Fenamate NSAIDs inhibit the NLRP3 inflammasome and protect against Alzheimer's disease in rodent models

Abstract

Non-steroidal anti-inflammatory drugs (NSAIDs) inhibit cyclooxygenase-1 (COX-1) and COX-2 enzymes. The NLRP3 inflammasome is a multi-protein complex responsible for the processing of the proinflammatory cytokine interleukin-1β and is implicated in many inflammatory diseases. Here we show that several clinically approved and widely used NSAIDs of the fenamate class are effective and selective inhibitors of the NLRP3 inflammasome via inhibition of the volume-regulated anion channel in macrophages, independently of COX enzymes. Flufenamic acid and mefenamic acid are efficacious in NLRP3-dependent rodent models of inflammation in air pouch and peritoneum. We also show therapeutic effects of fenamates using a model of amyloid beta induced memory loss and a transgenic mouse model of Alzheimer's disease. These data suggest that fenamate NSAIDs could be repurposed as NLRP3 inflammasome inhibitors and Alzheimer's disease therapeutics.

Figure 1. Fenamate NSAIDs inhibit IL-1β processing and release.

(a) iBMDMs were primed with LPS (1 μg ml⁻¹, 2 h) then pre-treated with NSAID at indicated concentration before stimulating with ATP (5 mM, 1 h). (b–d) Murine primary BMDMs from WT (b) or NLRP3^−/− (c,d) mice were primed with LPS (1 μg ml⁻¹, 4 h) and pre-treated with NSAID (100 μM, 15 min) before stimulating with monosodium urate (MSU) crystals (250 μg ml⁻¹, 4 h) (b), transfected ultrapure flagellin from Salmonella typhimurium (1 ng per 1,000 cells, 2 h) (c), or transfected DNA (0.66 ng per 1,000 cells, 4 h) (d). Supernatants were analysed by ELISA. Data are presented as mean % IL-1β release versus vehicle (DMSO) control+s.e.m (n=3 or 4). NS, not significantly different, *P<0.05, **P<0.01 determined by one-sample t-test versus hypothetical value of 100%.

Figure 2. Fenamate NSAIDs inhibit ASC speck formation and caspase-1 activation.

(a) iBMDMs stably expressing ASC protein conjugated to mCherry were primed with LPS (1 μg ml⁻¹, 2 h) then pre-treated with selected drug (100 μM, 15 min) before stimulation with ATP (5 mM, 30 min) under live microscopy. Formation of ASC specks (examples indicated by white arrows) was quantified (lower right) and presented as mean % specks counted versus vehicle+s.e.m (n=4). Scale bars are 20 μm. (b,c) THP-1 cells were primed with LPS (1 μg ml⁻¹, 4 h) and pre-treated with NSAID (200 μM, 15 min) before stimulating with nigericin (10 μM, 1 h). Supernatants were taken and analysed for IL-1β by ELISA (b) and the p10 active subunit of caspase-1 by western blot (c). ELISA data are presented as mean % IL-1β release versus vehicle (DMSO) control+s.e.m (n=3). *P<0.05, ***P<0.001 determined by one-sample t-test versus hypothetical value of 100%.

Figure 3. Fenamates inhibit NLRP3 via reversible blockade of the Cl⁻ channel VRAC.

(a) iBMDMs were primed with LPS (1 μg ml⁻¹, 4 h) and pre-treated with NSAID, reversible caspase-1 inhibitor YVAD-CHO or irreversible NLRP3 inhibitor 3,4-methylenedioxy-β-nitrostyrene (MNS) (100 μM, 15 min) before washing 3 × 5 min (or no wash) and stimulation with ATP (5 mM, 1 h). (b) Cells were pre-treated with drug as above before either media replacement without drug, media replacement with drug or no media replacement followed by ATP stimulation. (c–e) iBMDMs were primed as above. Currents were measured in the whole-cell configuration of the patch clamp technique. Voltage ramps were applied from −90 mV to +90 mV for a duration of 360 ms every 20 s. Non selective cation currents were evoked by 5 mM ATP and were measured in the absence (ATP) or presence of 100 μM flufenamic acid (c, n=8) or mefenamic acid (d, n=10). (f–j) iBMDMs were primed as above before whole-cell volume-regulated chloride currents (VRAC) were induced by hypotonic extracellular medium. Currents were induced by 360 ms-lasting voltage ramps from −90 mV to +90 mV every 20 s. Examples of current recordings measured in the absence or presence of 100 μM ibuprofen (f, n=6), diclofenac (g, n=5), flufenamic acid (h, n=6), or mefenamic acid (i, n=6). ***P<0.001, NS, not significantly different, determined by one-sample t-test versus hypothetical value of 100%. (K) iBMDMs were primed as above and pre-treated with flufenamic acid, mefenamic acid, Cl⁻ channel blockers NPPB or DIDS (100 μM, 15 min) before stimulation with ATP (5 mM, 1 h). (l) iBMDMs were primed as above and pre-treated with DCPIB (10 μM, 15 min) before stimulation with ATP (5 mM, 1 h). IL-1β in the supernatants (a,b,k,l) was quantified by ELISA and data are presented as mean % IL-1β release versus vehicle (DMSO) control+s.e.m or mean IL-1β release (n=4). *P<0.05, **P<0.01, ***P<0.001, NS, not significantly different determined by one-way ANOVA with Tukey's post hoc analysis (a) or one-sample t-test versus hypothetical value of 100% (b–k). ^###P<0.001 versus ATP (l).

Figure 4. Mefenamic acid is protective in rodent models of Alzheimer's disease through an anti-inflammatory mechanism.

(a,b) Female Lister hooded rats (200–230 g) received an acute unilateral intracerebroventricular injection of soluble Aβ_1–42 on day 0 (5 nmol in 10 μl) which was followed by 14 days (starting one day before surgery) of i.p. injection of mefenamic acid (5 mg kg⁻¹) or vehicle. Animals were then tested in the novel object recognition task on 14 d (a) and 35 d (b) post surgery. Discrimination index data are presented as mean+s.e.m (n=5–10). NS, not significantly different, ^###P<0.001 compared with vehicle/vehicle treated animals and **P<0.01 compared to Aβ1-42/vehicle group. (c–f) 13–14 month old transgenic mouse model of Alzheimer's 3 × TgAD and wild-type (WT) control were treated with vehicle or 25 mg kg⁻¹ day⁻¹ mefenamic acid. (c) On day 18 memory was assessed with the novel object recognition task; discrimination index data are presented as mean+s.e.m (n=8–10). ^###P<0.001 compared to vehicle/WT animals and **P<0.01 compared with vehicle/3 × TgAD mice. (d–f) Evaluation of Iba1 and IL-1β expressing microglia within the subicula of 3xTgAD and WT mice following vehicle or mefenamic acid treatment. Microglial activation (d) and IL-1β expression (e) were evaluated and presented as mean+s.e.m (n=8–10). ^###P<0.001 compared with vehicle/WT animals and **P<0.01 compared to vehicle/3 × TgAD mice. (f) Representative images of microglial activation and IL-1β co-localization of 3 × TgAD and WT subicula following vehicle or mefenamic acid treatment. Scale bars are 15 μm. Statistical analyses performed using two-way ANOVA followed by Sidak corrected post hoc analysis.