Introduction

Vangl2, a core PCP component, mediates Wingless-type (Wnt)/PCP signaling, and controls homeostasis, development, and repair of organs (Bailly, Walton, & Borg, 2018; Brunt et al., 2021; Hatakeyama, Wald, Printsev, Ho, & Carraway, 2014). Vangl2 has four transmembrane domains with both carboxyl termini and amino oriented toward the cytoplasm, which is phosphorylated in the endoplasmic reticulum (ER) and then transported to the cell surface and becomes stabilized, while the unphosphorylated Vangl2 is unstable and internalized to degrade via the lysosomal pathway (D. Feng et al., 2021). The shuttle of Vangl2 between cytoplasm and cell membrane results in its multifunction, including adhesion, membrane protrusive activity, migration, and bridging proteins (Hatakeyama et al., 2014). Indeed, Vangl2 inhibited matrix metalloproteinase 2 (MMP2) activity and affected cell adhesion to extracellular matrix proteins (Jessen & Jessen, 2017). Vangl2 also modulates glomerular injury by promoting MMP9 (Papakrivopoulou et al., 2018). Moreover, the abnormal function of Vangl2 results in various diseases, such as cancer, kidney glomerular injury, idiopathic pulmonary fibrosis and systemic dysplasia (Papakrivopoulou et al., 2018; Poobalasingam et al., 2017). Vangl2 is markedly down-regulated in patients with emphysema (Poobalasingam et al., 2017), while which level was up-regulated and amplified in breast, ovarian, and uterine carcinomas (Cancer Genome Atlas Research et al., 2013; Cerami et al., 2012; Gao et al., 2013). In addition, Vangl2 is bounded to p62 to promote breast cancer (Puvirajesinghe et al., 2016). Our previous study showed that Vangl2 prevents osteogenic differentiation in mesenchymal stem cells, resulting in osteogenic dysplasia (Gong et al., 2021). Vangl2 mediated downstream factors of Toll-like or interleukin (IL)-1 receptor, such as myeloid differentiation factor 88 (MyD88) (Gong et al., 2021), suggesting that Vangl2 may play roles in immune-related diseases, including autoimmune diseases. However, the function of Vangl2 in inflammatory diseases remains uncovered.

The NF-κB signaling is critical for the pathogenesis of a number of inflammatory diseases, such as inflammatory bowel disease (IBD), rheumatoid arthritis (RA), systemic lupus erythematosus (SLE) and sepsis (Liu, Zhang, Joo, & Sun, 2017). NF-κB activation relies on the pattern recognition receptors (PRRs) recognition of pathogen-associated molecular patterns (PAMPs), including Toll like receptor (TLR), NOD-like receptors (NLRs) and RIG-I-like receptors (RLRs) (Liu et al., 2017). LPS triggers TLR activation and the recruitment of adaptor proteins including MyD88. This in turn activates a series of downstream canonical NF-κB signaling cascade, resulting in the phosphorylation and degradation of IκB and the nuclear translocation of RelA/p65 and p50 to induce the transcription of several inflammatory cytokines, such as interleukin (IL)-1, IL-6, and tumor necrosis factor-α (TNF-α) (Funes, Rios, Escobar-Vera, & Kalergis, 2018; Locati, Curtale, & Mantovani, 2020). Since uncontrolled immune responses are detrimental to the host, inappropriate or excessive NF-κB activity contributes to the pathogenesis of various inflammatory diseases and cancer (Cartwright, Perkins, & C, 2016). Thus NF-κB signaling must be tightly regulated to maintain immune balance in the organism. In recent decades, extensive studies have focused on the mechanisms underlying the regulation of NF-κB signaling. Recent research demonstrated that NLRC5 strongly prevents NF-κB signaling pathway by interacting with IκB kinase (IKK) α/IKKβ and blocking their phosphorylation (Cui et al., 2010). COMMD1, PPARγ, SOCS1 and GCN5 were also shown to negatively regulate NF-κB signaling (Bartuzi, Hofker, & van de Sluis, 2013; Mao et al., 2009). Meanwhile, tripartite motif-containing protein 21 (Trim21) enhanced the interaction of p65 with IKK, which promotes p65 phosphorylation and downstream gene activation (Yang et al., 2021). However, the molecular mechanisms underpinning the regulation of NF-κB signaling are still elusive.

Autophagy, a conserved intracellular degradation pathway, decomposes cytoplasmic organelles and components, and acts as a defense mechanism to pathogen infection, playing a crucial role in nutrient recycling, stress response and cellular homeostasis (Ashrafi & Schwarz, 2013; Denton & Kumar, 2019). Recent evidence indicates that autophagy is highly selective when delivering specific substrates to autolysosomal degradation by virtue of a number of cargo receptors, including sequestosome 1 (SQSTM1/p62), optineurin (OPTN), nuclear dot protein 52 (NDP52/CALCOCO2) and neighbor of BRCA1 (NBR1) (Gatica, Lahiri, & Klionsky, 2018). Selective autophagy targets immune regulators for degradation, thus suppressing innate immune signaling, such as type I IFN and NF-κB signaling (Pradel, Robert-Hebmann, & Espert, 2020; Tong et al., 2012). Moreover, p62 protein has been identified as a novel Vangl2-binding partner (Puvirajesinghe et al., 2016), and our recent study has demonstrated that Vangl2 reduces lysosomal chaperone-mediated autophagy (CMA) activity by targeting LAMP-2A for degradation (Gong et al., 2021). However, whether and how Vangle2 is involved in the selective autophagic regulation of NF-κB signaling remains largely unknown.

In this study, we uncover a previously unrecognized role of Vangl2, as a “molecular brake”, in the negative regulation of NF-κB signaling to prevent excessive and potentially harmful immune responses during sepsis in both human patient samples and LPS induced mouse model. Induction of Vangl2 upon inflammation recruits an E3 ubiquitin ligase PDLIM2 to catalyze K63-linked ubiquitination on p65, thus promoting the recognition of p65 by the cargo receptor NDP52 and resulting in the selective autophagic degradation of p65. Our findings provide a potential target for the treatment of inflammatory diseases.

Results

Loss of Vangl2 promotes inflammatory responses in LPS-induced septic shock

To investigate the potential role of Vangl2 in inflammatory response, we first analyzed the expression of Vangl2 in peripheral blood mononuclear cells (PBMCs) from sepsis patients (Fig. S1 A) and found mRNA level of Vangl2 was increased in the sepsis patients compared to healthy control (HC) (Fig. 1 A). Moreover, the serum of the high Vangl2 expression group had lower levels of IL-6, white blood cell (WBC), and acute C- reactive protein (CRP) than the low Vangl2 expression group (Fig. S1 A). Expression of Vangl2 from database GSE156382 also confirmed that Vangl2 mRNA was induced during sepsis (Fig. S1 B). To further determine whether Vangl2 expression could be regulated in response to inflammatory stimulation, we treated mice with LPS to activate NF-κB pathway and detected Vangl2 mRNA in different tissues from LPS treated mice, and found the expression of Vangl2 significantly increased in secondary lymphoid organs including the spleen and lymph nodes after LPS stimulation, but not in other tissues (Fig. 1 B). Moreover, qPCR and western blot analyses revealed strong upregulation of Vangl2 at both the mRNA and protein levels after LPS stimulation in bone marrow (BM)-derived macrophages (BMDMs), neutrophils and peritoneal macrophages (pMACs) (Fig. S1, C-E), suggesting that Vangl2 is induced in immune organ tissues and cells in response to inflammation.

Figure 1.

To determine the function of Vangl2 during LPS-induced sepsis, we specifically ablated Vangl2 in myeloid cells by crossing Vangl2^flox/floxmice with mice expressing lysozyme proximal promoter (Lyz2-Cre). The resultant homozygous Vangl2^flox/flox × Lyz2- Cre mice were designated Vangl2^ΔM mice and selective deletion of Vangl2 in myeloid cells of Vangl2^ΔM mice were confirmed by PCR (Fig. S1 F). Although there were no significant differences in the size of spleen and lymph nodes between wild-type (WT) and Vangl2^ΔM mice (Fig. S1 G), myeloid-specific loss of Vangl2 increased the number of monocytes, macrophages and neutrophils in the spleen and bone marrow (Fig. S1, H and I). To gain further insight into the physiological function of Vangl2, we treated WT and Vangl2^ΔM mice with a high dose of LPS and monitored mice survival. We found that all Vangl2^ΔM mice died after LPS-induced septic shock within 20 hours, compared to only 20% of the WT mice, whereas the remaining WT mice survived for more than 50 hours (Fig. 1 C). Consistent with this observation, Vangl2^ΔM mice markedly increased the protein level of IL-1β in isolated CD11b⁺ splenocytes (Fig. 1 D), and mRNA levels of Il1b, Tnfa and Il6 in the spleen (Fig. 1 E). Meanwhile, Vangl2^ΔM mice showed markedly elevated serum amounts of proinflammatory cytokines such as IL-1β, TNF-α and IL-6 after LPS treatment, compared with WT mice (Fig. 1 F). Together, these data provide in vivo evidence that myeloid cell-specific deletion of Vangl2 in mice enhances the sensitivity and severity of LPS-induced septic shock and is associated with increased expression of proinflammatory cytokines.

Vangl2 negatively regulates NF-κB activation and inflammation in myeloid cells

To investigate the mechanisms by which Vangl2 prevents sepsis, we performed RNA- seq analysis to identify signal pathways involved in LPS-induced septic shock by comparing LPS-stimulated BMDMs from Vangl2^ΔM and WT mice. Results showed 907 up-regulated genes (including inflammatory cytokines) and 1092 down-regulated genes in response to Vangl2 deficiency in BMDMs after LPS stimulation (Fig. S2 A). Consistent with Fig. 1, gene ontology (GO) analysis further identified that these genes are involved in cellular immune responses, including “cellular response to lipopolysaccharide” (Fig. S2 B). Moreover, KEGG analysis revealed genes involved in TNF signaling pathway and cytokine-cytokine receptor interaction were highly enriched in Vangl2^ΔM BMDMs after LPS stimulation, suggesting that Vangl2 may regulate these signaling pathways and related cytokines release (Fig. S2 C).

To determine the function of Vangl2 in innate immune signaling in myeloid cells, we isolated BMDMs, neutrophils and pMACs from WT and Vangl2^ΔM mice, treated them with LPS, and performed immunoblot analysis with specific antibodies. We found that phosphorylations of p65 were enhanced and remained high level for a sustained period in Vangl2-deficient pMACs and neutrophils, compared with WT cells (Fig. 2, A-D). Consistent with this observation, Vangl2-deficient macrophages, and neutrophils showed markedly elevated pro-inflammatory cytokines such as TNF-α, IL-6 (Fig. 2 E and F and Fig. S2 D and E) and IL-1β (Fig. S2 F), after LPS treatment. Moreover, we detected enhanced p65 nuclear accumulation in LPS-induced Vangl2-deficient neutrophils, compared with WT neutrophils (Fig. 2, G and H).

Figure 2.

To further confirm the function of Vangl2 in regulating NF-κB signaling, we overexpressed Flag-tagged Vangl2 in A549 cells. We found that Vangl2 overexpression inhibited phosphorylation of p65 in A549 cells (Fig. 2, I and J). These results suggest that Vangl2 prevents LPS-induced NF-κB activation and proinflammatory cytokine production.

Vangl2 inhibits LPS-induced NF-κB activation by interacting with p65

To clarify the regulatory mechanism of Vangl2, we transfected Chinese hamster ovary (CHO) or 293T cells with NF-κB luciferase reporter vector, with or without the Vangl2 plasmid, then treated the cells with LPS, IL-1β, or TNF-α. We found that Vangl2 markedly inhibited NF-κB activation induced by LPS, IL-1β, or TNF-α in a dose- dependent manner (Fig. 3, A-C). Next, we sought to determine potential signaling molecules that mediated the NF-κB-luc reporter. NF-κB-luc activity was strongly activated by overexpression of MyD88, IRAK1, TRAF6, IKKα, IKKβ or p65, but all of these activities were inhibited when Vangl2 was co-transfected at increasing concentrations (Fig. 3 D and Fig. S3, A-F), suggesting that Vangl2 may block NF-κB activation at the very downstream signaling level of p65.

Figure 3.

To test this prediction, we transfected 293T cells with HA-tagged Vangl2 together with Flag-tagged IKKα, IKKβ, p65, TRAF6, IRAK1 or MyD88. Co-immunoprecipitation (co-IP) assay revealed that Vangl2 interacted strongly with IKKα, IKKβ, p65 and MyD88 (Fig. 3 E and Fig. S3 G, H). In addition, endogenous co-IP immunoblot analyses showed that Vangl2 was strongly associated with p65 upon LPS stimulation in BMDMs, but not with IKKα, IKKβ or MyD88 (Fig. 3 F and Fig. S3 I). Moreover, ZDOCK server predicted that Vangl2 may potentially interact with p65 by a hydrogen bond (Fig. S3 J). We further investigated the co-localization of Vangl2 and p65 by confocal microscopy and found a weak-co-localization of Vangl2 with p65 in unstimulated cells and the co-localization between Vangl2 and p65 was notably enhanced upon LPS stimulation (Fig. 3 G and Fig. S3 K). To determine how Vangl2 interacts with cytoplasmic p65, we isolated cytoplasm and membrane in LPS treated THP-1 cells. We found that Vangl2 interacted with p65 mainly in the cytoplasm, although most of Vangl2 located on the membrane (Fig. S3 L). Together, these data suggest that Vangl2 may interact with p65 in the cytoplasm to inhibit NF-κB signaling.

Vangl2 comprises an N-terminal cytoplasmic tail (NT), a transmembrane (TM) domain, a Prickle-binding domain (PkBD), and a C-terminal cytoplasmic tail (CT) (Nagaoka, Furuse, Ohtsuka, Tsuchida, & Kishi, 2019). To map the essential domains of Vangl2 that mediate its association with p65, we generated several deletion constructs of Vangl2. We found Vangl2 FL (full-length), T2 (ΔNT+TM), T3 (ΔCT), and T4 (ΔNT) interacted with the full-length p65, while Vangl2 T1 (ΔPkBD+CT) abrogated their association (Fig. 3 H and Fig. S3 M), indicating that the PkBD domain is important for the Vangl2-p65 interaction. Additionally, we constructed HA-tagged Vangl2 PkBD plasmid and co-IP assay revealed that Vangl2 FL and PkBD interacted with p65 (Fig. 3 I). Moreover, we observed that deletion of the PkBD domain (T1), but not other domains, of Vangl2 abolished the Vangl2-mediated inhibition of NF-κB-luc activity, and transfection of Vangl2 PkBD domain achieved similar inhibition on p65 induced NF-κB activation as Vangl2 FL did (Fig. 3 J). Collectively, these data suggest that Vangl2 suppresses NF-κB signaling by targeting p65 through Vangl2 PkBD domain interaction.

Vangl2 promotes the autophagic degradation of p65

Next we sought to study the physiological role of Vangl2-p65 interaction in the regulation of NF-κB signaling by transfecting 293T cells with Flag-tagged p65, together with increasing doses of Vangl2, and found that Vangl2 dramatically decreased the protein level of p65 in a dose-dependent manner (Fig. 4 A). To exclude the possibility that the downregulation of p65 protein by Vangl2 was caused by the inhibition of p65 transcription, qPCR results suggested that the abundance of p65 mRNA did not change in cells with increased expression of Vangl2 (Fig. 4 B). Moreover, Vangl2 facilitated the degradation of the p65 phosphorylation mutants (Fig S4A), and Vangl2 phosphorylation mutants were observed to degrade p65 (Fig S4B), indicating that Vangl2 mediated p65 degradation is independent of its phosphorylation status. Since activated p65 translocates to the nucleus, we next assessed whether Vangl2 regulates the degradation of p65 in the cytoplasm or nucleus, and found that the Vangl2 interacted with p65 mainly in the cytoplasm (Fig. S4 C) and mediated the degradation of p65 in the cytoplasmic fraction (Fig. 4 C), which is consistent with the result that Vangl2 and p65 co-localized in the cytoplasm (Fig. 3 G). To further assess whether Vangl2 regulates the degradation of endogenous p65, we found that Vangl2-deficient BMDMs stabilized the expression of endogenous p65 after LPS treatment (Fig. 4 D). After 12 h of LPS stimulation, the cycloheximide (CHX)-chase assay result showed that the degradation rate of p65 in Vangl2^ΔM BMDMs was slower than that of WT cells (Fig. S4 D). Together, these data suggest that Vangl2 promotes p65 protein degradation.

Figure 4.

To investigate whether Vangl2 degrades p65 through an autolysosome or proteasome pathway, 293T cells were transfected with p65, together with or without the Vangl2 plasmids, and treated with different pharmacological inhibitors. We found the degradation of p65 induced by Vangl2 was blocked by autolysosome inhibitor chloroquine (CQ) and bafilomycin A1 (Baf A1), or autophagy inhibitor 3-methyladenine (3-MA) (Fig. 4 E), but not by the proteasome inhibitor MG132 or caspase-1 inhibitor Z- VAD and VX-765 (Fig. 4 E and Fig. S4 E). Furthermore, Vangl2 significantly increased the degradation of p65 during rapamycin-triggered autophagy (Fig. 4 F), which suggested that Vangl2 promoted autophagic degradation of p65.

To further demonstrate Vangl2 mediated p65 degradation through autophagy, we transfected WT, ATG5, and Beclin1 knockout (KO) 293T cells with Vangl2, and found that the degradation of p65 triggered by Vangl2 was almost abrogated in ATG5 and Beclin1 KO cells (Fig. 4 G). The p65 turnover rates were markedly reduced in ATG5 and Beclin1 KO cells post CHX treatment (Fig. S4, F and G), suggesting that the impaired autophagy prevented p65 degradation. Consistently, NF-κB activation induced by p65 was rescued in ATG5 and Beclin1 KO cells expressing Vangl2, compared to WT 293T cells (Fig. 4, H and I). In addition, the colocalization of p65 and LC3B was enhanced following LPS stimulation (Fig. S4, H). Together, our data suggest that Vangl2 specifically promotes p65 degradation through autophagy.

Vangl2 promotes the recognition of p65 by cargo receptor NDP52

Accumulating evidence showed that cargo receptors play crucial roles in selective autophagic degradation by delivering substrates (X. He et al., 2019; Kirkin & Rogov, 2019; Wu et al., 2021). Since there is no research suggests that Vangl2 is a cargo protein, we hypothesized that Vangl2 might bridge p65 to the cargo receptors for autophagic degradation. To identify the potential cargo receptor responsible for Vangl2- mediated autophagic degradation of p65, we co-transfected 293T cells with Vangl2 and various cargo receptors, followed by Co-IP assay. Result suggested that Vangl2 strongly interacted with the cargo receptors p62 and NDP52, and slightly associated with NBR1 and Nix (Fig. 5 A). However, p65 only interacted with the cargo receptors p62 and NDP52 (Fig. 5 B), which is consistent with a recent finding that p62 protein is a Vangl2- binding partner (Puvirajesinghe et al., 2016). We next attempted to clarify whether p62 or NDP52 is involved in Vangl2-mediated autophagic degradation of p65. Interestingly, we found that Vangl2 promoted the association between p65 and NDP52, but did not affect p65-p62 complex (Fig. 5 C). Consistently, we found that Vangl2 mediated degradation of p65 was rescued in NDP52 KO cells, but not in p62 KO cells (Fig. 5 D). Endogenous co-IP immunoblot analyses also revealed that deficiency of Vangl2 remarkably attenuated the association of endogenous p65 and NDP52 (Fig. 5 E). Likewise, NDP52, but not p62, enhanced the association between p65 and Vangl2 (Fig. 5 F and Fig. S5 A). And Vangl2 failed to inhibit the activation of NF-κB signaling in NDP52 KO cells (Fig. 5 G), but not in p62 KO cells (Fig. S5 B). Furthermore, compared with WT cells, CHX-chase assay results showed that the degradation rates of p65 were reduced in NDP52 KO cells (Fig. 5 H), but not in p62 KO cells (Fig. S5 C). Taken together, these data suggest that Vangl2 mediates the NDP52-directed selective autophagic degradation of p65.

Figure 5.

Vangl2 increases the K63-linked poly-ubiquitination of p65

It has been well documented that ubiquitin chains attached to the substrates and served as a signal for the recognition by cargo receptors (Otten et al., 2021; Shaid, Brandts, Serve, & Dikic, 2013; Yin, Popelka, Lei, Yang, & Klionsky, 2020). The ubiquitin- associated (UBA) domain of NDP52 mostly recognized ubiquitinated substrates for degradation through autophagy (Johansen & Lamark, 2014; Yamano & Youle, 2020). We hypothesized that Vangl2 may affect the ubiquitin chains on p65 for subsequent NDP52-dependent degradation. To demonstrate this, we performed endogenous co-IP assay and found that the poly-ubiquitination of p65 was impaired in BMDMs from Vangl2^ΔM mice after LPS stimulation, compared to WT mice (Fig. 6 A). Moreover, Vangl2 specifically increased the total and K63-linked (K63-only ubiquitin mutant) poly- ubiquitination of p65 (Fig. 6 B). In contrast, Vangl2 did not affect other ubiquitin linkages (K11, K27, K33, or K48-only ubiquitin mutant) of p65 in an overexpression system (Fig. 6 C). To further substantiate the hypothesis that Vangl2 is involved in K63-linked poly- ubiquitinated p65, we conducted a knockdown experiment using siRNA to reduce Vangl2 expression. Our results demonstrated that the absence of Vangl2 inhibits K63- linked poly-ubiquitination of p65 (Fig. 6 D). Together, these results suggest that Vangl2 promotes the degradation of p65 by promoting the K63-linked ubiquitination of p65.

Figure 6.

Vangl2 recruits PDLIM2 to ubiquitinate p65

Although Vangl2 promotes K63-linked ubiquitination and degradation of p65, Vangl2 is not an E3 ubiquitin ligase. We hypothesized that Vangl2 might function as a scaffold protein to link p65 and its E3 ubiquitin ligase for ubiquitination or to block the interaction of deubiquitinase (DUB) with p65. To identify the potential E3 ubiquitinase or DUB responsible for Vangl2 mediated ubiquitination of p65, we further analyzed our RNA-seq data to identify E3 ubiquitin ligase or DUB involved in ubiquitination of p65 by comparing LPS-stimulated BMDMs from Vangl2^ΔM and WT mice. The result of differentially expression identified 88 down-regulated genes related to E3 ubiquitin ligase and 56 up- regulated genes related to DUB in response to Vangl2 deficiency in BMDMs after LPS stimulation (Fig. S6 A), which included PDZ-LIM domain-containing protein 2 (PDLIM2), Trim21 and DUB ubiquitin-specific peptidase 7 (USP7). Recent research showed that ubiquitin E3 ligases PDLIM2 and Trim21 ubiquitinated p65 via K63 linkage and enhanced the interaction of p65 with IKK (Healy & O’Connor, 2009; Jodo, Shibazaki, Onuma, Kaisho, & Tanaka, 2020; Yang et al., 2021), while DUB USP7 promoted NF-κB- mediated transcription (Mitxitorena et al., 2020). Here, we found that the mRNA level of PDLIM2 and USP7 decreased in Vangl2-deficient BMDMs, compared with WT BMDMs (Fig. S6, B-D). To investigate which E3 ubiquitin ligase or DUB is recruited by Vangl2, luciferase assay in Vangl2-expressing 293T cells transfected with scramble, Pdlim2, Usp7, or Trim21 siRNA (Fig. S6, E-G) suggested Vangl2-mediated inhibition of NF-κB activation and degradation of p65 were blocked by knocking down PDLIM2, but not USP7 or Trim21 (Fig. 7 A). Meanwhile, Pdlim2 knockdown in BMDMs also resulted in higher expression of Il6 and Il1b in response to LPS stimulation (Fig. 7, B and C), indicating that PDLIM2 plays a key role in promoting Vangl2-mediated p65 degradation.

Figure 7.

We next investigated whether Vangl2 promoted the association between p65 and PDLIM2 by co-IP experiments. Our results indicated that Vangl2 interacted with PDLIM2 (Fig. 7 D) and promoted the association between p65 and PDLIM2 (Fig. 7 E). Notably, PDLIM2 accelerated the degradation of p65 in the presence of Vangl2 (Fig. 7 F). Conversely, PDLIM2 deficiency markedly impaired Vangl2-mediated K63-linked ubiquitination of p65 (Fig. 7 G) and knockdown of Vangl2 inhibited K63-linked poly- ubiquitination of p65 mediated by PDLIM2 (Fig. 7 H). Taken together, these data suggest that Vangl2 functions as a potential adaptor to recruit E3 ubiquitin ligase PDLIM2 to p65 and promotes the K63-linked ubiquitination of p65 for its subsequent autophagic degradation.

Discussion

As a core PCP component, Vangl2 is widely known for its function in organ development, such as brain, tooth, tongue and kidney (Bailly et al., 2018; Hatakeyama et al., 2014), trafficking from the endoplasmic reticulum (ER) to the cell surface, and subsequently shuttles between the endocytic vesicles and cell surface (D. Feng et al., 2021). The function of Vangl2 mostly depends on its cellular localization (Hatakeyama et al., 2014) and is reported mainly through PCP/WNT signaling pathway (Jessen & Jessen, 2019). Activated Vangl2 exhibits extremely long cytoplasmic and intercellular branches and delivers Wnt to multiple cells to enhance Wnt/β-Catenin signaling (Brunt et al., 2021). During myocardial hypertrophy, Vangl2 aggravates myocardial hypertrophy by regulating Wnt/JNK signaling (Brunt et al., 2021; Jessen & Jessen, 2017, 2019) and the expansion of cardiomyocyte surface area (Brunt et al., 2021; Jessen & Jessen, 2017, 2019). However, our previous study showed that the lysosome-suppressing function of Vangl2 in osteoblast differentiation is not dependent on conventional PCP pathway (Gong et al., 2021), and we found Vangl2 did not affect the JNK pathway after LPS treatment (data no shown), suggesting that Vangl2 has additional functions in pathways besides PCP. In this study, we provided direct evidence that the expression of Vangl2 was increased during sepsis and upregulated significantly in immune organs (lymph nodes and spleen) upon LPS stimulation, which is consistent with our previous finding that Vangl2 regulated the downstream signaling of TLR or IL-1R. In addition, the present study showed that Vangl2 prevented the progression of sepsis and the accumulation of inflammatory cytokines through suppressing NF-κB pathway: Vangl2 inhibited LPS-induced NF-κB activation by delivering p65 to autophagosome for degradation. To the best of our knowledge, this is the first study proves that Vangl2 regulates NF-κB signaling and inflammatory responses, and could serve as a potential target for therapeutic purposes in diseases associated with NF-κB signaling.

As is well-known, Vangl2 phosphorylation is essential for its functions in the mammalian PCP pathway in multiple tissues. However, our findings indicate that Vangl2 phosphorylation mutants also promote p65 degradation, suggesting that Vangl2 regulates the NF-κB pathway independently of its conventional PCP pathway. The PCP pathways frequently intersect with developmentally important pathways. Previous research has demonstrated that Vangl2 deficiency can affect developmental processes. It would be importance to investigate whether Vangl2-dependent NF-κB is influenced by developmental pathways. Distinct from previous studies, our findings revealed no significant differences in spleen and lymph node size between wild-type (WT) and Vangl2^ΔM mice. However, myeloid-specific loss of Vangl2 resulted in increased numbers of monocytes, macrophages, and neutrophils in the spleen and bone marrow. This suggests that the ablation of Vangl2 in myeloid cells does not affect the development of immune system development, but rather affects the differentiation of specific cell types. Moreover, we found that Vangl2 is induced and functions under LPS stimulation. Collectively, these findings suggest that Vangl2-dependent NF-κB functions do not affect the development of the immune system, but rather influence the progression of inflammatory-related diseases.

The proteasome, lysosome and autolysosome pathways are the major systems that are utilized by eukaryotic cells to maintain the protein abundance and immune homeostasis (Deretic, 2021). Previous studies have shown that cellular levels of Vangl2 (D. Feng et al., 2021) and its binding partner Prickle2 (Nagaoka et al., 2019) are maintained by the proteasomal pathway. Our recent study showed that Vangl2 mediated osteogenic differentiation by limiting chaperone-mediated autophagy (CMA) in mesenchymal stem cells, suggesting a close relationship between Vangl2 and autophagy (Gong et al., 2021). Indeed, our study suggests Vangl2 promotes p65 degradation through autophagy pathway, but not proteasomal pathway. Importantly, we demonstrated that the degradation of p65 mediated by Vangl2-NDP52 complex is regulated by autophagy induction through rapamycin treatment and autophagy blockade by ATG5 KO or Beclin1 KO. Considering our findings and other reports, Vangl2 may play multi-functional roles in regulating different types of autophagy (i.e. CMA or selective autophagy).

Selective autophagy requires that cargo receptors recognize the labeling of cargoes with degradation signals and subsequently engaged with the LC3 localized in the autophagosome membrane (Shaid et al., 2013). Common cargo receptors mainly include p62/SQSTM1, CALCOCO2/NDP52, OPTN, NBR1 and TOLLIP (Johansen & Lamark, 2014; Yamano & Youle, 2020). A study in HEK293T cells showed that LRRC25 promotes the autophagic degradation of p65 through enhancing the interaction between p65 and p62 (Y. Feng et al., 2017). In addition, a recent study showed that Vangl2 interacts with p62, subsequently promoting breast tumors by activating JNK signaling (Puvirajesinghe et al., 2016). Thus, we hypothesized that Vangl2 may promote the autophagic degradation of p65 by recruiting the cargo receptor p62. However, our data suggested that Vangl2 markedly increased the p65-NDP52 interaction but not p65-p62. Strikingly, Vangl2 mediated autophagic degradation of p65 was abolished in NDP52 KO cells, but not in p62 KO cells. Thus, our findings identify that NDP52 is the new cargo receptor responsible for Vangl2-mediated selective autophagic degradation of p65.

Accumulating evidence has shown that cargo receptors mainly recognize ubiquitination modifications on the substrates and then promote degradation in auto- lysosome (Shaid et al., 2013). As expected, we observed Vangl2 promoted the ubiquitination of p65, thus enhancing the association between cargo protein and p65. K48- and K63-linked ubiquitination were mostly reported modifications on p65 (Kauppinen, Suuronen, Ojala, Kaarniranta, & Salminen, 2013; Korbecki, Bobinski, & Dutka, 2019). For example, E3 ubiquitin ligase RNF182 inhibited TLR triggered cytokine production by promoting K48-linked poly-ubiquitination of p65 (Cao, Sun, Chang, Sun, & Yang, 2019). Trim21 promoted K63-linked poly-ubiquitination of p65, but did not affect the stability of p65 (Yang et al., 2021). PDLIM7 cooperated with PDLIM2 to inhibit inflammation by promoting K63-linked poly-ubiquitination of p65 (Jodo et al., 2020). In this study, we revealed that Vangl2 specifically promoted K63-linked poly-ubiquitination of p65, but not other ubiquitin linkages by recruiting a previously unrecognized E3 ligase PDLIM2, which further explore the molecular mechanism by which regulates the stability of p65.—Although many studies focus on the regulation of p65, the location of p65 degradation has not been clarified. A study showed PDLIM2 degrades p65 through the proteasomal pathway in the nucleus (Jodo et al., 2020), while we found Vangl2 mediated autophagic degradation of p65 mainly happened in cytoplasm, but not in nucleus. Combined with previous studies, our study indicates that p65 undergoes different degradation pathway in distinct location of cells.

Autophagy is a fundamental biological process contributing to multiple life processes. Emerging evidence has suggested that crosstalk between autophagy and innate immune signaling plays critical roles in diseases with inflammatory components, including infections, cancer, autoimmunity, and metabolic disorders (Pradel et al., 2020; Wu et al., 2021). Recent study showed that the interplay between autophagy and type I IFN or NF-κB signaling drives or suppresses inflammatory responses during SARS-CoV- 2 infection (Hui et al., 2021). Moreover, autophagy related key genes, such as Atg5, Atg9, and ULK1, also play critical roles in inflammatory diseases (Y. He et al., 2018; Peng et al., 2019). Here, our discovery of Vangl2-PDLIM2-NDP52 complex in the regulation of p65-mediated NF-κB signaling could be a therapeutic target for the development of immunotherapy against infection and inflammation.

Based on our findings, we propose a working model that Vangl2 negatively regulates NF-κB signaling (Fig. 7 I). Vangl2 functions as an adaptor to recruit ubiquitin ligase PDLIM2 and increase K63-linked ubiquitination on p65, which promotes the recognition of p65 by cargo receptor NDP52 and the autophagic degradation of p65, resulting in suppressing the production of pro-inflammatory cytokines and ameliorating sepsis. Our findings provide a potential target for the treatment of inflammatory diseases.

Materials and Methods

Animal and sepsis model

All animal experiments were approved by the Southern Medical University Animal Care and Use Committee (SMUL20201010). The Vangl2^flox/flox mice and lysozyme-Cre (Lyz2- Cre) mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA; Jax no. 025174 and 004781, respectively). Vangl2^flox/flox mice were hybridized with Lyz2-Cre mice to obtain Vangl2^flox/floxLyz2-Cre (Vangl2^ΔM) mice with Vangl2 specific deficiency in myeloid cells. Co-housed littermate controls with normal Vangl2 expression were used as wild type (WT). For sepsis model, 10-weeks-old mice were intraperitoneally (i.p.) injected with LPS (25-30 mg/kg), and survival rate of mice was continuously observed. For other detection, mice with sepsis were sacrificed and samples were collected at indicated time points.

Isolation of immune cells

For bone marrow (BM)-derived macrophages (BMDMs), BM cells were collected from the femur and tibia, and then maintained in 20% L929 conditioned media with 1% penicillin-streptomycin and macrophage-colony stimulating factor (M-CSF) (St Louis, MO, USA) for 6 days, as previously described(Tan et al., 2019). Mouse peritoneal macrophages were acquired from ascites of indicated mice, which were administrated intraperitoneally (i.p.) with 4% (v/v) thioglycollate (BD) for three consecutive days before sacrifice. As for primary neutrophils, mice were administrated (i.p.) with 4% (v/v) thioglycollate (BD) for 4 h before sacrifice. Peritoneal cavities were flushed with PBS to obtain peritoneal macrophages or neutrophils. Cells were cultured in complete DMEM (Corning) supplemented with 100 U/ml penicillin (Sigma), 100 μg/ml streptomycin (Sigma) and 10% FBS (HyClone) for 6 h and then washed twice with PBS to remove adherent cells. For human PBMC isolation, 4 ml of whole blood was harvested from each sepsis patient and healthy control. The white membrane layer was collected after density gradient centrifugation at 400 g for 25 min and the PBMCs were filtered and prepared for the subsequent experiments after the removal of the red blood cells. For mouse CD11b⁺ splenocytes isolation, spleens were digested in 2% FBS-DMEM with 200 U/ml DNase I (Sigma, USA) and 1 mg/ml collagenase type II (Sigma, USA) at 37 °C for 30 min. Then tissues digested were filtered through a 70Dμm cell strainer and red cells were removed by ACK. The rest splenocytes were labeled with anti-mouse CD11b biotin antibodies (Biolegend). The mixture above was then incubated with streptavidin-paramagnetic particles (BD Biosciences) at 4 °C for 30 min. Purification of CD11b⁺ splenocytes was performed by DynaMag™ (Thermo Fisher Science). Isolated cells achieved a purity of ≥95% measured by FACS.

Cell isolation and culture

HEK293T, A549, CHO and Raw 264.7 cells were obtained from the American Type Culture Collection (ATCC). Above cell lines were maintained in complete DMEM, supplemented with 10% FBS, 100 U/ml penicillin and streptomycin. All cells were cultured at 37 °C with 5% CO₂.

Cell treatment

To test cytokines expression and signaling pathway activation, BMDMs, neutrophils and A549 cells were treated with LPS (200 ng/ml) for the indicated time. For dual luciferase assay, HEK293T cells were treated with LPS (250 ng/ml), IL-1β (40 ng/ml) or TNF-α (20 ng/ml) for 6 h. For protein degradation inhibition assays in HEK293T cells, CQ (50 μM), 3MA (10 mM), or bafilomycin A1 (Baf A1) (0.2 μM) was used to inhibit autolysosome- or lysosome-mediated protein degradation. MG132 (10 μM) was used to inhibit proteasome-mediated protein degradation. Z-VAD-FMK (50 μM) was used to inhibit Caspase-mediated protein degradation.

Luciferase and reporter assays

HEK293T cells were plated in 24-well plates and transfected with pRL-TK and plasmids encoding the NF-κB luciferase reporter, together with different plasmids following: Flag- MyD88, Flag-IRAK1, Flag-TRAF6, Flag-IKKα, Flag-IKKβ, Flag-p65 and an increasing doses of the HA-Vangl2 vector (250, 500, 1000 ng) or empty vector. In addition, CHO or HEK293T cells were administrated with or without LPS (1000 ng/ml), IL-1β (1000 ng/ml) or TNF-α (100 ng/ml) for 6-8 h after transfection with pRL-TK and plasmids encoding the NF-κB luciferase reporter. Then, cells were collected at 24 h post-transfection and luciferase activity was analyzed by Dual-Luciferase Reporter Assay Kit (Vazyme) performed with a Luminoskan Ascent luminometer (Thermo Fisher Scientific). The activity of Firefly luciferase was normalized by that of Renilla luciferase to obtain relative luciferase activity.

Enzyme-linked immunosorbent assay (ELISA)

IL-1β, IL-6 and TNF-α in cell supernatants and mice serum were measured using ELISA kit (#E-EL-M0037c, #E-EL-M0044c, #E-EL-M1084c, respectively; Elabscience Biotechnology) following the manufacturer’s instructions. Absorbance was detected at 450 nm by the Multiscan FC (Thermo Fisher, Waltham, MA, USA).

Statistical Analysis

The data of all quantitative experiments are presented as mean ± SD of at least three independent experiments. Curve data was assessed by GraphPad Prism 8.0 (USA). And comparisons between groups for statistically significant differences were analyzed with a two-tailed Student’s t test. The statistical significance was defined as p < 0.05.

Online supplemental material

Fig. S1 shows that the expression of Vangl2 during sepsis and LPS treatment.

Fig. S2 shows that Vangl2 defection promotes LPS-induced NF-κB activation and production of inflammatory cytokines.

Fig. S3 shows that Vangl2 interacts with p65 to inhibit NF-κB activation.

Fig. S4 shows that Vangl2 promotes p65 degradation by autophagic pathway.

Fig. S5 shows that Vangl2 promoted autophagic degradation of p65 is not mediated by cargo receptor p62.

Fig. S6 shows that expression of candidate E3 ubiquitin ligases in WT and Vangl2- deficient BMDMs after LPS treatment.

Table S1 shows reagents and antibodies used in this study. Table S2 shows primers sequences for quantitative RT-PCR.

Acknowledgements

We thank Dr. Jun Cui (Sun Yat-sen University) for providing ATG5, Beclin1, NDP52, and p62 KO HEK293T cells. We thank Experimental Education/Administration Center of School of Basic Medical Science of Southern Medical University for providing assistance. This work was supported by grants from National Natural Science Foundation of China (82171741, 81801579 and 82302731), Guangdong Basic and Applied Basic Research Foundation (2021A1515012140), and China Postdoctoral Science Foundation (2021M701622).

Author Contributions

J.L., J.Z., H.J. and Z.H. designed and performed the experiments. Y.Z., L.H., D.W., K.Z., P.T., and Z.G. provided assistance or technical support in some experiments. H. L., Y. X., J. Y., Q. X. and Z. S. collected samples. J. L. performed statistical analysis. J.L., J.Z., Z.H., C.P. X.B. and X.Y. performed data analysis and wrote the manuscript. X.Y. supervised the entire project.