Renal Physiology

Knockdown of endosomal/lysosomal divalent metal transporter 1 by RNA interference prevents cadmium-metallothionein-1 cytotoxicity in renal proximal tubule cells

Marouan Abouhamed, Natascha A. Wolff, Wing-Kee Lee, Craig P. Smith, Frank Thévenod


Chronic exposure to Cd2+ causes renal proximal tubular (PT) damage. Cd2+ reaches the PT mainly as cadmium-metallothionein 1 (CdMT-1) complexes that are filtered at the glomerulus and then internalized in part via endocytosis mediated by megalin and cubulin. Subsequently, Cd2+ is thought to be released in the cytosol to activate cell death pathways. The proton-coupled divalent metal transporter DMT1 also transports Cd2+ and is expressed exclusively in endosomes/lysosomes in rat PT cells. Using vector-based RNA interference with short-hairpin small-interfering RNAs (shRNAs) to downregulate DMT1 in the rat renal PT cell line WKPT-0293 Cl.2, we tested the hypothesis that endosomal/lysosomal DMT1 is involved in CdMT-1 nephrotoxicity. One out of 5 shRNAs tested (sh3) significantly reduced expression of DMT1 protein detected by immunoblotting and DMT1 mRNA as determined by RT-PCR by 45.1 ± 9.6 and 36.9 ± 14.4% (n = 5–6), respectively. Similarly, sh3 reduced perinuclear DMT1 immunostaining in transfected cells. Consistent with the assumed role of DMT1 in CdMT-1 cytotoxicity, sh3, but not the empty vector or sh5, significantly attenuated cell death induced by a 24-h exposure to 14.3 μM CdMT-1 by 35.6 ± 4.2% (n = 13). In contrast, neither fluorescently labeled metallothionein-1 (MT-1) uptake nor free Cd2+ toxicity was altered by the effective DMT1 shRNA (sh3), indicating that cellular uptake of metal-MT-1 complexes and Cd2+-induced cell death signaling are not affected by DMT1 knockdown. Thus we conclude that endosomal/lysosomal DMT1 plays a role in renal PT CdMT-1 toxicity.

  • DMT1
  • endocytosis
  • heavy metals
  • metallothionein

the nonessential metal cadmium (Cd2+) is an environmental pollutant of global concern. Every year thousands of tons of Cd2+ are released in the environment by industry (e.g., the manufacture of Cd2+-nickel batteries), and some enters the food chain because of contamination of water, plants, and nutrients (2, 12, 18). Because Cd2+ is not degraded in the environment, the risk of human exposure is increasing constantly (2). For the general population, i.e., nonoccupational exposure, the principal Cd2+ intake is through the consumption of contaminated foods such as root vegetables, shellfish, and offal, but also from inhalation of tobacco smoke. Because Cd2+ has a long biological half-life (10–25 yr) and low rate of excretion, the body becomes a “sink,” Cd2+ accumulates, and toxicity to vital organs, including the lungs, liver and kidneys, ensues (2, 12).

How does Cd2+ reach the kidneys? Following pulmonary or gastrointestinal absorption, Cd2+ binds to plasma albumin and is transported around the body. It primarily accumulates in the liver where it forms complexes with small peptides and proteins via sulfhydryl groups, such as with glutathione (GSH) or the metal-binding protein metallothionein-1 (MT-1), which has a high affinity for Cd2+ (dissociation constant ∼10−17 M; see Ref. 17). Cd2+ is then either secreted in the bile as CdGSH, where it may be recycled via the enterohepatic circulation, or it is released into the blood circulation mainly as cadmium-metallothionein 1 (CdMT-1) complexes (8), which redistribute to the kidneys (22).

CdMT-1 is filtered in the glomerulus because of its small molecular mass (∼7 kDa). The majority of Cd2+ accumulating in the kidney is localized in the epithelial cells lining the proximal tubule (PT), particularly in the S1 segment (10, 12, 19, 28), because the PT cells of the S1 segment possess transporters, metabolizing enzymes, and receptors for free Cd2+ as well as for the bound forms of Cd2+ (CdMT-1 or CdGSH; see Refs. 7, 19, 23, 26). Cd2+ gains entry in PT cells through mechanisms of ionic and/or molecular mimicry at the site of transporters of essential elements (e.g., Fe2+, Cu2+, Zn2+) and/or organic molecules (GSH, MT-1, transferrin; see Ref. 5). In the long run, Cd2+-induced nephrotoxicity can result in a general transport defect of the PT characterized by proteinuria, aminoaciduria, glucosuria, and phosphaturia (for review, see Ref. 25).

The uptake pathways for CdMT-1 in PT cells and the cellular processes underlying in vivo cadmium nephrotoxicity are targets of current research. We have previously shown that CdMT-1 uptake induces apoptosis of WKPT-0293 Cl.2 cells (9), an immortalized cell line derived from the S1 segment of rat renal PT (27). Subsequent work with WKPT-0293 Cl.2 cells showed that CdMT-1 is partly internalized via the polyspecific receptors for small protein ligands, megalin, and cubilin (26) that are expressed in the apical plasma membrane (PM) of PT cells (6, 24, 26). Using inhibitors that interfere with the function of acidic endocytotic compartments, we have shown that uptake of CdMT-1 and subsequent apoptosis was prevented (9). Furthermore, CdMT-1 colocalized with the endosomal/lysosomal markers rab5A and LAMP1. We therefore speculated that Cd2+ is released from the MT-1 moiety in late endosomes/lysosomes and subsequently transferred in the cytosol, possibly via an endosomal/lysosomal metal transporter (9, 19). Once in the cytosol, free Cd2+ is thought to induce reactive oxygen species and subsequent cell death (20).

Interestingly, the H+-coupled divalent metal transporter DMT1, which is highly expressed in the kidney cortex (11, 13), transports Fe2+ and Cd2+ even more effectively (3, 4). We recently determined the exact cellular location of DMT1 in rat kidney cortex sections and WKPT-0293 Cl.2 cells using an antiserum against an epitope common to all four known DMT1 isoforms (1). Immunostaining of rat kidney cortex and WKPT-0293 Cl.2 cells showed that the DMT1 protein is expressed in the membranes of late endosomes/lysosomes, but not in the PM. Moreover, uptake of transferrin (Tf), an iron-binding protein that is filtered by the glomerulus and endocytosed at the apical PM of PT cells (15) (for review, see Ref. 29), was only observed following application to the apical membrane of WKPT-0293 Cl.2 cells and subsequently colocalized with endosomal/lysosomal DMT1. We therefore proposed that Tf and (Cd)MT-1 share the same receptor-mediated endocytosis and trafficking pathways to late endosomes and lysosomes where they undergo proteolytic degradation and that DMT1 represents a candidate efflux pathway not only for free Fe2+, but also for Cd2+, from the acidic endosomal/lysosomal compartments to the cytosol (1).

In this study, we tested the hypothesis that, following internalization of CdMT-1 by PT cells, endosomal/lysosomal DMT1 mediates efflux of Cd2+ to cause cell death. DMT1 expression in cultured immortalized rat WKPT-0293 Cl.2 cells was significantly reduced using a vector-based short-hairpin RNA (shRNA) approach. Consequently, PT cell death induced by CdMT-1 was significantly decreased in DMT1 knockdown cells. We propose that, following receptor-mediated endocytosis of CdMT-1 complexes, DMT1 contributes to the release of Cd2+ from acidic endosomal/lysosomal compartments in the cytosol to induce cell death of PT cells.



MT-1 from rabbit liver, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), paraformaldehyde (PFA), protease inhibitor cocktail, and human serum Tf were obtained from Sigma (St. Louis, MO). Alexa Fluor 546 carboxylic acid succinimidyl ester was from Molecular Probes Europe (Leiden, The Netherlands). Slide-A-Lyzer Dialysis products were from Pierce Chemical (Rockford, IL). 2′-(4-Ethoxyphenyl)-5-(4-methyl-1-piperazinyl)-2,5′-bi-1H-benzimidazole, 3HCl (H-33342) was obtained from Calbiochem (San Diego, CA). CdCl2 was from Merck (Darmstadt, Germany).


The DMT1 gene can generate four different proteins that share complete identity except at the extreme amino- and carboxy-termini (1). The anti-DMT1 antibody used (termed DMT1-com) has previously been characterized (11) and was raised in rabbit to the peptide sequence MVLCPEEKIPDDGASGDHGDSC that is common to all four known isoforms. A dilution of 1:500 was used for immunoblotting and 1:150 for immunofluorescence labeling of WKPT-0293 Cl.2 cells. Secondary Cy3-conjugated donkey anti-rabbit IgG for immunostaining was obtained from Jackson Immuno Research (dilution 1:600). The horseradish peroxidase (HRP)-conjugated donkey anti-rabbit immunoglobulin was diluted 1:5,000 (Amersham Pharmacia Biotech).


Cell culture.

An immortalized cell line from the S1 segment of rat renal PT (WKPT-0293 Cl.2; see Ref. 27) was cultured as previously described (1, 21). Cells were passaged (passage no. <40) two times a week upon reaching confluency.

Generation of shRNA vector constructs.

Out of several candidate small-interfering RNA (siRNA) targets independently generated by two prediction programs (Dharmacon siDESIGN; Dharmacon, Chicago, IL, and GenScript siRNA Target Finder; GenScript, Piscataway, NJ), the following five were selected (positions within the rat DMT1 sequence, accession no. NM_013173): siRNA1: 5′-AGTGAGTTCTCCAACGGAATA-3′ (1506–1526), siRNA2: 5′-CGCTCGGTAAGCATCTCTAAA-3′ (1728–1748), siRNA3: 5′-TGCAGTGGTTAGCGTGGCTTA-3′ (1637–1657), siRNA4: 5′-TGAAGTGTGTCACCGTCAGTA-3′ (506–526), and siRNA5: 5′-GACCAGGTCTATTGCCATCAT-3′ (1346–1366).

Short hairpin (sh) constructs were designed, generated, and introduced in the pRNATin-H1.2/Hygro vector by GenScript. The vector also carries a green fluorescent protein (GFP) marker (coral GFP) that allows for assessment of transfection efficiency. The plasmids were subsequently transformed into DH5α competent cells (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions, and plasmid was purified for transfections using the QIAprep Miniprep kit (Qiagen, Hilden, Germany).

Transient transfection of WKPT-0293 Cl.2 cells with DMT1-shRNA vector constructs.

Initial experiments were performed to determine the transfection procedure with the highest transfection efficiency of empty pRNATin-H1.2/Hygro vector and included chemical reagents, such as Lipofectamine 2000 (Invitrogen) and Effectene (Qiagen), or electroporation technologies (Biorad Gene Pulser; Bio-Rad Laboratories, Munich, Germany; and Amaxa Nucleofector; Amaxa, Cologne, Germany). Transfection efficiency was monitored by scoring cells expressing GFP as a percentage of total number of cells. Only cells with bright green fluorescence were considered as positive. The highest transfection efficiency was obtained with Effectene.

For transient transfection of WKPT-0293 Cl.2 cells with DMT1-shRNA vector constructs, typically 9 × 104 cells per well were seeded in 24-well plates and grown for 24 h before transfection of pRNATin-H1.2/Hygro empty vector or DMT1-shRNA vector constructs with Effectene transfection reagent in serum-free growth medium according to the manufacturer's instructions. DNA (0.45 μg) was mixed with 3.2 μl Enhancer, 4 μl Effectene reagent, and 60 μl buffer. Immediately before transfection, 290 μl serum-free medium were added to the DNA transfection mixture, and the cells in each well were rinsed one time with serum-free medium and replaced with 345 μl of the DNA transfection mixture. Depending on the application, cells were transfected for 4–8 h, after which the transfection reagent was replaced with serum-containing medium, and the cells were incubated further for a total of 24–48 h posttransfection before assaying for DMT1 expression or CdMT-1 cytotoxicity.


Cells were harvested 48 h posttransfection by trypsin digestion and centrifugation at 400 g for 5 min at 4°C. Total RNA was extracted from six wells (∼2 × 106 cells) for each vector construct using the RNeasy Mini Kit (Qiagen) with on-column DNase digestion following the manufacturer's instructions. First-strand cDNA was synthesized with the Omniscript RT kit (Qiagen), using 2.0 μg RNA/20 μl reaction and oligo(dT) primer (Operon Biotechnologies, Huntsville, AL). All four DMT1 isoforms were simultaneously amplified from WKPT-0293 Cl.2 cDNA by PCR using actin as standard (HotStarTaq Master Mix kit from Qiagen), with 4 vol/100 vol and 2 vol/100 vol of RT reaction for DMT1 and actin, respectively. After the initial activation step at 95°C for 15 min, PCR of DMT1 (25 cycles) and actin (23 cycles) was performed under the following amplification conditions: 94°C for 30 s, 57.5°C for 20 s, and 72°C for 20 s. The following primers (Operon Biotechnologies, Cologne, Germany) were used: DMT1 (GenBank accession no. NM_013173): 1148 forward 5′-CAACTCTACCCTGGCTGTGG-3′ and 1536 reverse 5′-TCCTCCAGCCTATTCCGTTG-3′. The DMT1 primers selected amplify a nucleotide sequence common to the four isoforms of DMT1 (IRE+, IRE−, exon 1A, and exon 1B; see Ref. 1): actin (GenBank accession no. NM_007393): 886 forward 5′-TGGAATCCTGTGGCATCCATGAAAC-3′ and 1234 reverse 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′.

Control reactions for each primer pair were also performed without RT, or with water instead of cDNA template. PCR products were separated on 1.5% agarose gels containing 0.01% GelRed (Biotium, Hayward, CA), loading 10 μl PCR reaction per lane. Densitometric analysis was performed using the TINA version 2.09 program package. Optical densities of RT-PCR signals for actin and DMT1 were used for semiquantitative analyses of the data.


WKPT-0293 Cl.2 cells (9 × 104/well) were transfected as described earlier and subsequently incubated in serum-containing medium. Posttransfection (24 h), cells were harvested by trypsin digestion, washed two times in PBS at 4°C, washed in ice-cold homogenization buffer (12 mM HEPES + 300 mM mannitol, adjusted to pH 7.4 and to which 100 μl/ml protease inhibitor cocktail was added immediately before use), and sonicated on ice for 3 × 5 s at 10 Ampères using a Branson sonicator 250 (Branson Ultrasonics, Danbury, CT). 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 8% SDS-PAGE and transferred to 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. Following incubation with HRP-conjugated secondary antibody (1:5,000) for 1 h at 4°C, blots were developed using Western Lighting Plus chemiluminescence reagents (Perkin Elmer Life Sciences, Boston, MA), and signals were visualized on X-ray film. Densitometric analysis was performed using TINA version 2.09. The ratios between DMT1-shRNA transfected and control samples (transfected with empty pRNATin-H1.2/Hygro vector) were calculated for each individual experiment and expressed as a percentage of controls.

Immunofluorescence labeling of shRNA transfectants.

For immunofluorescence labeling of DMT-1 in shRNA-transfectants, WKPT-0293 Cl.2 cells were plated on cover slips and transfected as described above. After transfection, cells were grown for a total of 48 h before being further processed. The following steps were each separated by three washes with PBS for 5 min each. Cells were fixed with 4% PFA in PBS for 30 min, permeabilized with 1% SDS/PBS for 15 min, and then blocked with 1% BSA in PBS for 60 min. These procedures were carried out at room temperature (RT). Subsequently, the cells were incubated with primary anti-DMT1 antibody (1:150 diluted in PBS + 1% BSA) overnight at 4°C and with secondary Cy3-conjugated anti-rabbit antibody (1:600) for 1 h at RT (in PBS + 1% BSA), followed by counterstaining with 0.8 μg/ml H-33342 for 5 min. After three more washes in PBS for 5 min, the cells were rinsed in dH2O, and cover slips were mounted on glass slides with DAKO fluorescence mounting medium (Dako Cytomation, Carpinteria, CA). The cells were viewed using filters for Cy3 (red), fluorescein isothiocyanate (FITC; green), and DAPI (blue) with excitation/emission wavelengths of 545/610, 480/535, and 360/460 nm, respectively, using a mercury short-arc photooptic lamp HBO (103 W/2; OSRAM, Augsburg, Germany) as light source, which was connected to a Zeiss Axiovert 200M microscope (Carl Zeiss, Jena, Germany) equipped with a Fluar ×40, 1.3 oil immersion objective. Images were captured using a digital CoolSPAN ES CCD camera (Roper Scientific, Tuscon, AZ) and acquired, processed, and analyzed with MetaMorph software (Universal Imaging, Downingtown, PA). Cy3 (red), FITC (green), and DAPI (blue) images were merged using the Color Combine overlay function in the Metamorph software. To obtain a semiquantitiative estimate of the effect of sh3 on nuclear and perinuclear immunofluorescence labeling for DMT1, 50 nontransfected and transfected WKPT-0293 Cl.2 cells from four experiments were analyzed with the MetaMorph software. In nontransfected cells, perinuclear staining deviated across the cytosol, which occurred to a smaller extent in sh3-tranfected cells. Thus we evaluated integrated rather than mean intensity, since this should give a more accurate estimation of the sh3-induced fluorescence decrease. Integrated fluorescence intensity was used as well for analysis of nuclear staining. Background fluorescence of cell-free areas was subtracted before performing further analysis.

MTT cell viability assay.

CdMT-1 was prepared from MT-1 exactly as previously described (9). WKPT-0293 Cl.2 cells were seeded at 2.5 × 104 cells/well in 48-well plates. Postseeding (24 h), wells showing similar cell growth and distribution were transfected with DMT1-shRNA or empty vector as described above. After 24 h, the medium was aspirated and replaced with 100 μl modified serum-free culture medium, to which either 14.3 μM MT-1, CdMT-1 (100 μM Cd2+ + 14.3 μM MT-1), or 20 μM CdCl2 had been added, and the cells were incubated for another 24 h. We noticed a general increase in cell death in sh3-transfected MT-1-treated cells (i.e., when compared with MT-1-treated controls) in Tf-containing medium, suggesting that the beneficial effect of DMT1 knockdown on cell survival upon CdMT-1 treatment might be partly obscured by interference of DMT1 knockdown with Tf-mediated iron handling by the PT cells. The modification of the medium was based on the following rationale: 1) Tf and MT-1/CdMT-1 are known to share the same receptor-mediated uptake mechanism (26); 2) we reasoned that efficient silencing of endosomal/lysosomal DMT1 reduced iron supply to the cells through Tf endocytosis. Hence we replaced Tf through FeCl2 and ascorbic acid to allow for iron uptake in the cells via alternative transport routes. Preliminary experiments had shown 20 μM CdCl2 to be equally toxic as the selected concentration of CdMT-1 in cells transfected with empty vector when incubated for 24 h. All experiments with free Cd2+ were performed in parallel to CdMT-1/MT-1 experiments to ensure comparable experimental conditions. Cell viability was then assayed using the MTT test as described previously (16). Briefly, the cells were incubated for 3 h at 37°C in serum-free medium without phenol red containing 1 mg/ml MTT. After the medium was aspirated, the resulting formazan product was dissolved in isopropanol and measured at 560 and 690 nm.

Fluorescence measurement of Alexa Fluor 546-conjugated MT-1 uptake in WKPT-0293 Cl.2 cells and semiquantitative analysis of fluorescence intensity.

To obtain Alexa Fluor 546-conjugated MT-1, we essentially followed the protocol described by Klassen et al. (14). Briefly, Alexa Fluor 546 carboxylic acid succinimidyl ester was coupled to rabbit liver MT-1 according to the manufacturer's instructions (Molecular Probes). The conjugate was separated from uncoupled labeling reagent by microdialysis using Slide-A-Lyze dialysis cassettes (Pierce) with a molecular mass cut-off of 3.5 kDa.

Serum-containing culture medium was aspirated 24 h posttransfection and replaced with 200 μl serum-free medium containing 50 μM FeCl2 and 1 mM ascorbic acid, but without Tf. Cells were then incubated for a further 24 h with 7.2 μM Alexa Fluor 546-conjugated MT-1. All subsequent steps were performed at 22°C. Cells were rinsed three times for 7 min in PBS, fixed with 4% PFA/PBS for 30 min, washed three times again in PBS for 5 min, counterstained with H-33342 as described above, and washed a further four times in PBS for 5 min, followed by a final rinse for 5 min in dH2O. Cover slips were mounted on glass slides with DAKO fluorescence mounting medium. The cells were viewed using filters for Cy3 (red), FITC (green), and DAPI (blue) and processed and analyzed as described above.

For quantitative determination of Alexa Fluor 546-conjugated MT-1 uptake in WKPT-0293 Cl.2 cells transfected with DMT1-shRNA or empty vector, ca. 50 GFP-positive cells in three different experiments were analyzed with MetaMorph software. Cy3 images were used for analysis after threshold background subtraction. The mean fluorescence intensity of GFP-positive (i.e., transfected) cells was determined by outlining individual cells. The values obtained were corrected for endogenous cellular autofluorescence by subtracting the Cy3 mean fluorescence intensity of positive cells transfected with the DMT1-shRNA or empty vector incubated in the absence of Alexa Fluor 546-conjugated MT-1.

Statistical Analyses

Representative data or means ± SE are shown. Statistical analysis using unpaired Student's t-test was carried out with Sigma Plot 8.0 (Chicago, IL). For more than two groups, one-way ANOVA was used, assuming equality of variance with Levene's test and Tukey's post hoc test for pairwise comparison with SPSS 12.0. Results with P ≤ 0.05 were considered to be statistically significant.


Effect of DMT1-shRNA Vector Constructs on DMT1 Protein Expression in WKPT-0293 Cl.2 Cells

To determine the extent of DMT1 silencing by the five candidate siRNA sequences selected (siRNA 1–5; see Methods) at the protein level in WKPT-0293 Cl.2 cells, semiquantitative immunoblots were performed 24 h posttransfection. Figure 1A shows a representative immunoblot comparing WKPT-0293 Cl.2 cells transfected with either the control vector or the DMT1-shRNA vector constructs. Immunoblotting of PT cells with DMT1-com antiserum detected a broad band centered at ∼75 kDa, which was in agreement with the molecular mass of DMT1 previously reported in the WKPT-0293 Cl.2 cell line and rat kidney cortex (1, 11). DMT1 expression was not significantly affected by sh1, sh2, sh4, or sh5 compared with empty vector. In contrast, sh3 caused a significant knockdown of DMT1 expression. Densitometric analyses of six immunoblot experiments showed that DMT1 was silenced by 45.1 ± 9.6% by sh3 (P < 0.02; n = 6; Fig. 1B). Therefore, sh3 was selected for use in further experiments.

Fig. 1.

Expression of DMT1 protein in WKPT-0293 Cl.2 cells transfected with DMT1-short-hairpin RNA (shRNA) vector constructs. A: DMT1 immunoreactivity in WKPT-0293 Cl.2 cells transfected with empty pRNATin-H1.2/Hygro vector (vector) or DMT1-shRNA vector constructs (sh1-sh5) as described in Methods. After transfection (24 h), cell homogenates from WKPT-0293 Cl.2 cells (50 μg protein) were separated by SDS-PAGE on 8% acrylamide gels. Blots were incubated with anti-DMT1-com antibody (1:500). One out of 6 similar experiments is shown. MW, molecular mass. B: semiquantitative analysis of DMT1 protein immunoreactivity in PT cells transfected with empty vector or DMT1-shRNA vector constructs. Immunoblotting signals were evaluated by densitometry. The ratio of the optical density of DMT1 signals in shRNA-transfected cells over the corresponding empty vector signals was calculated and expressed as %empty vector. Graph shows means ± SE of 6 experiments. *Significant difference (P < 0.02) to empty vector using 1-way ANOVA.

We also investigated the impact of silencing of DMT1 by sh3 on intracellular expression and localization of the DMT1 protein in WKPT-0293 Cl.2 cells. Using immunofluorescence staining with DMT1-com antiserum, we confirmed intracellular DMT1 expression, as previously described (1). Again, strong intracellular staining with no indication of PM staining was observed. DMT1 labeling was punctate and localized primarily in perinuclear and nuclear compartments, as visualized by H-33342 nuclear marker (Fig. 2). We have previously demonstrated that the perinuclear staining is associated with late endosomes and lysosomes, where endocytosed protein-metal complexes, such as Tf or CdMT-1, are transported (1, 26). When transfected with sh3, DMT1 expression was decreased in individual positive cells identified by expression of the vector-encoded marker GFP (Fig. 2A). In particular, the perinuclear staining of DMT1 was decreased by DMT1 knockdown by ∼50% (Fig. 2, B and C and data not shown) compared with surrounding nontransfected cells (Fig. 2, B and C). In contrast, nuclear staining was not significantly affected by sh3 (Fig. 2, B and C and data not shown).

Fig. 2.

Immunofluorescence labeling of intracellularly expressed DMT1 in sh3-transfected and nontransfected WKPT-0293 Cl.2 cells. After transfection with sh3 vector, cells were grown for a total of 48 h before fixation, permeabilization, and incubation with primary polyclonal rabbit antibody DMT1-com (1:150). A: labeling was performed with secondary antibodies (1:600) conjugated to Cy3 (red), and nuclei were counterstained with H-33342 (blue). Successfully transfected cells were identified by GFP expression (green). B: in nontransfected cells, DMT1 staining was characteristically perinuclear (arrowheads), which has been shown to represent late endosomes and lysosomes. In sh3-positive (GFP-positive) cells, DMT1 labeling was reduced in perinuclear vesicular structures (indicated by arrows). C: in contrast, nuclear staining was not affected by sh3, as can be seen from DMT1 only images. Bars = 50 μm.

Effect of DMT1-shRNA Vector Constructs on DMT1 mRNA Expression in WKPT-0293 Cl.2 Cells

Conclusively, knockdown of DMT1 by sh3 was further confirmed by RT-PCR. WKPT-0293 Cl.2 cells were transfected with either empty vector (control) or sh3, and cells were harvested 48 h later. RT-PCR was executed with a primer pair that does not distinguish among the four DMT1 isoforms (IRE+, IRE−, exon 1A, exon 1B; see Ref. 1) or actin as control. In Fig. 3A, the first two lanes show PCR products for actin (349 bp), which display no difference between control and sh3-transfected cells. DMT1 PCR products (389 bp) were loaded in the next two lanes. In sh3-transfected cells, there is a considerable attenuation of DMT1 mRNA expression compared with empty vector, which confirms the observations at the protein level (Figs. 1 and 2). No bands were detected in RT control reactions without RT or with H2O (negative controls; data not shown). Semiquantitative densitometry analysis revealed a statistically significant reduction of DMT1 mRNA by sh3 as measured by RT-PCR by 36.9 ± 14.4% of the controls (n = 5; P < 0.05), whereas sh3 did not affect actin mRNA expression (Fig. 3B).

Fig. 3.

DMT1-sh3 vector transfection of WKPT-0293 Cl.2 cells reduces DMT1 mRNA expression. A: first-strand cDNA was synthesized from RNA isolated from empty pRNATin-H1.2/Hygro vector- or DMT1-sh3 vector-transfected WKPT-0293 Cl.2 cells. PCR was then performed with primers detecting all four splice variants (IRE+, IRE−, exon 1A, exon 1B) of rat DMT1 (389 bp) or rat actin (349 bp) as a housekeeping gene. Equal volumes of products were run on a 1.5% agarose gel stained with GelRed. B: semiquantitative analysis of DMT1 and actin RT-PCR products of WKPT-0293 Cl.2 cells transfected with empty pRNATin-H1.2/Hygro vector (vector) or DMT1-sh3 vector (sh3). RT-PCR signals were evaluated by densitometry. Optical densities (arbitrary units) of RT-PCR signals for actin and DMT1 were used for semiquantitative analyses of the mRNA expression. *P < 0.05; n = 5. NS, no significant difference between sh3-transfected cells and empty pRNATin-H1.2/Hygro vector-transfected cells using unpaired Student's t-test.

Effect of DMT1 Silencing on CdMT-1-induced Cell Death of WKPT-0293 Cl.2 Cells

Twenty four hours after transfection with sh3, sh5, or empty vector (pRNATin-H1.2/Hygro), cells were incubated with 14.3 μM CdMT-1 (100 μM Cd2+ and 14.3 μM MT-1) or 14.3 μM MT-1 in serum-free medium for 24 h. In cells that had been incubated with MT-1 only, no difference in cell viability was observed with empty vector, sh3, or sh5, as determined with the MTT assay (0.29 ± 0.03 absorbance units for empty vector vs. 0.27 ± 0.03 for sh3; n = 13; P > 0.72 and 0.22 ± 0.02 absorbance units for empty vector vs. 0.20 ± 0.02 for sh5; n = 6; P > 0.54). After 24 h exposure with 14.3 μM CdMT-1, 56.4 ± 3.3% of PT cells transfected with the empty vector remained viable relative to MT-1-treated cells. In contrast, when PT cells had been transfected with sh3, cell viability in the presence of CdMT-1 significantly increased to 71.2 ± 3.6% (P < 0.01; n = 13; Fig. 4). In other words, cell toxicity was reduced by ∼36%, suggesting that knockdown of late endosomal/lysosomal DMT1 by sh3 prevents toxicity of CdMT-1. The magnitude of sh3-induced reduction of cell death was comparable to the percentage of silencing of DMT1 expression at the protein and mRNA levels (see Figs. 1B and 3B).

Fig. 4.

Transfection with DMT1-sh3 vector constructs selectively reduces cadmium-metallothionein 1 (CdMT-1)-induced cell death of WKPT-0293 Cl.2 cells. Cells were transfected with empty pRNATin-H1.2/Hygro vector, DMT1-sh3, or DMT1-sh5 vector constructs for 24 h and incubated with 14.3 μM metallothionein (MT-1) or CdMT-1 (100 μM Cd2+, 14.3 μM MT-1) for an additional 24 h. Cell viability of PT cells was then assayed using the MTT test as described in Methods. The graph displays cell viability with CdMT-1 expressed as a percentage of the respective MT-1 controls. Means ± SE of 6–13 experiments are shown. Transfection with sh3 significantly increased viability of CdMT-1-treated WKPT-0293 Cl.2 cells, but not transfection with sh5 (n = 6). *Significant difference of sh3 (P < 0.01; n = 13) to empty pRNATin-H1.2/Hygro vector. NS, no significant difference of sh5 to empty pRNATin-H1.2/Hygro vector using unpaired Student's t-test.

In addition to sh3, we also tested the effect of sh5 on CdMT-1 toxicity. Upon addition of CdMT-1 for 24 h, cell viability of cells transfected with sh5 was 58.9 ± 5.7% relative to 51.9 ± 4.2% for cells transfected with empty vector (P > 0.38; n = 6; Fig. 4) compared with the respective MT-1-treated cells, i.e., CdMT-1-induced cell toxicity was not significantly affected by sh5 compared with empty vector transfected cells.

Effect of DMT1 Silencing on Alexa Fluor 546 MT-1 Uptake in WKPT-0293 Cl.2 Cells

Next we investigated MT-1 uptake in cells transfected with empty vector or DMT1-sh3 vector construct. Rabbit MT-1 was coupled to Alexa Fluor 546, and the uptake of Alexa Fluor 546-conjugated MT-1 was monitored by fluorescence microscopy. In previous studies, a significant uptake of MT-1 at 24 h had been observed with a concentration of 7.2 μM Alexa Fluor 546-conjugated MT-1 (26). As shown in Fig. 5A, the intracellular distribution of MT-1 in transfected cells was comparable to that in nontransfected cells. Moreover, in three different experiments, sh3 did not significantly affect uptake of Alexa Fluor 546-conjugated MT-1 in WKPT-0293 Cl.2 cells (Fig. 5B).

Fig. 5.

Effect of DMT1-sh3 vector transfection on Alexa Fluor 546-coupled MT-1 internalization by WKPT-0293 Cl.2 cells. A: cells were transfected with empty pRNATin-H1.2/Hygro vector (vector) or DMT1-sh3 vector construct (sh3). After transfection (24 h), they were incubated with 7.2 μM Alexa Fluor 546-coupled MT-1 for an additional 24 h. Internalized Alexa Fluor 546-coupled MT-1 was then detected by immunofluorescence microscopy. Successfully transfected cells were identified by GFP expression (green). Bars = 10 μm. B: 50 cells in 3 experiments were analyzed for vector- and sh3-transfected cells, and mean Alexa Fluor 546 fluorescence intensity per GFP-positive cell was calculated. The difference between pRNATin-H1.2/Hygro vector- and sh3-transfected cells was not significant (NS; n = 3 experiments) using unpaired Student's t-test. a.u., Arbitrary units.

Effect of DMT1 Knockdown on Cd2+-induced Cell Death of WKPT-0293 Cl.2 Cells

In preliminary experiments, the Cd2+ concentration that caused a similar extent of toxicity as CdMT-1 in empty vector-transfected PT cells after 24 h was determined. Following 24 h exposure with 20 μM Cd2+, 58.2 ± 7.1% of PT cells transfected with the empty vector were viable compared with non-Cd2+-treated controls, which were set to 100% (Fig. 6). When cells were transfected with sh3 for 24 h and incubated subsequently with 20 μM Cd2+ for 24 h, cell viability was 61.2 ± 8.9% of the respective controls (P > 0.79 compared with Cd2+-treated vector-transfected cells; n = 7; Fig. 6).

Fig. 6.

Effect of DMT1-sh3 vector transfection on cell death induced by Cd2+ in WKPT-0293 Cl.2 cells. Cells were transfected with empty pRNATin-H1.2/Hygro vector or DMT-sh3, as described in Methods. After transfection (24 h), cells were incubated for an additional 24 h in serum-free medium with or without 20 μM CdCl2 that was equipotent as 14.3 μM CdMT-1 (100 μM Cd2+ + 14.3 μM MT-1). Cell viability was then assayed using the MTT test.MTT absorbance in the respective controls without Cd2+ was 0.25 ± 0.04 (vector) and 0.27 ± 0.04 (sh3) and normalized to 100%. Means ± SE of 7 experiments are shown for Cd2+-exposed cells relative to Cd2+-free controls. The difference between pRNATin-H1.2/Hygro vector- and sh3-transfected cells was not significant using unpaired Student's t-test.


In the present study, we tested the possible involvement of DMT1 in CdMT-1 cytotoxicity, using shRNA knockdown of DMT1. Applying RNAi via shRNAs to eukaryotic cells poses many challenges. shRNAs show varying effectiveness in different cell types in a manner yet poorly understood: some cells respond well to shRNAs and show a robust knockdown, others show no such knockdown (even despite efficient transfection). We selected five different siRNA targets, which were independently generated by two prediction programs and were predicted to knockdown the DMT1 isoforms expressed in PT. In our hands, four out of five selected shRNA constructs did not demonstrate any significant knockdown of DMT1 at the protein level (Fig. 1). In contrast, sh3 significantly reduced DMT1 expression, both at the protein and mRNA level (Figs. 1 and 3). The reasons for these differences are, however, difficult to assess. Furthermore, the combined transfection of sh3 and sh5 did not enhance DMT1 knockdown (data not shown).

Transient transfection with sh3 significantly decreased DMT1 protein expression by ∼45%, as detected by immunoblotting (Fig. 1), and attenuated perinuclear DMT1 immunostaining (Fig. 2). Analysis of immunofluorescence labeling in the perinuclear and nuclear areas of transfected cells (Fig. 2) provided an approximation of the quantitative changes of DMT1 expression induced by sh3. In cells transfected with empty vector, nuclear staining was similar to nontransfected cells in the same microscopic field (data not shown). Nuclear fluorescence in sh3-transfected cells was marginally increased relative to nontransfected cells (∼2%). In contrast, perinuclear DMT1 staining was decreased in sh3-transfected cells by >50% in the sh3-transfected cells that were analyzed. Hence the putative nuclear DMT1 could represent a cross reaction with an epitope in a nuclear protein related to the peptide used to make the antibody (our prior work showed that the peptide antigen eliminated nuclear fluorescence as well; see Ref. 1). This nuclear protein could be associated with the nuclear membrane, since nuclear staining disappeared in mitotic cells (data not shown). Alternatively, an isoform of DMT1 might be expressed in the nucleus that turns over more slowly, perhaps because of decreased access to degradation, and is therefore less affected by sh3.

Because a significant decrease of DMT1 protein expression was detected using an antibody that is common to all four DMT1 isoforms expressed in renal PT, we validated the bulk effect of DMT1 knockdown by sh3 at the mRNA level by performing RT-PCR with primer sequences that do not distinguish among the four DMT1 isoforms. Based on the significant knockdown of DMT1 protein expression induced by sh3, we expected to detect a significant reduction of DMT1 mRNA using conventional RT-PCR. Indeed, expression of mRNA was also reduced by ∼35–40% (Fig. 3). Quantitative RT-PCR would be an attractive approach to determine the functional significance of the different DMT1 isoforms expressed in PT cells and their differential knockdown by sh3 (possibly because of differential turnover of mRNA). Future studies investigating the expression of the different DMT1 isoforms, both at the mRNA and protein level, may provide a deeper insight into the differential contribution of DMT1 isoforms in the process of endosomal/lysosomal DMT1-mediated divalent metal transport.

When we consider the end point of CdMT-1 uptake by PT, namely cell death, the decrease of DMT1 protein and mRNA expression varied between 35 and 45% relative to control cells transfected with empty vector. DMT1 knockdown significantly decreased CdMT-1-induced cytotoxicity by ∼35% (Fig. 4), which is consistent with a role for DMT1 in CdMT-1-induced cytotoxicity in PT cells, possibly by mediating transport of Cd2+ released from MT-1 out of endosomal/lysosomal compartments in the cytosol.

The observation that Alexa Fluor 546-conjugated MT-1 internalization was not affected by sh3 (Fig. 5) suggests that shRNA transfection and/or DMT1 knockdown, per se, did not interfere with receptor-mediated endocytosis of CdMT-1 via megalin/cubilin (and possibly other uptake pathways; see Ref. 26), which is therefore not the cause of cell death reduction observed with sh3 (Fig. 4). Moreover, the intracellular distribution of endocytosed MT-1 in sh3-transfected cells and nontransfected neighboring cells was comparable (see Fig. 5A), which also implies that sh3 did not affect the trafficking of endocytosed MT-1.

Interestingly, the toxicity of free Cd2+ was not altered by DMT1 knockdown (Fig. 6). This observation has several implications. First, it demonstrates that sh3 transfection does not interfere with the cell death signaling pathways mediating toxicity induced by Cd2+ (and CdMT-1), and, accordingly, the reduction of CdMT-1 toxicity in sh3-transfected cells must be the consequence of decreased DMT1 function and operate upstream of cell death signaling pathways. In addition, the reduction of PT toxicity observed with CdMT-1 but not with free Cd2+ in sh3-transfected cells (compare Figs. 4 and 6) supports immunocytochemical evidence for DMT1 expression in endosomes/lysosomes rather than in the PM of WKPT-0293 Cl.2 cells (1). Hence the control experiments shown in Figs. 5 and 6 nicely confirm our previous studies comparing Cd2+ and CdMT-1 toxicity in PT cells, in which we showed that inhibitors, which interfere with the function of acidic endocytotic compartments (chloroquine, LY-294002), prevented uptake of 109CdMT-1 or MT-1 immunoreactivity as well as CdMT-1-dependent apoptosis, but had no effect on 109Cd2+ uptake or Cd2+-dependent apoptosis (9). What can we infer from this study concerning the physiological role of DMT1 in the PT? It is tempting to speculate that endosomal/lysosomal DMT1 in PT cells could contribute to retrieval of filtered Tf-bound iron, thus preventing substantial body iron losses in the urine by contributing to the release of Fe2+ in the cytosol.

In summary, we provide evidence that significant knockdown of endosomal/lysosomal DMT1 by RNA interference in the rat renal PT cell line WKPT-0293 Cl.2 significantly attenuates CdMT-1-dependent cell death. The transient transfection with a specific shRNA construct resulting in DMT1 downregulation neither interfered with MT-1 internalization and trafficking nor with PT Cd2+ toxicity, indicating that the effect is specific. The results therefore support a role for DMT1 in CdMT-1-induced PT cytoxoxicity by mediating efflux in the cytosol of Cd2+ released from MT-1 in endosomal/lysosomal compartments and resulting in PT cell death. In conjunction with our previous report on the mechanisms of CdMT-1 and Tf uptake by PT cells and the colocalization of these protein-metal complexes with endosomal/lysosomal DMT1 (1, 26), the data also suggest that CdMT-1 uses physiological endocytosis pathways for Tf-bound iron as well as free divalent metal efflux from acidic endosomes/lysosomes by a process of molecular mimicry.


This study was supported by the Deutsche Forschungsgemeinschaft (TH 345/8-1 and 8-2) and start-up funds from the University of Witten/Herdecke to F. Thévenod as well as from the Wellcome Trust and Royal Society to C. P. Smith.


We thank Dr. U. Hopfer (Department of Physiology & Biophysics, Case Western Reserve University, Cleveland, OH) for providing the WKPT-0293 Cl.2 rat proximal tubule cell line.

Current address for M. Abouhamed: Institute for General Zoology and Genetics, Westfalian Wilhelms-University, 48149 Munster, Germany.


  • 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.


View Abstract