Vacuole Membrane Protein 1 (VMP1) Restricts NLRP3 Inflammasome Activation by Modulating SERCA Activity and Autophagy

Altered expression of vacuole membrane protein 1 (VMP1) has recently been observed in the context of multiple sclerosis and Parkinson’s disease (PD). However, how changes in VMP1 expression may impact pathogenesis has not been explored. Here, we report that genetic deletion of VMP1 from a monocytic cell line resulted in increased NLRP3 inflammasome activation and release of proinflammatory molecules. Examination of the VMP1 dependent changes in these cells revealed that VMP1 deficiency led to decreased SERCA activity and increased intracellular [Ca2+]. We also observed calcium overload in mitochondria in VMP1 depleted cells, which was associated with mitochondrial dysfunction and release of mitochondrial DNA into the cytoplasm and the extracellular environment. Autophagic defects were also observed in VMP1 depleted macrophages. Collectively, these studies reveal VMP1 as a negative regulator of inflammatory responses, and we postulate that decreased expression of VMP1 can aggravate the inflammatory sequelae associated with neurodegenerative diseases like PD.


Introduction
Vacuole membrane protein 1 (VMP1) was rst characterized in models of acute pancreatitis [1]. This initial study observed that VMP1 regulates autophagosome formation and autophagic ux, suggesting that it promotes cellular homeostasis [1]. VMP1 also facilitates lipid droplet formation, and most recently has been found to be a crucial host factor for SARS-CoV-2 and pan-coronavirus infection [1][2][3][4].
Interestingly, VMP1 modulates contacts between the endoplasmic reticulum (ER) and mitochondria, lipid droplets, and endosomes through its interaction with the sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA), such that the interaction between VMP1 and SERCA modulates cytosolic Ca 2+ levels close to these membrane contacts [5,6]. However, despite being rst identi ed within the context of an in ammatory disease, the role of VMP1 in regulating cellular in ammatory responses has not previously been investigated.
In ammatory responses are necessary and bene cial for controlling infections but need to be tightly regulated to prevent unnecessary damage to the host. Dysregulated or excessive in ammatory responses can contribute to several diseases including neurodegenerative diseases like Alzheimer's disease (AD) and Parkinson's disease (PD) and autoimmune disorders such as multiple sclerosis (MS) [7][8][9].
In ammatory responses are often mediated by the in ammasome, a multiprotein complex that is assembled and activated by numerous sterile and pathogen derived stimuli [10]. Notably, in ammasome activation can be regulated by changes in cytoplasmic ion concentrations and alterations in autophagy [11,12].
In ammasome activation consists of a ligand-sensing protein such as nucleotide-binding domain, leucine-rich repeats containing family, pyrin domain-containing-3 (NLRP3), an adaptor protein such as apoptosis-associated speck-like protein containing a CARD (ASC), and nally an effector caspase, most commonly caspase-1 [13,14]. Assembly of the in ammasome results in the activation of caspase-1, a protease which then cleaves proin ammatory cytokines such as proIL-1β and gasdermin d (GSDMD) which induces the formation of pores in the plasma membrane through which IL-1b and other in ammatory effectors are released [15][16][17]. Canonical in ammasome activating signals include reactive oxygen species (ROS) overproduction, K + e ux, mitochondrial dysfunction, ER stress, and lysosomal rupture [18][19][20][21]. Although Ca 2+ ux is not thought to be directly sensed and induce in ammasome activation, it is known that changes in cytoplasmic [Ca 2+ ] can induce K + e ux and mitochondrial dysfunction, which can both directly activate the in ammasome. Given the putative connection between VMP1 function and in ammasome activation, and the recent connection between VMP1 and in ammatory diseases, we hypothesized that dysregulation in VMP1 KO cells may exacerbate cellular in ammatory responses, particularly related to in ammasome activation [22].
Here we show that VMP1 restricts NLRP3 in ammasome activation and the release of proin ammatory cytokines in THP-1 cells, a commonly used model of macrophages and microglia. We show that VMP1 knockout (KO) leads to increased release of IL-1β and other in ammatory mediators following lipopolysaccharide (LPS) and alpha-synuclein (α-syn) bril treatment. In VMP1 KO cells, we observe elevated levels of cytosolic Ca 2+ in response to the activation of Ca 2+ dependent signaling pathways and a corresponding reduction of Ca 2+ stores in the ER. We also observed that mitochondria in VMP1 KO cells exhibit evidence of Ca 2+ overload and mitochondrial dysfunction, accompanied bythe exposure of mitochondrial DNA (mtDNA) in the cytoplasm and the extracellular environment. We also observe disruption of autophagic ux in VMP1 KO cells, which may limit the ability of cells to degrade damaged mitochondria and thereby exacerbate in ammatory responses. Collectively, these results reveal a previously unreported role for VMP1 in regulating cellular in ammatory responses and provides a mechanism by which changes in VMP1 expression during disease may in uence disease pathogenesis.

Results
VMP1 depletion exacerbates the release of proin ammatory molecules.
To determine whether VMP1 affects the secretion of proin ammatory cytokines and other molecules, we utilized CRISPR-Cas9 to knockout VMP1 in THP-1s, a human monocytic cell line that can be terminally differentiated into macrophage-like cells [4,23]. The knockout was veri ed by western blot (Fig. 1A). To activate in ammatory pathways, differentiated THP-1s were rst primed with LPS then ATP to activate the NLRP3 in ammasome. Treatment of VMP1 KO THP-1s with LPS and ATP resulted in increased IL-1β secretion compared to treated control cells, suggesting an increase in in ammasome activation in VMP1 KO cells (Fig. 1B). VMP1 mRNA expression levels did not change following LPS/ATP treatment (Fig. S1A), although pro-IL-1b mRNA levels were increased in VMP1 KO cells following LPS and ATP treatment ( Fig   S1B). We and others have also observed that a-syn brils can induce NLRP3 in ammasome activation via their ability to induce vesicular damage following endocytosis [24][25][26]. We therefore wanted to determine whether primed VMP1 KO macrophages treated with α-syn brils release more IL-1β than control cells. For these treatment conditions, supernatant was collected after 24 hrs, and IL-1β release was measured. After 24 hours in culture, there was elevated basal IL-1β release from VMP1 KO cells compared to control that was not apparent after a 30 min culture, suggesting that VMP1 KO cells exhibited basal in ammasome activation without the addition of in ammatory stimuli (Fig. 1C). Primed VMP1 KO cells treated with α-syn release more IL-1β than control cells (Fig. 1C). We also measured the release of galectin-3 (Gal3), another in ammatory protein that depends on GSDMD pores for its secretion [27]. VMP1 KO cells also had increased Gal3 release from cells treated with LPS and ATP (Fig. S1C) while we observed no signi cant change in the expression of gal-3 mRNA in VMP1 KO cells (Fig. S1D). Overall, these data suggest that both at a basal level as well as following activation by in ammatory stimuli, VMP1 KO cells release more proin ammatory molecules. VMP1 KO cells have increased NLRP3 in ammasome and caspase-1 activation.
To determine whether VMP1 KO cells have increased in ammasome activation, we employed a uorochrome-labeled inhibitors of caspases (FLICA) assay to measure active caspase-1. Treatment of VMP1 KO cells with LPS and ATP resulted in an increase in FLICA signal compared to control cells in both cell types (Fig. 1D). However, higher levels of FLICA staining were observed in VMP1 KO cells (Fig. 1D), revealing increased caspase-1 activation in these cells. Similar results were observed when caspase-1 activation was measured using a luciferase-based caspase-1 biosensor [28] (Fig. 1E). Collectively, these data suggest that VMP1 depletion increases caspase-1 activation following treatment with LPS and ATP.
We also observed an increase in LDH release from VMP1 KO cells, consistent with increased pyroptotic cell death, following LPS and ATP treatment, while we observed no increase in LDH release from control cells following addition of LPS and ATP (Fig. 1F). Collectively, these data suggested that in VMP1 KO sensitizes cells to in ammasome activating stimuli.
VMP1 depletion leads to changes in genes associated with signaling and degradative pathways.
To understand the changes in VMP1 KO cells that may prime them to respond to in ammasome activating stimuli, we used RNA-Seq to de ne changes in gene expression that occur in VMP1 KO cells. VMP1 KO cells exhibited altered gene expression of genes associated with Toll-like receptor (TLR) signaling and Nod-like receptor (NLR) signaling pathway, observing upregulation of both pathways ( Fig. 2A, 2B), consistent with our observation that VMP1 KO increased in ammasome activation and IL-1b release in these cells. Most TLRs were upregulated in untreated VMP1 KO cells compared to control cells suggesting that VMP1 KO cells were broadly primed to recognize most pathogen associated molecular patterns (PAMPs) ( Fig. 2A). Furthermore, the P2X7R was upregulated in VMP1 KO cells. ATP binds to and activates the P2X7R to promote K + e ux and downstream NLRP3 in ammasome activation, suggesting that VMP1 KO cells are also primed to respond to damage associated molecular patterns (DAMPs) (Fig. 2B). Several sensors that ultimately form the in ammasome complex as well as targets of caspase-1 cleavage were also upregulated in VMP1 KO cells (Fig. 2B). Upregulation of several TLR and NLR pathway genes highlight that VMP1 KO cells are primed to respond to various proin ammatory stimuli. We also observed alterations in the related Cytosolic DNA-sensing pathway when comparing LPS and ATP treated VMP1 KO cells to treated control cells (Fig. 2C). Interestingly, upregulation of gene expression in this pathway suggest that VMP1 KO cells may be more sensitive to exposed cytoplasmic mtDNA which may be increased in VMP1 KO cells, consistent with previous reports demonstrating that VMP1 mediates the selective degradation of damaged mitochondria by mitophagy [29]. We also observed changes in gene expression in genes associated with calcium signaling, including a downregulation of SERCA and upregulation of other genes associated with calcium signaling in VMP1 KO cells (Fig. 2D), consistent with the previously reported association of VMP1 and SERCA [30]. VMP1 KO cells also exhibited changes in genes associated with the autophagic/lysosomal pathway (ALP) (Fig S2A,  S2B), also consistent with prior reports of VMP1 modulating activity of this pathway. Collectively, these changes to in ammatory signaling pathways and other pathways, such as calcium and the ALP suggest plausible mechanisms by which the cellular response to in ammatory stimuli may be enhanced. LPS and ATP treatment leads to increased mitochondrial depolarization in VMP1 KO cells.
To characterize mitochondrial function in VMP1 KO cells, we developed an imaging approach where live differentiated THP-1s were incubated with MitoTracker Red CMXRos (MTRed) and MitoTracker Green FM (MTGreen). Others have used ow cytometry based methods to perform an assessment of mitochondrial function in cells [35]. To develop a conceptually similar assay that measures mitochondrial function at the level of individual mitochondria, we used uorescence microscopy and post-acquisition analysis to build surface masks around the MTGreen channel to allow us to measure the membrane potential of populations of individual mitochondria in these cells, as well as quantify the number of mitochondria per cell and the average volume of mitochondria ( Fig. 5A, S3). To assess the relative amounts of functional or dysfunctional mitochondria, the intensity max of the MTRed was plotted on the y-axis and the intensity max of the MTGreen was plotted on the x-axis. The vast majority of mitochondria in untreated control cells were functional (high MTRed signal), and with treatment, the number of MTRed positive mitochondria decreased and an increase in MTRed negative mitochondria was observed (Fig. 5B, 5D). The relative intensity of MTRed + mitochondria also decreased following LPS and ATP treatment (Fig. 5B,5D). Untreated VMP1 KO cells had similar mitochondria populations to control cells (Fig. 5C). However, LPS and ATP treatment induced a more pronounced decrease in the number of functional mitochondria and those mitochondria that remained MTRed + by our gating criteria had an even lower uorescence intensity than mitochondria from control cells following LPS and ATP treatment (Fig. 5E).
Data from three experiments revealed a signi cant decrease in functional mitochondria (Fig. 5F) and a corresponding increase in the number of dysfunctional mitochondria (Fig. 5G) in VMP1 KO cells following LPS and ATP treatment. A reduction in membrane potential was also observed in VMP1 KO cells when mitochondria were analyzed independently of MTGreen signal (Fig. 5H). Analysis of the total number of mitochondria per cell revealed that LPS and ATP treatment increased the number of mitochondria per cell, as did VMP1 KO, such that treatment of control THP-1s with LPS and ATP resulted in an increase in the number of mitochondria that was comparable to VMP1 KO cells. The average number of mitochondria per cell was highest for VMP1 KO cells treated with LPS and ATP (Fig. 5I). We observed a corresponding change in mitochondrial volume where control THP-1s have the largest and most variability in mitochondrial volume, with reduced total mitochondrial volume observed in the VMP1 KO cells (Fig. 5J). In this case, control and VMP1 KO cells had a similar total volume of mitochondria following LPS and ATP treatment (Fig. 5J). Previous work suggested that altered mitochondrial fusion can result in mitochondrial fragmentation and increased apoptotic cell death [36]. Our data support that there may be a defect in mitochondrial fusion in treated VMP1 KO cells given that there was an increase in the number of mitochondria, a reduction in average mitochondrial volume as well as increased cell death in VMP1 KO cells following LPS and ATP treatment ( Fig. 5I-J, 1F).
VMP1 KO cells have increased levels of exposed cytoplasmic mtDNA and extracellular mtDNA.
When mitochondria are dysfunctional, mitochondrial outer membrane permeabilization (MOMP) occurs, allowing for the extrusion of the mitochondrial inner membrane into the cytosol and release of mtDNA [37]. To determine whether there was increased exposure of cytoplasmic mtDNA in VMP1 KO cells, differentiated THP-1s were treated then incubated with MTRed and stained with anti-DNA antibodies under mild detergent conditions to prevent the staining of mtDNA in intact mitochondria but to detect cytoplasmically exposed mtDNA (Fig. 6A, S4A,B). Representative images showed that treatment of VMP1 KO cells with LPS and ATP led to an increase in cytoplasmic DNA puncta, but there also appeared to be an increase in DNA pucta attached to the glass coverslip outside of the boundary of the cell (Fig. 6A). Imaris software was utilized to quantify the number of DNA puncta inside and outside of the cell (Fig.   S5). The data demonstrated that there were relatively few DNA puncta inside of the cell in untreated control and VMP1 KO cells and control cells treated with LPS and ATP. This contrasts with the larger increase in the number of intracellular DNA puncta per eld of view for LPS and ATP treated VMP1 KO cells (Fig. 6B). Interestingly, it was observed that there was also an increase in DNA puncta on the coverslips outside of the cell in VMP1 KO cells treated with LPS and ATP (Fig. 6C). To validate this result, we isolated DNA from the culture supernatant and quanti ed cell-free mtDNA via qPCR using primers speci c to genes expressed in the small, circular mitochondrial genome. Using this assay, we detected mtDNA in the supernatant of LPS and ATP treated control cells, and a further increase in mtDNA in the supernatant from LPS and ATP treated VMP1 KO cells (Fig. 6D). There was no difference in the amount of mitochondrial DNA across conditions in the cell lysates even though there were more mitochondria in VMP1 KO cells (Fig. 6E, 5I). Overall, these data suggested that LPS and ATP treatment in VMP1 KO cells caused the exposure of cytoplasmic mtDNA and the release of extracellular mtDNA which is likely to contribute to the increased in ammatory responses observed in these cells.

VMP1 KO cells have disrupted autophagic ux.
Given that autophagy modulates in ammatory responses and that previous reports have suggested that VMP1 KO can impair autophagic ux, we wanted to assess autophagic processes in VMP1 KO macrophages under basal and in ammatory conditions [11,38,39]. Double staining or colocalization of microtubule-associated protein 1A/1B-light chain 3 (LC3) and lysosome-associated membrane protein 1 (LAMP1) suggested fusion of the autophagosome and lysosome which was primarily observed in control cells (Fig. 7A). The number of LAMP1 + puncta were about the same between control and VMP1 KO with a slight increase in control cells treated with LPS and ATP (Fig. 7B). There was a modest but signi cant increase in the number of LC3 + puncta in untreated VMP1 KO cells compared to untreated control cells (Fig. 7C). Interestingly, the number of LC3 + puncta with LPS and ATP treatment for both control and VMP1 KO cells decreased (Fig. 7C). Untreated control cells exhibited the highest level of colocalization between LC3 and LAMP1, suggesting fusion of the autophagosome and lysosome which was disrupted in treated control cells and untreated and treated VMP1 KO cells (Fig. 7D) which was consistent with previous reports [5,38]. Notably, while image analysis revealed what appeared to be modest changes in autophagic function in VMP1 KO cells, immunoblotting results demonstrated a marked increase in LC3II levels in VMP1 KO cells and an accumulation of p62 compared to control cells suggesting that autophagic ux was disrupted (Fig. 7E). Notably, there was an absence of detectable LC3I in VMP1 KO cells which is supported by previous reports that LC3II levels can increase even under conditions when autophagy is blocked (Fig. 7E) [40]. Taken together, these data suggested that autophagy was disrupted in VMP1 KO which is consistent with an accumulation of damaged mitochondria and extracellular mtDNA which is was hypothesized to be released through secretory autophagy.

Discussion
In this study, we used genetic deletion of VMP1, a protein previously identi ed for its critical role in regulating SERCA activity [41] and autophagic ux [5], to elucidate whether VMP1 regulates innate in ammatory responses. This study is the rst to characterize the role of VMP1 in innate immune responses and autophagic processes. Initial experiments demonstrated that VMP1 negatively regulates the release of IL-1β gal-3 in response to canonical NLRP3 agonists, as VMP1 KO led to increased release of IL-1β and gal-3 ( Fig. 1 and S1C). Increased release of IL-1β in response to LPS and ATP suggested this phenotype was due in part to increased in ammasome activation. VMP1 KO macrophages had exacerbated NLRP3 in ammasome activation in response to LPS and ATP (Fig. 1). The NLRP3 in ammasome is uniquely activated by a variety of seemingly unrelated signals, and our data showed that several signals were increased in VMP1 KO cells [42]. These signals included Ca 2+ mobilization, mitochondrial dysfunction, and release of mtDNA (Fig. 3, 5, and 6). Ca 2+ mobilization as a signal upstream of NLRP3 in ammasome activation is likely to be indirect, either leading to increased potassium e ux [43]  likely contributed to mitochondrial dysfunction indicated by the loss of membrane potential and release of mtDNA (Figs. 5 and 6). Previous work has shown that NLRP3 activators can trigger apoptosis leading to the loss of mitochondrial membrane potential and release of mtDNA into the cytosol which can then trigger NLRP3 in ammasome activation [47]. We believe that the combination of these signals resulted in increased NLRP3 in ammasome activation in VMP1 KO cells (Fig. 1D,E).
Previous studies have demonstrated that in VMP1 KO cells autophagy was impaired at autophagosome/lysosome fusion [5,38]. Our data was consistent with these ndings where we found that in VMP1 KO cells there was less colocalization between LC3 and LAMP1 in untreated cells as well as decreased levels of LC3-I and increased p62 in treated and untreated cells (Fig. 7). This defect in autophagy likely resulted in the persistence of dysfunctional mitochondria and increased NLRP3 in ammasome activation (Fig. 5, 1D,E) [11,[47][48][49]. Along these lines, previous work showed that caspase-1 activation resulted in the inhibition of mitophagy resulting in the release of mtDNA into the cytoplasm and more dysfunctional mitochondria [50]. Perhaps both VMP1 and caspase-1 activation regulate mitophagy such that with in ammatory stimulation and Ca 2+ overload almost all mitochondria in VMP1 KO cells lose membrane potential and release more mtDNA into the cytosol (Fig. 4, 5, and 6). Typically, with NLRP3 activation, p62 is recruited to damaged mitochondria which are then ubiquitinated through a Parkin-mediated mechanism for degradation, but mitophagy is defective in the absence of VMP1 expression which causes an accumulation of damaged mitochondria [51,52]. Additionally, another mechanism that restricts in ammasome activation involves the interaction between p62 and ASC which targets in ammasome components to autophagosomes for degradation although this response is likely defective in VMP1 KO cells [39]. Aside from promoting in ammasome activation, autophagy inhibition can elevate IL-1β release due to an increase in available proIL-1β for cleavage in the cytosol similar to what we observed (Fig. 1) [53]. As a result of impaired autophagic ux in VMP1 KO cells, we hypothesized that damaged mitochondria and mtDNA were released through secretory autophagy recently termed autophagic secretion of mitochondria [54]. Recent work suggested that mtDNA was primarily released from the cell due to membrane rupture, but a fraction of the mtDNA can also be released through GSDMD/gasdermin E (GSDME) pores upon pyroptotic/apoptotic cell death [55]. However, it seems more likely that in VMP1 KO cells due to diminished autophagosome fusion with the lysosome, there was increased fusion of the autophagosome containing damaged mitochondria and mtDNA with the plasma membrane to release these contents from the cell that otherwise could not be degraded (Fig. 6C,D) [56]. Cathepsin B which is released from damaged lysosomes had more activity in VMP1 KO cells and could increase NLRP3 in ammasome activation (Fig. 7F) [20,57]. Overall, our data suggest that VMP1 through its interactions with SERCA and its modulation of autophagy dampen in ammatory responses by keeping a number of proin ammatory signals in check including Ca 2+ mobilization, degradation of dysfunctional mitochondria, and the release of cytoplasmic and extracellular mtDNA.
Recently, a number of studies have found that VMP1 may be a host-factor utilized by viruses during infection [58-60], suggesting that VMP1 is a putative antiviral target. While our study does not preclude the possibility that VMP1 might be leveraged in this way, our ndings do suggest that targeting VMP1 interactions with viral proteins may be more desirable than inhibiting VMP1 activity more broadly, which may have negative impacts on cellular in ammatory responses that may occur following VMP1 inhibition. This is supported by recent reports that have observed decreased expression of VMP1 has been identi ed in the monocytes of patients with Primary-Progressive MS and in the peripheral blood mononuclear cells (PBMCs) of PD patients [61,62]. The progression of these neurologic diseases is characterized by dysregulated in ammatory responses that contribute to disease sequalae.
In summary, we showed that in response to in ammatory stimuli VMP1 KO cells release more IL-1β and gal-3 due at least in part to increased in ammasome activation compared to control cells. Following ATP stimulation, VMP1 KO cells had elevated levels of cytoplasmic [Ca 2+ ] which resulted in the loss of membrane potential in almost all mitochondria and the cytoplasmic release of mtDNA. Impaired autophagic ux in VMP1 KO cells prevented the cells from degrading damaged mitochondria and promoted the release of mtDNA further exacerbating in ammatory responses. Taken together, our ndings identify a novel role for VMP1 in modulating in ammatory responses. Our ndings support a potentially critical role for VMP1 in the progression of these diseases that perhaps can one day be targeted therapeutically.

Materials
Cell Culture, Differentiation, and Treatments. HEK293T and THP-1 cells were obtained from the American Type Culture Collection (ATCC). Cells were cultured with 5% CO 2 at 37°C in either DMEM or RPMI supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco),10 µg/mL cipro oxacin hydrochloride, 100 IU/mL penicillin, and 100 µg/mL streptomycin. THP-1s were differentiated by adding phorbol 12-myristate 13-acetate (PMA) (Sigma-Aldrich, P1585) at a concentration of 1 µg/mL for 48 h. The cells were then allowed to rest for 72 h prior to treatment. Unless otherwise noted, the cells were treated with 100 ng/mL lipopolysaccharides from Escherichia coli O55:B5 (Sigma-Aldrich, L4524) for 4 h. The media was changed and then some wells were treated with 5 mM adenosine 5-triphosphate disodium salt hydrate (ATP) (Sigma-Aldrich, A2383) for 30 min. To measure the response to α-syn, α-syn brils were obtained as described previously, and 1 µM α-syn was added to differentiated THP-1s for 24 hrs [63]. To test SERCA inhibition, 1 µm thapsigargin (Tocris, 1138) was added at the same time as any in ammatory stimuli including when it served as a control. The supernatant or cells were then collected for analysis.
Cloning and generation of stable cell lines. VMP1 knockout THP-1 cell lines were generated using a modi ed version of the LentiCRISPRv2 plasmid (Addgene plasmid number 52961, a gift from Feng Zhang) that has the puromycin resistance cassette replaced with a G418 resistance cassette to create LentiCRISPRv2-G418 [64]. The following oligonucleotide guide RNA sequence was annealed and cloned into LentiCRISPRv2-G418: VMP1 guide RNA 5'-CTTTTGTATGCCTACTGGAT-3' as described previously [4,23]. Cells transduced with the LentiCRISPRv2-G418 backbone served as a control for selection. To generate stable cell lines, lentivirus was prepared by transfecting equal amounts of VSV-G, psPAX2 (from Didier Trono, NIH AIDS Reagent program [catalog number 11348]) [65, 66], and LentiCRISPRv2-G418 (either the backbone or the clone containing the guide RNA of interest) using polyethylenimine (PEI) into HEK293T cells. Retrovirus was prepared by transfecting equal amounts of VSV-G, pCigB, and pMSCVpuro-Mito-Pericam using PEI into HEK293T cells. Viral supernatant was harvested 48 h posttransfection and ltered through 0.45-µm lters (Millipore). The concentrated supernatant was applied to THP-1 cells by spinoculation at 13°C for 2 h at 1,200 x g. Media was changed 24 h later. Forty-eight hours after transduction, geneticin (G418) (Gibco) was added to the cells at a concentration of 0.5 mg/mL. Following 3-4 weeks of selection, lymphocyte separation media was used to remove dead cells, and healthy cells were collected to validate the knockout by western blot. VMP1 KO cells were maintained in culture for at most 4 weeks following successful selection.
To measure caspase-1 activation, THP-1s were transduced with lentiviral vector prepared as described above with a caspase-1 biosensor containing the IQAD amino acid target sequence as described previously [28]. Sandwich ELISAs. Cell culture supernatants were analyzed using the following kits: Human IL-1 beta/IL-1F2 DuoSet ELISA (R&D Systems, DY201-05) for IL-1β and Human Galectin-3 DuoSet ELISA for gal-3 (R&D Systems, DY1154). The manufacturers' protocols were followed. Alternatively, gal-3 protein levels were also measured using an in-house sandwich ELISA described previously [70]. Brie y, mouse anti-LGALS3 B2C10 (Santa Cruz Biotechnology, SC-32790) was diluted in pH 9.6 carbonate buffer to a nal concentration of 1 µg/mL to coat 96-well Maxisorp ELISA plate (Nunc, 44-2402-22) at 4°C overnight on an orbital shaker. Between each step the wells were washed 3x-5x with PBS containing Tween-20. The wells were blocked 1:1 with RPMI supplemented with 10% characterized FBS and PBS for 2 h at RT. The culture supernatant was then added to the wells. The standard curve was generated by serially diluting recombinant gal-3 (Abcam, ab89487). Biotin conjugated rat anti-LGALS3 (M3/38; Millipore Sigma, 125402) was diluted to a nal concentration of 500 ng/mL in PBS with 1% bovine serum albumin (BSA) (Sigma-Aldrich, A7906) and added to the wells for 2 h on a rocker at RT. Then streptavidin HRP (ImmunoReagents, Ba-103-HRPX) was diluted to 1 µg/mL and incubated for 30 min at RT on a rocker.
The HRP signal was detected with the addition of 1 x 3,3',5,5-tetramethylbenzidine (Invitrogen, 00-4201-56) and then the reaction was quenched with 2 N sulfuric acid. The absorbance was read at 450 nm on a PowerWave XS plate reader (BioTek Instruments) with Gen5 software. The standard curve was t with a 4-Parameter Logistic (4PL) curve.
GAPDH was utilized as a housekeeping gene for normalization.
RNA-Sequencing and pathway analysis. Control or VMP1 KO differentiated THP-1s were left untreated or treated with 100 ng/mL LPS (4 h) then 5 mM ATP (30 min). Supernatant was collected and analyzed by gal-3 ELISA as described above to ensure that the sample phenotype was consistent with previous experiments. RNA was isolated using the NucleoSpin RNA Plus extraction kit. Part of the RNA was saved for qPCR to assess proIL-1β, and gal-3 gene expression as described above. RNA samples were submitted to the University of Chicago Genomics Facility to assess the concentration and quality of the RNA.
Samples that passed the check were then used for library preparation and whole genome sequencing using Illumina NovaSeq. Pathway analysis was performed following a previously published protocol [71]. Heatmaps were generated using the pheatmap function (RRID:SCR_016418) in R. Caspase-1 activation assays. For the FAM-FLICA caspase-1 activation assay, following treatment, the cells were incubated with FAM-FLICA caspase-1 (YVAD) substrate following the manufacturer's protocol (Immunochemistry Technologies, 97). Brie y, the FLICA substrate that was resuspended in DMSO was diluted in PBS 1:5. The diluted substrate was added to the wells at a nal concentration of 1:30. The plate was incubated at 37°C for 1 h. The cells were washed with apoptosis buffer from the kit and were allowed to sit for 10 min in the incubator. Hoechst dye was added at a dilution of 0.5% in apoptosis buffer. The cells were incubated with the dye for 15 min then the cells were washed once with 1x apoptosis buffer then fresh buffer was added prior to imaging. A 20x lens was used to take 10 images per well. Data were collected by z-stack imaging and were analyzed as maximum intensity projections (MIPs). Cells were imaged using z-stacks with 1 µm between each stack and a total of 5 z-stacks. A surface algorithm was built in Imaris around each individual cell, and the data are displayed as the intensity max of each individual cell for a given treatment.
To measure caspase-1 activation using the biosensor, following treatment, cells were lysed with 1x passive lysis buffer (Promega, E1941). Lysates were transferred to a white 96 well plate in duplicate or triplicate. Fire y luciferase substrate was added, and luminescence (relative light units) was quanti ed.
Lactate dehydrogenase assay. Lactate dehydrogenase (LDH) release was measured using a previously published protocol [70,72]. To measure LDH release, supernatant was collected 3 h after signal 2. was used to record changes in cytosolic [Ca 2+ ] cyt . Fluo-4 recordings were acquired in line-scan mode (3 ms per scan; pixel size 0.12 µm). All images were analyzed using ImageJ software (NIH, USA). The [Ca 2+ ] cyt was calculated by the following formula: [Ca 2+ ] cyt = (F 0 -F min )/(F max -F min ), where F 0 was the Fluo-4 uorescence; F max and F min were the uorescence level at 3 mM Ca 2+ /ionomycin and at the lowest baseline recording, respectively. The calcium-induced calcium release was calculated as the summation of the area under the curve for 150 seconds for ATP and 900 seconds for thapsigargin reported in arbitrary units.
THP-1s were differentiated in delta T dishes (Bioptechs, 04200417B). Prior to imaging, the media was replaced by Tyrode's solution (NaCl 135 mM; KCl 4 mM; CaCl 2 3 mM; MgCl 2 1 mM; glucose 10 mM; HEPES 10 mM; pH 7.4). Images were acquired using the 60x lens with the EMCCD camera. Cells were excited at 380 nm (DAPI excitation) or 495 nm (FITC excitation) and emission was recorded at 510 nm [73]. Baseline recordings were taken for approximately 30 sec prior to the addition of 5 mM ATP. Five points were taken per dish, and recordings were taken every 8 sec for about 5 min. All images were analyzed using ImageJ software (NIH, USA). The in ux rate was quanti ed for each cell by using the slope of the linear t of the uorescence change during 15 sec following the addition of ATP. The e ux rate was quanti ed for each cell using the slope of the linear t of a 50 sec period after calcium levels started to decline.
MitoTracker live cell imaging. To assess mitochondrial mass and mitochondrial membrane potential, THP-1s were differentiated in delta T dishes (Bioptechs, 04200417B) and treated as described above. Another spots algorithm was built around the 647 LC3 signal with an estimated diameter of 0.500 µm and area above 0.500 µm 2 .
Quanti cation of cellular and cell-free mtDNA using qPCR. Quantitative PCR was performed to measure both cellular and cell-free mtDNA as described in a published protocol [74]. Brie y, cells were treated as described above then the supernatant and cell pellets were collected. The supernatant was spun down at 1500 rpm for 10 min at 4°C to remove cell debris. The cell pellets were resuspended in 200 µL PBS. DNA was isolated following the manufacturer's protocol from the QIAamp DNA Mini Kit (Qiagen, 51304) except the samples were lysed with buffer AL, mixed by pulse-vortexing, and were incubated at 56°C for 10 min. The DNA was then placed in a bath sonicator for 5 min for supernatant or 10 min for cell lysate. After sonication, the concentration of DNA in each sample was determined and was adjusted to the same concentration. The standard curves were generated as described. The quantity of mtDNA in the supernatant is reported as the absolute copy number per µL. The quantity of mtDNA in the cell lysate is reported as the fold difference using the formula 2 (−ΔΔC t ) .
Statistical Analysis.       Data are from at least three independent experiments with at least 10 images per condition. E) LC3B and SQSTM1/p62 levels were probed by western blot. Statistical differences were calculated using one-way