The H+-coupled polyligand transport protein divalent metal transporter 1 (DMT1) plays a key role in mammalian iron homeostasis. It has a widespread pattern of expression including tissues associated with iron acquisition and storage. Interestingly, it is also highly expressed in the kidney, yet its function in this tissue is unknown. The aim of this study was to determine the cellular location of DMT1 in proximal tubule cells as a first step to determining the role of this protein in the kidney. To do this we performed RT-PCR and immunostaining experiments using rat kidney and the S1 proximal tubule-derived WKPT-0293 Cl.2 cell line. RT-PCR revealed that mRNAs encoding all four DMT1 splice variants were present in RNA extracted from rat kidney cortex or WKPT-0293 Cl.2 cells. Immunostaining of rat kidney cortex or WKPT-0293 Cl.2 cells showed that DMT1 protein was expressed intracellularly and was not present in the plasma membrane. Expression of DMT1 partially colocalized with the late endosomal/lysosomal proteins LAMP1 and cathepsin-L. Using immunogold labeling, DMT1 was shown to be expressed in the membranes of late endosomes/lysosomes. Uptake of Alexa Fluor 546-transferrin was only observed following application to the apical membrane of WKPT-0293 Cl.2 cells. Within these cells, Alexa Fluor 546-transferrin colocalized with DMT1. In conclusion, renal proximal tubular cells express DMT1 in the membranes of organelles, including late endosomes/lysosomes, associated with processing of apically sequestered transferrin. These findings have implications for renal iron handling and possibly for the handling of nephrotoxic metals that are also DMT1 ligands, including Cd2+.
- renal tubular transport
divalent metal transporter 1 (DMT1), encoded by the SLC11a2 gene, has recently emerged as a key protein responsible for whole body and cellular iron balance (7, 13, 14, 16, 17). In the duodenum, DMT1 is expressed in the apical enterocyte membrane and mediates proton-coupled transport of ferrous iron from the intestinal lumen across the apical membrane and into the enterocyte. Iron then travels across the cell and is exported across the basolateral membrane via the iron-regulated transporter protein 1 (Ireg1; also known as ferroportin 1 or metal transporter protein 1), which, in turn, donates iron to plasma transferrin (1, 9, 22). The transmembrane copper-dependent feroxidase hephestin is also required for this process (32). DMT1 is also critical for cellular iron balance because it mediates export of iron from endocytic vesicles (15, 28). Following endocytosis of the transferrin/transferrin receptor complex, vesicles are acidified by a vacuolar H+-ATPase. This causes iron to dissociate from transferrin and also provides the outwardly directed proton gradient that energizes export of iron out of the vesicle via DMT1. This function has been well characterized in erythroid cells, where efficient transport of iron is constantly required for de novo heme synthesis. It has been also shown in other tissues and several cell lines of different origin and therefore seems to be ubiquitous. This can explain the widespread distribution of DMT1 (2).
Alternative splicing of the DMT1 gene produces proteins that differ with respect to their NH2 termini and COOH termini (14). Splicing in of exon 1A introduces an in-frame ATG start codon, and consequently variants containing exon 1A have 29–31 additional NH2-terminal amino acids compared with splice variants containing exon 1B (18). The number of amino acids added by the addition of exon 1A is species dependent and in rat exon 1A encodes 31 amino acids. COOH-terminal variants differ, dependent on the alternative splicing of two terminal exons: one contains an iron-response element (IRE) in the non-coding 3’-untranslated region (UTR), and this transcript and encoded protein are termed DMT1 IRE+; the alternative exon does not contain an IRE in the 3’-UTR and is accordingly known as the non-IRE form. Differences in the coding sequence of these exons result in two proteins with different COOH termini. Taking into account the different permutations resulting from alternate 5’ splicing, four isoforms can be generated and these are designated 1A-IRE+, 1B-IRE+, 1A-IRE−, and 1B-IRE−. These are depicted in Fig. 1. The existence of isoforms suggests that the protein may have additional, not yet described functions and subcellular locations.
Interestingly, DMT1 is highly expressed in the kidney cortex (6, 12, 17, 33). RT-PCR analysis of whole kidney indicates that transcripts containing exon 1A, exon 1B, and both 3’ exons, IRE+ and non-IRE, are present in the kidney (18), although exactly which of the four mRNA variants shown in Fig. 1 are present in the kidney has not been determined.
Using an affinity-purified antisera targeting an epitope common to all four known DMT1 isoforms (see Fig. 1), we have previously shown that DMT1 is expressed in rat kidney in an intracellular compartment within proximal tubular cells, although the identity of this compartment was unclear (12).
Studies in epithelial cell lines have identified the specific intracellular locations of DMT1. DMT1 in the Sertoli cell line TM4 was mainly found to be localized to recycling endosomes and to a lesser extent to the plasma membrane, although DMT1 was not detected in lysosomes (15). This localization is congruent with a role for DMT1 in export of iron out of recycling endosomes. In subsequent studies, stable transfection of LLC-PK1 with DMT1 IRE− revealed its presence in plasma membranes as well as in sorting and recycling endosomes (31). In the larynx carcinoma cell line HEp-2 and polarized Madin-Darby canine kidney (MDCK-II) cells, green fluorescent protein (GFP)-tagged DMT1 IRE+ localized to late endosomes and lysosomes, whereas GFP-tagged DMT1 IRE− was found to localize to subapical early endosomes, but not plasma membranes (27). Using a panel of peptide-targeted antisera against IRE+, IRE−, and 1A variants, Knöpfel and co-workers (19) reported plasma membrane expression of DMT1 isoforms in rat duodenum. Interestingly, immunostaining of both apical and basolateral plasma membranes was reported using an IRE+ targeted antiserum.
As a first step in determining the cellular function of DMT1 in the renal proximal tubule, we determined the exact cellular location of DMT1 in rat kidney cortex sections and WKPT-0293 Cl.2 cells, an immortalized cell line derived from the S1 segment of rat proximal tubule (34). To do this we used an affinity-purified antisera (DMT1-com) targeting an epitope common to all four known DMT1 isoforms (see Fig. 1) and a spectrum of complementary techniques including immunoperoxidase staining, immunogold electron microscopy, and immunofluorescence microscopy. We conclude that DMT1 is expressed in the late endosomal/lysosomal compartment associated with processing of luminally reabsorbed transferrin. These findings have implications for the renal handling of transferrin-bound iron and other metal-protein complexes possibly by receptor mediated endocytosis.
MATERIALS AND METHODS
An immortalized cell line from the S1 segment of rat proximal tubule (WKPT-0293 Cl.2) (34) was cultured as previously described (30, 34). Cells were passaged (passage number <60) twice a week on reaching confluency.
RT-PCR of rat kidney cortex.
Total RNA was extracted from snap-frozen dissected rat kidney cortex using the RiboPure method (Ambion). RT-PCR was performed with 1 μg of total RNA. cDNA was synthesized with oligo(dT) and Superscript II RT (Invitrogen, Paisley, UK). Forward and reverse primers spanning several introns were designed using Primer Express (Applied Biosystems, Warrington, UK) based on the GenBank sequences AF008439, AF029757, and CB728894. The primers were as follows: DMT1 IRE+ exon 1A isoform sense (5′-cgtccgatggggaagaagca-3′, AF008439 nucleotides 6–25) and antisense (5′-GCGGTCTCTAATTCTGCAATTC-3′, AF008439 nucleotides 2294–2314); DMT1 IRE− exon 1A isoform sense (5′ cgtccgatggggaagaagca-3′, AF029757 nucleotides 1656–1675) and antisense (5′-TGTTCAGGAGGTAGATTTCA-3′); DMT1 IRE+ exon 1B isoform sense (5′-CTCCTGGGATATGGGGTCGC-3′, CB728894 nucleotides 9–28) and antisense (5′-GCGGTCTCTAATTCTGCAATTC-3′); and DMT1 IRE− exon 1B isoform sense (5′-CTCCTGGGATATGGGGTCGC-3′) and antisense (5′-TGTTCAGGAGGTAGATTTCA-3′). Amplification was for 35 cycles of denaturation at 94°C for 30 s, annealing at 52–55°C for 30 s, and extension at 72°C for 2 min.
RT-PCR of WKPT-0293.Cl2 cells.
Total RNA was isolated from WKPT-0293.Cl2 cells using an RNeasy Mini Kit (Qiagen). RNA was treated with DNase I (Qiagen) for 15 min at room temperature. One microgram of RNA was reverse-transcribed with Omniscript RT (Qiagen), using oligo(dT) primers (Operon) according to the manufacturer’s instructions. RT-PCR was carried out with HotStar master mix (Qiagen). Thermal cycling was as follows: 1 cycle of initial HotStar polymerase activation at 95°C for 15 min followed by 1 cycle of extension at 72°C for 10 min, then 35 cycles at 94°C for 30 s, 54°C for 30 s, and then 72°C for 2 min for the following primer pairs: DMT1 IRE+ exon 1A isoform sense (5′-cgtccgatggggaagaagca-3′) and antisense (5′-GCGGTCTCTAATTCTGCAATT C-3′) and DMT1 IRE− exon 1A isoform sense (5′ cgtccgatggggaagaagca-3′) and antisense (5′-GGCACAAAAGGGCTTAGAGA-3′, AF029757 nucleotides 1738–1757). Amplification was for 35 cycles of 95°C for 30 s, 60°C for 50 s, and 72°C for 2 min for the following primer pairs: DMT1 IRE+ exon 1B isoform sense (5′-CGTGTCAGGAGGTGGTGGAG-3′) and antisense (5′-GCGGTCTCTAATTCTGCAATTC-3′); DMT1 IRE- exon 1B isoform sense (5′-CGTGTCAGGAGGTGGTGGAG-3′) and antisense (5′-GGCACAAAAGGGCTTAGAG-3′); and actin (GenBank accession no. NM007393; sense 5′-TGGAATCCTGTGGCATCCATGAAC-3′ and antisense 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′).
The DMT1 gene can generate four different proteins that share complete identity except at the extreme NH2 and COOH termini (Fig. 1). The anti-DMT1 antibody used (termed DMT1-com) was raised in rabbits to a peptide sequence, MVLCPEEKIPDDGASGDHGDSC, that is common to all four known isoforms and has previously been characterized (15). A dilution of 1:500 was used for immunoperoxidase staining of rat cortical sections. For immunofluorescence staining of WKPT-0293 Cl.2, the antibody was used at a 1:100 dilution. The monoclonal anti-cathepsin L clone 33/2 (catalog no. C2970, mouse ascites fluid, Sigma) and diluted 1:1200. Mouse monoclonal antibody to early endosome antigen 1 (EEA1; catalog no. 610456, BD Biosciences) was diluted 1:250. Secondary antibodies were Alexa Fluor 488 chicken anti-rabbit IgG (green) from Molecular Probes (catalog no. A21441, dilution 1:600), FITC-conjugated donkey anti-mouse IgG (green) from Jackson ImmunoResearch (catalog no. 715–095-151, dilution 1:50), and Cy3-conjugated donkey anti-rabbit IgG (red) from Jackson ImmunoResearch (catalog no. 711–165-152, dilution 1:600).
Male Wistar rats were killed by cervical dislocation after being anesthetized, and the kidney cortex was dissected and homogenized by Polytron in homogenization buffer (pH = 7.4) comprising 12 mM HEPES+300 mM mannitol, to which 1 μg/ml pepstatin, 2 μg/ml leupeptin, and 1 μg/ml polymethylsulfonyl fluoride were added immediately before use. The homogenate was spun at 1,000 g (relative centrifugal force) for 10 min at 4°C to get rid of the nuclear pellet consisting of unbroken tissue, whole cells, cell nuclei, and large debris. The postnuclear supernatant was then centrifuged at 17,000 g for 30 min at 4°C to yield a crude membrane pellet that was resuspended in homogenization buffer. The protein concentration was determined by the Bradford method according to the manufacturer’s instruction (Bio-Rad).
Aliquots of membrane fraction containing 20 μg of protein were mixed with Laemmli buffer and incubated for 15 min at 65°C before being subjected to 8% SDS-PAGE and transferred onto nitrocellulose transfer membrane (Schleicher & Schuell Bioscience, Dassel, Germany). The membranes were blocked in Tris-buffered saline-Tween 20 (TBST; 150 mM Tris·HCl, 1.5 M NaCl, 0.1% Tween 20) containing 5% fat-free milk for 30 min at room temperature and washed twice in TBST. They were then treated with previously affinity-purified DMT1-com antibody diluted 1:500 and incubated by rocking overnight at 4°C. Peptide-blocking experiments were also performed by incubating DMT1-com with 50 μg/ml of the target immunizing peptide.
The anti-DMT1 antibodies were removed by washing with TBST for 10 min twice. The peroxidase-conjugated goat anti-rabbit immunoglobulins (Dako) diluted 1:5,000 in TBST with 5% fat-free milk powder was used as a secondary antiserum. Signals were visualized using ECL Plus and Hyperfilm ECL (Amersham Pharmacia Biotech) according to the manufacturer’s protocol.
To obtain homogenate from WKPT-0293 Cl.2 cells, confluent monolayers were scraped with a rubber policeman, washed in homogenization buffer, as described above and sonicated on ice for 3 × 5 s at 10 A. Aliquots of homogenate containing 50 μg of protein were mixed with Laemmli buffer and incubated for 15 min at 65°C before being subjected to 7.5% SDS-PAGE and transferred onto polyvinylidene difluoride membranes overnight at 4°C. Blots were blocked with 3% nonfat dry milk and incubated overnight at 4°C with primary anti-DMT1-com (1:500) antiserum or with antiserum that had been preincubated with 50 μg/ml of antigenic peptide for 60 min at room temperature. Following incubation with horseradish peroxidase-conjugated secondary antibody (1:10,000) for 1 h at 4°C, blots were developed using Western Lighting Plus chemiluminescence reagents (PerkinElmer Life Sciences, Boston, MA), and signals were visualized on X-ray film.
Tissue preparation and immunohistochemistry.
Male Wistar rats were anesthetized, and kidneys were perfusion fixed with 3% paraformaldehyde (PFA)/PBS via the descending aorta. Kidneys were removed and postfixed in 3% PFA/PBS before being embedded in paraffin wax. Paraffin wax sections (5 μm) were cut using a Leica RM2135 rotary microtome (Leica Microsystems) and mounted onto Superfrost microscope slides.
Deparaffinized and dehydrated kidney sections were incubated in 1% H2O2 in methanol for 30 min to block endogenous peroxidase activity. Masked antigens were retrieved by microwaving the sections in TEG buffer (10 mM Tris, 0.5 mM EGTA, pH = 9.0) for 7 min at 800 W and 5 min at 400 W. Sections were left to cool down before treatment with 50 mM NH4Cl in PBS for 30 min. Sections were blocked with 1% BSA and incubated overnight at 4°C in a humidified chamber with DMT1 antibody diluted in 0.1% BSA with 0.3% Triton X-100 in PBS. Sections were rinsed with 0.1% BSA and then incubated with peroxidase-conjugated goat anti-rabbit immunoglobulins, diluted 1:200, for 1 h at room temperature. After rinsing in 0.1% BSA, the reaction was visualized using 3,3’diamino-benzidine-tetrahydrochloride (DAB). Sections were counterstained with Mayer’s hematoxylin, dehydrated, and coverslips were mounted with Eukitt (Kindler, Freiburg, Germany). Immunostaining was examined using a Zeiss Axiophot light microscope (Carl Zeiss), and photographs were taken using a Spot RT Color digital camera (Diagnostic Instruments).
Immunogold electron microscopy.
Rat kidneys were perfused retrograde through the descending aorta with 2 % formaldehyde in 0.1 M cacodylate buffer, pH 7.4, and postfixed in 1% formaldehyde in 0.1 M cacodylate buffer, pH 7, for 1 h. Cryosections (70–90 nm) were preincubated for 5 min with 0.1% SDS in 1% BSA, 0.05 M glycine in 0.01 M PBS. Then sections were washed three times and preincubated for 30 min in the same solution but without SDS. Subsequently, the sections were incubated with DMT1-com antibody (1:50) overnight and 10-nm gold anti-rabbit IgG conjugate (1:50) for 1 h. The sections were analyzed using a Philips P2 electron microscope equipped with an electronic camera.
Immunofluorescence staining of WKPT-0293 Cl.2 cells.
Cells (5 × 104) were grown for 48 h on glass coverslips. Unless otherwise indicated, all staining steps were performed at room temperature. Cells were washed three times for 1 min in PBS, fixed with 4% PFA/PBS for 30 min, washed three times in PBS for 5 min, permeabilized with 1% SDS/PBS for 15 min, washed three times in PBS for 5 min, blocked for 1 h in 1% BSA, washed for 1 min in PBS, incubated with primary antibody (DMT1-com diluted 1:100) overnight at 4°C, washed three times in PBS for 5 min, and then incubated with conjugated secondary antibody for 1 h. After three more washes in PBS for 5 min, cells were counterstained with 2′-(4-ethoxyphenyl)-5-(4-methyl-1-piperazinyl)-2,5′-bi-1H-benzimidazole, 3HCl (H-33342; 1 μg/ml), washed four times in PBS for 5 min, washed once in H2O, and then the coverslips were mounted onto glass slides with Dako fluorescence mounting medium (catalog no. S30238). Peptide-blocking experiments were also performed by incubating DMT1-com with 250 μg/ml of the target immunizing peptide. The cells were viewed using filters for Cy3, FITC, and DAPI with excitation/emission wavelengths of 545/610, 480/535, and 360/460 nm, respectively, using a mercury short-arc photo optic lamp HBO (103 W/2, OSRAM) as a light source, connected to a Zeiss Axiovert 200M microscope (Carl Zeiss) equipped with Fluar ×20, 0.75 UV and Fluar ×40, 1.3 oil-immersion objectives. Images were captured using a digital CoolSPAN ES CCD camera (Roper Scientific) and acquired, processed, and analyzed with MetaMorph software (Universal Imaging). Cy3 (red), FITC (green), and DAPI (blue) planes of focus were selected, and images were merged using the Color Combine function in the MetaMorph software.
Internalization of Alexa Fluor 546 transferrin conjugate in WKPT-0293 Cl.2 cells.
To study apical and basolateral uptake of transferrin, 5 × 104 WKPT-0293.Cl2 cells were grown to confluence under the conditions described above on 2-compartment filters with 24 wells on each plate (catalog no. 3470, Costar Transwell-Clear, Corning) with a 0.4-μm pore size and a 0.33-cm2 surface area. Measurements of transepithelial electrical resistance (TEER) were performed using an epithelial voltameter EVOM with an STX2 electrode (World Precision Instruments). Cells (1 × 104) were seeded in each well and allowed to grow for 2 days before the start of TEER measurements. The electrical resistance readings were undertaken at room temperature. Actual resistance readings were corrected for the blank resistance of the filters by subtracting the resistance of the filters without the cells. The blank reading was 7.7 ± 1.4 Ω (means ± SE; n = 17) and was subtracted from the resistance reading. The resistance values were converted into Ω × cm2. Experiments with transferrin were started 5–6 days after seeding, when TEER reached a stable value of 103.3 ± 3.4 Ω × cm2 (means ± SE; n = 10) (data not shown). Human serum transferrin coupled to Alexa Fluor 546 (catalog no. T23364, Molecular Probes; 50 μg/ml serum-free cell culture medium) was then added to either the apical or basolateral compartment of the Transwell-Clear culture dish for 5–15 min at 37°C, washed once for 3 min on ice in PBS, incubated with 50 mM EGTA (pH 7.3 with NaOH) in PBS for 60 min on ice (this procedure allowed removal of basolaterally applied transferrin that had accumulated in the intercellular space), washed once for 1 min at 4°C in PBS, fixed with 4% PFA/PBS for 30 min, washed three times in PBS for 5 min, and washed once in H2O. Transwell-Clear filters were mounted onto glass slides, embedded in Dako fluorescence mounting medium, and a glass coverslip was fixed on top of the Transwell-Clear filter. The cells were viewed using a filter for Cy3 as described above.
For colocalization studies, 5 × 104 cells were grown for 48 h on glass coverslips. The cell culture medium was aspirated and replaced by serum-free medium containing 25 μg/ml Alexa Fluor 546 transferrin conjugate and incubated for 20 min at 37°C or 60 min at 37°C. Coverslips were washed seven times for 3 min in ice-cold PBS, fixed with 4% PFA/PBS for 30 min, and processed further for colocalization studies with antibodies against DMT1, EEA1, or cathepsin L as described above.
To test for the presence of DMT1 mRNA splice variants in the kidney cortex, we performed RT-PCR on rat kidney cortex RNA using the specific primer sets shown in Fig. 1. cDNAs corresponding to DMT1 1A-IRE+ (2.3 kb), DMT1 1A-IRE− (1.8 kb), DMT1 1B-IRE+ (2.3 kb), and DMT1 1B-IRE− (1.8 kb) were amplified from kidney cortex RNA (Fig. 2). No bands were detected in reactions without cDNA (negative controls). The sizes of these products were as predicted from corresponding GenBank sequences. Subsequent nucleotide sequencing and Southern blot analysis confirmed that the products were correct (data not shown). In the WKPT-0293 Cl.2 cell line, RT-PCR products corresponding to DMT1 1A-IRE+ (2.3 kb), DMT1 1A-IRE− (1.8 kb), DMT1 1B-IRE+ (2.3 kb), and DMT1 1B-IRE− (1.8 kb) were also detected.
These results show that mRNAs encoding the four DMT1 splice variants are present in rat kidney cortex and also in the cell line WKPT-0293 Cl.2 derived from the S1 segment of rat proximal tubule.
Immunoblotting and immunohistochemistry of rat kidney and WKPT-0293 Cl.2 cells.
Immunoblotting of rat kidney cortex or WKPT-0293 Cl.2 cells with DMT1-com antiserum detected a broad band centered at 70 kDa (Fig. 3), which was in accord with the molecular mass of DMT1 previously reported (12). We used this antiserum to examine the distribution of DMT1 in the rat kidney using light microscopy immunohistochemistry. DMT1-com stained the majority of proximal tubules in the cortex. Staining was intracellular and punctate, suggesting localization of DMT1 in a vesicular compartment (Fig. 4). Importantly, apical brush-border or basolateral membranes were not immunoreactive. There was no staining following preincubation with the corresponding immunizing peptide (Fig. 3).
Immunogold electron microscopy.
To identify precisely the subcellular localization of DMT1, we employed immunogold electron microscopy on rat kidney cortex sections using DMT1-com antiserum (Fig. 5). The labeled structures ware identified as late endosomes/lysosomes based on morphological criteria. The labeling was localized specifically in the membrane of these structures. There was no labeling of apical or basolateral membrane.
Immunofluorescence of WKPT-0293 Cl.2 cells.
We continued our study of WKPT-0293 Cl.2 cells using immunofluorescence with a DMT1-com antibody (Fig. 6A). Again, we observed strong intracellular staining with no indication of plasma membrane staining. DMT1 labeling was punctuate, partially comprising perinuclear and nuclear compartments as visualized by the nuclear marker H-33342 (Fig. 6A). Both the nuclear and perinuclear staining were specific, because preincubation of the DMT1-com antiserum (1:100 dilution) with an excess of 250 μg/ml peptide antigen abolished immunostaining (data not shown). To get better insight into subcellular localization of DMT1 in the cells, we performed colocalization studies of DMT1 with markers of the endosomal/lysosomal trafficking pathway (Fig. 6B). There was only moderate colocalization of DMT1 with EEA1, a marker of early endosomes. In addition, DMT1 showed some colocalization with the lysosomal protease cathepsin L (Fig. 6B) and the late endosomal/lysosomal marker LAMP1 (data not shown), suggesting localization of DMT1 in late endosomal and lysosomal membranes.
Internalization of transferrin in WKPT-0293 Cl.2 cells and colocalization with DMT1.
Growing WKPT-0293 Cl.2 cells on permeable supports enables access to both apical and basolateral membranes. To determine whether transferrin enters the cell across the apical, basolateral or both membranes, we applied Alexa Fluor 546 transferrin to either the apical or basolateral membranes of confluent WKPT-0293 Cl.2 monolayers (Fig. 7). Interestingly, there was no detectable uptake of Alexa Fluor 546 transferrin observed after application to the basolateral compartment. In contrast, application of transferrin to the apical chamber resulted in the rapid uptake of transferrin within 5 min of exposure. After 15-min internalization, the fluorescence signal was more pronounced. Accordingly, we investigated the localization of DMT1 within intracellular compartments through which transferrin was trafficked. After 20-min incubation at 37°C, Alexa Fluor 546 transferrin partially colocalized with EEA1 and cathepsin L (Fig. 8), whereas after 60-min incubation, transferrin colocalized mainly with cathepsin L (Fig. 8). When cells exposed to Alexa Fluor 546 transferrin were stained for DMT1, colocalization increased with incubation time.
DMT1 is a key transporter for the duodenal absorption of dietary ferrous iron and for iron export out of endocytotic vesicles as part of transferrin receptor-mediated cellular iron acquisition (7, 15, 28). Its latter role explains its ubiquitous expression throughout the body. Interestingly, the kidney expresses relatively high levels of DMT1 compared with other tissues, and the reason for this is currently unknown (6, 12, 17, 33). DMT1 expression is most concentrated in the kidney cortex, particularly in the proximal tubule, and to a lesser extent in the distal tubule and cortical collecting duct (12, 33). As a first step in determining the cellular function of DMT1 in the renal proximal tubule, we determined which DMT1 splice variants are present in the kidney cortex and in the rat epithelial cell line WKPR-0293 Cl.2, derived from the S1 segment of kidney proximal tubule (34), and investigated the cellular location of DMT1 using a spectrum of complementary techniques.
Previous studies have reported the presence of the four variant exons (exons 1A, 1B, IRE+ and IRE−) in the kidney but stopped short of identifying which splice variants were present (18). Analysis of kidney cortex RNA and RNA extracted from WKPT-0293 Cl.2 cells showed both RNA pools contained all four splice forms of DMT1 mRNA (see Fig. 1). This indicates that the four DMT1 protein isoforms are likely to be expressed in the kidney cortex and WKPT-0293 Cl.2 cells. We did not microdissect individual nephron segments and perform RT-PCR; therefore, we cannot specify if there is differential expression of DMT1 transcripts in nephron segments that are present in the kidney cortex, namely, proximal tubules, distal tubules, and cortical collecting ducts. However, the finding that all four splice variants are also present in proximal tubule-derived WKPT-0293 Cl.2 cells argues in favor of the four isoforms being present in vivo in proximal tubule cells.
Immunoblotting of protein homogenates with DMT1-com antisera identified a prominent band with a molecular weight of ∼70 kDa in both kidney cortex and WKPT-0293 Cl.2 cells (Fig. 3), which is representative for DMT1 protein. Immunoperoxidase staining of rat kidney sections localized DMT1 to proximal tubules and gave a pattern of staining similar to that we have previously obtained using immunofluorescence (12, 33). Intracellular structures were strongly stained, whereas there was no detectable signal associated with plasma membranes. In the WKPT-0293 Cl.2 cells, staining was also intracellular, with no plasma membrane labeling. These data along with our previous work strongly suggest that in the rat DMT1 is responsible for neither apical nor basolateral membrane translocation of iron or its other divalent metals ligands (17). However, the situation may be different in the mouse because Canonne-Hergaux and colleagues (6) reported apical plasma membrane staining of mouse proximal tubules. Clearly, further examination of the mouse is called for, possibly with a battery of DMT1 antisera targeting independent epitopes.
The electron microscopy immunogold method enabled us to gain further insight into the ultrastructural location of DMT1. With our DMT1-com antiserum, immunogold labeling was observed in the membrane of large intracellular vesicles identified by their morphology as late endosomes/lysosomes.
In WKPT-0293 Cl.2 cells, in addition to vesicular labeling we also observed perinuclear and nuclear staining. Whereas the former observation was consistent with DMT1 labeling in rat kidney tissue, the latter finding was apparent only in cultured cells and was observed both with immunofluorescence (Fig. 6A) and immunohistochemistry (not shown). The reason for this difference is unclear; however, there is a precedent for localization of DMT1 to the nucleus: in neural cell lines, Kuo and colleagues (21) demonstrated that DMT1 IRE+ immunoreactivity was limited to the plasma membrane and to cytosolic vesicles, whereas, DMT1 IRE− staining localized to the nucleus. What functional role DMT1 may play in the nucleus and why nuclear staining is only apparent in cultured cells remain to be determined.
To gain further insight into compartmental distribution of DMT1 in proximal tubule cells, we performed colocalization studies with DMT1-com and specific markers of the endosomal and lysosomal pathway, including the early endosomal marker EEA1 and the lysosomal markers LAMP1 and cathepsin L. All these markers showed partial overlap with DMT1. LAMP1 and cathepsin L showed the highest degree of overlap, indicating that DMT1 is present in late endosomes/lysosomes in WKPT-0293 Cl.2 cells. Previously, we reported partial overlap with LAMP1 and DMT1 in rat proximal tubules (12). When taken together with the immunogold experiments, we conclude that DMT1 is expressed in lysosome membranes in proximal tubule cells and we speculate that it may play a role in translocation of iron or other divalent ligands across these membranes.
This conclusion is consistent with our finding of the high degree of signal overlap between DMT1 and Alexa Fluor 546-labeled transferrin. Transferrin was clearly seen to have a high degree of overlap with DMT1, indicating that DMT1 is associated with an intracellular compartment involved in the processing of transferrin. Furthermore, growing WKPT-0293 Cl.2 cells on permeable supports enabled the polarization of transferrin uptake to be investigated. These experiments showed that transferrin uptake by WKPT-0293 Cl.2 cells was asymmetric in that detectable amounts of transferrin only entered the cells when presented to the apical membrane. In contrast, no detectable internalization occurred when the protein was applied to the basolateral membrane. This may suggest that in vivo transferrin is reabsorbed from the primary urine into proximal tubule cells and that little, if any, can be taken up from the blood. The time course of transferrin uptake by WKPT-0293 Cl.2 cells nicely matches the kinetics of apical transferrin uptake into proximal tubules in the anesthetized rat, where transferrin was found to accumulate in lysosomes over a time period of 15 min (25).
Our findings in WKPT-0293 Cl.2 cells are consistent with our previous observation that under physiological conditions considerable amounts of transferrin are filtered by the glomeruli, reabsorbed in the proximal tubule via cubilin mediated endocytosis, and degraded in lysosomes (20). Together with the presence of an efficient iron transporter in the late endosomal/lysosomal membrane reported in this study, these data suggest that apical endocytosis of transferrin followed by transport of iron out of late endosomes/lysosomes may constitute a physiological mechanism of iron acquisition by the proximal tubular cells. From studies over the past decade, it has become evident that proximal tubular endocytosis not only facilitates removal of protein from the tubular fluid but also, more importantly, the uptake of vital substances serving important functions in the kidney and other organs. Among the plasma carrier proteins, retinol binding protein, transcobalamin-vitamin B12, and vitamin D binding protein have been studied intensively (8, 23). In contrast to other tissues, where the classic transferrin cycle has been described, uptake of transferrin across the proximal tubule basolateral membrane has never been demonstrated in vivo and therefore the relative contribution of this mechanism to overall iron acquisition by these cells is unknown. We postulate that iron retrieved in the course of apical endocytosis of filtered transferrin by cubilin-mediated endocytosis followed by its gating via DMT1 may constitute a significant pool of iron delivered to proximal tubular cells.
The handling of transferrin by WKPT-0293 Cl.2 cells may also shed some light on the processing of other luminally endocytosed metal-protein complexes by proximal tubular cells and on the possible role of DMT1 in this process. DMT1 is known to function as a H+-coupled transporter of a range of divalent transition metal ions, including cadmium (Cd2+) and lead (3, 4, 17, 26). In the circulation, both metals may be bound more or less avidly by proteins and subsequently endocytosed by proximal tubular cells (for a review, see Ref. 5). For instance, Cd2+ binds with high affinity to metallothioneins (MT), the Cd-MT complex is filtered by the kidney glomerulus, and handling of the filtered complex by the proximal tubule is well documented in vivo (10, 24) and in vitro (11); the available evidence suggests that Cd-MT is luminally endocytosed by proximal tubular cells and subsequently degraded in an acidic endosomal/lysosomal compartment (11). The expression of DMT1 within endosomal/lysosomal compartments that are associated with processing of apically reabsorbed transferrin, as described in this report, suggests that DMT1 could play a significant role in proximal tubule toxicity induced by uptake of Cd-MT complexes (29, 30) or other toxic DMT1 ligands.
We conclude that, in renal proximal tubular cells, DMT1 is expressed in the membranes of late endosomes/lysosomes, the subcellular organelles associated with processing of luminally reabsorbed transferrin. These findings have implications for the renal handling of iron and possibly nephrotoxic metals that are also DMT1 ligands, including Cd2+.
This study was supported by the Deutsche Forschungsgemeinschaft (TH 345/8–1 and 8–2 to F. Thévenod) and start-up funds from the University of Witten/Herdecke (F. Thévenod) and the Wellcome Trust (C. P. Smith) and Royal Society (C. P. Smith).
The authors sincerely thank Prof. Søren Nielsen (University of Aarhus, Aarhus, Denmark) for expert help with immunoperoxidase staining. We thank Dr. U. Hopfer (Dept. of Physiology and Biophysics, Case Western Reserve University, Cleveland, OH) for providing the rat proximal tubule cell line.
↵* M. Abouhamed and J. Gburek contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society