Itaconate and fumarate derivatives exert a dual inhibitory effect on canonical NLRP3 activation in macrophages and microglia

The NLRP3 inflammasome is a multi-protein complex that regulates the protease caspase-1 and subsequent interleukin (IL)-1β release from cells of the innate immune system, or microglia in the brain, in response to infection or injury. Derivatives of the metabolites itaconate and fumarate, dimethyl itaconate (DMI), 4-octyl itaconate (4OI) and dimethyl fumarate (DMF), limit both expression of IL-1β, and IL-1β release following NLRP3 inflammasome activation. However, the direct effects of these metabolite derivatives on NLRP3 inflammasome responses in macrophages and microglia require further investigation. Using murine bone marrow-derived macrophages, mixed glia and organotypic hippocampal slice cultures (OHSCs), we demonstrate that DMI and 4OI pre-treatment limited IL-1β, IL-6 and tumor necrosis factor production in response to lipopolysaccharide (LPS) priming, as well as inhibiting subsequent NLRP3 inflammasome activation. DMI, 4OI, DMF and monomethyl fumarate (MMF), another fumarate derivative, also directly inhibited biochemical markers of NLRP3 activation in LPS-primed macrophages, mixed glia and OHSCs, including ASC speck formation, caspase-1 activation, gasdermin D cleavage and IL-1β release. Finally, DMF, an approved treatment for multiple sclerosis, as well as DMI, 4OI and MMF, inhibited NLRP3 activation in macrophages in response to the phospholipid lysophosphatidylcholine, which is used to induce demyelination, suggesting a possible mechanism of action for DMF in multiple sclerosis through NLRP3 inhibition. Together, these findings reveal the importance of immunometabolic regulation for both the priming and activation steps of NLRP3 activation in macrophages and microglia. Furthermore, we highlight itaconate and fumarate derivatives as a potential therapeutic option in NLRP3-driven diseases, including in the brain. Summary statement We show that itaconate and fumarate derivatives inhibit both the priming and activation steps of NLRP3 inflammasome responses in macrophages and microglia, revealing the importance of immunometabolic NLRP3 regulation.

or DMF (125 µM, 20 h). LPS (1 µg ml -1 , 4 h) was then added to the wells to induce priming (n=4). 146 (Ai) Cell lysates were probed by western blotting for NRF2, pro-IL-1β and NLRP3 protein, and (Aii-147 iv) densitometry was performed on each independent experiment (expressed relative to Veh+LPS 148 treatment). (B) WT BMDMs were treated as above, followed by nigericin (10 µM, 60 min; n=4). 149 Supernatants were assessed for (Bi) IL-6, (Bii) TNF and (Biii) IL-1β content by ELISA. (C) ASC-150 citrine BMDMs were treated as above, followed by nigericin (10 µM, 90 min; n=3). ASC speck 151 formation was measured over a period of 90 min. Image acquisition began immediately after addition suggested to be independent of NRF2 (Hooftman et al., 2020;Swain et al., 2020). MMF treatment 180 limited NLRP3 activation in LPS-primed BMDMs, although it was not as potent as DMF (Fig 2Di,  181 Dii). We confirmed that exogenous itaconate treatment also inhibited NLRP3 activation in LPS-primed 182 BMDMs, although much higher doses were required because it is less cell permeable (Supplementary 183 Figure 5) (Swain et al., 2020). To determine whether the inhibitory effects of itaconate and fumarate 184 derivatives were relevant in human macrophages, LPS-primed human MDMs were treated with DMI, 185 4OI and DMF prior to nigericin stimulation. 4OI and DMF alone appeared to induce IL-1β release, 186 although slight increases in cell death were observed for these treatments, which may have allowed 187 passive release of unprocessed pro-IL-1β (Figure 2Ei). Both DMI and DMF reduced nigericin-induced 188 IL-1β release, whereas 4OI did not significantly reduce IL-1β release at this dose ( Figure 2Ei). No 189 inhibition of cell death was observed ( Figure 2Eii)  observed; similarly, 4OI did not increase LPS-induced NRF2 levels (Figure 3Ai, Aii). IκBζ protein 228 levels were not measured. Despite this, both DMI and 4OI reduced pro-IL-1β production in mixed glial 229 cultures, without affecting NLRP3 protein levels (Figure 3Ai-iv). IL-1β release upon subsequent 230 stimulation with nigericin was also inhibited by both derivatives, and 4OI modestly reduced cell death 231 (Figure 3Av, Avi). Given that the inhibition of pro-IL-1β production and mature IL-1β release was 232 comparable between DMI and 4OI in mixed glia, only 4OI was used in OHSCs. NRF2 accumulation 233 could not be reliably detected in OHSCs upon LPS priming (data not shown). 4OI reduced the 234 production of pro-IL-1β in response to LPS priming, but did not affect NLRP3 production (Figure 3Bi-235 iii). 4OI pre-treatment strongly inhibited IL-1β release in response to LPS and nigericin treatment and 236 reduced cell death (Figure 3Biv, Bv). These data suggested that itaconate derivatives were able to limit 237 the priming of inflammasome responses, and that this may be a relevant mechanism to regulate 238 microglial inflammatory gene expression. 239 (1 µg ml -1 , 3 h) was then added to the wells to induce priming (n=3). (Ai) Cell lysates were probed by 243 western blotting for NRF2, pro-IL-1β and NLRP3 protein, and (Aii-iv) densitometry was performed 244 on each independent experiment (expressed relative to Veh + LPS treatment). (Av, Avi) WT mixed 245 glia were treated as above, followed by nigericin (10 µM, 60 min; n=3 Given that 4OI appeared to inhibit mature IL-1β release more strongly than it inhibited pro-IL-1β 256 production in OHSCs, we investigated whether itaconate and fumarate derivatives could directly limit 257 NLRP3 inflammasome activation in mixed glia and OHSCs. LPS-primed mixed glial cultures were 258 treated with DMI and 4OI prior to nigericin stimulation, and this reduced IL-1β release but did not 259 inhibit cell death (Figure 4Ai  interaction with cysteine 548 on murine NLRP3, preventing NEK7 binding (Hooftman et al., 2020). 345 This mechanism is plausible, given that DMI, 4OI and DMF are electrophiles that modify cysteine according to a recent pre-print (Kuo et al., 2020b). These studies suggest that itaconate production 363 could be an endogenous, protective response to limit ischaemic damage. Our previous report showed 364 increased levels of IL-1β and NLRP3 expression after ischaemic stroke, but NLRP3 deficiency or 365 inhibition did not improve stroke outcome (Lemarchand et al., 2019). This could suggest that 366 endogenous itaconate production within the brain in response to ischaemia, whilst too late to inhibit 367 inflammatory cytokine production, may be able to limit NLRP3 inflammasome activation. It is 368 important to note that detection of increased NRF2 levels in response to DMI and 4OI treatment was 369 not reliable in the microglial models employed in this study, either due to a lower amount of NRF2 370 signalling, questioning the relevance of NRF2 activation in the brain, or due to the instability of NRF2 371 protein during OHSC sample processing steps such as water-bath sonication. It should also be noted 372 that the extent of NLRP3 inhibition mediated by the itaconate and fumarate derivatives in the OHSCs 373 was lower than in the BMDM and mixed glial assays. It is possible that higher doses or longer treatment 374 times of the metabolite derivatives would result in greater NLRP3 inhibition, given that we have 375 previously shown that MCC950 is able to potently inhibit IL-1β release and ASC speck formation in 376 OHSCs (Hoyle et al., 2020). 377 We demonstrate that DMF, an approved clinical treatment for relapsing-remitting multiple sclerosis 378 it may exert similar inhibitory effects to DMF on the priming response. Importantly, here we confirm 398 that MMF can inhibit NLRP3 activation in response to both nigericin and LPC stimulation in 399 macrophages, suggesting that NLRP3 inhibition could indeed be a relevant in vivo mechanism for DMF 400 treatment. While DMF exhibited toxic effects during the pre-treatment experiments, as its dose was 401 matched to that of 4OI, it is likely that titration of the DMF concentration would result in reduced 402 toxicity but similar potency. Given that DMI and 4OI, which both also drive NRF2 accumulation, 403 exerted similar inhibitory effects on nigericin-and LPC-induced NLRP3 activation, it is possible that 404 these itaconate derivatives may also offer therapeutic potential in the treatment of multiple sclerosis. 405 Indeed, DMI was recently shown to be protective in a mouse model of multiple sclerosis (Kuo et al., 406 2020a). 407 We have revealed a two-pronged immunometabolic mechanism of NLRP3 regulation by limiting both 408 NLRP3 priming and canonical activation, suggesting that treatment with derivatives of metabolites 409 such as itaconate and fumarate may represent a viable therapeutic strategy in NLRP3-driven diseases, 410 although further work is required to confirm this. Future work should also aim to establish whether 411 itaconate and fumarate derivatives inhibit NLRP3 activation through the same or differing 412 mechanisms, either confirming this as a promising therapeutic target for drug design, or revealing novel 413 targets. centrifuged at 500 × g for 10 min and the pellet was resuspended in fresh culture medium before being 452 incubated in a flask at 37°C, 90% humidity and 5% CO2. After 5 days, the cells were washed, and fresh 453 medium was applied. The medium was subsequently replaced every 2 days. On day 12 of the culture, 454 the cells were seeded at 2 × 10 5 cells ml -1 in 24-well plates and incubated for a further 2 days prior to 455 use. 456

instructions. 518
Cell death assays 519 Cell death was assessed by measuring lactate dehydrogenase (LDH) release into the supernatant using 520 a CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) according to the manufacturer's 521 instructions. Cell death in OHSCs was assessed by adding propidium iodide (25 µg ml -1 ; Sigma, 522 P4864) to the culture medium for the final 30 min of the inflammasome activation protocol followed 523 by widefield microscopy. 524 Live imaging of ASC speck formation 525 ASC-citrine-expressing primary BMDMs were used to perform live imaging of ASC speck formation. 526 For itaconate derivative pre-treatment assays, cells were seeded at 1 × 10 6 cells ml -1 in 96-well plates 527 and incubated for 1 h, and were then treated with vehicle (DMSO), DMI, 4OI or DMF (125 µM, 20 h). 528 LPS (1 µg ml -1 , 4 h) was then added to the wells to induce priming. The medium was replaced with 529 optimem, and nigericin (10 µM) was added to activate the NLRP3 inflammasome. For assays where 530 itaconate derivative treatments were added after LPS priming, cells were seeded overnight at 1 × 10 6 531 cells ml -1 in 96-well plates. Cells were then first primed with LPS (1 µg ml -1 , 4 h). The medium was 532 replaced with optimum containing vehicle, DMI, 4OI or DMF (125 µM, 15 min) prior to addition of 533 nigericin (10 µM). Image acquisition began immediately after nigericin treatment. Images were 534 subsequently acquired every 10 min for a further 90 min using an IncuCyte ZOOM ® Live Cell Analysis 535 system (Essen Bioscience) at 37 °C using a 20X/0.61 S Plan Fluor objective. Speck number was 536 quantified using IncuCyte ZOOM ® software, and was assessed for each treatment at the final time 537 point of 90 min. 538

Snapshot widefield microscopy 549
Images were collected on a Zeiss Axioimager.M2 upright microscope using a 5X or 20X Plan 550 Apochromat objective and captured using a Coolsnap HQ2 camera (Photometrics) through 551 Micromanager software (v1.4.23). Specific band-pass filter sets for DAPI and FITC were using to 552 prevent bleed-through from one channel to the next. 553 Image processing analysis 554 Analysis was performed using FIJI (ImageJ) on images acquired from the same region of up to three 555 separate OHSCs (from the same insert) per treatment, and these values were averaged for each 556 biological repeat. ASC speck formation was quantified on 20X widefield microscopy images by 557 subtracting background (50 pixel rolling ball radius), manually setting thresholds and analysing 558 particles with the following parameters: size 1-10 μm 2 , circularity 0.9-1.0. To quantify PI uptake, 559 images were acquired on a widefield microscope using a 5X objective, background was subtracted (5.0 560 pixel rolling ball radius) and thresholds for images were automatically determined using the default 561 method. The total area of PI-positive signal was measured in the whole field of view, and was then 562 normalised to the total area of DAPI signal. 563

Data analysis 564
Data are presented as the mean ± standard error of the mean (SEM) together with individual data points 565 where possible. Data were analysed using repeated-measures one-way or two-way analysis of variance 566 (ANOVA), or mixed effects model, with Dunnett's or Sidak's post-hoc test using GraphPad Prism 567 (v8). Transformations or corrections were applied as necessary to obtain equal variance between groups 568 prior to analysis. Statistical significance was accepted at *P<0.05. 569

5
Acknowledgments 570 ASC-citrine mice were a kind gift from Douglas Golenbock (University of Massachusetts Medical 571 School) and Te-Chen Tzeng (Bristol Myers-Squibb, Cambridge, USA). We thank Dr Kevin Stacey 572 (University of Manchester) for the routine isolation of human PBMCs. 573

6
Competing interests 574 The authors declare that the research was conducted in the absence of any commercial or financial 575 relationships that could be construed as a potential conflict of interest. The data that support the findings of this study are available from the corresponding author upon 583 reasonable request. Brain. Behav. Immun. S0889-1591, 30209-9. 599