Myeloid NCOA4 deficiency augments spleen iron levels.
Prior research has demonstrated that whole-body NCOA4 KO mice exhibit decreased systemic iron but increased iron levels in specific tissues, including the spleen (Bellelli et al., 2016). Notably, myeloid cells, particularly monocytes and macrophages, contribute to spleen function by clearing aged or damaged red blood cells in the red pulp (Lewis et al, 2019). Under certain conditions like infections, alterations in myeloid cell activity in the spleen can impact overall immune system function and blood homeostasis (Bronte and Pittet, 2013). Our data reveals reduced mRNA levels of Ncoa4 (Fig. S1A), but not the macrophage marker adhesion G protein-coupled receptor E1 (Adgre1, encoding the F4/80 protein, Fig. S1B) or the neutrophil marker Ly6g (Fig. S1C), in spleen tissues of KO mice compared to wildtype (WT, Ncoa4F/F) mice. Interestingly, the mRNA level of the Tfrc (Fig. S1D) was increased in the spleen tissues of KO mice compared to WT mice. Perl’s iron staining further confirms elevated iron levels in the splenic red pulp from KO mice compared to WT mice (Fig. S1E, S1F). This evidence underscores the crucial role of myeloid NCOA4 in regulating splenic iron levels.
Myeloid NCOA4 depletion increases cellular iron, oxidative stress and ferroptosis levels in BMDM cells.
Two prominent features of ferroptosis are elevated iron levels and heightened intracellular oxidative stress, both of which are regulated by NCOA4 (Li K et al, 2022). Therefore, we explored the impact of myeloid NCOA4 KO on intracellular iron, oxidative stress, and cell death in BMDM. To further verify NCOA4 deletion in myeloid cells, we initially analyzed the mRNA and protein levels of NCOA4 in BMDM of KO and WT mice. qPCR analysis demonstrated that the mRNA expression of Ncoa4 was significantly decreased (Fig. 1A). Immunoblot analysis revealed a significant decrease in the protein expression of NCOA4, along with an increase in FTH1, in BMDM from myeloid NCOA4 KO compared to their WT littermate control mice (Fig. 1B). This indicates impaired ferritinophagy function. Interestinlgy, FerroOrange staining indicated a notable rise in cellular iron levels (Fig. 1C), and MitoFerroGreen along with DAPI staining demonstrated increased mitochondrial iron and cell death in BMDM from myeloid NCOA4 KO mice (Fig. 1D-1F). Our qPCR analysis further demonstrated that the mRNA expression of iron uptake transporters Slc11a2 and Tfrc, but not iron storage Fth1 and iron exporter Slc40a1, were increased in BMDM cells from myeloid NCOA4 KO compared to their WT littermate control mice (Fig. S2A-S2D). These results support the concept that iron uptake mechanisms are upregulated in NCOA4 KO mice.
Our results also showed that myeloid NCOA4 KO mice exhibit elevated lipid peroxidation levels indicated by BODIPY C11 staining in BMDM (Fig. 1G-1J). However, our qPCR analysis demonstrated that the mRNA expression of antioxidant genes, including Nrf2, Nqo1, and Ho-1, was not changed in BMDM cells from myeloid NCOA4 KO mice compared to their WT littermate controls (Fig. S2E-S2G). Interestingly, the mRNA levels of the pro-inflammatory cytokine Il1β, but not Tnfα and Il6, were also increased (Fig. S2H-S2J). The mRNA levels of the pro-inflammatory M1 macrophage marker Nitric oxide synthase 2 (Nos2) (Fig. S2K) and the anti-inflammatory cytokine Interleukin 10 (Il10) (Fig. S2L) did not show a significant change. These data indicate a mild increase in ferroptosis in the BMDM from KO mice at the basal level.
Myeloid NCOA4 deficiency in mice increases susceptibility to Salmonella-induced colitis.
A recent publication demonstrated that myeloid FTH1 KO mice experience deteriorated cellular iron handling, exacerbating Salmonella infection by triggering hyperinflammation (Haschka et al., 2021). Thus, we conducted further studies on the role of myeloid NCOA4 in controlling infections with the intracellular pathogen Salmonella. Our preliminary findings indicate that myeloid NCOA4 KO mice, compared to WT, experience similar body weight loss (Fig. 2A) and histopathological changes (Fig. S3A, S3B) but demonstrate shortened colon length (Fig. 2B, 2C), impaired Salmonella clearance (Fig. 2D), and elevated expression of proinflammatory cytokines including Il1β (Fig. 2E), Tnfα (Fig. 2F), Il6 (Fig. 2G) and Cxcl1 (Fig. 2H). These results highlight a pivotal role for NCOA4 in not only controlling bacterial spread but also in regulating proinflammatory responses within the colon. Importantly, this function is distinct from its well-established role in ferritinophagy, underscoring the multifaceted nature of NCOA4 in the context of colonic homeostasis and immune regulation.
Our results further showed that Salmonella infection in WT mice activates NRF2 signaling, as evidenced by increased protein levels of its downstream target antioxidant proteins, including NQO1, HO-1, and FTH1, in the mouse colon tissues (Fig. 2I). Conversely, myeloid NCOA4 KO mice exhibit diminished NRF2 signaling molecules, including NRF2, NQO1, HO-1, and FTH1 (Fig. 2I), in Salmonella-treated mouse colon tissues. In WT mice, Salmonella increased levels of a key negative regulator of the NRF2 antioxidant signaling pathway KEAP1 (Fig. 2I). However, these increases were not observed in the KO mice. Consistently, the mRNA levels of Nrf2 and Nqo1 were significantly decreased in the Salmonella-treated colon tissues from the KO mice compared to WT mice (Fig. S3C, S3D). Analysis of colon tissues treated with Salmonella revealed increased mRNA levels of Tfrc (Fig. S3E), reduced mRNA levels of Ncoa4 (Fig. S3F), and increased mRNA levels of Nos2 (Fig. S3G), but no changes were observed in the mRNA levels of Il10 (Fig. S3H). Furthermore, the antimicrobial peptides regenerating islet-derived 3 beta (Reg3b) (Fig.S3I) and calcium-binding protein A8 (S100a8) (Fig.S3J) but not the siderophore Lipocalin 2 (Lcn2) (Fig.S3K). These results suggest that the deficiency of myeloid NCOA4 enhances the suppression of the NRF2 signaling pathway, leading to increased oxidative stress in Salmonella-treated mouse colon tissues.
NCOA4 suppresses oxidative stress by directly binding to KEAP1, thereby stabilizing NRF2.
The molecular basis for the interaction between NCOA4 and NRF2 is still missing so far. Using Ubibrowser, a comprehensive database for proteome-wide known and predicted ubiquitin ligase/deubiquitinase-substrate interactions in eukaryotic species (Wang et al, 2022), we found that NCOA4 is a high confidence E3 ligase substrate of KEAP1 (Fig. 3A). Our reciprocal co-immunoprecipitation assay confirmed that NCOA4 directly interacts with KEAP1 (Fig. 3B, 3C). Furthermore, our data highlight a distinctive interaction between endogenous NCOA4 and KEAP1 in human cells (Fig. S4A), suggesting the physiological relevance of the KEAP1-NCOA4 interaction.
Under iron-replete conditions, NCOA4 binding by HERC2, an E3 ubiquitin ligase, is increased, leading to proteasomal degradation of NCOA4 (Mancias et al., 2015). Given that KEAP1 is also a E3 ubiquitin ligase, we hypothesized that it causes NCOA4 ubiquitination and proteasomal degradation. Indeed, KEAP1 overexpression reduced NCOA4 levels, but the reduced NCOA4 was not restored by the proteasomal inhibitor MG132 (Fig. 3D). Ectopic expression of KEAP1 decreases the expression of IKKβ via autophagic degradation but not proteasomal degradation (Kim et al, 2010). Moreover, treatment with the autophagic flux inhibitor Bafilomycin A1 stabilizes NCOA4 in H4 neuroglioma cells (Goodwin et al., 2017). Thus, we further tested the effect of autolysosomal inhibition via chloroquine on KEAP1 mediated NCOA4 degradation. Interestingly, the KEAP1 overexpression reduced NCOA4 was restored by chloroquine (Fig. 3E), suggesting that KEAP1 mediates autolysosomal degradation of NCOA4.
KEAP1 is a multidomain homodimeric protein which has five distinct domains (Fig. 3F, Dinkova-Kostova et al, 2012, Shibata et al, 2008): (i) NTR: N-terminal region (amino acids 1 − 60); (ii) BTB: broad complex, Tramtrack, Bric- ábrac (amino acids 61 − 179)—the domain through which KEAP1 dimerizes; (iii) IVR: intervening region (amino acids 180 − 314) which is a particularly cysteine-rich region containing 8 cysteine residues among its 134 amino acids; (iv) Kelch domain (amino acids 315 − 598)—the domain through which KEAP1 binds to NRF2; and (v) CTR: C terminal region (amino acids 599 − 624). To determine the exact NCOA4 binding motifs in the KEAP1 protein, we have co-transfected different KEAP1 mutant plasmids (deltaN, deltaBTB, deltaIVR, deltaKelch and 180–625 aa, Lau et al, 2010) with HA-NCOA4. We found that the IVR and Kelch domains of KEAP1 protein are essential for NCOA4 binding (Fig. 3G). Furthermore, KEAP1 overexpression reduced the protein expression levels of NRF2, whereas co-transfection with NCOA4 partially restored the levels of NRF2 (Figure S4B). These results support our hypothesis that NCOA4 and NRF2 are competing for the same KEAP1 binding domain.
Activation of NRF2 protects myeloid NCOA4 KO mice from Salmonella-induced colitis.
To further assess the role of NRF2 signaling in NCOA4 deficiency-enhanced colitis, NCOA4 KO and WT mice were induced with colitis using Salmonella. Subsequently, mice were treated with either vehicle or daily oral gavage of 500 mg/kg NRF2 activator Oltipraz (Ramos-Gomez et al, 2001) through intraperitoneal injection (Kim et al, 2019; Silva et al, 2022). Activating NRF2 didn’t influence mouse body weight loss (Fig. 4A) but significantly increased the colon lengths in myeloid NCOA4 KO mice (Fig. 4B). Furthermore, NRF2 activation significantly decreased pro-inflammatory cytokines including Il1β (Fig. 4C) and Tnfα (Fig. 4D), indicating protective effects in myeloid NCOA4 KO mice. In addition, activation of NRF2 resulted in the restoration of NRF2 signaling proteins (NRF2, NQO1, FTH1 and KEAP1) as well as NCOA4 in colon tissues obtained from myeloid NCOA4 KO mice treated with Salmonella (Fig. 4E). This suggests an on-target drug effect and highlights a competitive relationship between NRF2 and NCOA4.
Antioxidants Tempol and curcumin protect against Salmonella-induced colitis in mice lacking myeloid NCOA4.
We have demonstrated that dextran sulfate sodium-induced acute colitis is diminished in mice treated with Tempol (Xue et al., 2017). Here we conducted further investigations into the role of Tempol in controlling Salmonella-induced colitis. Tempol didn’t influence mouse body weight loss (Fig. S5A). However, Tempol significantly protected myeloid NCOA4 KO mice but not WT mice with Salmonella-induced colitis, as evidenced by increased colon length (Fig. 5A) and reduced levels of proinflammatory cytokines including Il1β (Fig. 5B), Tnfα (Fig. 5C) and Il6 (Fig. 5D). In addition, Tempol also led to the restoration of NRF2 signaling proteins (NRF2, NQO1 and HO-1) in colon tissues obtained from myeloid NCOA4 KO mice treated with Salmonella (Fig. 5E). Interestingly, the protein levels of FTH1 were not significantly changed (Fig. 5E). This suggests Tempol can reduce oxidative stress and alleviate colitis caused by Salmonella infection in mice lacking myeloid NCOA4.
Considering the synthetic nature of Tempol and aiming for future clinical therapeutic and prophylactic applications with enhanced safety, we investigated the potential of using the naturally occurring antioxidant curcumin. Curcumin (C21H2OO6) is a lipophilic substance of a polyphenol nature, obtained from rhizomes of turmeric (Curcuma longa L.). It has antibacterial and antioxidant properties, and because it is a natural compound and low cost, curcumin exhibits great promise as a therapeutic agent to develop new treatments for bacterial infections and oxidative stress-related disease. However, the role of curcumin in controlling Salmonella is a subject of controversy (Leyva-Diaz et al, 2021). Previous publications have demonstrated the antioxidant effect of curcumin (Jakubczyk et al., 2020). However, one report suggests that curcumin can enhance the pathogenicity of Salmonella by increasing its robustness, possibly through the upregulation of genes involved in resistance against microbial peptides (Marathe et al., 2010). We hypothesize that specific delivery of curcumin into myeloid cells might mitigate potential side effects. First, we performed in vitro studies with RAW264.7 cells stimulated with lipopolysaccharide (LPS), an inflammatory stimulant, to compare the effects of free curcumin and nanoparticles conjugated with curcumin (nano-curcumin). Our results showed an increase in the protein levels of NCOA4 following stimulation with LPS concentrations higher than 1µg/mL (Fig. S5B). While both free and nano-curcumin were not able to reduce the mRNA expression levels of Tnfα in LPS challenged RAW264.7 cells (Fig. S5C), nano-curcumin have a slightly stronger effect on inhibiting LPS increased Il1β mRNA levels (Fig. S5D). Next, we treated mice with salmonella and the next day with free or nano-curcumin through intraperitoneal injection. Myeloid cell-targeting nano-curcumin didn’t have a significant improvement on the body weight (Fig. S5E). However, nano-curcumin provided significantly better protection in mice with Salmonella-induced colitis, as indicated by increased colon length (Fig. 5F). Interestingly, both curcumin and nano-curcumin similarly reduced the levels of proinflammatory cytokines in myeloid NCOA4 KO but not WT mice (Fig. 5G-5I). Additionally, nano-curcumin also led to the restoration in NRF2 signaling proteins (NRF2, NQO1, and HO-1) in colon tissues from myeloid NCOA4 KO mice (Fig. 5J). Consistently, the protein levels of FTH1 were not significantly changed (Fig. 5J). This suggests nano-curcumin can reduce oxidative stress and alleviate colitis caused by Salmonella infection in mice lacking myeloid NCOA4.
Together, antioxidants may alleviate colitis severity in myeloid NCOA4-deficient individuals but could impede macrophages' ability to combat Salmonella in WT mice due to the direct antimicrobial functions of ROS (Herb and Schramm, 2021). This underscores the need for a nuanced approach in considering antioxidant therapy based on the specific immune context.
A low iron diet and ferroptosis inhibition protect mice from Salmonella-induced colitis enhanced by myeloid NCOA4 depletion.
Salmonella relies on iron for growth in monocyte-macrophage system cells (Nairz, et al, 2014). A low iron diet didn’t have a significant improvement on the body weight (Fig. S6A). However, the protective impact of the low-iron diet is evident in increased colon length (Fig. 6A) and decreased levels of proinflammatory cytokines including Tnfα (Fig. 6B), Il1β (Fig. 6C) and Nos2 (Fig. 6D) in both myeloid NCOA4 KO and WT mice. Additionally, the low iron diet effectively decreased the mRNA levels of Slc11a2 (Fig. 6E). These findings emphasize the crucial role of dietary iron levels in influencing colitis severity, proposing a potential avenue for modulating colitis outcomes through dietary interventions that target iron availability.
Recent research has demonstrated that ferroptosis sensitivity is modulated by NCOA4, given the central role of NCOA4-mediated ferritinophagy in regulating intracellular iron levels (Santana-Codina et al., 2018). A ferroptosis inhibitor Ferrostatin-1 didn’t have a significant improvement on the body weight (Fig. S6B). However, our data reveal that colon length was reduced in myeloid NCOA4 KO mice compared to WT mice 7 days after Salmonella infection, an effect rescued by Ferrostatin-1 (Fig. 6F). Similarly, proinflammatory cytokine Tnfα (Fig. 6G), Il6 (Fig. 6H) and iron uptake transporter Tfrc (Fig. 6I) were elevated in colons from myeloid NCOA4 KO mice compared to WT mice 7 days after Salmonella infection, with ferroptosis inhibitor Ferrostatin-1 also rescuing these effects. Thus, these data suggest that the heightened susceptibility to Salmonella-induced colitis in myeloid NCOA4 knockout mice is attributable to ferroptosis.
Myeloid cell-specific NCOA4 overexpression protects mice from Salmonella-induced colitis.
To assess if myeloid NCOA4 is sufficient to protect mice from Salmonella-induced colitis, we generated myeloid cell specific NCOA4 OE mice. qPCR analysis confirmed that NCOA4 is increased in the spleens of OE mice compared to WT mice (Fig. S7A). Also, qPCR and immunoblot analysis confirmed that NCOA4 mRNA (Fig. 7A) and protein (Fig. 7B) expression levels are significantly increased in BMDM cells from OE mice than WT mice. Interestingly, we observed increased mRNA expression of Nrf2 (Fig. S7B), Nqo1 (Fig. S7C), and Hmox1 (Fig. S7D) in BMDM cells from OE mice compared to WT mice. However, there was no significant change in the pro-inflammatory cytokine Il1β (Fig. S7E).
Although there is no significant change in body weight (Fig. S7F), the OE mice exhibit protection from Salmonella-induced colitis, as indicated by longer colon lengths (Fig. 7C, 7D). Interestingly, analysis of colon tissues treated with Salmonella showed decreased expression of Tfrc (Fig. 7E), Slc11a2 (Fig. 7F), Nos2 (Fig. 7G), Tnfα (Fig. 7H) and Cxcl1 (Fig. 7I), confirming that NCOA4 OE diminishes inflammation due to Salmonella infection by downregulating iron levels. Further analysis confirmed increased protein levels of NRF2, NQO1, HO-1, FTH-1 and KEAP1 in these colon tissues (Fig. 7J). These results demonstrate that NCOA4 OE manifests less severe colitis and oxidative stress.