USP22 suppresses the NLRP3 inflammasome by degrading NLRP3 via ATG5-dependent autophagy

ABSTRACT

The NLRP3 inflammasome is involved in a diverse range of inflammatory diseases. The activation of inflammasomes must be tightly regulated to prevent excessive inflammation, and the protein ubiquitination system is reported to be one of the ways in which inflammasome activation is regulated. However, the deubiquitination regulatory mechanisms of inflammasome activation remain elusive. Here, we demonstrated that USP22 (ubiquitin specific peptidase 22) promotes NLRP3 degradation and inhibits NLRP3 inflammasome activation. USP22 deficiency or in vivo silencing significantly increases alum-induced peritonitis and lipopolysaccharide-induced systemic inflammation. Mechanistically, USP22 inhibits NLRP3 inflammasome activation via the promotion of ATG5-mediated macroautophagy/autophagy. USP22 stabilizes ATG5 via decreasing K27- and K48-linked ubiquitination of ATG5 at the Lys118 site. Taken together, these findings reveal the role USP22 plays in the regulation of NLRP3 inflammasome activation and suggest a potential therapeutic target to treat NLRP3 inflammasome-related diseases.

Abbreviations: ATG5: autophagy related 5; ATP: adenosine triphosphate; CASP1: caspase 1; IL18: interleukin 18; IL1B/IL-1β: interleukin 1 beta; LPS: lipopolysaccharide; NLRC4: NLR family, CARD domain containing 4; NLRP3: NLR family, pyrin domain containing 3; PYCARD/ASC: PYD and CARD domain containing; TNF/TNF-α: tumor necrosis factor; USP22: ubiquitin specific peptidase 22.

Introduction

The inflammasomes play an important role in the process of many inflammatory diseases. There are four main members of the inflammasomes, such as NLRP1, NLRP3, NLRC4 and AIM2, among which NLRP3 inflammasome is the most studied [1]. The NLRP3 inflammasome is a molecular platform for the activation of cysteine protease CASP1, which leads to the cleavage and secretion of the proinflammatory cytokines IL1B/IL-1β (interleukin 1 beta) and IL18 [1,²]. The NLRP3 inflammasome plays critical roles in host immune responses to infections and sterile injuries [3,4]. Under physiological conditions, the activation and inactivation of the NLRP3 inflammasome are tightly regulated to maintain immune homeostasis. However, in pathological states, excessive or persistent NLRP3 inflammasome activation can lead to a wide variety of inflammatory, metabolic, degenerative, and aging-related diseases, such as gout, atherosclerosis, type 2 diabetes, and Alzheimer disease [1,^5,6]. Although the NLRP3 inflammasome has been extensively investigated, endogenous mechanisms involved in the negative regulation of the NLRP3 inflammasome remain largely unknown [7].

The activation of the NLRP3 inflammasome requires two signals. The first signal is provided by endogenous molecules or microbial molecules such as lipopolysaccharide (LPS), which induce the expression of NLRP3 and pro-IL1B via the activation of the NFKB/NF-κB signaling pathway; the second signal is trigged by the presence of a variety of substances, including ATP, bacterial and fungal toxins and components, viral RNA, and particulate matter [8–11]. The NLRP3 inflammatory response has been reported to be regulated at both the transcriptional and post-translational modification level of NLRP3, as well as other inflammasome components [12,13]. Importantly, autophagy has also been reported to play a protective role in the negative regulation of NLRP3 inflammasome activation. The autophagic removal of NLRP3 inflammasome activators, such as intracellular DAMPs, NLRP3 inflammasome components, and cytokines, can reduce NLRP3 inflammasome activation and inflammatory responses [14–16].

Deubiquitination enzymes (DUBs) are a family of proteases that remove ubiquitin tags from proteins of interest undergoing proteasomal degradation. Approximately 100 DUBs have been identified in human cells [17]. DUBs are classified into distinct families according to their catalytic domain organization, including ubiquitin-specific peptidases (USPs), ubiquitin C-terminal hydrolases (UCHs), ovarian tumor proteases (OTUs), Machado-Joseph disease proteases (MJDs), and COPS5/Jab1/Mov34/Mpr1 metalloproteases (JAMMs) [18]. USP22 (ubiquitin specific peptidase 22) belongs to the USP family of DUBs; it is thought to play roles in cell-cycle progression, chemotherapy resistance, and metastasis. A high level of USP22 expression may be an important factor in tumor progression and could serve as a molecular marker of tumors [19,20]. Recently, USP22 was also found to be involved in host antiviral responses, deubiquitinating and stabilizing KPNA2 following viral infection to facilitate nuclear translocation of IRF3 [21]. However, it is unclear whether USP22 is involved in inflammatory responses, and this remains to be investigated.

In this study, we investigated the role that USP22 plays in the regulation of the NLRP3 inflammasome. We used both in vitro and in vivo assays to explore this. Our findings suggest that USP22 does indeed play a role in regulating the activation of the NLRP3 inflammasome. We speculate that this may represent a potential therapeutic target to treat NLRP3 inflammasome- or inflammation-related diseases, such as sepsis, gout, and atherosclerosis.

Results

USP22 silencing promotes activation of the NLRP3 inflammasome

NLRP3 is a core protein of the NLRP3 inflammasome, so we hypothesized that DUBs might regulate the activity or stability of NLRP3 and thereby facilitate or inhibit NLRP3 inflammasome activation. To test this hypothesis, we screened for DUBs that affected the stability of NLRP3 by co-transfection of individual DUBs and HA-tagged NLRP3 into 293 T cells, followed by immunoblot assays. The results showed that USP22 might be a candidate DUB that promotes the degradation of NLRP3 (Fig. S1). To elucidate the possible role of USP22 in the activation of the NLRP3 inflammasome, USP22 was efficiently downregulated in mouse peritoneal macrophages and human THP-1 cells by specific small interfering RNA (siRNA) (Fig. S2A and B). The immunoblot assays also indicated that the NLRP3 protein level was increased in USP22-silenced macrophages stimulated by LPS at the indicated hours (SI Appendix, Fig. S2C). USP22 silencing had no effect on the expression of Nlrp3 and Il1b mRNA (Fig. S3A-B). To assess whether USP22 regulates the NLRP3 inflammasome activation, mouse peritoneal macrophages (PMs) were primed with LPS for 16 h and then stimulated with NLRP3 activator nigericin. USP22 silencing had no effect on the protein level of pro-CASP1, pro-IL1B, or PYCARD (Figure 1(a)). USP22 silencing also had no effect on the expression of NLRP3, IL1B or CASP1 mRNA in LPS plus nigericin stimulated mouse PMs and human THP-1 cells (Fig. S3C-H). However, the protein level of NLRP3 was elevated in USP22-silenced macrophages, and the NLRP3 inflammasome activation was strengthened, as more cleaved CASP1/p10 and IL1B/p17 was detected in the supernatant of USP22-silenced macrophages treated with the NLRP3 inflammasome activator nigericin (Figure 1(a)). Furthermore, IL1B secretion was significantly increased in both mouse peritoneal macrophages and THP-1 cells (Figure 1(b)); IL18 secretion was also elevated in mouse peritoneal macrophages (Figure 1(c)). These results indicated that USP22 negatively regulated NLRP3 inflammasome activation.

Figure 1.

USP22 silencing promotes NLRP3 inflammasome activation. (a) Immunoblots of IL1B, pro-CASP1, and cleaved CASP1 in the supernatants or cell lysates of USP22-silenced mouse primary macrophages or PMA-pretreated human THP-1 cells, primed with LPS, and then stimulated with nigericin (Nig.). (b) ELISA analysis of IL1B in supernates (SNs) from mouse primary macrophages and human THP-1 cells. (c) ELISA analysis of IL18 in SNs from mouse primary macrophages and human THP-1 cells. (d) USP22-silenced or control THP-1 macrophages were treated with palmitate at the indicated concentrations for 24 h. Immunoblots show the NLRP3, IL1B, pro-CASP1 and cleaved CASP1 in the cell lysates. And (e) IL1B in SNs from THP-1 cells silenced of USP22 were analyzed by ELISA. (f) qPCR analysis of the expression of Usp22 in peritoneal exudate cells (PECs) recovered 6 h after alum injection in Usp22-silenced mice and control mice. (g) Immunoblot analysis of NLRP3 expression in peritoneal exudate cells recovered from Usp22-silenced mice and control mice 6 h after alum injection and quantization of NLRP3. (h and i) Flow cytometry analysis of the inflammatory cell subset in peritoneal exudate cells recovered 6 h after alum injection (n = 6). (J) ELISA analysis of IL1B secretion in the lavage fluid from Usp22-silenced mice or control mice (n = 6–8). Data shown are representative of three separate experiments (A, D, and G) or are means ± SEM of three independent experiments (B, C and E). * p < 0.05, ** p < 0.01.

The activation of NLRP3 inflammasome included two steps, priming and triggering (assembly). We first examined whether USP22 regulates LPS-induced priming of inflammasome activation. Our results showed that Usp22 silencing had no influence on activation of NFKB and MAPK signaling pathways (Fig. S3I). Usp22 silencing also had no influence on either mRNA or protein levels of TNF/TNF-α and IL6 (Fig. S3J-M). These results suggest that USP22 silencing did not affect the LPS-induced priming. Defective mitochondria are potential sources of reactive oxygen species (ROS), which was reported as one of the critical mediators in triggering NLRP3 inflammasome assembly [22]. Our results showed that Usp22 silencing had no effect on nigericin-induced mitochondrial fission and clustering (Fig. S4A). Besides, Usp22 silencing also had no influence on ROS production (Fig. S4B and C). Taken together, these results indicated that USP22 negatively regulates NLRP3 protein stability and thus controls the activation of the NLRP3 inflammasome.

To further investigate the role USP22 plays in NLRP3 inflammasome activation, we exposed cultured THP-1 cells to palmitate, as a model of hyperlipidemia conditions [23]. The palmitate was used to stimulate THP-1 macrophages for 24 h. An immunoblot assay also indicated that the level of NLRP3 protein was increased in USP22-silenced THP-1 cells stimulated with palmitate at the indicated concentrations (Figure 1(d)). Furthermore, an ELISA assay showed that IL1B secretion was also significantly elevated following USP22 silencing (Figure 1(e)). These results suggested that USP22 silencing could increase NLRP3 inflammasome activation in response to palmitate challenge.

Given that USP22 inhibited NLRP3 inflammasome activation in vitro, we further examined the biological effect of USP22 in vivo by knocking down Usp22 in a mouse peritonitis model. The Usp22 and control-siRNA oligo complex with in vivo jetPEI significantly reduced the endogenous Usp22 mRNA level in the peritoneal cavity (Figure 1(f)). Then, the mice transfected with siRNA oligos were stimulated with alum to trigger peritonitis. Silencing of Usp22 enhanced the expression of NLRP3 in peritoneal exudate cells (PECs) (Figure 1(g)). We analyzed the peritoneal exudate cells and neutrophils in the lavage samples. The total number of peritoneal exudate cells recruited upon alum challenge was markedly increased in mice with silencing of Usp22; neutrophils were also significantly increased in the Usp22-siRNA transfection group (Figure 1(H,i)). We also detected IL1B and TNF in the lavage samples, with the results showing that IL1B was significantly increased in Usp22-silenced mice, which indicated more intense inflammation resulting from Usp22 silencing in vivo (Figure 1(j)). These findings further demonstrated that USP22 suppressed the expression NLRP3 and inhibited the activation of NLRP3 inflammasome activation in the Usp22-silencing mouse peritonitis model.

Usp22 deficiency promotes activation of the NLRP3 inflammasome

To further investigate the potential functions of USP22 in regulating the NLRP3 inflammasome, we used Usp22-deficient (Cre-Esr usp22^fl/fl) mice, obtained through the partial deletion of exon 2 of the murine Usp22 locus, induced by tamoxifen²¹. We utilized Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice and generated Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl peritoneal macrophages and BMDMs, followed by LPS+nigericin or LPS+alum stimulation. Again, the knockout of Usp22 markedly elevated nigericin and alum-induced CASP1 and IL1B cleavage, in both peritoneal macrophages and BMDMs (Figure 2(a,b)). The secretion of IL1B and IL18 was also increased in Usp22-deficient macrophages (Figure 2(c–f)). To date, MCC950 is the most potent and most specific small-molecular inhibitor of NLRP3 activation known [24]. When selective blockade of NLRP3 inflammasome activation using MCC950, the secretion of IL1B was reduced in usp22 knockout macrophages (Figure 2(g)), which indicating USP22 inhibits NLRP3 activity to suppress the secretion of IL1B. In contrast, Usp22 deletion did not affect TNF secretion (Figure 2(h)). We also examined whether USP22 specifically regulated NLRP3 inflammasome activation and found that deletion of Usp22 expression had no effect on either flagellin-induced NLRC4 inflammasome activation or poly(dA:dT) transfection-induced AIM2 inflammasome activation (Figure 2(i.j)). Taken together, our data demonstrated that Usp22 deficiency specifically enhanced NLRP3 inflammasome activation.

Figure 2.

Knockout of Usp22 promotes activation of the NLRP3 inflammasome. (a) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were stimulated with LPS for 16 h and activated with nigericin for 30 min or alum for 6 h. Immunoblot analysis of the expression of indicated proteins. (b) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl BMDMs were stimulated with LPS for 16 h and activated with nigericin for 30 min or alum for 6 h. Immunoblot analysis of the expression of the indicated proteins. (c) ELISA analysis of IL1B in the supernatants of usp22-knockout mouse primary macrophages. (d) ELISA analysis of IL1B in the supernatants of usp22-knockout BMDMs. (e) ELISA analysis of IL18 in the supernatants of usp22-knockout mouse primary macrophages. (f) ELISA analysis of IL18 in the supernatants of usp22-knockout BMDMs. (g) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were pretreated with MCC950, then stimulated with LPS and nigericin. ELISA analysis of IL1B in the supernatants. (h) ELISA analysis of TNF in the supernatants of usp22-knockout mouse primary macrophages. (i) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl BMDMs were stimulated with LPS for 16 h, then transfected with flagellin (0.5 μg/mL) for 1 h; ELISA analysis of IL1B in the supernatants. (j) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl BMDMs were stimulated with LPS for 16 h, then transfected with poly(dA:dT) (2 μg/mL) for 1 h; ELISA analysis of IL1B in the supernatants. Data shown are representative of three separate experiments (A, B) or are means ± SEM of three independent experiments (C-J). * p < 0.05, ** p < 0.01, *** p < 0.001. ns, not significant.

Usp22 deficiency enhances alum-induced peritoneal inflammation and LPS-induced systemic inflammation by inhibition of the NLRP3 inflammasome

To confirm the role played by USP22 in regulating NLRP3 inflammasome activation in vivo, we intraperitoneally injected the Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice with alum, which induced inflammatory cell-infiltration and peritonitis in an NLRP3-dependent manner. The results showed that Usp22 knockout significantly increased the number of total peritoneal exudate cells and neutrophils recruited (Figure 3(a,b)). The total protein concentration in peritoneal exudate was also markedly increased in usp22-deficient mice (Figure 3(c)). The secretion of IL1B, but not TNF, was significantly elevated in Cre-Esr usp22^fl/fl mice compared with its secretion in control Cre-Esr Usp22^fl/+ mice (Figure 3(d,e)). NLRP3 expression in peritoneal exudate cells was also increased in usp22-deficient mice (Figure 3(f)). Therefore, these findings suggest that USP22 inhibited the activation of the NLRP3 inflammasome in vivo via regulation of the NLRP3 protein.

Figure 3.

Usp22 knockout exacerbates LPS-induced systemic inflammation and alum-induced peritoneal inflammation. (a-f) 8-week-old Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl female mice (n = 6/group) were intraperitoneally injected with alum (1 mg/mouse); 6 h later, the mice were euthanized for the following analyses. (a and b) Flow cytometry analysis of the inflammatory cell subset in peritoneal exudate cells recovered 6 h after alum injection. (c) Protein concentration in the peritoneal lavage fluids. (d and e) ELISA analysis of IL1B and TNF secretion in the peritoneal lavage fluids. (f) Immunoblot analysis with the indicated antibodies of peritoneal exudate cells isolated from Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice 6 h after alum injection. (g) 8-week-old Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl female mice (n = 10/group) were intraperitoneally injected with LPS (20 mg/kg), and survival of the mice was monitored for 48 h. (h-q) 8-week-old Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl female mice (n = 6/group) were intraperitoneally injected with LPS (10 mg/kg). (h and i) Flow cytometry analysis of the inflammatory cell subset in peritoneal exudate cells recovered 4 h after LPS injection. (j-l) ELISA analysis of IL1B, IL18, and TNF secretion in the peritoneal lavage fluids. (m-o) ELISA analysis of IL1B, IL18, and TNF secretion in the serum. (p) H&E staining of left-lung samples. (q) Immunoblot analysis with the indicated antibodies of peritoneal exudate cells recovered from Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice 4 h after LPS injection. Data shown are means ± SEM (A-E and G-O). * p < 0.05, ** p < 0.01, *** p < 0.001. ns, not significant.

Septic shock is an inflammatory disorder, so we sought to determine whether USP22 could suppress the inflammatory response in septic shock via inhibition of the NLRP3 inflammasome. The Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice were challenged with LPS to induce septic shock, and the results showed that Usp22 deficiency increased the mortality of mice with endotoxic shock (Figure 3(g)). Flow cytometry analysis further showed that the total number of peritoneal exudate cells and neutrophils recruited upon LPS challenge was markedly increased in usp22-knockout mice (Figure 3(h,i)). An ELISA assay also indicated that the deletion of Usp22 markedly promoted the secretion of IL1B and IL18 in peritoneal lavage fluids but had no effect on TNF (Figure 3(j-l)). Similar results were obtained for serum (Figure 3(m-o)). Moreover, deletion of Usp22 exacerbated endotoxemia-induced lung injury, as more inflammatory cell infiltration and pulmonary alveolar epithelium damage in usp22-deficient mice. (Figure 3(p)). In addition, NLRP3 expression in peritoneal exudate cells was increased in usp22-deficient mice (Figure 3(q)). These results suggested that USP22 can suppress NLRP3 inflammasome activation and protect against septic shock in mice by promoting NLRP3 degradation in vivo.

USP22 interacts with NLRP3 at the LRR domain

To examine the endogenous association between NLRP3 and USP22, we performed co-immunoprecipitation experiments in mouse peritoneal macrophages with LPS stimulation or not. Immunoprecipitation of USP22 pulled down NLRP3 (Figure 4(a)), while immunoprecipitation of NLRP3 pulled down USP22 (Figure 4(b)). In addition, USP22 was partially co-localized with NLRP3 in the perinuclear region, and this interaction was markedly enhanced by LPS and nigericin stimulation (Figure 4(c)). These results indicated that USP22 mainly interacts with NLRP3 in the cytoplasm and regulates the amount of NLRP3 protein. To explore which domain(s) of NLRP3 were responsible for its interaction with USP22, we used three NLRP3 truncated mutants: Flag-NLRP3-ΔPYD, Flag-NLRP3-ΔNACHT, and Flag-NLRP3-ΔLRR, which were co-transfected with MYC-USP22 in 293 T cells. Coimmunoprecipitation experiments with these truncated mutants revealed that NLRP3-ΔLRR lost the ability to interact with USP22, whereas Flag-NLRP3-ΔPYD and Flag-NLRP3-ΔNACHT did not (Figure 4(d)). These findings suggested that USP22 interacts with NLRP3 at the LRR domain.

Figure 4.

USP22 interacts with NLRP3 at the LRR domain. (a) Co-IP analysis of the interaction between NLRP3 and USP22 in mouse peritoneal macrophages following immunoprecipitation with anti-USP22. (b) Co-IP analysis of the interaction between NLRP3 and USP22 in mouse peritoneal macrophages following immunoprecipitation with anti-NLRP3. (c) Immunofluorescence staining for USP22 (green) and NLRP3 (red) in mouse peritoneal macrophages. Scale bars represent 10 μm. (d) 293 T cells were co-transfected with MYC-USP22 and Flag-NLRP3 or its truncated mutants, immunoprecipitated with anti-Flag, and immunoblotted using the indicated antibodies. Data shown are representative of three separate experiments.

USP22 promotes NLRP3 degradation via lysosomal degradation

As we had determined that USP22 interacts with NLRP3, we next investigated how USP22 regulates the stability of NLRP3. To do this, we co-transfected 293 T cells with HA-NLRP3 and various concentrations of USP22; the results showed that USP22 promoted NLRP3 degradation in a dose-dependent manner (Figure 5(a)). USP22 also accelerated the degradation of NLRP3 following cotreatment with a protein synthesis inhibitor, cycloheximide (CHX) (Figure 5(b)), indicating that USP22 post-translationally regulated the stability of the NLRP3 protein. We next investigated the mechanisms underlying the regulation of NLRP3 stability by USP22. Using two different protein degradation inhibitors, chloroquine (CQ) and MG132, we confirmed that USP22 promoted NLRP3 degradation in a lysosome-dependent manner, as only CQ could rescue the degradation of NLRP3 (Figure 5(c)). According to whether depending ATG5/ATG7 or not, autophagy is classified into canonical autophagy and alternative autophagy [25]. We silenced ATG5 or ATG7 in 293 T cells by siRNA and then co-transfected 293 T cells with HA-NLRP3 and Flag-USP22; the results showed that silencing ATG5 or ATG7 also could rescue the degradation of NLRP3, indicating that USP22 regulates NLRP3 stability by canonical autophagy (Fig. S5A). Autophagy has three types, including macro-autophagy, micro-autophagy and chaperone-mediated autophagy [26]. VER-155008 is an inhibitor of HSPA8 that can inhibit chaperone-mediated autophagy [27]. Our results showed that the accelerated degradation of NLRP3 in USP22-overexpressing cells was rescued by CQ, 3-MA and Baf-A1, but not by VER-155008, indicating USP22 promotes NLRP3 degradation independent of chaperone-mediated autophagy (Fig. S5B).

Figure 5.

USP22 post-translationally regulates NLRP3 stability. (a) 293 T cells were co-transfected with HA-NLRP3 and various concentrations of Flag-USP22, then immunoblot assays were performed using the indicated antibodies. (b) 293 T cells were co-transfected with HA-NLRP3 and Flag-USP22 and incubated with CHX for the indicated times. HA-tag and Flag-tag antibodies were used to detect the expression of relevant proteins. The quantification of relative NLRP3 levels is shown in the lower panel. (c) 293 T cells were co-transfected with HA-NLRP3 and Flag-USP22 for 24 h, incubated with the lysosome inhibitor CQ or proteasome inhibitor MG132 for 4 h, then immunoblot analyses were performed using the indicated antibodies. (d) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were incubated in EBSS conditions with or without CQ for 3 h, the immunoblotted with the indicated antibodies. (e) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were treated with rapamycin at the indicated concentration for 12 h, the immunoblotted with the indicated antibodies. (f) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were incubated in EBSS conditions with or without CQ for 3 h, and immunofluorescence staining for LC3B (green) and DAPI (blue). Scale bars represent 10 μm. (g) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were treated with rapamycin (250 nM) for 12 h, and immunofluorescence staining for LC3B (green) and DAPI (blue). Scale bars represent 10 μm. (h) Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl primary macrophages were stimulated with LPS for 16 h, and with or without rapamycin (250 nM) for 12 h, the immunoblotted with the indicated antibodies. (i) The production of IL1B was analyzed by ELISA, Data shown is mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. (J) 293 T cells were co-transfected with HA-NLRP3, Flag-USP22, and Flag-USP22-C185S, then immunoblotted with the indicated antibodies. Data shown are representative of three separate experiments.

To further confirm the role of USP22 in autophagy, we obtained the peritoneal macrophages from Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice and found that knockout of Usp22 significantly reduced the amount of the autophagy marker LC3-II in autophagy-induced conditions with Earle’s balanced salt solution (EBSS) treatment or rapamycin treatment (Figure 5(d,e)). We then monitored the autophagic flux by detecting the LC3 using immunofluorescent staining. The results showed that under EBSS-treated conditions or rapamycin treated conditions, depletion of USP22 remarkably reduced the formation of LC3B puncta (Figure 5(f,g)). These results suggest that USP22 positively regulates autophagy. Previous studies had reported that autophagy deficiency could induce aberrant activation of inflammasome [28]. Consistently, we also observed reduced NLRP3 and IL1B production in Rapamycin-treated macrophages, indicating the regulation of inflammasome by autophagy (Figure 5(h,i)). We also found that the enzymatically inactive mutant USP22 (C185S) could partly rescue the degradation of NLRP3, indicating that the enzyme activity of USP22 is required to regulate the stability of NLRP3 (Figure 5(j)). These results suggested that USP22 functionally promotes NLRP3 degradation via lysosomal degradation.

USP22 interacts with and stabilizes ATG5

It has been reported that autophagy-related proteins, such as SQSTM1/p62, ATG5, and ATG7, play important roles in regulating NLRP3 inflammasome activation [23,29,30]. We found that the expression of ATG5 and SQSTM1/p62 was decreased in Usp22-silencing mouse peritoneal macrophages, while the expression of ATG7 was not affected (Figure 6(a)). In addition, coimmunoprecipitation analysis showed that ATG5 interacted with USP22 and NLRP3 in mouse peritoneal macrophages (Figure 6(b)). In contrast, there was no relationship between SQSTM1/p62 and either USP22 or NLRP3 (Fig. S6A and B). Consistent with these observations, ATG5 also co-localized with USP22 and NLRP3 in mouse peritoneal macrophages (Figure 6(c) and Fig. S6C). Silencing of Atg5 or Usp22 decreased the level of the protein LC3-II, which suggested that ATG5 and USP22 promote autophagy in mouse peritoneal macrophages (Fig. S7A and B). Autophagy was consistently elevated by the overexpression of USP22 (Fig. S7C and D). A coimmunoprecipitation assay indicated that there was an interaction between USP22 and ATG5 (Figure 6(d)). To better understand the role of ATG5 in the process of NLRP3 inflammasome activation, we used specific small interfering RNA (siRNA) to silence the expression of Atg5 in mouse peritoneal macrophages; we found that the expression of PYCARD, pro-CASP1, and pro-IL1B were unaffected by Atg5 silencing (Figure 6(e)). However, the expression of NLRP3 was elevated in Atg5-silenced macrophages, while cleaved IL1B and CASP1 in the supernatant were increased in Atg5-silenced macrophages (Figure 6(e)). Furthermore, IL1B secretion was significantly increased in Atg5-silenced macrophages (Figure 6(f)). Overexpression of Atg5 also promoted the degradation of NLRP3 (Figure 6(g)). These results suggested that USP22 interacts with and stabilizes ATG5, which may account for its inhibitory function on NLRP3 inflammasome activation.

Figure 6.

USP22 interacts with and stabilizes ATG5. (a) Immunoblot analysis of the expression of ATG5, ATG7, and SQSTM1/p62 in the cell lysates of Usp22-silenced mouse peritoneal macrophages with or without LPS stimulation. (b) Co-IP analysis of the interaction between NLRP3 and ATG5 in mouse peritoneal macrophages following immunoprecipitation with anti-NLRP3. (c) Immunofluorescence staining for USP22 (green) and ATG5 (red) in mouse peritoneal macrophages. Scale bars represent 10 μm. (d) 293 T cells were transfected with Flag-ATG5 and MYC-USP22, immunoprecipitated with anti-Flag, and then immunoblotted using the indicated antibodies. (e) Immunoblots of IL1B, pro-CASP1, and cleaved CASP1 in the supernatants or cell lysates of Atg5-silenced mouse primary macrophages, primed with LPS, and then stimulated with nigericin (Nig.). (F) IL1B in SNs from macrophages silenced of Atg5 were analyzed by ELISA. (g) 293 T cells were transfected with HA-NLRP3 and Flag-ATG5, then immunoblotted with the indicated antibodies. Data shown are representative of three separate experiments (A-E, G) or are means ± SEM of three independent experiments (F). ** p < 0.01.

USP22 promotes NLRP3 degradation through ATG5-dependent autophagy

To further investigate the function of ATG5 in USP22-mediated NLRP3 lysosomal degradation, we co-transfected 293 T cells with MYC-USP22 and various concentrations of ATG5. The results showed that USP22 promoted the stability of ATG5 in a dose-dependent manner (Figure 7(a)). We next found that the proteasome inhibitor MG132 stabilized ATG5 and that the addition of USP22 did not further stabilize ATG5. However, we did not observe such effects upon treatment with the autolysosome inhibitor CQ (Figure 7(b)). Using cotreatment with a protein synthesis inhibitor, CHX, we further demonstrated that USP22 stabilized ATG5 (Figure 7(c)). As USP22 is a DUB and stabilizes ATG5 via proteasome-mediated degradation, we next investigated whether USP22 participates in the ubiquitination of ATG5. We found that ATG5 ubiquitination was markedly decreased in 293 T cells overexpressing USP22 (Figure 7(d)). As different types of polyubiquitin linkages mediate distinct biological functions, we expressed Flag-tagged ATG5 in 293 T cells together with various ubiquitin mutants (K6, K11, K27, K29, K33, K48, and K63), all of which contained just one indicated lysine available for poly-linkage. Notably, overexpressing USP22 mainly decreased K27- and K48-linked ubiquitination of ATG5 (Figure 7(e)). We compared the protein sequence of ATG5 in various sepsis samples and found five conserved lysine residues (K51, K118, K129, K136, and K171) in ATG5 (Figure 7(f)). We generated mutants in which lysine was replaced with arginine. The K118R substitution rather than other point mutations abolished USP22-mediated deubiquitination of ATG5 (Figure 7(g)). Importantly, the knockout of Atg5 could partly rescue the degradation of NLRP3 mediated by USP22 (Figure 7(h)), which indicated that USP22 promotes the degradation of NLRP3 via ATG5-mediated autophagy and that K118 of ATG5 is an important site for the ubiquitination and the stability of ATG5.

Figure 7.

USP22 promotes the degradation of NLRP3 via ATG5-mediated autophagy. (a) 293 T cells were co-transfected with expression plasmids for Flag-ATG5, various concentrations of MYC-USP22; then, immunoblot analysis of the expression of Flag-ATG5 and MYC-USP22 in 293 T cells was performed. (b) 293 T cells were co-transfected with Flag-ATG5 and MYC-USP22 for 24 h, incubated with the lysosome inhibitor CQ or the proteasome inhibitor MG132 for 4 h, then immunoblot analyses using the indicated antibodies were performed. (c) 293 T cells were co-transfected with Flag-ATG5 and MYC-USP22 and incubated with CHX for the indicated times. MYC-tag and Flag-tag antibodies were used to detect the expression of relevant proteins. (d) Immunoblot analysis of lysates from 293 T cells transfected with HA-tagged ubiquitin (HA-Ub) and Flag-ATG5, with or without MYC-USP22, followed by IP with anti-Flag M2 beads, then probed with anti-HA tag. (e) Immunoblot analysis of lysates from 293 T cells transfected with HA-tagged ubiquitin (HA-Ub), HA-tagged K6-linked ubiquitin (HA-K6-Ub), HA-tagged K11-linked ubiquitin (HA-K11-Ub), HA-tagged K27-linked ubiquitin (HA-K27-Ub), HA-tagged K29-linked ubiquitin (HA-K29-Ub), HA-tagged K33-linked ubiquitin (HA-K33-Ub), HA-tagged K48-linked ubiquitin (HA-K48-Ub), or HA-tagged K63-linked ubiquitin (HA-K63-Ub) and Flag-ATG5, with or without MYC-USP22, followed by IP with anti-Flag M2 beads, then probed with anti-HA. (f) Conserved lysine residues of ATG5 in the indicated species. (g) Immunoblot analysis of lysates from 293 T cells transfected with HA-tagged ubiquitin, MYC-USP22, Flag-ATG5 or indicated mutant Flag-ATG5, followed by IP with anti-Flag M2 beads, then probed with anti-HA. (h) Atg5 knockout or control 293 T cells were co-transfected with HA-NLRP3 and Flag-USP22, then immunoblotted with the indicated antibodies. Data shown are representative of three separate experiments. (i) A model showing how USP22 negatively regulates NLRP3 inflammasome via ATG5-mediated autophagy.

Discussion

NLRP3 inflammasome overactivation has been reported to play an important role in the molecular mechanism of a diverse range of diseases, including autoinflammatory syndromes and metabolic diseases [1,6]. In our study, we found that USP22 deficiency or silencing enhanced the expression of NLRP3, activation of CASP1, and release of IL1B and IL18 in both mouse macrophages and human THP-1 cells. However, USP22 deficiency had no effect on either NLRC4 or AIM2 inflammasomes, indicating that USP22 may specifically regulate the activation of the NLRP3 inflammasome. Moreover, in our animal model, mice lacking USP22 exhibited an increased inflammatory response in animals with alum-induced peritoneal inflammation and LPS-induced systemic inflammation, as indicated by increased IL1B and IL18 secretion and greater recruitment of immune cells. Mechanistically, we found that USP22 interacts with NLRP3 and promotes NLRP3 degradation in an ATG5-dependent autophagic pathway (Figure 7(i)). Therefore, our results revealed a role for USP22 in regulating NLRP3 inflammasome activation and thus suggest a potential therapeutic target for the treatment of NLRP3 inflammasome-related diseases.

USP22 has long been associated with poor prognosis of cancer and neurological disorders; however, little is known about the role of USP22 in inflammation-related diseases [31–33]. Recently, a study showed that USP22 promotes IRF3 nuclear translocation and antiviral responses by deubiquitinating the importin protein KPNA2, indicating a previously unknown function of USP22 in host antiviral responses [21]. While USP22 is localized in both the cytoplasm and the nucleus, we found that USP22 could be transferred from the nucleus to the cytoplasm following LPS stimulation (data not shown). USP22 increased the protein levels of NLRP3 but did not affect the mRNA levels of NLRP3, suggesting an essential role of cytoplasmic USP22 in suppressing NLRP3 inflammation activation.

Autophagy is a metabolic pathway that plays an important role in maintaining the dynamic balance of cells [34]. Many studies have shown that autophagy negatively regulates the activation of the NLRP3 inflammasome. The mechanism of autophagy inhibition of the NLRP3 inflammasome may related to the reduction of PYCARD [35], NLRP3 [36], and the clearance of mitochondrial reactive oxygen species (ROS) [22]. Our results also revealed that USP22 promotes the degradation of NLRP3 via ATG5-mediated autophagy. USP22 deubiquitinates and stabilizes ATG5, which promotes NLRP3 lysosomal degradation. However, further investigations are necessary to understand the precise mechanism and the role ATG5 plays in the regulation of the NLRP3 inflammasome.

In conclusion, our findings demonstrate that USP22 promotes NLRP3 degradation via ATG5-mediated autophagy and acts as a negative regulator in NLRP3 inflammasome assembly and activation. Thus, our findings also suggest USP22 may be a potential target for the treatment of NLRP3 inflammasome-driven inflammatory diseases.

Materials and methods

Mice

C57BL/6 j mice (8-weeks old, female) were purchased from Vital River Company (219). The mice were housed in a specific pathogen-free sterile facility under controlled conditions (12 h light/dark cycle; 25 ± 2°C). The procedures for the animal experiments strictly conformed to the Chinese National Institute of Health Guide for Care and Use of Laboratory Animals, and approval was obtained from the Animal Care Committee of Shenzhen University, Shenzhen. Cre-Esr Usp22^fl/+ mice were kindly provided by Dr. Bo Zhong (Wuhan University, Wuhan, China)²¹. To achieve conditional knockout of Usp22, 8-week-old Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl mice were intraperitoneally injected with tamoxifen (80 mg/kg, dissolved in corn oil) once per day for 5 days; 7 days later, the mice were challenged with LPS or alum.

Reagents and antibodies

LPS, tamoxifen, and 4-OHT were purchased from Sigma-Aldrich (L2630, T5648, H6278). Phorbol 12-myristate 13-acetate (PMA), nigericin, and ATP were purchased from InvivoGen (tlrl-pma, tlrl-nig, tlrl-atpl). Alum was obtained from ThermoFisher Scientific (77,161).

Anti-IL1B antibody was purchased from Cell Signaling Technology (12,703). Anti-USP22 antibody was purchased from Abcam (ab195289). Anti-NLRP3, anti-CASP1, and anti-PYCARD/ASC antibodies were purchased from AdipoGen Life Sciences (AG-20B-0006, AG-20B-0042/AG-20B-0044, AG-25B-0006). Anti-HA-tag, anti-MYC-tag, and anti-ACTB antibody were purchased from Proteintech (51,064-2-AP, 16,286-1-AP, 66,009-1-Ig). Anti-DDDK-tag was purchased from MBL (M185-3).

Cell culture

To obtain mouse peritoneal macrophages, mice were intraperitoneally injected with 3% thioglycolate (Millipore, 70,157). Four days later, peritoneal exudate macrophages were harvested and seeded on plates. The next day, the cells were replaced to fresh medium and the adherent monolayer cells were used as mouse peritoneal macrophages. Bone marrow cells were isolated from mouse femurs. The bone marrow cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum (Gibco, 10,099,141), 1% streptomycin and penicillin (Gibco, 15,140,122), and CSF2/M-CSF (10 ng/mL; Peprotech, AF-300-25) for bone marrow-derived macrophage (BMDM) differentiation. Primary cells for Cre-Esr Usp22^fl/+ and Cre-Esr usp22^fl/fl were treated with 4-OHT for 2 days, allowed to recover for 1 day, then stimulated with LPS. THP-1 and 293 T cells were purchased from the American Type Culture Collection (CRL-3216, TIB-202). THP-1 macrophages were differentiated by treatment with 100 nM PMA for 24 h.

RNA interference assay

The siRNAs were synthesized and transfected with jetPRIME (PolyPlus Transferion, 101,000,046), according to the manufacturer’s instructions. The siRNA oligos were synthesized by GenePharma (Shanghai, China). The siRNA sequences used were as follows: Control-siRNA, 5′-UUCUCCGAACGUGUCACGUTT-3′; Mm-Usp22-siRNA, 5′-CCUCUACUGUGUCUUCUUUTT-3′; Hs-USP22-siRNA, 5′-GCUGUUUCACAAAGAAGCAUAUUCA-3′; Mm-Atg5-siRNA, 5′-ACCGGAAACTCATGGAATA-3′; Hs-ATG5-siRNA, 5′-CAUCUGAGCUACCCGGAUA-3′; Mm-Atg7-siRNA, 5′-GCAUCAUCUUCGAAGUGAATT-3′.

Enzyme-linked immunosorbent assay

Supernatants from cell cultures and sera were collected and the concentration of IL1B (ThermoFisher Scientific, KAC1211 and BMS6002), TNF (ThermoFisher Scientific, BMS607-3 and BMS223-4), and IL18 (MultiSciences (LiankeBio), 70-EK218 and 70-EK118) were determined by ELISA according to the manufacturer’s instructions.

Preparation of cytoplasm and nuclear proteins

The experiments were performed as previously described [21].

Immunoprecipitation and immunoblot analysis

The cells were lysed in cell lysis buffer (Cell Signaling Technology, 9803) containing 2 mM Tris-HCl, pH 7.5, 15 mM NaCl, 0.1 mM Na₂EDTA, 0.1 mM EGTA, 0.1% Triton X-100, 0.25 mM TSP, 0.1 mM β-glycerophosphate, 0.1 mM Na₃VO₄, and 1 μg/mL leupeptin, then mixed with 1% protease and phosphatase inhibitor cocktail (Bimake, B14001). Cell lysates were subjected to SDS-PAGE, and immunoblot analysis was performed with the indicated antibodies. For immunoprecipitation assays, cells were lysed in Nonidet P-40 lysis buffer (Beyotime Biotechnology, P0013 F) with 1% protease and phosphatase inhibitor cocktail (Bimake, B14001). The lysates were immunoprecipitated with IgG (Beyotime Biotechnology, A7028) or the indicated antibodies, and the precipitates were washed five times with Nonidet P-40 lysis buffer and detected by immunoblot analysis with indicated antibodies.

In vivo siRNA transfection

In vivo siRNA transfection was carried out using in vivo jetPEI (PolyPlus Transferion, 101,000,030), according to the manufacturer’s instructions. Briefly, mice were intraperitoneally injected with transfection mixture consisting of Usp22-siRNA or control-siRNA dissolved in in vivo jetPEI with an N:P ratio of 7. The interfering efficacy was determined by quantitative PCR (qPCR) and western blotting.

Immunofluorescence labeling and confocal microscopy

Immunostaining was performed as previously described [37]. Briefly, the cells were washed three times with PBS (Solarbio Life Science, P1020), fixed with 4% paraformaldehyde in PBS for 10 min, and then permeabilized with 0.2% Triton X-100 in PBS for 15 min. Coverslips with adherent cells were blocked with 1% FBS in PBS, incubated with the appropriate primary antibodies (4°C, overnight), and subsequently visualized by Alexa Fluor 488 donkey anti-rabbit or Alexa Fluor 555 donkey anti-mouse (ThermoFisher Scientific, A-21206 and A-31570). The stained cells were visualized under a Zeiss LSM 880 microscope using a 63× objective.

In vivo LPS challenge

C57BL6j mice were intraperitoneally injected with LPS (20 mg/kg) and monitored for the following 48 h for lethality. For the regular endotoxic shock model, the mice were intraperitoneally injected with LPS (10 mg/kg). Four hours following LPS administration, the mice were euthanized, serum was collected, and the lung samples and abdominal cavity were washed with PBS. Peritoneal lavage fluid was collected and analyzed by flow cytometry. The levels of IL1B, IL18, and TNF in serum and peritoneal lavage fluid were determined by ELISA. The lung samples were fixed and sectioned for hematoxylin and eosin (H&E) staining.

In vivo peritonitis

For the mouse peritonitis models, C57BL6j mice were intraperitoneally injected with 1 mg alum (ThermoFisher Scientific, 77,161) dissolved in 0.2 mL sterile PBS; 6 h later, the mice were euthanized. Peritoneal lavage fluid was collected and analyzed by flow cytometry. The levels of IL1B and TNF were determined by ELISA.

Statistical analysis

All experiments were independently performed at least three times in triplicate. The results were expressed as the mean ± SD or mean ± SEM. Comparisons between two groups were analyzed using the Student’s t-test. Mouse survival was analyzed using the Kaplan–Meier method. All statistical analyses were performed using GraphPad Prism 6. p-value <0.05 was considered to be a statistically significant difference.

Funding Statement

This work was supported by the National Natural Science Foundation of China (No. 32000668, U1801283, 31870908, 32100700), the Guangdong Province Basic and Applied Basic Research Fund (No. 2019A1515110146), the Guangdong Provincial Science and Technology Program (No. 2019B030301009) and SZU Top Ranking Project (No. 86000000210) to Weilin Chen; Guangdong Science and Technology Department; Basic and Applied Basic Research Foundation of Guangdong Province; Shenzhen University.

Disclosure statement

All the authors declare that there are no conflicts of interest.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2022.2107314