% SDS polyacrylamide gels. The protein samples were transferred to PVDF membranes (Millipore, Billerica, MA) and probed with antibodies as follows: rabbit monoclonal antibody to caspase-1 (1:1000, ab179515; Abcam, Cambridge, MA) for recognizing procaspase 1 and the cleavage fragment, p10 (p12); rabbit anti-caspase-3 (1:1000, 9662; Cell Signaling Technology, Danvers, MA) for detecting full-length caspase-3 and the large cleaved fragment from caspase-3; rabbit anti-mature IL-1β (1:1000; Abcam), mouse anti-IL-1β (1:1000; Cell Signaling Technology), mouse anti-TXNIP (1:800; Santa Cruz Biotechnology), rabbit anti-ASC (1:1000; Cell Signaling Technology), rabbit anti-NLRP3 (1:1000; Cell Signaling Technology), rabbit anti-Nrf2 (1:1000; Abcam), rabbit anti-TrX (1:1000; Abcam), rabbit anti-ASK1 (1:1000; Cell Signaling Technology), rabbit anti-phospho-ASK1 (Ser83, 1:800; Santa Cruz Biotechnology), rabbit anti-phospho-APP (Thr668, 1:1000; Cell Signaling Technology), rabbit anti-BACE1 (1:1000; Sigma-Aldrich), mouse anti-GAPDH (1:10,000; Kangchen Biotech, Shanghai, China), and rabbit anti-lamin B1 antibody (1:1000; Abcam). After incubation with the appropriate HRP-conjugated secondary antibodies, the immunological complexes were visualized with enhanced chemiluminescence (Amersham Biosciences).

Immunohistochemistry and confocal laser scanning microscopy

Four human postmortem brain tissues (AD: serial numbers T4304, T4339, and 235-95; normal: P535-00) were obtained from the New York Brain Bank. Another two human brain tissues (normal) were obtained from Fengtian Hospital of China (one from a 59-year old and the other from a 63-year old). Samples were incubated with rabbit anti-TXNIP antibody (1: 100; Abcam) for 16 h at 4°C. After thorough rinses, the samples were then treated with biotinylated IgG (1: 200) for 1 h and incubated with streptavidin peroxidase for 30 min. Following treatment with TBS containing 0.025% 3,3-diaminobenzidine and 0.0033% H₂O₂ for 5 min, the sections were dehydrated, cleared, mounted, and examined with an Olympus microscope (Model DP71). Coronal frozen mouse brain sections (6 μm) were first blocked with 3% normal bovine serum albumin at room temperature for 30 min. The samples were then incubated with a mixture of primary antibodies of either rabbit anti-mature IL-1β (1:200; Abcam) or mouse anti-Aβ (1:500; Sigma) and goat anti-Iba1 (1:100; Abcam), rabbit anti-Aβ (1:100; Abcam) and mouse anti-GFAP (1:200; Cell Signaling Technology), rabbit anti-TXNIP (1:100; Abcam) and mouse anti-GFAP (1:200; Cell Signaling Technology), rabbit anti-TXNIP and goat anti-Iba-1 (1:100; Abcam), rabbit anti-TXNIP and mouse anti-NeuN (1:100; EMD Millipore), rabbit anti-Nrf2 (1:100; Abcam) and mouse anti-NeuN (1:100) overnight at 4°C. After rinsing, the samples were incubated with the appropriate secondary antibodies for 2 h at room temperature. The images were acquired using a confocal laser scanning microscope (Model SP2; Leica) and analyzed with Nikon EclipseNet and ImageJ software (National Institutes of Health, Bethesda, MD). Dystrophic astrocytes and microglia were determined by calculating the debris score according to a subjective scale of 0-3 μm with NIH Image software. Polarization index of microglia to Aβ plaques was assessed as previously described (72).

IP analyses

IP assays were performed as previously described (67). Protein from mouse brain tissue or cell lysates was incubated with primary antibodies against TXNIP, NLRP3, or TrX at 4°C for 12 h. Immune complexes were captured with TrueBlot IP beads (eBioscience, Hatfield, United Kingdom) at 4°C for 4 h. After rinsing three times with IP lysis buffer, the bound proteins were boiled for 5 min and then eluted before performing Western blot analysis. Proteins bands were resolved and blotted with antibodies anti-NLRP3, anti-ASC, or anti-ASK1 at 4°C for 12 h.

Chromatin IP assay

ChIP was performed as previously described (67). Briefly, mouse brain tissues were treated with 4% paraformaldehyde for chromatin crosslinking. After neutralization in 0.125 M glycine, the samples were homogenized by sonication and were lysed in SDS lysis buffer (pH 8.1) containing protease inhibitors (Roche Molecular Biochemicals). One-third of the protein samples were saved as the input genomic DNA control. The remaining lysates were mixed with an antibody against Nrf2 (1:40; Abcam), MondoA (1:20; Proteintech), or nonimmune IgG (negative control). The IPs containing protein-DNA complexes were washed and then captured with protein A-agarose beads. After thorough rinsing, the precipitates were incubated with elution buffer (pH 8.0). To reverse crosslinking, the input tissue and elutes were added to 20 μL of 5 M NaCl and heated at 65°C for 4 h (27). The protein-DNA complexes were treated with 20 μg/mL of proteinase K (1 h, 45°C). DNA was purified with a PCR purification kit (Qiagen, Valencia, CA). Real-time PCR was performed using samples of ChIP DNA with the primers for amplifying TXNIP ARE or TXNIP ChoRE-a.

Cell culture and treatment

Human glioblastoma A172 and human SH-SY5Y neuroblastoma cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA). A172 cells were cultured in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 medium (Life Technologies) at a ratio of 1:1 supplemented with 10% fetal bovine serum (FBS; Hyclone, UT) and 1% penicillin/streptomycin (Thermo Fisher Scientific). SH-SY5Y cells stably transfected with the human APPswe or empty vector (neo) (74) were grown in DMEM containing 10% FBS and 40 μg/mL gentamicin. Cells were incubated in a fully humidified, 5% CO₂ incubator at 37°C. When cells were ∼80% confluent, the cells were synchronized by overnight exposure to fresh media without serum or antibiotics. The cells were then treated with LPS, Dl-NBP, and Nrf2 siRNA using Lipofectamine 2000, trigonelline, or adenoviral gene transfer (Ad-TXNIP) as indicated. Untreated cells were used as vehicle control. The cells and culture supernatants were harvested for analyses.

Measurement of Aβ1–40 and Aβ1–42 levels

Soluble Aβ1–40 and Aβ1–42 levels in the culture medium of human APPswe and neo cells were detected by sandwich enzyme-linked immunosorbent assay (ELISA). The samples were loaded onto 96-well plates, and the levels of Aβ1–40 and Aβ1–42 were measured according to the manufacturer's instructions of the ELISA kit (Invitrogen, Carlsbad, CA). The absorbance was measured at 450 nm.

β-secretase and γ-secretase activity analyses

The activity of β-secretase and γ-secretase in the lysates of APPswe or neo cells was determined using specific peptides that were conjugated to the reporter molecules EDANS and DABCYL (R&D Systems, Minneapolis, MN). Briefly, after the indicated intervention, cells were collected and treated with extraction buffer. After homogenization, the supernatants were harvested. The samples were mixed with the reaction buffer and substrate, and were incubated in the dark (2 h, 37°C). The absorbance was measured using a fluorescent microplate reader.

Nrf2 DNA binding activity assessment

First, following the indicated treatment, nuclear Nrf2 was extracted from A172 cells with a nuclear extract kit (Active Motif, Inc., Carlsbad, CA) in accordance with the manufacturer's instructions. The samples were then treated with buffer A containing 40 μL of 10% Nonidet P (NP-40). Centrifugation was performed, and the pellets were resuspended in a prechilled, hypertonic, nuclear extraction buffer B containing 50 mM HEPES, 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 0.5 mM dithiothreitol, 0.1 mM PMSF, and 20% glycerol (pH 7.9). The supernatant containing the nuclear proteins was obtained, and the Nrf2 DNA binding activity of nuclear extracts was determined with the TransAM ELISA kit (Active Motif).

Adenoviral gene transfer for TXNIP overexpression

An adenovirus vector encoding TXNIP was generated for adenoviral gene transfer (Ad-TXNIP) as previously described (62). Viruses were propagated in HEK293 cells, purified, and the stock titer was 10⁹ infectious units/mL. A recombinant adenovirus encoding green fluorescent protein (Ad-GFP) was used as an internal control. After preliminary dose/response experiments for TXNIP expression, 1000 μL of stock in 1.75 mL of media was chosen for the infection in A172 cells, which were plated at a cell density of 1 × 10⁵/cm². The A172 cells were infected for 3 h and then recovered for 24 h in media without serum or antibiotics.

Determination of IL-1β secretion in the culture medium

The levels of proinflammatory cytokines IL-1β in the supernatants of the pharmacologically or chemically treated cells and the control cells were determined using the IL-1β ELISA kit (eBioscience, CA) in accordance with the manufacturer's instructions. The total amount of protein in the medium was used as the loading control.

Quantitative real-time PCR

Total RNA from mouse brain tissue and cultured cells was isolated using TRIzol (Invitrogen). A total of 2 μg of RNA from each sample was reverse transcribed using the Prime Script RT Reagent Kit (Takara, Otsu, Japan). Each gene amplification was performed using SYBR Green PCR Master mix (Applied Biosystems, Inc. [ABI], Foster City, CA) with an ABI 7300 Sequence Detection System. The following primers were used: mouse TXNIP: 5′-ATGGCCAGACCAAAGTGTTC-3′ (sense) and 5′-GGCTGTCTTGAGAGTCGTCC-3′ (antisense); mouse TrX: 5′-CAACAGCCAAAATGGTGAAGCTG-3′ (sense) and 5′-AGGTTTTAAACAGCTG-3′ (antisense); mouse NLRP3: 5′-AGCCTTCCAGGATCCTCTTC-3′ (sense) and 5′-CTTGGGCAGCAGTTTCTTTC-3′ (antisense); mouse ASC: 5′-GAAGCTGCTGACAGTGCAAC-3′ (sense) and 5′-GCCACAGCTCCAGACTCTTC-3′ (antisense); mouse β-tubulin: 5′-CTGGGCTAAAGGCCAC-3′ (sense) and 5′-AGACACTTTGGGCGAG-3′ (antisense); human TXNIP: 5′-TGGTGGATGTCAATACCCCT-3′ (sense) and 5′-ATTGGCAAGGTAAGTGTGGC-3′ (antisense); human TrX: 5′-AGCAGCCAAGATGGTGAAGCAGA-3′ (sense) and 5′-GCTCCAGAAAATTCACCCACC-3′ (antisense); human NLRP3: 5′-AACAGCCACCTCACTTCCAG-3′ (sense) and 5′-CCAACCACAATCTCCGAATG-3′ (antisense); human ASC: 5′-ACTCATTGCCAGGGTCACAGAAGTG-3′ (sense) and 5′-GCTTCCTCATCTTGTCTTGGCTGGT-3′ (antisense); and human β-tubulin: 5′-TCTGTTCGCTCAGGTCCTTT-3′ (sense) and 5′-TTCATGATGCGATCAGGGTA-3′ (antisense).

Statistical analysis

All data are presented as mean ± standard error of the mean. Statistical significance of the differences between groups was determined using one-way or two-way ANOVA. The genotype × treatment interaction, together with main effects of genotype and treatment, was analyzed. p < 0.05 was considered statistically significant.