Myeloid Ikaros–SIRT1 signaling axis regulates hepatic inflammation and pyroptosis in ischemia-stressed mouse and human liver

Background & Aims: Although Ikaros (IKZF1) is a well-established transcriptional regulator in leukocyte lymphopoiesis and differentiation, its role in myeloid innate immune responses remains unclear. Sirtuin 1 (SIRT1) is a histone/protein deacetylase involved in cellular senescence, inflammation, and stress resistance. Whether SIRT1 signaling is essential in myeloid cell activation remains uncertain, while the molecular communication between Ikaros and SIRT1, two major transcriptional regulators, has not been studied. Methods: We undertook molecular and functional studies to interrogate the significance of the myeloid Ikaros–SIRT1 axis in innate immune activation and whether it may serve as a homeostatic sentinel in human liver transplant recipients (hepatic biopsies) and murine models of sterile hepatic inflammation (liver warm ischemia-reperfusion injury in wild-type, myeloid-specific Sirt1-knockout, and CD11b-DTR mice) as well as primary bone marrow-derived macrophage (BMM) cultures (Ikaros silencing vs. overexpression). Results: In our clinical study, we identified increased post-reperfusion hepatic Ikaros levels, accompanied by augmented inflammasome signaling yet depressed SIRT1, as a mechanism of hepatocellular damage in liver transplant recipients. In our experimental studies, we identified infiltrating macrophages as the major source of Ikaros in IR-stressed mouse livers. Then, we demonstrated that Ikaros-regulated pyroptosis – induced by canonical inflammasome signaling in BMM cultures – was SIRT1 dependent. Consistent with the latter, myeloid-specific Ikaros signaling augmented hepatic pyroptosis to aggravate pro-inflammatory responses in vivo by negatively regulating SIRT1 in an AMPK-dependent manner. Finally, myeloid-specific SIRT1 was required to suppress pyroptosis, pro-inflammatory phenotype, and ultimately mitigate hepatocellular injury in ischemia-stressed murine livers. Conclusion: These findings identify the Ikaros–SIRT1 axis as a novel mechanistic biomarker of pyroptosis and a putative checkpoint regulator of homeostasis in response to acute hepatic stress/injury in mouse and human livers.


Lay summary
This report describes how crosstalk between Ikaros and SIRT1, two major transcriptional regulators, influence acute hepatic inflammation in murine models of liver ischemia-reperfusion injury and liver transplant recipients. We show that the myeloid Ikaros-SIRT1 axis regulates inflammasome-pyroptotic cell death and hepatocellular damage in stressed livers. Thus, the Ikaros-SIRT1 axis may serve as a novel checkpoint regulator that is required for homeostasis in response to acute liver injury in mice and humans.

Introduction
Orthotopic liver transplantation (OLT) has become the standard of care for end-stage liver disease and hepatic malignancies. 1 By leading to early allograft dysfunction and failure, hepatic ischemia-reperfusion injury (IRI) represents a risk factor for acute and chronic OLT rejection and contributes to the shortage of donor organs. 2 Despite compromising clinical outcomes, the mechanisms accounting for liver IRI are not well understood.
Liver IRI represents an innate immune continuum of proinflammatory cytokine release and hepatocyte death. Our group was one of the first to document that macrophage Toll-like receptor 4 (TLR4) signaling triggers a sterile hepatic inflammatory cascade in OLT recipients. 3 TLR4-mediated canonical activation of the inflammasome complex was associated with a more recently identified inflammatory cell death termed pyroptosis. 4 The inflammasome cleaves caspase-1, resulting in proteolytic cleavage of Gasdermin D (GSDMD), the executor of pyroptosis needed for IL1b and IL18 secretion. 5 Its N-terminal domain (GSDMD-N) then translocates to the plasma membrane, causing water influx, cell swelling/lysis, and release of mature IL1b/IL18, which exacerbates immune cell recruitment. 6 We have reported on the TLR-primed inflammasome as a critical signaling platform that detects stressors (e.g., damage-associated molecular patterns) in hepatic IRI 7,8 while others targeted GSDMD to focus on pyroptosis in innate immune cell 9 and Kupffer cell 10,11 death. However, little is known about how recruited macrophages might trigger pyroptosis in acute liver inflammation. Ikaros family zinc finger protein 1 (IKZF1/Ikaros), a conserved nuclear transcriptional factor well-studied in hemato-lymphoid lineage development/differentiation, 12 serves as a repressor of inflammatory genes 13 and regulator of the inflammatory response in mature T cells. 14 By controlling self-antigen mediated B-T cell interactions and restraining TLR signaling, Ikaros acts as a "guardian" preventing autoimmunity/promoting lymphocyte homeostasis. 15 Besides its role in macrophage NF-jB chromatin binding and gene expression in response to lipopolysaccharide (LPS), 16 the function of Ikaros in myeloid cell innate immunity remains elusive. Sirtuin 1 (SIRT1; silent mating type information regulation 2 homolog 1), a member of the class III histone/protein deacetylases, is involved in cellular senescence, inflammation, and stress resistance. While macrophage SIRT1-deficiency triggers NF-jB activation and pro-inflammatory gene programs, 17 we have reported that pharmacologic SIRT1 overexpression promotes an anti-inflammatory phenotype and attenuates sterile inflammation in IR-stressed mouse livers. 18,19 Consistent with SIRT1 orchestrating macrophage phenotypes in the formation of murine abdominal aortic aneurysms, 20 we previously identified that SIRT1-driven M2-type macrophage polarization promoted hepatic rejuvenation in mouse and human OLT. 18,21 However, whether SIRT1 regulates myeloid cell inflammasome activation and pyroptosis to promote the anti-inflammatory phenotype in IR-stressed livers remains unknown. Although Ikaros signaling may control the AMP-activated protein kinase (AMPK) metabolic pathway 22,23 through SIRT1 and AMPK regulation, 24 the molecular communication between Ikaros and SIRT1, two major transcriptional gene regulators, has not been studied.
To interrogate the significance of the Ikaros-SIRT1 myeloid axis in the activation of hepatic innate immunity induced by IR stress, we conducted molecular/functional studies encompassing primary mouse macrophage cultures, mouse models of sterile hepatic inflammation, and human OLT recipients.

Patients and methods
Clinical liver transplant study The study was approved by the UCLA Institutional Research Board (IRB #13-000143). Patients provided informed consent before they participated in the study. As specified by UCLA protocols, we performed a retrospective analysis of 55 adult patients who underwent OLT (May 2013-August 2015) and received routine standard of care and immunosuppressive therapy. Recipients who underwent re-transplantation were excluded from the study. Donor livers, procured from donation after brain death or cardiac death, were perfused and stored in UW solution (Niaspan; Bristol-Meyers Squibb, Princeton, NJ). Protocol Tru-Cut needle biopsies were obtained from the left liver lobe about 2 h after portal reperfusion (before abdominal closure) and snap-frozen. Hepatic biopsies were screened by quantitative reverse-transcription PCR (qRT-PCR) with GAPDH normalization for IKZF1, TLR4, Cathepsin G, CD80, CD86, CXCL10, and western blots with b-actin normalization for SIRT1 and cleaved caspase-1 expression. Recipient blood samples were evaluated for serum alanine aminotransferase (sALT)/serum aspartate aminotransferase (sAST) levels.

Animals
The study was approved by the UCLA Animal Research Committee (ARC #1999-094). C57BL/6 male mice: wild-type (WT), CD11b-DTR (Jackson Laboratory, Bar Harbor, ME), FLOX and myeloid-specific SIRT1-deficient (mSirt1-KO) (NIEHS, Research Triangle Park, NC) were used. Animals were housed under pathogen-free conditions and received human care outlined in the Guide for Care and Use of Laboratory Animals (National Academies Press, 2011).
Liver IRI mouse models Mice were anesthetized, injected with heparin (100 U/kg), and an atraumatic clip was used to interrupt the hepatic artery/portal venous blood supply to the left/middle liver lobes. 25 After 60 min of partial (75%) warm ischemia, the clamp was removed, and mice were sacrificed at 6 h of reperfusion. The sALT levels were measured with Infinity ALT Liquid Stable Reagent (Thermo Scientific, Rockford, IL).
To focus on macrophage-specific Ikaros function, CD11b-DTR mice were first treated with diphtheria toxin (DT; 25 ng/g i.v. at day -1) to deplete native CD11b+ cells, 26 and then infused with in vitro generated siRNA-Ikaros vs. siRNA-Control macrophages (5×10 5 cells i.v.) at 1 h before hepatic IR-insult. DT treatment does not trigger hepatotoxicity in WT mice. 26 For further details regarding the materials and methods used, please refer to the CTAT table and supplementary information.

Results
The pro-inflammatory function of Ikaros in human liver transplantation We aimed first to retrospectively evaluate Ikaros expression and its correlation with hepatocellular function in liver transplant recipients. Fifty-five human OLT biopsies, collected about 2 h after portal reperfusion, were classified into low (n = 28) vs. high (n = 27) IKZF1 gene expression groups (Fig. 1A). As shown in Table S1, the IKZF1-high expression cohort had a significantly higher BMI (p = 0.0441). At the same time, there was no correlation between IKZF1 levels and recipient sex, race, height, ABO compatibility, or preoperative serum AST/ALT levels. In addition, we found no correlation between IKZF1 grouping and donor/graft variables, including sex, height, BMI, or warm/cold graft ischemia times (Table S2). Patients with low hepatic IKZF1 expression showed improved early OLT function, evidenced by lower sALT (p <0.05) compared to the IKZF1high expression group (Fig. 1B). The latter cases were characterized by increased hepatic TLR4 (innate activation marker; p = 0.0010), Cathepsin G (neutrophil marker; p <0.0001), and pro-inflammatory profile (CD86, CD80, CXCL10; p <0.0001) (Fig. 1C).
Ikaros expression by recruited macrophages in IR-stressed WT mouse liver We next screened for Ikaros, SIRT1, and Casp1-p20 expression in a mouse model of partial liver warm ischemia (60 min). By 6 h post-reperfusion, the peak of hepatocellular damage in this model, 27 we observed increased levels of Ikaros in IR-stressed livers associated with enhanced SIRT1 and Casp1-p20, compared with controls ( Fig. 2A). The hepatic induction of IKZF1 was accompanied by increased levels of Cd80, Cd86, and Cxcl10 mRNA (Fig. 2B). Then, we attempted to identify the dominant Ikaros-producing cell in our model. Indeed, hepatic Ikaros colocalized with the CD11b surface marker by immunofluorescence (IF) staining (Fig. 2C)  parenchymal cells (Fig. 2D). The expression of Ikaros by CD11b+F4/80+ liver-resident cells was comparable between Sham and IR-stressed groups (Fig. S1D). Furthermore, extending the hypoxia time (from 60 min to 90 min) increased the hepatic recruitment of Ikaros-expressing CD11b+ cells (Fig. S2). These results identify infiltrating macrophages as the major contributor to Ikaros expression in IR-stressed mouse liver.

Ikaros silencing suppresses inflammasome activation and upregulates SIRT1 in BMMs
To focus on macrophage-specific Ikaros function, we next investigated primary mouse bone marrow-derived macrophage (BMM) cultures. LPS stimulation (100 ng/ml for 6 h) increased mRNA levels of Izkf1 and inflammation markers (Fig. 3C), and subsequently translation of the Ikaros protein. This was accompanied by upregulation of Casp1-p20, as assessed by IF and western blots, respectively (Fig. 3A,B and S3A). Then, we employed the small-interfering RNA (siRNA) approach to test whether Ikaros signaling regulates macrophage inflammation in vitro. BMMs transfected with siRNAs to suppress the Ikaros gene (siIkaros) or scrambled siRNAs (siControl), were stimulated with LPS. We found that Ikaros silencing increased SIRT1 expression, accompanied by decreased Casp1-p20 and IL1b-p17 levels ( Fig. 3D and S3B). At the same time, we observed an increase in Sirt1 mRNA (p <0.01), which correlated with decreased pro-inflammatory cytokine programs (Tnfa, Il1b, Cxcl10; p <0.05) compared with siControl-BMM (Fig. 3E). To determine whether overexpression of Ikaros influences macrophage inflammation, we cultured mouse BMMs transfected with Ikzf1 mRNA or GFP mRNA as a control. Unlike Ikaros-silenced BMMs, those transfected with Ikzf1 mRNA displayed increased Ikaros, CXCL10, and pro-inflammatory markers (iNos, p-STAT1, CD86) but decreased Sirt1 mRNA compared to controls (Fig. 3F/G/S3C). These results highlight that the regulatory function of Ikaros in macrophage inflammation is negatively associated with SIRT1 signaling.

Myeloid SIRT1 deficiency exacerbates liver IRI and promotes inflammasome-pyroptosis activation in vivo
Having documented the significance of the Ikaros-SIRT1 axis in macrophage activation in vitro, we then asked whether myeloidspecific SIRT1 regulates macrophage function in a murine liver IRI model. Groups of mSirt1-deficient (KO) and Sirt1-proficient (FLOX) mice were subjected to 60 min hepatic warm ischemia (Fig. 6A) Fig. 6G-I). Moreover, qRT-PCR analysis showed an increased pro-inflammatory gene expression program (Tnfa, Il1b, Cxcl10, Ly6G, and Ifnc) (Fig. 6J), while Bcl-xL protein levels associated with hepatoprotection were reduced in mSirt1-KO recipients (Fig. 6G,H). Furthermore, a profound suppression of phosphorylated NLRP3 at Ser295 (p-NLRP3) was detected selectively in mSirt1-KO, despite comparable total NLRP3 levels in both FLOX and mSirt1-KO livers (Fig. S8A-C). These results are consistent with the homeostatic function of myeloid SIRT1, evidenced by suppressed pyroptosis, diminished pro-inflammatory cell activation/infiltration, and mitigated hepatocellular injury in IR-stressed livers.

Discussion
In this study, we identified Ikaros-SIRT1 signaling as a regulator of the canonical macrophage inflammasome-pyroptosis pathway, leading to distinct signatures of sterile inflammation in mouse and human IR-stressed livers. To dissect the impact of the Ikaros-SIRT1 axis on macrophage activation, we analyzed mouse BMM cultures (under Ikaros silencing), coupled with screening of IR-stressed livers in WT, myeloid-specific Sirt1-KO, and macrophage-repopulated CD11b-DTR mouse recipients. Parallel screening of hepatic biopsies from 55 liver transplant recipients underlines the role of Ikaros-SIRT1 regulation in the pro-inflammatory myeloid profile and inflammasome-pyroptosis pathway in human liver stress/injury. As Ikaros was expressed primarily by liver-infiltrating CD11b+ macrophages (Fig. 2C,D), we screened for Ikaros function in BMM cultures (Fig. 4). Ikaros silencing impaired BMM commitment toward the pro-inflammatory phenotype and suppressed the inflammasome-pyroptosis pathway, with simultaneous upregulation of SIRT1. Notably, SIRT1 ablation abrogated the Ikarosstimulating effect on caspase-1-GSDMD processing, a prerequisite for pyroptosis. These in vitro data highlight the novel function of Ikaros-SIRT1 in regulation of the inflammasome.
To focus on the Ikaros-SIRT1 axis in BMM activation/ recruitment in vivo, we utilized the CD11b-DTR mouse system, in which: i) adoptively transferred Ikaros-proficient BMMs recreated liver damage in otherwise IRI-resistant CD11b-deficient recipients; ii) Ikaros-silenced BMMs ameliorated liver IRI/ pyroptosis, compared with Ikaros-proficient controls; and iii) SIRT1 was readily expressed by adoptively transferred liverinfiltrating Ikaros-silenced CD11b+ BMMs. In addition to comparable SIRT1-pyroptosis signatures in vitro and in vivo, these data document the divergent impact of Ikaros signaling upon hepatic inflammation. Together, we propose that Ikaros signaling drives activation of the inflammasome and exacerbates liver IRI in conjunction with negative SIRT1 regulation. Indeed, disruption of myeloid-specific SIRT1 unmasked increased pyroptosis and aggravated liver IRI (Fig. 6). However, as mSirt1-KO mice underwent Cre-mediated Sirt1 deletion under the control of M lysozyme, we cannot exclude a possibility that SIRT1 deficiency in liver-resident Kupffer cells may have also contributed to hepatic IRI or pyroptosis in liver-recruited macrophages. Of note, IR-stressed livers without BMM reconstitution showed upregulated Casp1-p20 levels compared to Sham controls (Fig. S6). Likewise, although it remains controversial whether bioactive IL1b derives from circulating monocytes or cirrhotic liver tissue itself, recent data point towards the importance of its serum levels as a prognostic marker. 28,29 The importance of macrophage recruitment for the SIRT1-pyroptosis axis was supported by enhanced SIRT1 levels and reduced hepatic pyroptosis in siIkaros BMM-transferred DTR test mice. We have shown Kupffer cell frequency drastically decreases after IR stress, with a simultaneous hepatic influx of circulating CD11b+ macrophages. 30 Notably, we have detected upregulated Ikaros levels in liver-infiltrating CD11b+ cells, coinciding with enhanced pyroptosis (GSDMD) but negatively associated with SIRT1, in a murine LPS/D-Galactosamine hepatitis model (Fig. S7). These findings validate the role of the Ikaros-SIRT1 axis in the mechanism of macrophage pyroptosis and innate immune-driven liver inflammation cascade.
We have reported on the activation of multiprotein inflammasome complex in the context of caspase-1 induced IL1b production in the mechanism of liver IRI. 7,8 Although caspase-1 may induce apoptosis rather than pyroptosis in certain cell types, 31 accumulating evidence points to GSDMD, the substrate of caspase-1, as a critical mediator of pyroptosis. 6 Indeed, in the current study, upregulated caspase-1 in SIRT1-deficient BMMs was mediated by GSDMD cleavage both in vitro and in vivo. Recently, caspase1-GSDMD processing was demonstrated in innate immune cells but not in hepatocytes, indicating that selective blockade of inflammasome-activated pyroptosis in the liver infiltrate might mitigate IRI. 9 Our current study also showed that the myeloid Ikaros-SIRT1 axis regulated caspase-1-GSDMD processing to secrete IL1b and IL18, i.e. cytokines central to inflammasome activation and pyroptosis. In agreement with the ability of bound IL18 to reduce liver infiltration by neutrophils and CD3+ cells, 32 IFNc and CD4 gene expression showed a similar trend in our study, albeit these failed to reach statistical significance. Thus, we may link the pro-inflammatory profile in cultured BMM/IR-stressed liver with upregulated caspase-1-GSDMD processing and IL1b/IL18 secretion. However, caspase-1 may exert a pro-inflammatory function independent of IL1b/IL18, 33  (C) wild-type BMMs transfected with siControl or siIkaros RNAs were pre-treated with Compound C (10 lM, 2 h). Lysates were probed by western blots for Ikaros, these cytokines might occur without macrophage death. 34,35 Hence, further studies are needed to clarify the relationship between pro-inflammatory signaling and caspase-1-GSDMD processing in the pyroptotic pathway in liver IRI. In addition to IL1b, IL1a was recently found to activate the inflammasome platform during cirrhosis. 29 In agreement with the distinct cleavage mechanism and IL1b/IL18 release, 36,37 our LPS/ATP-stimulated BMM cultures showed the release of pro-IL1a, but not of its mature (C terminal) form. Future experiments need to address the exact role of IL1a processing and signaling in sterile inflammation in response to IR stress. SIRT1 signaling extends lifespan and protects cells from various environmental stresses. 38 We reported that treatment with resveratrol, a small molecule SIRT1 activator, promoted myeloid protective function in mouse liver IRI. 18,19 In the present study, we contrasted myeloid-specific SIRT1 signaling to demonstrate, for the first time, that SIRT1 activation suppressed pyroptosis in IR-stressed livers. This is consistent with its effects on macrophage inflammasome function 39 when the myeloid AMPK/SIRT1 axis suppressed colitis through NLRP3 inflammasome and caspase-1 activation. Although SIRT1 induction has been linked with macrophage inflammasome activation, the regulatory mechanism of SIRT1 transcription has not been studied. Here, we report a new molecular mechanism by which Ikaros negatively regulates SIRT1 in a mouse model of hepatic inflammation and human OLT.
Crosstalk between immunity and metabolism is increasingly recognized, 40 and Ikaros signaling was shown to affect the AMPK metabolic pathway in a leukemia study. 23 Notably, both pharmacological and genetic inhibition of AMPK in our study point towards Ikaros-SIRT1 function being AMPK dependent. Although SIRT1 and AMPK are metabolic stress sensors, which may regulate each other in a positive feedback loop, 41,42 putative regulation of AMPK by SIRT1 remains unclear. 43 Indeed, consistent with published data from Sirt1 KO mice, 44 mSirt1-deficient BMMs showed higher expression of p-AMPKa than FLOX counterparts in our study. In this context, p-AMPKa upregulation alone does not necessarily suppress inflammasome activation under stress. Taken together, as part of an orchestrated molecular communication network, Ikaros regulates macrophage activity via the AMPK-SIRT1 axis. With published data deriving primarily from lymphocyte differentiation/developmental studies, there is an unmet need to further dissect Ikaros-SIRT1 signaling pathways in activated macrophages.
The question arises as to how the Ikaros-SIRT1 axis may regulate the canonical inflammasome-pyroptosis pathway. First, Ikaros negatively associated with phosphorylation of NLRP3 at Ser295 in our study, which was shown by others to inhibit inflammasome assembly. 45 Notably, Ikaros knockdown did not affect total NLRP3 levels despite suppressing NLRP3 activation in our LPS/ATP macrophage culture system. Thus, consistent with recent findings, 46 the Ikaros-SIRT1 axis may control inflammasome activation through post-translational modifications. In contrast to SIRT1, sirtuin 2 regulates the acetylation of NLRP3. 47 Thus, phosphorylation of NLRP3 may represent the molecular target for the Ikaros-SIRT1 axis. Second, Ikaros negatively regulated iNos (Fig. S4), which was shown to suppress inflammasome activation. 48,49 Furthermore, in line with unchanged total NLRP3 levels after Ikaros knockdown in our study, others suggested post-translational modulation as a mechanism by which iNos regulates inflammasome signaling. 48,50 On the other hand, the Ikaros-SIRT1 axis failed to correlate with anti-inflammatory IL10 signaling (data not shown), which was reported by others to regulate NLRP3 function. 51 In our clinical arm, it was remarkable to find that increased Ikaros levels were accompanied by significantly increased inflammatory markers and cleaved caspase-1 but decreased SIRT1 expression in human OLT biopsies (Fig. 1). Thus, Ikaros might represent a novel regulator of the macrophage inflammasomepyroptosis pathway in humans, while both Ikaros and SIRT1 might serve as putative therapeutic targets in liver transplantation. One might envision that prospective transplant patients may be conditioned, if needed, with Ikaros-suppressing or SIRT1-inducing compounds before the surgical procedure. Consistent with our reports on genetically modified BMM transfers in IRI-OLT models, 26,52 native BMMs can be manipulated ex vivo to suppress Ikaros or enhance SIRT1 expression and then be autografted back into patients.
Human OLT recipients with Ikaros-high levels had significantly higher BMI than the Ikaros-low clinical cohort. Additionally, the ratios of metabolic diseases such as diabetes mellitus and hypertension were more common in the Ikaroshigh patient group, although differences failed to reach statistical significance (Table S1). As the number of obese patients undergoing OLT increases, 53 several studies focused on pretransplant obesity as a risk factor, while recipient BMI as a prognostic factor for worsened OLT outcomes remains controversial. 54 The relationship between obesity and macrophage activation was studied in non-alcoholic fatty liver disease and non-alcoholic steatohepatitis. 55 In these conditions, distant macrophage differentiation in adipose tissues is regulated through cytokines, damage-associated molecular patterns, or free fatty acids. 56 Furthermore, peripheral blood-derived macrophages from individuals with type 2 diabetes displayed enhanced caspase-1 activation. 57 Therefore, patients with increased Ikaros expression might present activated inflammasome and pyroptotic circulating monocyte/macrophage profiles. Despite apparent disparities in pro-inflammatory parameters and IR-mediated hepatocellular damage, other factors, such as patient/graft survival or early allograft dysfunction, did not show significant differences. First, as the number of liver biopsies was limited (n = 55), a prospective study in a larger patient cohort is warranted. Second, the predictive value of Ikaros for post-OLT clinical outcomes remains uncertain and needs to be examined.
In summary, our translational study documents a novel regulatory function of the Ikaros-SIRT1 axis via AMPK signaling in macrophage activation. It may serve as a therapeutic target and as a checkpoint regulator of homeostasis in response to acute hepatic stress/injury in mouse and human livers.