ABCA1 deficiency contributes to podocyte pyroptosis priming via the APE1/IRF1 axis in diabetic kidney disease

Decreased ATP Binding Cassette Transporter A1 (ABCA1) expression and caspase-4-mediated noncanonical inflammasome contribution have been described in podocytes in diabetic kidney disease (DKD). To investigate a link between these pathways, we evaluated pyroptosis-related mediators in human podocytes with stable knockdown of ABCA1 (siABCA1) and found that mRNA levels of IRF1, caspase-4, GSDMD, caspase-1 and IL1β were significantly increased in siABCA1 compared to control podocytes and that protein levels of caspase-4, GSDMD and IL1β were equally increased. IRF1 knockdown in siABCA1 podocytes prevented increases in caspase-4, GSDMD and IL1β. Whereas TLR4 inhibition did not decrease mRNA levels of IRF1 and caspase-4, APE1 protein expression increased in siABCA1 podocytes and an APE1 redox inhibitor abrogated siABCA1-induced expression of IRF1 and caspase-4. RELA knockdown also offset the pyroptosis priming, but ChIP did not demonstrate increased binding of NFκB to IRF1 promoter in siABCA1 podocytes. Finally, the APE1/IRF1/Casp1 axis was investigated in vivo. APE1 IF staining and mRNA levels of IRF1 and caspase 11 were increased in glomeruli of BTBR ob/ob compared to wildtype. In conclusion, ABCA1 deficiency in podocytes caused APE1 accumulation, which reduces transcription factors to increase the expression of IRF1 and IRF1 target inflammasome-related genes, leading to pyroptosispriming.

www.nature.com/scientificreports/ and IL1β, thus leading to NFκB activation 9 . Non-canonical pyroptosis is initiated by caspase-4/5 in human or caspase-11 in mice, and eventually causes downstream canonical inflammasome formation 10 . Non-canonical pyroptosis was recently found to mediate podocyte injury in DKD 11 . The main inducers of non-canonical pyroptosis known today are lipopolysaccharide (LPS) and activation of TLR4 signaling 9 . Indeed, activation of TLR4 signaling was shown to enhance pyroptosis in tubular cells in experimental DKD 12 . However, the mechanisms leading to non-canonical pyroptosis priming in podocytes remain unclear.
In this study, we demonstrate that ABCA1 deficiency in DKD contributes to the increase in non-canonical pyroptosis-related genes such as caspase-4/11, gasdermin D, caspase-1 and IL1β but not NLRP3. We also show that ABCA1 deficiency-induced priming of non-canonical pyroptosis is mediated by accumulated APE1, subsequent activation of transcription factors via reduction by APE1 and the enhancement of IRF1 transcription. Therefore, we propose the ABCA1/APE1/IRF1 axis to be a novel inducer of non-canonical pyroptosis in DKD.

Results
Non-canonical pyroptosis-related genes are upregulated in ABCA1 KD podocytes. We previously described that decreased ABCA1 expression in podocytes precedes podocyte injury in DKD and sensitizes podocytes to it 5,6 . Pyroptosis was recently identified as a novel mechanism of podocyte injury in DKD. However, the link between ABCA1 deficiency and pyroptosis has not been investigated 8 . To examine whether ABCA1 deficiency contributes to pyroptosis in podocytes, we measured mRNA levels of genes involved in pyroptosis in ABCA1 knockdown podocytes (siABCA1). mRNA expression of caspase-4, GSDMD, caspase-1 and IL1β was significantly increased in siABCA1 podocytes compared to control (p = 0.005 for caspase-4, 0.004 for GSDMD, 0.008 for caspase-1 and 0.01 for IL1β), whereas NLR family pyrin domain-containing protein 3 (NLRP3) expression decreased (Fig. 1a), in contrast to the previous reports indicating that NLRP3 expression is increased in podocytes treated with high glucose and in experimental DKD 8,13 . At the protein level, caspase-4 and GSDMD were significantly increased in siABCA1 compared to siCO podocytes (p = 0.007 for pro caspase-4 and 0.0001 for FL-GSDMD), while caspase-1 expression was increased but at a level that did not reach statistical significance (p = 0.08) (Fig. 1b). Although IL1β protein in cell lysates did not increase in siABCA1, we found increased IL1β precursor in the supernatants of siABCA1 podocytes when compared to siCO (Fig. 1c). Caspase-4 and GSDMD also increased and caspase-1 had the tendency to increase in their non-cleaved inactive forms in siABCA1 podocytes (Fig. 1a,b). These data demonstrated that ABCA1 knockdown in podocytes primes cells for non-canonical pyroptosis but is not sufficient to activate pyroptosis, which is consistent with our previous observation that ABCA1 deficiency contributes to DKD progression but is not sufficient to cause podocyte injury by itself 6 .
TLR4 does not mediate IRF1 upregulation in ABCA1 KD podocytes. TLR4 signaling contributes to non-canonical pyroptosis priming 9 . ABCA1 deficiency in macrophages was found to enhance TLRs signaling by increasing TLRs (including TLR4) trafficking to plasma membrane lipid rafts 12,19 . To explore whether TLR4 signaling contributes to IRF1 and its downstream gene expression changes we observed in siABCA1 podocytes, we treated podocytes with TAK-242, a TLR4 inhibitor. The mRNA levels of IRF1 and caspase-4 were not significantly influenced by the inhibition of TLR4 (Fig. 4). These data suggest that TLR4 signaling does not contribute to ABCA1 deficiency-induced non-canonical pyroptosis priming in podocytes in our system.
Reducing ability of APE1 mediates IRF1 upregulation in ABCA1 KD podocytes. In addition to its role in reverse cholesterol transport, ABCA1 also has a role in exporting APE1 7 . APE1, also known as Ref-1 has two functions 20 . First, APE1 reduces transcription factors to activation form and oxidizes itself. Among the transcription factors reduced by APE1 are HIF1α, NFκB, STAT3, AP1 and p53 21  www.nature.com/scientificreports/ www.nature.com/scientificreports/ apurinic/apyrimidinic endonuclease to nick the abasic site of DNA to prep it for repair. To examine the amount of APE1 in whole cell and nuclei where it functions, we conducted Western blotting for APE1 in whole cell lysate and nuclear fractions of control and siABCA1 podocytes and found that APE1 was increased in siABCA1 podocytes compared to control both in whole cell lysate and nuclear fractions (Fig. 5a,b). Next, we used APX3330, a specific-inhibitor of oxidation-reduction function of APE1, to address the question if the reduction of transcription factors by APE1 contributes to the priming of non-canonical pyroptosis in siABCA1 podocytes. The APE1 www.nature.com/scientificreports/ redox inhibitor offset the increase of siABCA1-induced mRNA transcription of IRF1 and caspase-4, suggesting a role for APE1 in this process (Fig. 5c).

NFκB is involved in IRF1 upregulation induced by ABCA1 downregulation.
NFκB is a transcription factor that plays an important role in pyroptosis priming 9 . Several signaling pathways activate NFκB, including reactive oxygen species or TLR4 signaling [22][23][24] . NFκB can directly induce IRF1 transcription in immune cells while IRF1 can indirectly affect the regulation of NFκB 25,26 . To investigate whether NFκB is involved in the upregulation of IRF1 in siABCA1 podocytes, we used siRNA knockdown for RELA, the NFκB p65 subunit, in siCO and siABCA1 podocytes (Fig. 5d). We found a significant statistical interaction with siABCA1 and siRELA in IRF1 and caspase-4 (p_int = 0.02 for IRF1 and 0.006 for caspase-4), which indicated the role of NFκB in siABCA1-induced upregulation of IRF1 and caspase-4 (Fig. 5e). However, we were unable to confirm that the binding of NFκB to binding sites in IRF1 was increased with siABCA1 podocytes (Fig. 5f).  We previously demonstrated that decreased ABCA1 expression in podocytes plays an important role in DKD progression 5,6 . Immunofluorescence staining demonstrated increased APE1 expression in glomeruli of BTBR ob/ ob compared to WT mice (Fig. 6a). In support, we found increased mRNA levels of IRF1 and caspase-11, which is compatible with the activation of the ABCA1/APE1/IRF1 axis in vivo as well as in vitro in DKD (Fig. 6b).

Discussion
In this study, we demonstrated that ABCA1 deficiency in podocytes, which we previously showed to contribute to the progression of DKD, induced non-canonical pyroptosis priming in association with nuclear APE1 accumulation and subsequent activation of transcription factors by redox reaction (Fig. 7). Among the non-canonical pyroptosis-related genes upregulated in siABCA1 podocytes, caspase-4/11, GSDMD and IL1β were identified to be regulated by IRF1, while caspase-1 was not regulated by IRF1 but increased by APE1 redox reaction.
Although the function of ABCA1 has been largely attributed to cholesterol efflux, ABCA1 is also involved in other cellular export mechanisms 27 . ABCA1 participates in the secretion of phospholipids such as cardiolipin, lysophosphatidylcholine and phosphatidylserine, extracellular vesicles and proteins associated with anti-inflammatory response and apoptotic cell clearance [28][29][30][31] . Recently, a role for ABCA1 in APE1 export has been described 7 .
APE1 mainly has two functions, an endonuclease and a redox function 20,32 . In its redox function, APE1 is oxidized and a transcription factor is reduced through a thiol/sulfide exchange and activated by facilitated DNA binding. Those transcription factors include NFκB, Hypoxia Inducible Factor (HIF), Signal Transducer and Activator of Transcription 3 (STAT3), Activator Protein 1 (AP-1) and cAMP-response element binding protein (CREB). So far, in canonical and non-canonical pyroptosis priming, NFκB was shown to be activated via phosphorylation of inhibitor of NFκB (IκB) kinases by TLR4 signaling 22,33,34 . The canonical pyroptosis pathway involving NLRP3 contributes to DKD but a contribution of non-canonical pyroptosis was also recently reported 11,35 . This study demonstrates that ABCA1 deficiency in podocytes, which we previously demonstrated to play an important role in DKD progression 5,6 , leads to APE1 accumulation and activation of transcription factors by redox reaction of APE1, not by TLR4, in the absence of NLRP3 increase. Thus, this pathway could represent a new mechanism of non-canonical pyroptosis priming in DKD. Our data suggest that APE1 activates several transcription factors and that the combination of their activation could lead to a distinct pattern of gene expression in siABCA1 podocytes. However, further investigation to elucidate the exact details of this process are warranted. It can also be speculated that the involvement of APE1 in mitochondrial function via the regulation of mitochondrial RNA may contribute to DKD 32 .
Priming of pyroptosis in the absence of an activation or cleavage of pyroptosis-related proteins occurred in siABCA1 podocytes, consistent with our previous observation in vitro and in vivo that ABCA1 downregulation itself is not sufficient to cause injury 6 . Nevertheless, ABCA1 induction has been explored as a treatment for DKD 36 . APE1 has been investigated as a treatment target of malignancy, ocular diseases and inflammatory bowel disease 20,32 , and this study indicates the possibility that it could be a target of DKD as well as IRF1.
We have previously published that oxidized cardiolipin is suggested to be implicated in the sensitization of siABCA1 podocytes to DKD. 6 Some relationship between oxidized phospholipids (oxPLs) and non-canonical pyroptosis are described so far. Oxidized 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine (oxPAPC) can bind caspase-11 and activate it like LPS 37 . On the other hand, although we do not know specifically whether oxidized cardiolipin induces pyroptosis, some other oxPLs are demonstrated to sensitize bone marrow-derived macrophages to N-GSDMD-induced cytotoxicity 38 . It is possible that oxPLs and non-canonical pyroptosis mutually affect each other in ABCA1 downregulated podocytes.
In summary, we demonstrated that ABCA1 deficiency in podocytes, which contributes to DKD progression, induces the expression of non-canonical pyroptosis-related genes such as caspase-4/11, GSDMD, caspase-1 and IL1β via nuclear APE1 accumulation and subsequent reduction of transcription factors. Thus, non-canonical pyroptosis priming by ABCA1/APE1/IRF1 axis could represent a novel pathway to be targeted for the treatment of patients with DKD.   39 . Human podocytes were then thermoshifted and differentiated for 14 days at 37 °C in RPMI medium 10% FBS and 1%penicillin/streptomycin. ABCA1 siRNA knockdown (siABCA1) and non-targeting siRNA control (siCO) podocytes were generated and validated as previously described 40 . On day 10-14 of differentiation, podocytes were collected and analyzed. For APE1 inhibition,

Animals. All animal studies were approved by the Institutional Animal Care and Use Committee (IACUC)
at the University of Miami. All experiments were performed in accordance with the guidelines and regulations of IACUC at the University of Miami. Mice were sacrificed at 20 weeks by isoflurane inhalation. The authors complied with the ARRIVE guidelines. All the samples used for this manuscript including mRNA of glomeruli and frozen sections of kidneys of wildtype (wt) and BTBR ob/ob mice were collected and stored for our previous paper 6 and were newly analyzed. The detailed methods are described previously 6 . 9 mice (5 females and 4 males) each were included in wt and ob/ob group. For immunofluorescence, 3 mice from each group which show typical pathological findings of normal or diabetic kidney disease in paraffin sections were selected and tissue in OCT from those 6 mice were freshly cut and analyzed. For qPCR, all the samples were analyzed, but because of the shortage of one sample, only 8 samples were analyzed for caspase-1 and IL1β in ob/ob group.   www.nature.com/scientificreports/ Hercules, A, USA). Western blot analysis was performed using following primary antibodies: caspase-4 ($4450, rabbit, 1:1000), caspase-1 (#3866, rabbit, 1:1000 all from Cell Signaling Technology Inc.), IL1β (AF-201-NA, goat, 1:1000, R and D systems), GSDMD (ab210070, rabbit, 1:1000, Abcam PLC), N-GSDMD (ab215203, rabbit, 1:1000, Abcam PLC), APE1 (NB100-116, mouse, 1:1000, NOVUS biologicals), GAPDH (CB1001, mouse, 1:10,000, Sigma-Aldrich), beta actin (A3854, mouse, 1:10,000, Sigma-Aldrich); or secondary antibodies: antimouse IgG HRP (Promega, W402B, 1:10,000), anti-rabbit IgG HRP (Promega, W401B, 1:10,000), anti-goat IgG HRP (Promega, V8051, 1:10,000). Signal was detected with Radiance ECL (Azure Biosystems Inc., Dublin, CA, USA) or WesternBright ECL (Advansta Inc., San Jose, CA, USA) using Azure c600 Imaging System. Protein from supernatants were isolated using Spin-X UF Concentrator (Corning Inc., Corning, NY, USA) and the values were normalized to the protein concentration of cells in the medium.
Nuclear fractions were obtained using a Cell Fractionation Kit (ab109719, Abcam PLC, Cambridge, UK) according to the manufacturer's instructions.
LPS electroporation. LPS wea electroporated into podocytes using program T-020 in Nucleofector 2b (Lonza, Valais, Switzerland) and Amexa Basic Nucleofector Kit for Primary Mammalian epithelial cells (VPI-1005) according to manufacturer's instruction. LPS-EB ultrapure (Invivogen, San Diego, CA, USA) was added to differentiated podocytes at a concentration of 3.0 ng/400,000 cells. The supernatants were collected at 4 h after electroporation and protein from them were collected using Amicon Ultra (Merck, Darmstadt, Germany) and analyzed with western blotting. Samples were normalized to mL/number of cells in the medium. Z-VAD-fmk (Selleck Biotech, Tokyo, Japan) was added 1 h prior to, during and after electroporation at a concentration of 20 nM.
Chromatin Immunoprecipitation (ChIP). The binding of NFκB to the IRF1 promoter was assessed using the chromatin immunoprecipitation assay. Podocytes were cross-linked for 10 min using 1% paraformaldehyde and sonicated into fragments. The fragmented DNA forming a complex with NFκB was immunoprecipitated with an anti-NFκB rabbit polyclonal antibody (Cell Signaling Technology, Inc. Danvers, MA, USA) and Dynabeads M-280 Sheep Anti-Rabbit IgG (Invitrogen, St. Louis MO, USA). The precipitated DNA was used as a template for PCR reactions with the primers listed in Table 1. Five NFκB binding sites in IRF promoter region (− 500 to + 1500 from the transcription starting point) were predicted by FIMO database and the one nearest to the transcription starting site was selected, which was consistent with the publicly available ChIP-seq data which showed a strong peak near the transcription starting site in B-lymphocytes (GSM935478, GSM935526, GSM935285, GSM5529, GSM935531, GSM935273 and GSM935279) 25 . Neph3 (also known as kirre like nephrin family adhesion molecule 2) promoter region has known NFκB-binding site, and was used as a positive control 41 . Homeobox A13 (HOXA13) is the negative-control primer, which was designed for the promoter regions of chromosome 20.
Images were acquired by laser scanning confocal microscopy using a Leica SP5 Inverted microscope, × 20 objective (Leica Microsystems CMS GmbH, Germany). Measuring of cell fluorescence was performed using Image J software.

Statistical analysis.
All the data are reported as mean ± standard error of mean and as individual values in the dot plots. Two groups were compared using independent sample t-test and a two-way ANOVA test was used to analyze the effect of two independent factors on outcomes and investigate if there is an interaction effect between the two factors on the outcomes. A p-value < 0.05 was considered statistically significant for all tests. GraphPad Prism, version 7.0 (GraphPad Software Inc., San Diego, CA, USA) was used for data analyses.

Data availability
The data underlying this article are available in the article. The datasets referred to in the current study are available in the the Gene Expression Omnibus (GEO) repository, GSM935478, GSM935526, GSM935285, GSM5529, GSM935531, GSM935273 and GSM935279.