Hephaestin and ceruloplasmin facilitate iron metabolism in the mouse kidney

Multicopper ferroxidases (MCFs) play an important role in cellular iron homeostasis. However, the role of MCFs in renal metabolism remains unclear. We used Hephaestin (Heph) and Ceruloplasmin (Cp) single or double (Heph/Cp) knockout (KO) mice to study the roles of MCFs in the kidney. Renal iron levels and the expression of iron metabolism genes were examined. The non-heme iron content both in the renal cortex and medulla of Heph/Cp KO mice was significantly increased. Perls’ Prussian blue staining showed iron accumulation on the apical side of renal tubular cells in Heph/Cp KO mice. A significant increase in ferritin protein expression was also observed in the renal medulla and cortex of Heph/Cp KO mice. Both DMT1 and TfR1 protein expression were significantly decreased in the renal medulla of Heph/Cp KO mice, while the expression of DMT1 protein was significantly increased in the renal cortex of these animals. Significant increase in proteinuria and total urinary iron was observed in the double knockout mice, and this was associated with compromised structural integrity. These results suggest that KO of both the HEPH and CP genes leads to kidney iron deposition and toxicity, MCFs could protect kidney against a damage from iron excess.

Multicopper ferroxidases (MCFs) are known to play a central role in iron nutrition and homeostasis in organisms ranging from yeast to humans 1 . The vertebrate MCFs, Ceruloplasmin (CP), Hephaestin (HEPH), and Zyklopen (ZP), are hypothesized to facilitate iron transport in diverse tissues by oxidizing ferrous iron to the ferric form so that it can bind to the circulating ferric iron carrier transferrin (TF). In these reactions, electrons are transferred from ferrous iron to the type I copper sites of MCFs. The electrons are then transferred to the MCF type II/type III copper site, where molecular oxygen is reduced to water 2 . Our previous work demonstrated that HEPH is located in a supra-nuclear compartment and on the basolateral membrane of intestinal enterocytes 3 . Apical iron administration can lead to the mobilization of HEPH from intracellular sites to the basolateral membrane 4 . HEPH is most strongly expressed in the small intestine, but it has also been found to be expressed in the kidney 5 . CP is mainly found as a soluble serum protein originating from the liver, but it is also found as a glycosylphosphatidylinositol (GPI)-linked protein in astrocytes and the kidney 6 . Mutations in the Cp gene lead to iron accumulation in multiple tissues in humans 7 and mice 8 .
The role of MCFs and a range of other proteins involved in iron metabolism in iron transport in the kidney remains poorly understood [9][10][11] . In general, over 99% of the iron filtered by the glomeruli is reabsorbed 9 . Wareing et al. suggested that iron was reabsorbed by late proximal tubule segments and early distal tubule segments 12 . Plasma TF (Mr 78,000) can pass through the glomerulus filter 13,14 and be reabsorbed in the proximal tubule via transferrin receptor 1 (TFR1). TFR1 is highly expressed in the proximal tubules of the renal cortex, and is also found on the apical membrane of collecting tubules and distal tubules in the medulla 10 . The TF-TFR1 complex is internalized into endosomes via cubilin-mediated endocytosis, and can be targeted to lysosomes for degradation 15 . Divalent metal-ion transporter 1 (DMT1) is localized to the endosomes and lysosomes within proximal tubule cells, organelles associated with the processing of apically sequestered TF 16 . It is predicted that iron is released from TF at these sites and exported into the cytosol by DMT1 16 . These findings suggest that some TF normally enters the glomerular filtrate, but it is retrieved by specific receptor-mediated uptake in the kidney tubular system. Iron released into the cytoplasm of renal tubular cells is subsequently exported across the basolateral membrane via ferroportin1 (FPN1) 17 . We hypothesize that this is the step which is likely to involve the MCFs as HEPH in the gut and CP in other tissues have been shown to increase the efficiency of FPN1-mediated iron transport 8,18,19 . DMT1 is also found on the apical membrane of distal tubules, where it can resorb iron from the tubular fluid.
Iron accumulation in the kidney is found in a number of diseases, including Fanconi syndrome, Dent's disease and chronic renal disease 14,20,21 . High kidney iron is generally considered to be deleterious to renal function [20][21][22] due to its propensity to catalyze the formation of reactive oxygen species 23 , which have the capacity to cause cellular injury. Recent clinical studies indicate that chelation of iron in the kidney has beneficial effects on the course of chronic kidney disease 24 .
In this study we have used Cp and Heph single, and Cp/Heph double knockout mice to examine the role of these MCFs in the kidney. These studies showed that knockout of both the Heph and Cp genes, but not either alone, leads to kidney iron deposition and toxicity. These findings suggest that MCFs may play an important role of protecting kidney against damage from iron excess and that either oxidase is able to compensate for the loss of the other.

Results
Iron status at 6 months of age. To evaluate the iron status of the Heph/Cp KO mice, hematological data were obtained from mice at six months of age. Heph/Cp KO and Heph KO mice, but not Cp KO, were anemic based on reductions in hemoglobin (Hb), mean cell volume (MCV), red cell number and hematocrit (Table 1). Heph/Cp KO mice were significantly more anemic than Heph KO animals. Iron levels were significantly lower in the plasma of Heph/Cp KO, Cp KO, and Heph KO mice compared to WT control mice, while the hepatic iron concentration was significantly higher in Heph/Cp KO and Cp KO mice compared to WT control and Heph KO mice ( Table 2). In the kidney, the iron concentration was only significantly increased in Heph/Cp KO mice compared to the other groups, and not in either of the single gene KOs (Table 2). We then examined the non-heme iron content in the renal cortex and medulla, and found that non-heme iron levels were significantly higher in both the renal cortex and medulla of Heph/Cp KO mice compared to the other genotypes examined (Table 2).
Iron distribution and ferritin expression in the renal medulla and cortex. Iron accumulation was observed by Perls' Prussian blue staining in the renal medulla of Heph/Cp KO mice, but no iron was observed in the corresponding region in the kidneys of WT, Cp KO or Heph KO mice (Fig. 1A). Some iron accumulation was also observed in the renal cortex of Heph/Cp KO mice, but not in mice with other genotypes (Fig. 1B). At higher magnification (1000X, Fig. 1A,B right hand panels), iron was observed intracellularly, but on the apical side of renal tubular cells and the Loop of Henle in Heph/Cp KO mice. Consistent with this elevated iron, western blotting showed that ferritin protein expression was increased in the renal medulla (Fig. 1C,E) and cortex (Fig. 1D,F) of Heph/Cp KO mice, but not in mice of any of the other genotypes All levels were normalized to tubulin and then normalized to ferritin expression in WT mice.
Expression of iron transport proteins in the renal medulla and cortex. Both HEPH and CP protein were detected in the renal medulla and cortex of WT mice ( Fig. 2A,B). As expected, no CP protein expression was detected in the kidneys of Heph/Cp KO or Cp KO mice, and no HEPH protein expression in Heph/Cp KO or Heph  . Fpn1 and Dmt1-IRE mRNA expression levels were not significantly different between any of the mouse groups examined (Fig. 4E,G,I,K). Dmt1 + IRE mRNA expression in the renal medulla, however, was significantly lower in Heph/Cp KO mice compared to WT mice (Fig. 4F), and Dmt1 + IRE mRNA expression in the renal cortex was significantly higher in Heph/Cp KO mice compared to all other groups (Fig. 4H). Tfrc mRNA expression in the renal medulla was significantly lower in Heph/Cp KO and WT mice compared to Heph KO and Cp KO mice, and its expression in Heph/Cp KO medulla was significantly lower than that of WT mice (Fig. 4J).
In contrast, Tfrc mRNA expression in the renal cortex was significantly higher in Heph/Cp KO mice compared to the other groups (Fig. 4L).
Urinary iron, protein content and HE, PAS staining in the renal medulla and cortex. Renal function was assessed by testing total urine iron and protein concentrations. Total iron content was measured in the urine of four group mice. Data were normalized to urine CR and expressed as microgram iron per milligram CR. The urinary iron and creatinine (CR) ratios were significantly increased in Heph/Cp KO mice compared to the other groups, and were significantly decreased in Heph KO mice compared to the other groups (Fig. 5A). The urinary protein and CR ratios were significantly increased in Heph/Cp KO mice compared to the other groups (Fig. 5B). The main renal injury location was in renal tubules in Heph/Cp KO mice. Morphological examination was performed by HE and PAS staining of renal medulla and cortex in four group mice, the results showed that renal tubular cells of medulla were dilated and nucleus shedding was obvious, and there was structural injury to the renal medulla in Heph/Cp KO mice. The number of renal tubular cells of medulla was lower in Heph/Cp KO mice compared to other three groups (Fig. 5C,E), but Morphological examination showed that there are no significant histological changes in the renal cortex of four group mice (Fig. 5D,F).

Discussion
Our current study demonstrated that the iron concentration increased in the liver and decreased in the serum in Cp single and Heph/Cp double KO mice, and only increased in the kidney of Heph/Cp double KO mice. In Heph KO mice, however, there was a reduction in iron levels in plasma, liver and kidney (Table 2), as well as decreased urinary iron excretion. These data suggest that Heph KO mice have a systemic iron deficiency. Heph/Cp double KO mice had a low plasma iron content with increased total urinary iron, suggesting decreased renal iron reabsorption, which in turn probably induced renal injury. Previous research showed that mice with a combination of Cp loss and sla mutant Heph had high iron concentrations in the liver, brain, heart and pancreas 25 . The kidney was not examined. CP is mainly found as a soluble serum protein originating from the liver, but it is also found as a membrane-bound GPI-linked protein. The GIP-Cp first described to be present on the surface of astrocytes 26 , and current reported on the plasma membrane of spleen and kidney 6 . Our previous work demonstrated that HEPH is located in a supra-nuclear compartment and on the basolateral membrane of intestinal enterocytes 3 . Apical iron administration can lead to the mobilization of HEPH from intracellular sites to the basolateral membrane in MDCK cell line 4 . But the location of CP and HEPH in the kidney is still unclear. Our work shows that the deletion of either CP or HEPH has little impact on the kidney, but deletion of both MCFs leads to kidney iron deposition and this in turn impacts kidney function. These results suggest that there is redundancy in the role played by MCFs in renal iron homeostasis, with either single oxidase being able to sustain normal kidney iron metabolism when the other is lost. CP and HEPH are MCFs that are known to work in concert with FPN1 to increase the efficiency of cellular iron export 27 . Existing data suggest that HEPH is the major oxidase involved in iron release from intestinal enterocytes 18 , while CP facilitates iron export from other tissues 8,19 . Our current work demonstrates that Cp, Heph, and Fpn1 mRNA and protein are all present in the mouse kidney, consistent with previous studies 5, 27 , and iron accumulation in the renal tubular cells and those in the Loop of Henle of Heph/Cp KO mice suggests that the oxidases are functional in this organ. Perls' staining showed iron accumulation in both the renal medulla and cortex, but iron levels were highest in the medulla. The levels of ferritin protein expression in the medulla and cortex were consistent with elevated intracellular iron in these regions, and maximum iron accumulation in the medulla. The greater number of tubular cells (including distal tubular cells and those of the loop of Henle) in the renal medulla than in the renal cortex may explain this difference in iron load. Previous reports have suggested  that most iron is reabsorbed by the proximal tubules, but the distal tubules also play a role in this process 12 . Our current data support the idea that the loop of Henle and the distal tubule also re-absorb iron. Compared with the renal medulla, the renal cortex had more glomerular and proximal tubular cells, the protein level of DMT1 was up-regulated in renal cortex of Heph/Cp KO mice may help to increase the iron transport and re-absorb when iron accumulated in the renal medulla of Heph/Cp KO mice, which may explain why the DMT1 protein was increased in the Cp/Heph KO mice cortex. Consistent with this, the cortex of the Heph/Cp KO mice had increased transferrin receptor and Dmt1 mRNA levels.
This region is likely to be particularly important in conceiving iron that is not re-absorbed by the tubular cells of the cortex it induces part of cortex proximal tubular cells to up-regulated DMT1 protein levels to reabsorb urinary iron, which may explain why the DMT1 protein was increased in the Cp/Heph KO mice cortex. Consistent with this, the cortex of the Heph/Cp KO mice had increased transferrin receptor and Dmt1 mRNA levels.
A small percentage of TF-bound iron is filtered through the glomeruli of the kidney 14,15 , and this iron is then reabsorbed by the proximal tubules 16 , Loop of Henle, and distal tubular cells 12 , likely still bound to TF. If these cells are unable to export iron across their basolateral membrane, intracellular iron accumulation would be expected. The apical supranuclear location of HEPH in MDCK and Caco2 cells, as well as results from colocalization studies in Cos7 cells, suggests that HEPH can localizes to recycling endosomes 4 . However, in vivo, in polarized mouse intestinal enterocytes, HEPH is present both on the lateral surfaces of enterocytes and also in the apical supranuclear region 3 , so precisely where it acts is unclear. If similar machinery is present in the kidney to enable vectorial transport of iron, relocalization of HEPH from a supranuclear compartment to the basolateral membrane may play a role in moving iron from the nephron lumen into the circulating blood, likely in conjunction with FPN1. Fpn1 mRNA and protein levels in the kidney did not vary significantly between mice of different genotypes, suggesting that FPN1 protein levels are insensitive to kidney iron accumulation. Nevertheless, the efficiency of export can still be enhanced by MCFs.
Double knockout of Heph and Cp resulted in changes in the expression of other genes involved in iron metabolism. In the renal medulla, where ferritin protein expression and iron levels were significantly increased, Dmt1 + IRE and Tfrc mRNA levels, and DMT1 and TFR1 protein expression were significantly reduced. Recent studies suggest that DMT1 is localized to the late endosomal and lysosomal membranes of proximal tubule cells, where it facilitates the uptake of TF-bound iron 16,28 . Dmt1 + IRE and Tfrc mRNA levels can respond to the local iron status of the renal medulla, presumably through IRP-IRE-mediated mechanisms 29,30,31 . Wareing et al. 28 have reported that rats fed an iron deficient diet had significantly increased renal DMT1 expression, while rats fed an iron-enriched diet showed a significant decrease in renal DMT1 expression. Thus renal DMT1 expression is sensitive to dietary iron content. Dmt1 + IRE mRNA contains a single IRE in its 3′ -untranslated region, similar to those in the Tfrc mRNA, and this element confers iron-dependent mRNA stabilization 30 . We hypothesize that lack of Heph/Cp results in reduced iron export from renal cells, which in turn results in cellular iron overload and decreased Dmt1 and Tfrc expression.
We propose that the role played by HEPH or CP protein in the kidney is similar to that of other cells types. Under physiological conditions, iron reabsorption is mediated through the uptake diferric TF by TFR1 and/or cubilin on the apical surface of the epithelial cells in the renal tubules 10,15 . Transferrin-bound iron is transported into endosomes, where ferric iron dissociates from transferrin and is reduced to the ferrous state, possibly by STEAP3 32,33 . Iron is then exported from the endosomes into the cytoplasm by DMT1 16 . This iron can be utilized by cell itself, or exported back into the circulation via FPN1, a process in which HEPH and CP participate. Lack of HEPH and CP would be expected to block this final iron export step. Our current result has showed that iron reuptake is stuck in the apical side of renal tubular cells. It could be explained that a defective iron export at the basolateral membrane could block the mobilization of HEPH (maybe CP) from intracellular sites to the basolateral membrane, and cause stopping of iron transport at the apical side. For understanding a detailed process, the further study is needed.
Excess intracellular iron is toxic and prolonged exposure can lead to the generation of reactive oxygen species and tissue damage. Our demonstration that mice lacking both MCFs showed significant increase urinary iron excretion, proteinuria, as well as morphological abnormalities in the renal tubules suggests considerable structural and functional damage to the kidney. Together, these findings suggest that MCFs play an important role in protecting the kidney against injury.
In summary, our work has provided further insights into iron transport pathways in the kidney and has shown that copper-dependent ferroxidases play an important role in the homeostasis of this essential organ.

Methods
Mice. Heph KO and Cp KO mouse strains have been previously described 8,34,35 . The mice were originally obtained from the laboratory of Dr Greg Anderson at the QIMR Berghofer Medical Research Institute (Brisbane, Australia) and were bred and maintained at the Medical School of Nanjing University 36 . Heph −/− (Heph KO) mice were bred to Cp −/− (Cp KO) mice to generate Heph −/− /Cp −/− (Heph/Cp KO) mice. All mice were on the C57BL/6J genetic background and were allowed unlimited access to a standard rodent unpurified diet containing approximately 180 mg/kg iron provided by Jiangsu Province Collaborative Medical Bioengineering Co., Ltd 37 . All animal studies were carried out in accordance with NIH guidelines, as described in the Guide for the Care and Use of Laboratory Animals of the NIH, and were approved by the Institutional Animal Care and Use Committee of Nanjing University.
Tissue preparation. Male mice from each litter were euthanized at 6 months of age after anesthesia with an intra-peritoneal injection of 3% sodium pentobarbital. Blood was collected by cardiac puncture using potassium EDTA tubes (catalogue no. 367841, Becton Dickinson and Company, Franklin Lakes) and the liver and kidney were removed. Whole blood was used for hematological analysis and centrifuged to provide plasma for iron measurement as previously described 38 . A piece of liver and the whole left kidney were dried by heating at 80-100 °C for 2 hours and used for subsequent iron analysis. Four mice in each genotype group were perfused via the heart, first with phosphate buffered saline (PBS), then with 4% Paraformaldehyde (PFA). Collected tissues were immediately fixed in 4% PFA solution. Tissues were embedded in paraffin and sectioned for future immunostaining experiments. The kidney was dissected to separate the cortex and medulla. Samples of each were snap frozen in liquid nitrogen and then stored at − 80 °C until required for RNA, protein, and iron concentration analyses.
Tissue iron staining and iron status measurements. For immunostaining experiments, tissue was embedded in paraffin, sectioned and stained for ferric iron using Perls' Prussian blue staining as previously described 35 . Other sections were detected with H&E or periodic acid Schiff (PAS) staining following the standard procedure. Blood hemoglobin and hematocrit levels were measured by an Automated Hematology Analyzer XS series (XS-800i Sysmex Corporation; Japan) in the Clinical Laboratory at Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School. The total concentration of iron in the dried tissues and plasma was measured using an Atomic Absorption Spectrometer (180-80, Hitachi; Japan) at the Modern Instrumental Analysis Center of Nanjing University after nitric acid digestion 38 . The concentration of non-heme iron in the renal cortex, renal medulla and total kidney was measured as previously described by a modification of the bathophenanthroline assay described by Torrance and Bothwell 39 . Data are expressed as μ mol Fe/g dry weight or μ mol Fe/L plasma.
Total RNA extraction and qRT-PCR analysis. Total RNA was isolated from mouse tissues using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Three micrograms of total RNA were reverse transcribed using a Transcriptor First stand cDNA Synthesis Kit (Roche Applied Science) and qRT-PCR was used to measure the expression of Dmt1 (the transcripts with and without an iron responsive element; Dmt1 + IRE and Dmt1-IRE respectively), Fpn1, Heph, Cp, and Tfrc in different parts of the kidney. All primers (Table 3) were designed by Primer3 software using the default settings 40 . Expression levels were normalized to that of the housekeeping gene GAPDH. Two-step PCR was monitored in real time by the FastStart Universal SYBR-Green Master (Roche Applied Science) according to the manufacturer's instructions on an Applied Biosystems 7300 Real-Time PCR System instrument (Life Technologies, Shanghai, China). Fluorescence curves were analyzed with the LightCycler software (version 3.5.3). Automated calculation of crossing points was carried out by the second-derivative maximum method. Each reaction showed a single specific peak in the corresponding melting curve.

Measurement of Urine iron, protein and Creatinine.
Base-line urine was collected in all group mice.
Total iron content in the urine of all group mice was measured using an atomic absorption spectrometer as previously described. Urinary protein concentration was measured using the bicinchoninic acid assay (Bioworld  Technology CO., MN) and urine creatinine (CR) levels were measured using the creatinine assay kit (Jian Cheng Biological Engineering Institute, Nanjing, China) according to the manufacture's instruction. Total urine iron content was normalized to urine CR and the urine protein excretion rate was expressed as the ratio of the urinary protein to urine CR.
Statistical analysis. Unless otherwise indicated, values are presented as means ± SEM. One-way ANOVA with Tukey's test for multiple comparisons was used to compare means for mouse tissue iron concentrations (i.e., liver, kidney, and plasma) and protein levels (i.e., L-ferritin, HEPH, CP, FPN1, DMT1 and TFR1) among the four groups. One-way ANOVA with Dunnett's test for multiple comparisons was used to compare means for gene expression (i.e., Heph, Cp, Fpn1, Dmt1 + IRE, Dmt1-IRE, and Tfrc) with WT control expression. Differences were considered statistically significant at P < 0.05. All statistical analyses were performed using GraphPad Prism 6 Software (GraphPad Software, San Diego, CA).