- Split View
-
Views
-
Cite
Cite
Haizhen Wang, Xue Jiang, Jieyu Wu, Linqiang Zhang, Jingfei Huang, Yuru Zhang, Xiaoju Zou, Bin Liang, Iron Overload Coordinately Promotes Ferritin Expression and Fat Accumulation in Caenorhabditis elegans, Genetics, Volume 203, Issue 1, 1 May 2016, Pages 241–253, https://doi.org/10.1534/genetics.116.186742
- Share Icon Share
Abstract
The trace element iron is crucial for living organisms, since it plays essential roles in numerous cellular functions. Systemic iron overload and the elevated level of ferritin, a ubiquitous intracellular protein that stores and releases iron to maintain the iron homeostasis in cells, has long been epidemiologically associated with obesity and obesity-related diseases. However, the underlying mechanisms of this association remain unclear. Here, using Caenorhabditis elegans, we show that iron overload induces the expression of sgk-1, encoding the serum and glucocorticoid-inducible kinase, to promote the level of ferritin and fat accumulation. Mutation of cyp-23A1, encoding a homolog of human cytochrome P450 CYP7B1 that is related to neonatal hemochromatosis, further enhances the elevated expression of ftn-1, sgk-1, and fat accumulation. sgk-1 positively regulates the expression of acs-20 and vit-2, genes encoding homologs of the mammalian FATP1/4 fatty acid transport proteins and yolk lipoproteins, respectively, to facilitate lipid uptake and translocation for storage under iron overload. This study reveals a completely novel pathway in which sgk-1 plays a central role to synergistically regulate iron and lipid homeostasis, offering not only experimental evidence supporting a previously unverified link between iron and obesity, but also novel insights into the pathogenesis of iron and obesity-related human metabolic diseases.
DUE to its essential roles in numerous cellular functions across nearly all living organisms, including oxygen transport, electron transport, DNA synthesis, and enzyme catalysis, exploring how iron is stored and regulated has been a growing focus in numerous fields. Dietary iron is absorbed primarily by duodenal enterocytes via the divalent metal-ion transporter 1 (DMT1) after it is reduced at the apical membrane. Subsequently, it is either stored in ferritin, which dynamically regulates iron sequestration, storage, and release or it is transported from enterocytes into the blood stream via the basolateral transporter ferroportin (Fleming and Ponka 2012). Consequently, the total amount of iron in the body is determined by its intake and storage, which are all finely regulated by many factors and signaling pathways at various levels (Fleming and Ponka 2012; Hubler et al. 2015), many of which not entirely understood.
Human epidemiology studies revealed that elevated levels of ferritin may be an indication of systemic iron overload (Cook et al. 1974; Zimmermann 2008) and are positively associated with obesity (Wenzel et al. 1962; Gillum 2001; Iwasaki et al. 2005) and obesity-related diseases such as diabetes, hypertension, dyslipidaemia, or nonalcoholic fatty liver disease (Jehn et al. 2004; Bozzini et al. 2005; Mascitelli et al. 2009; Dongiovanni et al. 2011; Kim et al. 2011; see Zafon et al. 2010 for review). Precisely why elevations in ferritin levels or systemic iron overload are associated with these conditions is not entirely clear, with potential explanations ranging from excess iron causing oxidative stress, endoplasmic reticulum (ER) stress, inflammation, and dysfunction of adipose tissue (Hubler et al. 2015; Nikonorov et al. 2015).
When including other environmental or dietary factors, the relationship between obesity, its comorbidities, and iron becomes more complex. Iron is required for the adipogenesis of the 3T3-L1 preadipocytes (Moreno-Navarrete et al. 2014a), and it increases the rate of adipocyte lipolysis (Rumberger et al. 2004). In the mouse liver, iron significantly upregulates the transcripts of seven enzymes in the cholesterol biosynthesis pathway, resulting into cholesterol accumulation. This occurs independently of the conserved lipogenic regulator SREBP2 (Graham et al. 2010). The combination of iron and lipid-rich diet may exacerbate this situation because lipids also cause oxidative and ER stress, as well as inflammation. Collectively, this helps explain why dietary iron supplementation concurrent with a high-fat diet (HFD) greatly increases adiposity in rats (Tinkov et al. 2013), as well as hepatic fat accumulation in the liver of mice (Choi et al. 2013). The association seems to also hold true in reverse. Several studies found that reduction of iron by several different methods led to amelioration of adiposity and improvement of obesity and obesity-related diseases (Zhang et al. 2005; Fleming and Ponka 2012; Tajima et al. 2012; Moreno-Navarrete et al. 2014a,b). Unfortunately, the mechanisms of this association between iron and lipid accumulation or obesity remain largely unclear.
Though human clinical studies would likely shed a great deal of light on this relationship, such studies are not always practical or even possible. Alternatively, Caenorhabditis elegans may offer an ideal model due to the highly conserved nature of many proteins involved in iron homeostasis, including iron uptake (SMF-3, a homolog of DMT1), storage (FTN-1 and FTN-2, encoding ferritin), and export (FPN-1.1, FPN-1.2, FPN-1.3, encoding ferroportin), as well as potential orthologs for DCYTB ferrireductase and hephaestin multicopper oxidase (Anderson and Leibold 2014). Moreover, C. elegans FTN-1 and FTN-2 are more similar to human FTH than to FTL, and both FTN-1 and FTN-2 contain ferroxidase active-site residues (Gourley et al. 2003). Under iron overload, the expression of ftn-1 gene and protein, and to a lesser extent ftn-2, are induced; in contrast, the expression of SMF-3 is suppressed to reduce iron uptake (Gourley et al. 2003; Kim et al. 2004). ftn-1 RNAi (Kim et al. 2004) or ftn-1(ok3625) deletion mutant (Valentini et al. 2012) are iron sensitive and have reduced lifespans when exposed to high iron.
While these studies have investigated how iron affects growth, development, and longevity in C. elegans, whether or not it also coordinately regulates lipid metabolism is entirely unclear. In this study, we used C. elegans to identify novel genes and unravel complex pathways involved in iron metabolism and lipid metabolism. Our results show that iron overload by the established dietary supplementation with ammonium ferric citrate (FAC) (Gourley et al. 2003) coordinately promotes the expression of ferritin and fat accumulation in C. elegans. Specifically, iron overload induces the serum and glucocorticoid-inducible kinase sgk-1 to positively regulate acs-20 and vit-2, facilitating lipid uptake and accumulation.
Materials and Methods
C. elegans strains
C. elegans strains were maintained under standard culture conditions on NGM agar with Escherichia coli OP50, unless otherwise specified. The wild-type strain was N2. The strains used in this study were as follows: acs-20(tm3232)IV, age-1(hx546)II, akt-1(ok525)V, cyp-23A1(gk253)II, daf-2(e1370)III, ftn-1::GFP{XA6900, pha-1 (e2123ts)III;qaEx01[ftn-1::△pes-10::GFP-his,pha-1(+)]}(Romney et al. 2008), cyp-23A1(gk253)II;ftn-1::GFP{XA6900, pha-1(e2123ts)III;qaEx01[ftn-1::△pes-10 ::GFP-his,pha-1(+)]}, sgk-1(ft15)X, sgk-1(ok538)X, sgk-1(ok538)X;cyp-23A1(gk253), sgk-1::gfp(MQD862, N2;Psgk-1::sgk-1::gfp [hqIs150]) (Hertweck et al. 2004; Zhu et al. 2015), rsks-1(ok1255)III, vit-2(ok3211)X, vit-3(ok2348)X, vit-2::gfp{RT130,unc-199(ed3);pwIs23[Pvit-2::vit-2::gfp, unc-199(+)]}(Van Rompay et al. 2015), cyp-23A1(gk253);vit-2::gfp {RT130, unc-199(ed3);pwIs23 [Pvit-2::vit-2::gfp,unc-199(+)]}, and acs-20(tm3232);vit-2::gfp {RT130,unc-199(ed3); pwIs23[Pvit-2::vit-2::gfp,unc119(+)]}.
Construction of Psgk-1::gfp extrachromosomal transgenic strain
The Psgk-1::gfp transgenes were created using methods as previously described by Frokjaer-Jensen et al. (2008), and the transgenic strain was created by microinjection. Briefly, an extrachromosomal transgenic strain was made by injection into the EG4322 (ttTi5605;unc-119(ed3). The injection mix consisted of 50 ng/μl pCFJ151 inserted with target DNA fragment, pJL43.1 (Pglh-2::transposase), 5 ng/μl pCFJ104 (Pmyo-3::mCherry), and 2.5 ng/μl pCFJ90 (Pmyo-2::mCherry). DNA mixtures were injected into the gonads of young adult C. elegans. The extrachromosomal transgenic strain was assigned kunEx126[unc-119(ed3);Psgk-1::gfp+unc-119(+)]. The primers used for amplification of Psgk-1::gfp are Psgk-1F: GGATATCTGGATCCACGAA CTCCGGTAACTTACTCATTTTC and Psgk-1R2: GTCGACCTGCAGGCATGCAACC TCACCATTCTCGACTCTG; and GFP+unc-54 3′UTR-F: TTGCATGCCTGCAGGTCG AC and GFP+unc-54 3′UTR-R: CCAGAGCTCACCTAGGTATCTGCCGACTAGTAG GAAACAGT.
Supplementation of FAC
FAC supplementation was performed as described previously (Romney et al. 2008). In brief, synchronized L1 worms were placed and cultivated on NGM plates supplemented with 2.5 mg/ml or 5 mg/ml FAC. The pH value of FAC medium was adjusted to 7.0. L4 worms or young adults were harvested for further analysis.
RNAi screen regulators of ftn-1::GFP
Synchronized ftn-1::GFP(XA6900) L1 worms were seeded on NGM plates supplemented with FAC (5 mg/ml) or not. The performance of RNAi by feeding was followed as we described previously (Zhang et al. 2013). Worms were collected at ∼60 hr, at young adult stage, and GFP fluorescence was visualized under a fluorescence microscope (BX53, Olympus). RNAi of a specific gene displaying changed expression of FTN-1::GFP was considered a potential regulator of ftn-1::GFP.
Visualization of GFP fluorescence
At least 20 GFP worms were picked and mounted on an agarose pad and anesthetized using 10 mM sodium azide. GFP fluorescence was visualized under an Olympus BX53 fluorescence microscope.
Quantitative PCR analysis
Total RNA and cDNA preparation and quantitative PCR (qPCR) performance were followed as described previously (Ding et al. 2015).
Nile Red staining of fixed nematodes
L4s or young adult nematodes were washed off of growth plates and then fixed and stained with Nile Red as done previously (Brooks et al. 2009; Liang et al. 2010). Images were captured using identical settings and exposure time for each image, unless specifically noted.
Lipid extraction and analysis
Lipid extraction and analysis of young adult nematodes was determined as described previously (Watts and Browse 2006; Shi et al. 2013). To determine the triacylglycerol (TAG) and phospholipids, lipid extraction, thin-layer chromatography (TLC), and gas chromatography (GC) were performed as described (Watts and Browse 2006; Shi et al. 2013).
Dietary supplementation of oleic acid [C18:1(n-9)]
Fatty acid supplementation was carried out by adding sodium oleic acid (OA) (NuChek Prep) at a final concentration of 0.2 mM to NGM media as we previously described (Shi et al. 2013). Synchronized L1 worms were cultivated on NGM agar plates supplemented with OA. Young adults were washed off the plates, and fatty acid composition of adult nematodes was determined as previously described (Watts and Browse 2002; Liang et al. 2010).
Fluorescent fatty acid probe analysis
Analysis of fluorescent fatty acid probe was modified from Spanier et al. (2009). The 4,4-difluoro-5-methyl-4-bora-3a,4a-diaza-3-indacene-dodecanoic acid (BODIPY 500/510 C1, C12; Sigma) was dissolved in DMSO to a final concentration of 2.4 μM and then attenuated to 20 nM by M9 buffer. Briefly, synchronized L1 worms were cultivated on NGM plates and harvested at L4 stage, then immediately incubated in 20 nM BODIPY-12 solution for 10 min, and then visualized under a fluorescent microscope (BX53, Olympus).
Quantification of lipid droplet size
Quantification of the diameter of lipid droplets (LDs) was followed as described by Shi et al. (2013). In brief, young adult worms were fixed and stained by Nile Red and photographed. A 50 × 50 μm square was selected in the middle of the intestine for each worm and all of the LDs in this square were measured by the software CellSens Standard (Olympus). Roughly 10 worms were measured for each worm strain or treatment.
Lipid droplet purification
LDs of C. elegans were isolated as described by Zhang et al. (2012). In brief, 4 × 105 worms were harvested and washed by M9 buffer three times, and by buffer A. Next, worms were homogenized by a homogenizer (Cole-Parmer LabGENR700) and centrifuged at 1000 × g for 30 sec. The supernatants were homogenized again by a nitrogen cell disruption vessel (4639 45-ml cell disruption vessel; Parr Instrument, Moline, IL) at 500 psi for 30 min on ice. A 9-ml supernatant was loaded into a sw40 tube with 3 ml buffer B on the top and centrifuged at 10,000 × g for 1 hr at 4°. The LD fraction was transferred to another 1.5-ml tube and washed with 200 μl buffer B three times.
Western blot analysis
Young adults with one to three eggs were harvested, and their tissues ground and homogenized at 4° with an extraction buffer. The tissue homogenates were then centrifuged, and the supernatants were used for Western blot analysis. The LD protein and lipid samples were prepared as previously described (Zhang et al. 2012). Proteins were separated by 12% SDS gel (Bio-Rad) and then transferred to polyvinylidene difluoride membranes. The primary antibodies were sc8334 (Santa Cruz Biotechnology) rabbit anti-GFP antibody at 1:200 dilution, O00141 (Enzo) rabbit anti-SGK-1 antibody at 1:1000 dilution. Anti-DHS-3, anti-CAV-1, and anti-BIP antibodies were used as previously described by Zhang et al. (2012). The secondary antibody was goat antirabbit IgG from Beyotime at 1:5000 dilution. The membrane was exposed to chemiluminescent substrate (Pierce), and the film (Carestream, Xiamen, China) was then developed.
Statistical analysis
Data are presented as mean ± SEM, except when specifically indicated. Statistical analysis performed included t-test or analysis of variance (ANOVA) followed by least significant difference multiple comparisons using SPSS20.0 (IBM SPSS Statistics, Armonk, NY). All figures were made using GraphPad Prism 5 (GraphPad Software, La Jolla, CA).
Data availability
The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article.
Results
Iron induces sgk-1 expression to upregulate ferritin expression and fat accumulation
To determine whether iron overload promotes fat accumulation in C. elegans, we applied FAC to NGM plates. Consistent with previous reports (Gourley et al. 2003; Romney et al. 2008), FAC induced the expression of FTN-1::GFP (Figure 1A) and ftn-1 messenger RNA (mRNA) (Figure 1B) in wild-type N2 . Concurrently, FAC promoted fat accumulation, as indicated by an increased size of LDs and increased abundance of large-sized LDs (Figure 1C and Supplemental Material, Figure S1, A and B, and Figure S2). Consistently, isolated LDs from N2 as well as two reported LD markers DHS-3::GFP (Zhang et al. 2012) and mRuby::DGAT-2 (Xu et al. 2012) further confirmed that FAC indeed increases LD size (Figure 1D). Furthermore, lipid quantitation by TLC/GC shows that the relative percentage of triacylglycerol (TAG/total lipids, %) significantly increased in N2 worms under FAC treatment (5.0 mg/ml) compared to no FAC (0 mg/ml) (Figure 1E). Since FAC at 5.0 mg/ml resulted in obvious changes of both gene expression and fat accumulation, we used this concentration in most of the subsequent experiments. The increased fat accumulation induced by FAC is not likely to be a result of citrate supplementation, since dietary supplementation of ammonium ferric sulfate (FAS) also promotes expression of FTN-1::GFP as well as fat accumulation (Figure S3). Altogether, these findings suggest that iron overload indeed concurrently promotes ferritin expression and fat accumulation.
To investigate the underlying genetic mechanisms and signaling pathways required for this synergetic effect, we performed a selective genetic screen using RNAi knockdowns as well as mutant strains corresponding to ∼2000 genes, including transcription factors, kinases, cytochrome P450 family, and lipid metabolic genes. We screened RNAi knockdowns or mutations that led to changes in GFP fluorescence of the ftn-1::GFP(XA6900) strain after iron (FAC) exposure. We chose ftn-1, rather than ftn-2, because it shows the greatest response to iron in C. elegans (Gourley et al. 2003; Kim et al. 2004). In strains that failed to induce ftn-1::GFP after iron exposure, we further screened with Nile Red staining of fixed nematodes to determine if fat accumulation was also affected. We found that the elevated expression of FTN-1::GFP (Figure 1A) as well as ftn-1 mRNA (Figure 1B) induced by FAC was suppressed in the sgk-1(ok538) mutant strain. The sgk-1 gene encodes an ortholog of the serum and glucocorticoid-induced kinase in mammals (Hertweck et al. 2004). Furthermore, in the absence of FAC, compared to N2, the expression of ftn-1 mRNA was decreased in sgk-1(ok538) mutant, but was increased in an integrated sgk-1::gfp (MQD862, N2;Psgk-1::sgk-1::gfp) strain (Figure 1B) (Hertweck et al. 2004; Zhu et al. 2015). In the presence of FAC, like in N2, the expression of ftn-1 mRNA was increased in both gain-of-function mutant sgk-1(ft15) (Jones et al. 2009) and sgk-1::gfp (MQD862, N2;Psgk-1::sgk-1::gfp) worms (Figure 1B). These results indicate that sgk-1 regulates the expression of ftn-1.
Interestingly, we found that both the sgk-1(ok538) deletion mutants as well as the sgk-1(ft15) gain-of-function mutants had slightly increased fat content than N2 in the absence of iron (Figure 1E), consistent with previous reports (Jones et al. 2009; Soukas et al. 2009). However, when grown on FAC, only the sgk-1(ft15) gain-of-function mutants showed increased fat content (Figure 1E) and LD size (Figure 1C, Figure S1, A and C, Figure S2), indicating that SGK-1 activity is necessary for the induction of fat accumulation.
Since the sgk-1(ok538) deletion mutant was able to suppress the elevated expression of ftn-1 as well as the increase in fat content under iron overload, we asked whether the upregulation of ftn-1, as well as increased fat accumulation, were due to the induced expression of sgk-1. We found that FAC induced the level of sgk-1 mRNA in N2 worms by twofold (Figure 1F) and also led to increased intensity of GFP fluoresence in sgk-1::gfp and Psgk-1::gfp worms (Figure 1G). The increased GFP fluorescence was observed using the integrated sgk-1::gfp (MQD862, N2;Psgk-1::sgk-1::gfp) translational fusion strain (Hertweck et al. 2004; Zhu et al. 2015) (Figure 1G, left), as well as with a transcriptional reporter containing 2.2 kb of upstream sgk-1 promoter sequence fused to gfp, Psgk-1::gfp {kunEx126, [unc-119(ed3);Psgk-1::gfp+ unc-119(+)]} (Figure 1G, right). Interestingly, the two reporter strains showed different tissue expression. The translational SGK-1::GFP expresses in some head and tail neurons, with diffuse expression in the intestine (Hertweck et al. 2004), but the promoter-only fusion shows expression only in the intestine (Figure 1G, Figure S4), implying that the coding region of sgk-1 may contain regulatory sequences necessary for neuronal expression. Nevertheless, FAC led to increased expression of sgk-1 mRNA as well as increased fluorescence intensity in both strains (Figure 1, F and G), suggesting that FAC induces the transcriptional expression of sgk-1. Conversely, restriction of iron by BP (2,2′-bipyridyl, 0.1 mM), an iron chelator (Breuer et al. 1995; Romney et al. 2008, 2011), also represses the expression of SGK-1::GFP (MQD862, N2;Psgk1::sgk-1::gfp) (Figure S5). Altogether, these studies show that iron overload induces the expression of sgk-1, which promotes the expression of ftn-1 and fat accumulation.
cyp-23A1 represses sgk-1 to regulate ferritin expression and fat accumulation
We identified a mutant strain, cyp-23A1(gk253), predicted to carry a loss-of-function mutation in a gene encoding a member of the cytochrome P450 family, that greatly enhanced the expression of FTN-1::GFP (Figure 2A) and ftn-1 mRNA (up to 13.6-fold) (Figure 2B) compared to N2 under FAC treatment, suggesting that CYP-23A1 activity represses ferritin expression. Since iron overload induced the expression of sgk-1, we asked whether the exacerbated effects in the cyp-23A1(gk253) mutant strain could be due to misregulation of sgk-1. We found that the expression of sgk-1 mRNA (Figure 2C), as well as the intensity of the SGK-1::GFP (MQD862, N2;Psgk1::sgk-1::gfp) (Figure 2D), was higher in cyp-23A1(gk253) mutants than in N2 in both the absence and presence of FAC. In addition, Western blot analysis using anti-SGK-1 antibody also showed that the level of SGK-1 protein was significantly higher in cyp-23A1(gk253) mutants than in N2, both in the presence and absence of FAC (Figure 2, E and F). Remarkably, the highly induced expression of ftn-1 mRNA (Figure 2B) as well as the increased intensity of FTN-1::GFP (Figure 2A) in the cyp-23A1(gk253) mutant was abolished in sgk-1(ok538) background. Therefore, these findings indicate that CYP-23A1 represses sgk-1 expression, which eventually leads to decreased ftn-1 expression.
Next, we asked whether cyp-23A1 also affects fat accumulation. In the absence of FAC, the cyp-23A1(gk253) mutant displayed normal lipid droplet size similar to N2 (Figure 2G, Figure S1, A and F, Figure S2), although TLC/GC analysis revealed slightly higher fat content compared to N2 (Figure 2H). Similar to N2, iron overload increased fat accumulation (Figure 2H) and lipid droplet size (Figure 2G, Figure S1, A and F, Figure S2) in cyp-23A1(gk253) mutants, but was dependent on sgk-1 since both lipid droplet size and fat content did not increase in the sgk-1(ok538);cyp-23A1(gk253) double mutant strain (Figure 2, G and H, Figure S1, A and G, Figure S2). Collectively, these data show that CYP-23A1 appears to repress sgk-1 to regulate the level of ferritin and fat accumulation under iron overload.
The insulin/IGF-1 signaling pathway was previously reported to work with hypoxia signaling to regulate iron homeostasis in C. elegans (Ackerman and Gems 2012). However, RNAi knockdown of two reported transcriptional factors daf-16, encoding the FOXO homolog in the insulin/IGF-1 signaling pathway, and hif-1, encoding hypoxia-inducible factor-1 (Romney et al. 2011; Ackerman and Gems 2012), did not suppress the expression of ftn-1 mRNA in cyp-23A1(gk253) mutant background under FAC treatment, even though cyp-23A1(gk253) mutant successfully reversed the high expression of ftn-1 mRNA in hif-1RNAi worms under no FAC (Figure S6A). Additionally, the repression of sgk-1 expression by CYP-23A1 also appears to be independent of the transcription factors DAF-16 and HIF-1 (Figure S6B).
Iron facilitates lipid uptake via sgk-1
The capacity of lipid accumulation ultimately depends on the homeostasis of lipid synthesis, lipolysis, uptake, and transport. Previous studies contend that SGK-1 may participate in insulin/IGF-1 or target of rapamycin (TOR) signaling pathways to regulate lipid metabolism (Jones et al. 2009; Soukas et al. 2009; Lang and Stournaras 2013). However, we found that several mutants of genes in both pathways, e.g., daf-2(e1370), age-1(hx546), akt-1(ok525), daf-16(mu86), and rsks-1(ok1255) (Figure S7A), are responsive to iron-promoted fat accumulation, implying that they may not be directly involved in iron-induced fat accumulation. Furthermore, our analysis shows that FAC treatment has both positive and negative effects on gene expression of both lipid synthesis and lipid degradation genes, thus it is not clear from this analysis to what extent the balance of de novo lipid synthesis and degradation is disrupted by FAC (Figure S8, A–C).
To test the hypothesis that iron may enhance the production of reactive oxygen species (ROS) and induce oxidative damage that may impair the function of mitochondria and lipid oxidation (Levi and Rovida 2009), we examined the expression of genes encoding superoxide dismustases, sod-1, sod-2, sod-3, and sod-4, but found no significant gene expression changes (Figure S8D). We performed dihydroethidium (DHE) staining, which is widely used to evaluate reactive oxygen species production, and observed no significant changes after FAC treatment (Figure S8E). Finally, the relative oxygen consumption rate did not differ in N2, sgk-1(ok538), cyp-23A1(gk253), sgk-1(ok538);cyp-23A1(gk253) worms between FAC treatment and no FAC (Figure S8F). Taken together, these results do not support the hypothesis that iron overload is inducing oxidative stress or increased oxygen consumption in C. elegans.
The yeast Ypk1, the ortholog of human sgk-1, was previously found to be required for efficient fatty acid uptake (Jacquier and Schneiter 2010). Since increased lipid accumulation by iron overload is dependent on sgk-1, we posited that iron-induced lipid accumulation may be due to increased lipid uptake under regulation of SGK-1. After conducting staining of BODIPY-C12 fatty acid to indicate lipid uptake in C. elegans (Spanier et al. 2009), we found that iron indeed promoted lipid uptake as more fluorescence accumulated in the whole intestine of N2, cyp-23A1(gk253), sgk-1(ft15) worms at FAC 5.0 mg/ml compared to control (FAC 0 mg/ml) (Figure 3, A and B). However, in the sgk-1(ok538) and sgk-1(ok538);cyp-23A1(gk253) worms, fluorescence of BODIPY-C12 did not increase with FAC concentration (Figure 3A), which was mostly aggregated in the intestinal lumen, suggesting that iron-promoted lipid uptake depends on SGK-1.
To confirm these results, we further supplemented the C. elegans diet with OA, [C18:1(n-9), 0.2 mM]. Consistent with our previously report (Shi et al. 2013), C. elegans incorporated C18:1(n-9) in all six tested strains in the absence of iron (Figure 3, C and D). Additionally, iron overload enhanced an even higher level of C18:1(n-9) in N2, sgk-1::gfp (MQD862, N2;Psgk-1::sgk-1::gfp), and sgk-1(ft15) worms than under no FAC (Figure 3, C and D), further confirming that iron facilitates lipid uptake. Consistent with the results of dietary BODIPY-C12, under iron overload the level of C18:1(n-9) dramatically increased in cyp-23A1(gk253) worms, but remained unchanged in the sgk-1(ok538) mutant background (Figure 3C). Thus, dietary supplementation of BODIPY-C12 fatty acid and uptake of oleic acid support our hypothesis that SGK-1 is required for lipid uptake under iron overload.
The fatty acid transporter ACS-20 is required for iron-induced lipid uptake
The observation that iron overload facilitates lipid uptake to promote lipid accumulation raised the question of which gene products are responsible for lipid uptake in response to FAC. Generally, dietary lipids are taken up by cells by lipid transporters. The C. elegans genome contains a number of genes with homology to known lipid transporters that may display similar or distinct functions (Zhang et al. 2013). We tested several strains with mutations in putative lipid transporter genes including the acs-20(tm3232) strain, which carries a deletion in the acs-20 gene that encodes an ortholog of mammalian fatty acid transport proteins FATP1/4 (Kage-Nakadai et al. 2010) (www.wormbase.org). We found that lipid droplet size (Figure S1, A and H), fat storage indicated by postfix Nile Red staining (Figure 4A) and LipidTox Red staining (Figure S2), as well as quantitated by TLC/GC (Figure 4B), were unchanged in the acs-20(tm3232) mutants compared to N2 after iron overload. Thus, acs-20 is necessary for iron-promoted fat accumulation.
We next investigated whether the increased fat accumulation was due to activated acs-20. The expression of acs-20 mRNA was slightly, but significantly, increased in N2 at FAC 2.5 mg/ml (Figure 4C). However, at FAC 5.0 mg/ml, the relative expression of acs-20 mRNA was significantly increased to 3.9-fold in cyp-23A1(gk253) mutants compared to N2 (Figure 4C). Meanwhile, the increased expression of acs-20 in both N2 and cyp-23A1(gk253) mutant was significantly reduced in sgk-1(ok538) mutant background (Figure 4C), while the acs-20(tm3232);cyp-23A1(gk253) double mutants showed no response to iron-induced fat accumulation (Figure 4, A and D, Figure S2) and increased LD size (Figure S1, A and I). These findings suggest that sgk-1 upregulates acs-20 to promote fat accumulation under iron overload.
We then tested whether ACS-20 accounts for fatty acid transport under iron overload. Dietary supplement with BODIPY-C12 fatty acid showed that the fluorescence did not obviously increase with FAC concentration at 5.0 mg/ml in acs-20(tm3232) single and acs-20(tm3232);cyp-23A1(gk253) double mutants, unlike in N2 and the cyp-23A1(gk253) mutants (Figure 4E). Consistent with this finding, dietary supplementation of C18:1(n-9) further confirmed that unlike N2 and cyp-23A1(gk253) mutants, both acs-20(tm3232) and acs-20(tm3232);cyp-23A1(gk253) worms did not displayed increased levels of C18:1(n-9) in response to iron overload (Figure 4F), suggesting that the acs-20 mutants significantly suppressed the increased fatty acid uptake in both N2 and cyp-23A1(gk253) mutants. These results further confirmed that the fatty acid transporter ACS-20 accounts for iron facilitated lipid uptake under SGK-1 regulation.
VIT-2 translocates dietary lipids to lipid droplets
When taken up into cells, lipids are translocated into intracellular organelles for further utilization or storage in lipid droplets. As mentioned above, iron overload promotes lipid uptake for storage. We hypothesized that intracellular lipid transporters may be necessary to translocate lipids to lipid droplets for storage, so we searched for such an intracellular lipid transporter under iron overload. C. elegans genome contains six genes, vit-1–6, predicted to encode vitollogenins, which are homologous to the protein moieties of mammalian lipoproteins such as LDL (Baker 1988), which may have intracellular as well as intercellular lipid transport activity (Spieth and Blumenthal 1985; MacMorris et al. 1994; Grant and Hirsh 1999). We found that the relative mRNA expression of vit-2 and vit-3 transcripts was almost undetectable in sgk-1(ok538) mutant compared to N2, even under iron overload (Figure 5A). Meanwhile, we crossed the vit-2::gfp (RT130, pwIs23) transgenic strain into sgk-1(ok538) mutant, and found that VIT-2::GFP was undetectable by fluorescent microscopy. These results indicate that sgk-1 positively regulates the expression of vit-2 and vit-3.
Next, we treated vit-2::gfp (RT130, pwIs23) worms with FAC and found that the fluorescent intensity of VIT-2::GFP (Figure 5B) and the expression level of VIT-2 protein (Figure 5C) was markedly increased at FAC 5.0 mg/ml as compared to the control. The expression of VIT-2::GFP is also brighter in cyp-23A1(gk253) mutant, but not in acs-20(tm3232) mutant compared to N2 under both FAC treatment and no FAC (Figure S9). VIT-2 is a yolk lipoprotein, but it has also been reported to be a C. elegans lipid droplet-associated protein (Zhang et al. 2012; Vrablik et al. 2015). After isolating lipid droplets and other cellular fractions of vit-2::gfp worms treated with FAC 0 and 5.0 mg/ml, Western blot analysis using anti-GFP antibody showed that the level of VIT-2 protein was significantly increased in LD, total membrane (TM), and postnuclear supernatant (PNS) under FAC 5.0 mg/ml compared to FAC 0 mg/ml (Figure 5, D and E). Taken together, these results demonstrated that iron overload activates the expression of VIT-2, and this may promote lipid translocation to lipid droplets.
Since iron overload was able to activate the SGK-1-dependent expression of vit-2 and vit-3 (Figure 5A), we consequently hypothesized that worms depleted of vit-2 or vit-3 might be resistant to iron overload, similar to sgk-1(ok538) mutants. Indeed, compared to N2, lipid abundance did not increase in vit-2(ok3211) or vit-3(ok2348) mutants (Figure 5, F and G, Figure S2) in response to iron overload. Additionally, unlike N2, the lipid droplet size of vit-2(ok3211) did not change under iron overload (Figure 5H, Figure S1, A and J). However, both vit-2(ok3211) and vit-3(ok2348) mutant still displayed increased level of oleic acid [C18:1(n-9)] (Figure 5I). Taken together, these results suggest that VIT-2/3 may function to transport dietary lipids to lipid droplets for storage in response to iron.
Discussion
The correlation between iron overload and obesity is not mechanistically understood. Using a genetic screen to reveal genes required for the induction of ferritin as well as changes in lipid metabolism upon exposure to iron overload, we found that the serum and glucocorticoid-inducible kinase SGK-1 plays a central role to synergistically regulate iron and lipid homeostasis.
SGK1 belongs to the family of AGC kinases, is regulated by numerous different factors, and participates in a variety of cellular and physiological processes (Lang et al. 2009; Lang and Stournaras 2013). The transcriptional expression of SGK-1 has been previously reported to be downregulated by iron restriction in rat intestine (Marzullo et al. 2004) and kidney of chronic kidney disease in a rat model (Naito et al. 2012). Here, we consistently observed that iron restriction by the iron cheater BP represses the expression of SGK-1::GFP (MQD862, N2;Psgk1::sgk-1::gfp) in C. elegans (Figure S5). Meanwhile, iron overload induces expression of both the transcriptional and translational SGK-1::GFP reporters (Figure 1G), as well as the level of sgk-1 mRNA (Figure 1F, Figure 2C) and protein (Figure 2, E and F) in N2 and in cyp-23A1 mutants. Additionally, the sgk-1(ok538) mutation successfully suppressed the induced expression of ferritin in both N2 and in the cyp-23A1(gk253) mutants. Our present findings, along with work by others, offer consistent evidence of an evolutionarily conserved function of SGK-1 in response to iron to regulate iron homeostasis.
In mammals, SGK-1 has been implicated into glucose and lipid metabolism, as well as several metabolic diseases (Lang et al. 2009; Lang and Stournaras 2013). A common SGK1 gene variant is associated with increased blood pressure, obesity, and type 2 diabetes (Schwab et al. 2008). Upregulation of intestinal SGK1 expression may partially stimulate the Na+-coupled glucose transporter SGLT1, consequently leading to fat deposition and obesity (Fujita et al. 1998; Dieter et al. 2004; Lang and Stournaras 2013). SGK1 is expressed in white adipose tissue and its levels are induced during the conversion of preadipocytes into fat cells by direct phosphorylation of Foxo1 (Di Pietro et al. 2010). In C. elegans, several reports showed that SGK-1 regulates development, stress response, longevity, lipid metabolism (Hertweck et al. 2004; Jones et al. 2009; Soukas et al. 2009), and more recently, membrane trafficking (Zhu et al. 2015). Here we show that the gain-of-function mutant sgk-1(ft15) (Figure 1F), sgk-1::gfp (MQD862, N2;Psgk-1::sgk-1::gfp) (Figure 1F), and cyp-23A1(gk253) mutants (Figure 2C) displayed increased expression of sgk-1 and correspondingly higher fat content than wild-type N2 in both the absence and the presence of iron. These findings provide strong evidence that SGK-1 plays a crucial role in regulating lipid metabolism from C. elegans to mammals, which in turn further supports the usefulness of C. elegans to serve as a model to investigate the role of sgk-1 in lipid metabolism.
Lipid uptake and translocation are known to require lipid transporters. The C. elegans genome contains 22 genes known to encode acyl-CoA synthetase/fatty acid transporter (Zhang et al. 2013), of which ACS-20 is required for incorporation of exogenous very long chain fatty acids into sphingomyelin (Kage-Nakadai et al. 2010). While ACS-22 may functionally overlap with ACS-20 (Kage-Nakadai et al. 2010), we found that only ACS-20 is required for the increased fat accumulation under iron overload, suggesting that ACS-20, like SGK-1, regulates lipid hemostasis under iron overload. Our results also showed that iron has the ability to induce the transcriptional expression of acs-20, especially in cyp-23A1(gk253) mutants (Figure 4C). Interestingly, the increased expression of acs-20 mRNA in N2 and cyp-23A1(gk253) mutants was positively correlated to the high absorption of BODIPY-C12 fatty acid and oleic acid [C18:1(n-9)] (Figure 4F). In contrast, the acs-20(tm3232) mutants dramatically repressed the highly increased fat accumulation (Figure 4, B and D) as well as lipid uptake (Figure 4F) of both N2 and the cyp-23A1(gk253) mutant. These findings strongly suggest that ACS-20 may function as fatty acid transport to uptake lipids to store under iron overload. Since the increased expression of acs-20 in both N2 and cyp-23A1(gk253) mutant relied on sgk-1, we further propose that iron induces the expression of sgk-1, which in turn further activates acs-20 expression to uptake lipid into intestine for storage (Figure 6). This process may also explain the uncharacterized function of yeast Ypk1 in the regulation of fatty acid uptake, though clearly more follow-up is needed for confirmation.
C. elegans vit-1–6 belong to the family of vitellogenin genes that encode the protein moiety of yolk lipoproteins (Spieth and Blumenthal 1985; MacMorris et al. 1994; Grant and Hirsh 1999). While previous work has shown that yolk lipoproteins are synthesized in the intestine and transferred to developing oocytes using receptor-mediated endocytosis (Grant and Hirsh 1999), the roles of lipoproteins in intracellular lipid transfer have not previously been characterized. In our study, we found that vit-2 and vit-3 expression were dramatically repressed in the sgk-1(ok538) mutants (Figure 5A), indicating that SGK-1 is required for normal expression of vitellogenin genes. Furthermore, we found that vit-2(ok3211) and vit-3(ok2348) mutants were not responsive to iron-induced fat accumulation (Figure 5, F and G, Figure S2). Although the reduced fat accumulation may be due to reduced yolk production, an alternative interpretation is that vitellogenins play a role in intracellular lipid transport. VIT-2 was recently identified as a lipid droplet-associated protein (Zhang et al. 2012; Vrablik et al. 2015), and here we found that iron promoted the accumulation of VIT-2 protein in LDs (Figure 5, D and E). Meanwhile, the expression of VIT-2::GFP is also brighter in cyp-23A1(gk253) mutant, but not in acs-20(tm3232) mutants compared to N2 regardless of treatment with FAC (Figure S9). However, unlike the sgk-1(ok538) and acs-20(tm3232) mutants, the vit-2(ok3211) mutants still responded to iron-promoted oleic acid uptake (Figure 5I), suggesting that VIT-2 may function as a lipid transporter to facilitate intracellular lipid translocation to lipid droplets for storage under conditions of iron overload.
Cytochromes P450 are hemoproteins containing a heme cofactor and are known to be involved in numerous processes, such as lipid metabolism and detoxification (Pikuleva and Waterman 2013). Within 77 of 81 tested cytochrome P450 genes, we found that only cyp-23A1, like sgk-1, synergistically, but oppositely, regulated iron and lipid metabolism. CYP-23A1 is a homolog of the human gene CYP7B1 (Kuwabara and O’Neil 2001), which hydroxylates carbons 6 and 7 of the B-ring of oxysterols and steroids (Stiles et al. 2009). Mutations in the CYP7B1 gene are a cause of neonatal hemochromatosis (OMIM:231100) (Jimenez-Sanchez et al. 2001), characterized by hepatic failure in the newborn period and heavy iron staining in the liver. Remarkably, the C. elegans cyp-23A1(gk253) mutant also consistently displayed high expression of FTN-1::GFP and ftn-1 mRNA under iron overload, indicating CYP23A1, like CYP7B1, has an evolutionarily conserved role regulating iron homeostasis. Interestingly, sgk-1(ok538) mutants successfully suppressed the highly elevated expression of FTN-1::GFP and ftn-1 mRNA, as well as fat accumulation in the cyp-23A1(gk253) mutants under iron overload, which may offer a new approach to investigate CYP7B1-related diseases.
In summary, this study indicates that iron overload has the capacity to synergistically upregulate the level of ferritin and fat accumulation in an intact organism, C. elegans, thereby offering experimental evidence supporting a previously unverified link between iron and obesity. Although more research is necessary to elucidate all of the molecular mechanisms that participate in this complex process, by taking the advantage of genetics in C. elegans, we were able to characterize a novel response to iron overload, which induces the expression of sgk-1 to facilitate lipid uptake via fatty acid transporter ACS-20, and to translocate fats to lipid droplets for storage via VIT-2/3 (Figure 6). In this pathway, SGK-1 plays a central role in coordinating the regulation of ferritin levels and lipid accumulation in response to iron overload. Since SGK1 knockout mice display no obvious physiological defects (Wulff et al. 2002), targeting of SGK1 may hold promise in treating diseases related to iron and lipid metabolism.
Acknowledgments
We express our deep gratitude to Jennifer L. Watts for helpful comments on the manuscript; Mengqiu Dong for kindly providing the sgk-1::gfp (MQD862, N2;Psgk1::sgk-1::gfp) worm strain; and Shouhong Guang for EG4322, pCFJ151, pJL43.1, pCFJ104, and pCFJ90 constructs. Some strains were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440). This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB13030600), the National Natural Science Foundation of China (31460268, 31160216, 31171134, and U1202223), the Yunnan Natural Science Foundation (2013FA042 and 2011FZ179), the Yunnan Provincial Science and Technology Department (2014HB022), the Yunnan Overseas High-Level Talents Program (2015HA040 to B.L.), and the State Key Laboratory of Genetic Resources and Evolution (GREKF13-03).
Author contributions: B.L., H.W., X.Z., and X.J. conceived and designed the experiments, H.W., X.J., X.Z., J.W., L.Z., and Y.Z. performed the experiments; B.L., H.W., X.J., and X.Z. analyzed the data; H.W., X.J., L.Z., X.Z., J.H., and B.L. contributed reagents/materials/analysis tools; and B.L., H.W., and X.Z. wrote the paper. The authors declare that they have no competing interests.
Footnotes
Communicating editor: O. Cohen-Fix
Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.116.186742/-/DC1.
Literature Cited
Author notes
These authors contributed equally to this work.