Pharmacologic Targeting of PDIA1 Inhibits NLRP3 Inflammasome Assembly and Activation

The NLRP3 inflammasome is a cytosolic protein complex that regulates innate immune signaling in response to diverse pathogenic insults through the proteolytic processing and secretion of pro-inflammatory cytokines such as IL-1β. Hyperactivation of NLRP3 inflammasome signaling is implicated in the onset and pathogenesis of numerous diseases, motivating the discovery of new strategies to suppress NLRP3 inflammasome activity. We sought to define the potential for the proteostasis regulator AA147 to inhibit the assembly and activation of the NLRP3 inflammasome. AA147 is a pro-drug that is metabolically converted to a reactive metabolite at the endoplasmic reticulum (ER) membrane to covalently modify ER-localized proteins such as protein disulfide isomerases (PDIs). We show that AA147 inhibits NLRP3 inflammasome activity in monocytes and monocyte-derived macrophages through a mechanism involving impaired assembly of the active inflammasome complex. This inhibition is mediated through AA147-dependent covalent modification of PDIA1. Genetic depletion or treatment with other highly selective PDIA1 inhibitors similarly blocks NLRP3 inflammasome assembly and activation. Our results identify PDIA1 as a potential therapeutic target to mitigate NLRP3 inflammasome-mediated pro-inflammatory signaling implicated in etiologically diverse diseases.


INTRODUCTION
Inflammasomes are multi-protein complexes involved in orchestrating local and systemic innate immune responses to diverse insults. [1][2][3] While multiple inflammasome complexes have been identified, activation of the NLR family pyrin domain containing 3 (NLRP3) inflammasome is perhaps the most clearly understood. 3,4 The NLRP3 inflammasome can be activated by a multitude of pathogen-or damage-associated signals which generally involve both a priming and an activating stimulus. 2,3,5 The priming signal induces the NF-kB dependent expression of NLRP3 inflammasome signaling components, while a secondary activation signal then initiates oligomerization of individual components into the functional NLRP3 inflammasome complex (Fig. S1A). 3,6,7 Upon assembly and activation, the NLRP3 inflammasome stimulates the maturation and release of pro-inflammatory cytokines such as the highly potent cytokine IL-1β. 1,2,8 Secreted cytokines then promote inflammatory responses that can neutralize threats, although this can also contribute to inflammation-mediated tissue damage. 9,10 Pathologic IL-1b signaling caused by NLRP3 inflammasome hyperactivation has been linked to diverse diseases including cryopyrin-associated periodic syndromes (CAPS) 11,12 , type II diabetes mellitus 9,13,14 , as well as cardiac and cerebral ischemia and reperfusion (I/R) injury. [8][9][10]15,16 Considering the central importance of the NLRP3 inflammasome in the pathogenesis of disease, therapeutic strategies to inhibit the activation and signaling of this complex has emerged as a highly attractive approach to mitigate inflammation-related pathologies. 7,[17][18][19] In fact, inhibiting the NLRP3 inflammasome and downstream signaling has been shown to reduce tissue damage and improve outcomes in a variety of animal disease models, as well as in human patients. 10,13,16,19,20 Monoclonal antibodies that suppress IL-1β signaling downstream of inflammasome activation are clinically approved to treat several inflammatory diseases and have shown benefits in numerous animal models of inflammatory disorders. 10,13,20,21 Further, numerous pharmacologic strategies have been developed to inhibit NLRP3 inflammasome assembly and activation. 18 Compounds like MCC950 have been shown to bind the ATP binding pocket of NLRP3 to inhibit inflammasome assembly and mitigate pathologic inflammasome signaling in numerous mouse models of inflammatory disease. 15,[22][23][24][25] Other compounds that bind the NLRP3 ATP binding site, such as OLT1177, have also been tested in clinical trials for inflammatory diseases such as gout and osteoarthritis. 26,27 Despite the promise of these approaches, no direct NLRP3 inflammasome inhibitor is currently approved as a treatment, necessitating the continued development of new pharmacologic strategies to block NLRP3 inflammasome signaling. Intriguingly, numerous electrophilic compounds have also been found to covalently modify cysteine residues within the NLRP3 inflammasome and inhibit complex assembly. The anti-inflammatory compound oridonin inhibits NLPR3 inflammasome activity through a mechanism predicted to involve selective modification of NLRP3 Cys279 . 28 Similarly, compound RRx-001, which was originally identified as an anticancer compound, is predicted to inhibit inflammasome assembly through covalent modification of NLRP3 Cys409 . 29 Further, itaconate, an immunomodulatory metabolite, is predicted to block inflammasome assembly and activity in mice through modification of NLRP3 Cys548 . 30 More recent screens have identified additional electrophilic compounds that target multiple cysteines residues on NLRP3 to inhibit inflammasome assembly and activity, indicating that NLRP3 is a major electrophilic sensor within the cell. 31 We previously identified compound AA147 as a highly-selective regulator of endoplasmic reticulum (ER) proteostasis. 32 We showed that the 2-amino-p-cresol moiety of AA147 is metabolically activated at the ER membrane to a reactive electrophile that selectively modifies proteins localized near the ER membrane, such as protein disulfide isomerases (PDIs). 33 AA147-dependent PDI modification can both directly influence ER proteostasis for destabilized proteins such as immunoglobulin light chains 34 and induce adaptive ER proteostasis remodeling through activation of the ATF6 signaling arm of the unfolded protein response. 32 Previous results suggest that the NLRP3 inflammasome assembles at the ER membrane. 35 Considering the electrophilic sensing activity of NLRP3 31 , we predicted that local, metabolic activation of AA147 to a reactive electrophile could potentially inhibit NLRP3 inflammasome assembly and activity through direct targeting of NLRP3 cysteine residues. 31,36 Supporting this prediction, we previously showed that AA147 covalently modifies the electrophilic sensor KEAP1 to activate NRF2 signaling in neuronal cell models through a similar mechanism. 36 Here, we found that AA147 is a potent and effective inhibitor of NLRP3 inflammasome assembly and activation. However, this inhibition is independent of both AA147-dependent ATF6 activation. Instead, AA147 blocks inflammasome assembly and activation through a mechanism involving covalent modification of PDIA1a highly abundant ER PDI. Genetic depletion of PDIA1 mimics AA147-depenent inflammasome inhibition. We also demonstrate that highly selective PDIA1 inhibitors show potent activity against NLRP3 inflammasome activation. These results identify PDIA1 as a novel target for inhibiting inflammasome signaling that could be potentially exploited for the continued development of next generation inflammasome inhibitors.

AA147 reduces NLRP3 assembly and activity
We initially sought to define the potential for the metabolically activated pro-drug AA147 to inhibit NLRP3 inflammasome assembly. We employed a recently described, high content imaging-based screening assay of inflammasome assembly that utilizes the human monocyte cell line THP1 expressing the inflammasome subunit ASC fluorescently tagged with GFP (ASC-GFP) under control of an NFkB-regulated promoter. 31 Inflammasome assembly can be monitored in these cells by following the formation of ASC-GFP puncta, or 'specks', that accumulate in response to treatment with both a priming stimulus (e.g., LPS) and an activating stimulus (e.g., nigericin, ATP). Treatment with both LPS and nigericin robustly increases ASC-GFP speck formation in undifferentiated THP1 cells (Fig. 1A,B). Pre-treatment with the known NLRP3 inflammasome inhibitor MCC950 23 inhibited ASC-GFP speck formation. Interestingly, pre-treatment with AA147 similarly blocked ASC-GFP speck formation in undifferentiated THP1 cells, indicating that this compound inhibits inflammasome assembly.
Inflammasome activation leads to the increased secretion of the pro-inflammatory cytokine IL-1b. 7 Thus, we sought to determine the extent to which AA147 blocks this activity downstream of inflammasome assembly.
Initially, we monitored secretion of IL-1b from monocyte THP1 cells using HEK-Blue IL-1b reporter cells (Invivogen) -a cell line that expresses the IL-1b receptor and an NFkB-responsive secreted alkaline phosphatase (SEAP) reporter. Treatment of HEK-Blue IL-1b cells with conditioned media prepared on THP1 cells stimulated with LPS and nigericin increased secretion of SEAP, reflecting increased IL-1b secretion (Fig.   1C). Treatment of THP-1 cells with MCC950 blocked this effect, confirming that the increase in SEAP corresponds to NLRP3 inflammasome-dependent increases in IL-1b secretion from THP1 cells. AA147 similarly blocked IL-1b secretion from undifferentiated THP1 cells stimulated with LPS and nigericin (Fig. 1C). Similar results were observed in THP1 monocyte-derived macrophages stimulated with LPS and ATP, as measured by both the HEK-Blue IL-1b reporter assay (Fig. S1B) and an enzyme-linked immunoassay (ELISA) of conditioned media (Fig. 1D). Further, AA147 reduced IL-1b in conditioned media prepared on RAW264.7 mouse macrophages stimulated with LPS and ATP, as measured by immunoblotting (Fig. S1C). These results indicate that AA147 reduces NLRP3 inflammasome assembly and activity.

AA147-dependent inhibition of inflammasome activation is independent of ATF6 activation or NFkB inhibition.
AA147 was originally identified as an activator of the ATF6 signaling arm of the unfolded protein response. 32 Thus, we sought to determine the potential for AA147 to inhibit inflammasome activity through an ATF6dependent mechanism. Surprisingly, we did not observe significant increases of the ATF6-target BIP in immunoblots prepared from THP1 cells treated for 16 h with AA147, indicating that this compound did not significantly increase ATF6 signaling in these cells (Fig. S2A). Further, co-treatment with the ATF6 inhibitors Ceapin-A7 37 or the site 1 protease inhibitor PF-429242 38 did not decrease AA147-dependent reductions of IL-1b secretion in monocyte THP1 cells stimulated with LPS and ATP ( Fig. 2A,B). These results indicate that AA147 does not reduce NLRP3 inflammasome-dependent IL-1b secretion through an ATF6-dependent mechanism.
AA147 could also influence inflammasome activation through suppression of NFkB-dependent expression of inflammasome components and pro-IL-1b. AA147 modestly reduced expression of IL1B in THP1 monocyte derived macrophages stimulated with LPS ( Fig. 2C), although this reduction was not detectable in naïve THP1 cells which showed a less dynamic response to LPS (Fig. S2B). Further, AA147 did not impact the expression of other NLRP3 inflammasome subunits in differentiated THP1 cells (e.g., ASC, NLRP3, CASP1) (Fig.   S2C). We also did not observe changes in the protein levels for NLRP3 subunits or pro-IL-1b in differentiated THP1 cells stimulated with LPS and treated with AA147 (Fig. 2D). This suggests that AA147 does not significantly impact expression of inflammasome subunits or pro-IL-1b in these cells.

Genetic depletion of PDIA1 inhibits inflammasome assembly and activation.
AA147 activates ATF6 through a mechanism involving metabolic activation and direct covalent modification of a subset of ER protein disulfide isomerases (PDIs). 33 Previous results showed that AA147-dependent inhibition of PDIs could promote adaptive ER remodeling independent of ATF6 to influence secretion of destabilized, amyloidogenic immunoglobulin light chains. 34 Thus, we sought to determine if NLRP3 inflammasome activation and assembly could also be explained by AA147-dependent PDI inhibition. We used an alkyne-containing analog of AA147 (AA147 alk ) 33 to confirm that AA147 covalently modified PDIs in monocyte THP1 cells (Fig. S3A). We then shRNA-depleted the predominant PDIs modified by AA147, including PDIA1, PDIA3, PDIA4, PDIA5, and PDIA6 33,34 , in THP1 cells stimulated with LPS and ATP or nigericin and then monitored IL-1b secretion using the HEK-Blue IL-1b reporter assay. We confirmed efficient knockdown of individual PDIs by qPCR (Fig. S3B). While depletion of all tested PDIs show modest inhibition of IL-1b secretion from THP1 cells stimulated with LPS and either nigericin or ATP, the most significant effects were observed upon knockdown of PDIA1 and PDIA5 (Fig.   3A). We specifically focused on PDIA1, as this PDI is highly modified by AA147 across multiple cell types, whereas PDIA5 has not been identified as a robust target across cell types. 33,34,36 Additionally, CRISPR-mediated ablation of PDIA1 diminishes secretion of pro-inflammatory cytokines in other cell types, implicating PDIA1 as a potential regulator of inflammatory signaling. 39 We confirmed that PDIA1 protein was reduced in THP1 cells stably depleted of PDIA1 (Fig. S3C).
Importantly, we also showed that PDIA1-depletion did not increase expression of the UPR target gene BiP, indicating that reduced activity of PDIA1 does not globally induce ER stress in these cells. As with AA147 pretreatment, PDIA1 depletion reduced ASC-GFP speck formation in THP1 cells stimulated with LPS and nigericin (Fig. 3B). Further, we found that PDIA1 depletion inhibited autoproteolysis of caspase 1 (Fig. 3C,D) another marker of inflammasome assembly and activation. 7,40 PDIA1 depletion showed a modest reduction in IL1B expression in LPS stimulated THP1 monocytes, although no effect was observed for the NLRP3 inflammasome subunits NLRP3 or ASC (Fig. S3C). However, we did not observe significant reductions in pro-IL-1b in THP1 cells depleted of PDIA1 (Fig. 3E). These results indicate that PDIA1 depletion inhibits NLRP3 inflammasome assembly and activation through a mechanism analogous to that observed with AA147.

Selective, pharmacologic PDIA1 inhibitors block inflammasome assembly and activation.
The above results suggest that AA147 inhibits NLRP3 assembly and activation through covalent targeting of PDIs, most notably PDIA1. This suggests that other PDIA1 inhibitors should similarly inhibit NLRP3 inflammasome assembly and activation. To test this, we monitored ASC-GFP speck formation in monocyte THP1 cells treated with the highly selective, covalent PDIA1 inhibitor KSC-34 41 and stimulated with LPS and nigericin.
This indicates that these inhibit inflammasome activity through a PDIA1-dependent mechanism.

DISCUSSION
Pharmacologic interventions which suppress the NLRP3 inflammasome activity show significant promise in mitigating the harmful effects of persistent or excessive inflammation. 7,16,17 Here, we show that the metabolically activated electrophile AA147 can suppress pro-inflammatory IL-1b signaling by inhibiting assembly of the NLRP3 inflammasome. Although AA147 was initially characterized as a preferential activator of the ATF6 arm of the UPR 32 , the inhibition of IL-1b secretion was not dependent on ATF6 signaling. Instead, we show that AA147 impedes inflammasome activity through the covalent modification of PDIs, specifically PDIA1. Genetic ablation of PDIs, most notably PDIA1, showed similar results to AA147. Importantly, knockdown of PDIA1 inhibited inflammasome assembly, caspase 1 proteolytic autoactivation, and IL-1b secretion -highlighting the extent to which PDIA1 is involved in NLRP3 inflammasome signaling. Furthermore, selective and structurally distinct inhibitors of PDIA1 similarly reduced assembly of the inflammasome complex as well as IL-1b secretion.
Overall, these results indicate that indirectly suppressing inflammasome activation by targeting proteins which regulate NLRP3 inflammasome oligomerization, such as PDIA1, may represent a novel strategy for reducing the pathologic impacts of inflammasome overactivation.

PDIA1 regulates pathologic inflammatory signaling through multiple mechanisms. Previous studies have
shown that extracellular PDI is an important regulator of innate immune activity as it can modulate the extent of neutrophil adhesion by impacting the surface expression of molecules including L-selectin 42 and integrins [43][44][45][46] .
Blocking antibodies have shown that extracellular PDIs may also be involved in the release of tissue factor and IL-1b from macrophages, although a mechanism for the latter has not been established. 47 PDIA1 has also been implicated in regulating the secretion of other cytokines, including IL-6 and TNFa. 39 Here, we used structurally distinct pharmacologic inhibitors of PDIA1 as well as shRNA-mediated silencing to show that PDIA1 can also regulate the processing and release of the potent pro-inflammatory cytokine IL-1b by preventing NLRP3 inflammasome assembly. These studies suggest that PDIA1 activity may be a central regulator of inflammatory signaling across myeloid cell lineages through a diverse range of intracellular and extracellular signaling mechanisms. Therefore, further exploration of the potential of PDI inhibition to mitigate damaging inflammation may present new opportunities to improve patient outcomes across a wide range of disorders.
Alternating passages of THP1 cells were treated with 200 μg/ml Hygromycin Gold (Invivogen cat. ant-hg-1) as recommended by Invitrogen. ASC-GFP THP1 cells were passaged with 100 μg/ml Zeocin (Fisher Scientific cat. AAJ67140XF). Selection media was fully removed prior to inflammasome stimulation. THP1 cells were differentiated into macrophages by treating 1M cells /mL with 0.5 µM phorbol 12myristate 13-acetate (PMA; Sigma; cat P8139) for 3 h on Day 0. After the initial media change following PMA, media was refreshed daily for two additional days. On Day 2, compounds were added to media for pre-treatment.
The following day, THP1 derived macrophages were stimulated with inflammasome priming and activation stimuli; compound was readded to media with each media change after initial pre-treatment.

Compounds and Treatments
Ultra-pure TLR4 specific lipopolysaccharide derived from E. coli strain 0111:B4 (LPS; Invivogen cat. tlrl-3pelps) at a dose of 1µg/mL was used to induce NF-kB transcription activity for inflammasome priming across all experiments. Inflammasome assembly was induced by administering cells with 5mM ATP (Jena Biosciences cat. NU-1010) or nigericin (Invivogen cat. tlrl-nig) with dose, as indicated.
AA147 and AA147 alk were previously reported 33  Inflammasome assembly assay ASC. GFP THP1 cells were plated at a density of 20,000 cells / well in a 384 well plate and treated with compounds for 16 h. Cells were then primed with LPS (1µg/mL) for either 3 h or 16 h as noted. Subsequently, cells were stimulated with nigericin (10 µM) for 3 hours. Nuclei were stained using 5 µg/mL Hoechst 33342 (Thermo cat. H3570). Images were captured using Cellomics Cell Insight imaging reader (Thermo). An air ×10 lens was used to capture one image per well. For image analysis, stained nuclei were enumerated as representing the total number of live cells captured. ASC-GFP speck formation was calculated using an algorithm within the Cell Insight software, which factors both signal intensity and area size. Numerical results from the analyzed images were later exported for analysis using Prism 9 (GraphPad, San Diego, CA)

IL-1b assays
Undifferentiated THP1 cells were plated in fresh media at a density of 750K/mL and compounds were concurrently administered, as noted. The following day, cells were stimulated using LPS (1µg/mL) for 3 h and then ATP (5 mM) or nigericin (1 µM) were added directly to the cells. Cells were incubated at 37o for 24 h following the addition of the secondary stimulus. Cells and conditioned media were centrifuged at a low speed to remove suspended monocytes. This media was changed and ATP (5 mM) (Jena Biosciences cat. NU-1010) was added for 24 h prior to collection of conditioned media. IL-1β levels were detected in conditioned media using SDS-PAGE followed by immunoblotting.

ASC-GFP Flow Cytometry
ASC-GFP THP1 cells were plated at a density of 750K cells/mL were plated in 100 L in a 96 well plate.
Cells were pre-treated for 16 h with AA147. The following day, cells were stimulated with LPS (1µg/mL) for 3 h and immediately analyzed via flow cytometry on a NovoCyte 3000 (Acea

Bioluminescent caspase 1 cleavage assay
The caspase 1 glo (Promega; cat. G9951) was followed according to manufacturer's instructions. Briefly, undifferentiated THP1 cells were plated at a density of 750K cells/mL and treated for 16 h with LPS (1µg/mL) in a flask. The following day, cells were removed from LPS media, resuspended in fresh media at a density of 1M cells /mL, and 25 µL of the suspension was plated in a 384 well dish. 20 µM Nigericin was added to individual wells for 45 minutes. Subsequently, caspase 1 reagent as well as the proteasome inhibitor MG132, and ac-YVAD were added to respective treatments and or controls. Reactions were incubated in the dark at room temperature for a minimum of 1 prior to luminescence being read on a Tecan F200 Pro microplate reader.

Rhodamine labeling
Rhodamine-azide labeling reactions were performed in a 50 µL reaction containing 1.

SDS-Page and Immunoblotting
Briefly, cells were lysed in lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 20 mM Hepes pH 7.4, 1 mM EDTA, 1% Triton X-100, and protease inhibitor cocktail (Pierce cat. A32955)). The total protein concentration in cellular lysates was normalized using the Bio-Rad protein assay. Lysates were then denatured with 1 × Laemmli buffer + 100 mM dithiothreitol (DTT) and boiled before being separated by SDS-PAGE. Samples were transferred onto nitrocellulose membranes (Bio-Rad cat. 1620112). Membranes were then incubated overnight at 4 °C with primary antibodies with dilutions, as noted. Membranes were washed in TBS-T, incubated with the speciesappropriate IR-Dye conjugated secondary antibodies diluted at 1:10,000, and analyzed using the Odyssey Infrared Imaging System (LI-COR Biosciences). Quantification was carried out with LI-COR Image Studio software.

ACKNOWLEDGEMENTS.
We thank John Griffin and Laura Healy at Scripps Research for experimental advice related to the work described in this manuscript. This work was funded by the National Institutes of Health (DK107604 to RLW).