INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CADMIUM IS AN IMPORTANT environmental pollutant for humans. The Cd²⁺ concentrations in blood, urine, hair, and nails of people living in highly polluted areas were found to be increased (2-4, 9, 12, 17, 24, 30). Because of its widespread use (e.g., in batteries or as an anticorrosive agent), cadmium can be found almost everywhere and is also present in many food chains. The heavy metal is taken up via the lungs and the gastrointestinal tract, binds to metallothioneins, and accumulates in the kidneys with a half-life of 16-33 yr. Cadmium is cytotoxic, and, due to its accumulation in the kidney, it causes tubular dysfunctions and, in many cases, chronic renal damage (8). In Japan, whole rural populations were affected by "Itai-itai" disease caused by a massive cadmium pollution of their environment (3, 24, 31). Cadherins and catenins (19) are important molecules in epithelial cells during early development, induction of cell polarity, and also during ischemic injury (19, 32). The mechanisms of cadmium toxicity are still incompletely understood in terms of cell physiology and cell polarity (1, 8, 13, 27, 31). In models of the proximal tubules, cadmium interferes with the calcium-dependent adhesion molecules, the cadherin-catenin complex. To further elucidate the mechanisms of cadmium toxicity, we investigated the effect of different cadmium concentrations on LLC-PK₁ and Madin-Darby canine kidney (MDCK) cells and in vitro models of proximal and distal renal tubular cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture conditions. For these experiments, MDCK cells were used as a model of collecting duct cells, and LLC-PK₁ cells were used as a model of proximal tubule cells. Both cell lines can be grown continuously and are well characterized (5, 6, 20). The passages 75-90 of MDCK and 175-185 of LLC-PK₁ were studied. Both cell lines were grown as described in our previous studies (33). In brief, cells were cultured at 37°C, 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS, Seromed Biochrom, no. S0115), 1,000 mg/l L-glutamine (Seromed Biochrom, no. K0282), 1 ml HEPES buffer (GIBCO-BRL, Life Technologies; no. 043-05300), and 10,000 U/10,000 µg/ml penicillin/streptomycin (Seromed Biochrom, no. A2213) to confluence (~5-7 days), respectively. For the determination of transepithelial resistance (TER), lactate dehydrogenase (LDH) release, and light microscopy, the cells were placed in serum- and antibiotic-free medium 48 h prior to the experiment. For the determination of proliferation, a detailed description is given under Determination of cell proliferation (below). For the experiments, MOPS buffer with various concentrations of Cd²⁺ and Ca²⁺ was used. This buffer was chosen because it does not affect the solubility properties of Cd²⁺.

Solutions. MOPS buffer contains (in mM) 10 MOPS (Sigma, no. M-8899), 125 NaCl (Sigma, no. S5886), 5 KCl (Merck, no. 4936), 1 MgCl₂ · 6H₂O (Merck, no. 5832), and 0.1, 1, or 2 CaCl₂ · 2H₂O (Merck, no. 2382). Trypsin-EDTA was from Life Technologies (GIBCO-BRL no. 043-05300). PBS without calcium, magnesium, and sodium bicarbonate was from Life Technologies (GIBCO-BRL, no. 041-04190), 1 mM CdCl₂ stock in H₂O.

Determination of transepithelial resistance. For the measurements of TER, the cells were grown under conditions described above on collagen-coated, two-compartment filters with six wells on each plate (Costar Transwell, no. 3425 E06A4913) with 0.45-µm pore size and a 3-cm² surface area. The measurements were made with an EVOM system (World Precision Instruments, no. 18212-04A) with Ag-AgCl₂ electrodes (Millicell-E RS, no. MERSSTX 01, 1 each; Millipore) at 0, 6, and 24 h at a specially developed stand for precise measurement of TER. Special attention was given to the geometry. The distance between the two electrodes was kept constant, and the introduction of the electrodes was done with a hand-separated lever gear (14). Actual resistance reading was corrected for the blank resistance of the filters by subtracting the resistance of the filters without cells or after the cell experiments by destruction of the monolayer with trypsin-EDTA (there was no difference between the two methods). The blank reading was 85 ± 1 ohms and was subtracted from the resistance reading. The accuracy of the measurement was ±1 ohm. The results were converted into ohms × centimeters squared. The electrical resistance is temperature dependent. Therefore, all electrophysiological experiments were undertaken at 23°C (room temperature with a thermostat). TER correlates with the integrity of the cell-cell contacts, probably reflected by the zonula occludens (7, 21, 25).

Determination of LDH. After every measurement of the TER (0, 6, 24 h), 1 ml buffer of the apical compartment was removed and replaced with the same solution. After centrifugation to remove all cells or cell detritus, LDH concentration was assessed. The LDH release corresponds to the integrity of the cell membrane (5). It should be noted that the determination of LDH in the apical compartment does not mean that the damage occurs on the apical side only. In contrast, since LDH is almost freely diffusible, apical and basolateral concentrations are similar. Determination in the apical compartment only is because of practicability. LDH was measured using the LDH kit from Sigma Diagnostics.

Morphology. Cell cultures were monitored using an inverted microscope (Zeiss Telaval 31). For photographs (Fig. 5), cells were fixed with a series of increasing concentrations of a special ethanol solution containing glutaraldehyde (Merck, no. 111-30-8), veronalacetate (Merck, no. 6318), and osmic acid (Sigma, no. 00631) and embedded in Epon (1,2-epoxipropane). After hardening of the resin, the blocks were cut (0.5 µm) with a microtome (Ultracut E; Reichert & Jung, Linz, Austria) and photographed (Reichert & Jung Polyvar with integrated equipment for photographs; objectives: Plan Apo ×100/1.32 oil, Plan Apo Oel Ph ×40/1.00 oil, Plan ×25/0.45 oil) on Ektachrome 160-T film. Cadmium-treated cells were always compared with the appropriate time-control cell cultures.

Immunofluorescence and confocal microscopy. The cells were grown as described on small sterile glass plates and incubated with cadmium. This procedure was different from the other experiments. This was necessary, since fixation and transfer of cells from the transwell filters onto slides for immunofluorescence was not feasible (this is different from the above-mentioned procedure for light microscopy, where an Epon embedding was used). Immunofluorescence localization of the cadherin-catenin complex was performed to study the effect on adhesion molecules known as zonula adherens (belt desmosomes; Refs. 11, 18). Cells were fixed with methanol and incubated with either anti--catenin (M12K, MPI, rabbit) or DECMA-1 against E-cadherin (Sigma, no. U3254) as first antibodies, diluted 1:1,000 in PBS containing 1% FCS. The incubation with the second antibody [FITC-conjugated goat anti-rabbit IgG (1:25) or dichloro-triazinylaminofluorescein (DTAF)-conjugated goat anti-rat IgG (1:25); both from Dianova] was followed by four to six washing steps with 1% PBS/FCS. Thereafter, cells were embedded (Dako C563) and subjected to confocal microscopy for better localization of the specific fluorescence. An inverted confocal laser scanning microscope was used (Leica Lasertechnik CLSM-Fluorovert). The cells were mounted, and 1-µm sections of the monolayer were investigated. Thus nine sections were obtained from each glass plate. For analysis, the section close to the center of the monolayer was used. The neighboring sections had comparable fluorescent patterns. All experiments were undertaken in parallel with appropriate controls.

Rhodamine-phalloidin staining. Cells were grown on glass slides and treated as described above. After the different experimental periods, the cells were fixed in 4% formaldehyde. For staining, cells were first reconstituted in PBS and exposed to rhodamine phalloidin (Molecular Probes) in a final concentration of 0.5 µM for 30 min in the dark. Cells were washed five times with PBS. Slides were mounted using 50% glycerol/50% PBS and finally bordered with nail polish. Evaluation and microphotographs were done using a fluorescence microscope (Reichert & Jung Polyvar).

Cell extraction and fraction. Cell extraction and fraction were performed as described by others before (10). Treated cell layers were rinsed twice with PBS and solubilized in cytoskeleton (CSK) buffer [0.5% Triton X-100, 50 µM NaCl, 300 µM sucrose, 10 µM PIPES, pH 6.8, 3 µM MgCl₂, 1 µM Pefabloc (Boehringer-Mannheim), 0.1 mg/ml DNase, and 0.1 mg/ml RNase] at 4°C for 20 min. A rubber policeman was used to scrape cells from the dishes, and cells were transferred into clean tubes. After centrifugation, soluble supernatants were carefully collected. Pellets were resuspended in SDS buffer (1% SDS, 10 µM Tris · HCl, pH 7.5, 5 µM EDTA, and 2.5 µM EGTA), boiled for 2 min to solubilize, and diluted with CSK buffer.

Dot blot for -catenin and E-cadherin. Determination was done as described in our previous studies (33). In brief, 2.5 µl per sample were dropped on a nitrocellulose filter and dried at 70°C for 2 h. Filters were blocked overnight with milk (Glücksklee, containing 10% fat). Thereafter, the first antibody against -catenin or E-cadherin diluted 1:1,000 in PBS containing 1% FCS was added for 90 min. Washing steps with PBS containing 0.1% Tween 20 were repeated five times. The second antibody was goat anti-rabbit (1:1,000) or goat anti-rat (1:1,000) labeled with peroxidase. After an incubation time for 60 min, five further washing steps were performed, and the development with 3'-3-diaminobenzidine started thereafter.

For analysis, the dot blots were scanned, and the digitized images were semiquantified, using a standard containing -catenin or E-cadherin with a commercially available software system (ZeroDscan, Scanalytics).

Determination of cell proliferation. For the measurements of proliferation, the bromdesoxyuridine (BrdU) labeling and detection kit III (Boehringer, Mannheim, Germany) was used (for further details, see study protocol of the BrdU kit). In brief, the cells were grown on 96-well flat ground plates (Nunc). The cells were not preincubated with serum-free DMEM for 48 h. They were incubated with MOPS buffer with and without Cd²⁺ for the different time periods. Thus the cells were FCS free for up to 6 h only. After treatment, the cells were incubated with BrdU for 18 h. The incorporation of BrdU was determined using a specific antibody against BrdU labeled with peroxidase. After the development of the color reaction, the intensity was determined by an ELISA Reader (Titertek Multiscan MCC/340; Labsystems). The BrdU incorporation is proportional to the proliferation rate of the cells determined by [³H]thymidine incorporation (unpublished results from our laboratory).

ATP levels. For detection of intracellular ATP levels, the luciferin-luciferase assay (ATP Bioluminescence Assay Kit CLS II no. 1699695; Boehringer) was performed as described by the manufacturer. In brief, cells were washed twice with PBS, collected in a tube, centrifugated, and resuspended in boiling Tris-EDTA buffer (100 mM Tris, 4 mM EDTA, pH 7.75). After boiling for another 2 min, samples were centrifugated, and the supernatant was transferred to a fresh tube for ATP determination. Light emission was determined, using a luminometer (Bachoefer). Protein content was determined using the bicinchoninic acid method (Pierce). Concentration of ATP per milligram total protein was calculated and expressed.

Number of experiments, data processing, and statistics. All experiments were undertaken on the same filters in duplicate. For the TER and LDH measurements, 12 experiments (n = 24) with six parallel controls (n = 12) were performed. The BrdU experiments were repeated 3 × 24 (n = 72) for every concentration. The immunostaining was performed three times with three parallel controls, respectively. ATP levels were determined four times.

Statistical significance was tested with the Friedman and Wilcoxon tests, as appropriate.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Determination of TER. The absolute values of TER under control conditions in MDCK and LLC-PK₁ cells are given in Table 1. Each experiment was compared with its own control well. The control experiments did not show significant variation within the study.

TER: apical vs. basolateral application of Cd²⁺ in MDCK cells. It has been shown before that cadmium damages LLC-PK₁ cells more severely after application from the basolateral than from the apical side of the monolayer (22). In our experiments with MDCK cells, the same phenomenon was observed. The effect of Cd²⁺ given to the apical or basolateral compartment is shown in Fig. 1. TER decreased by more than 80%, compared with controls after a dose of Cd²⁺ (10 µM) given into the basolateral compartment for 24 h. The cells treated with the same dose in the apical compartment showed no significant changes of the TER. After 6 h under these conditions, there were no changes in TER, Cd²⁺ being in the apical or the basolateral compartment, respectively.

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Influence of the site of Cd2+ addition on transepithelial resistance (TER) of Madin-Darby canine kidney (MDCK) cells. Cd2+ (10 µM) is applied to the apical or basolateral compartment. Concentration of extracellular Ca2+ is 0.1 mM. Data are means ± SE. *** P < 0.001 (against 0 h), significant fall in TER after 24 h when Cd2+ was applied to the basolateral compartment.

Effect of extracellular Ca²⁺ concentration in Cd²⁺ toxicity on MDCK cells. The effect of cadmium on LLC-PK₁ cells (22) has been reported to be dependent on the extracellular concentration of calcium. In Fig. 2, the results of our experiments with MDCK cells are shown. The lower the extracellular Ca²⁺ concentration, the more pronounced is the damage by cadmium. In each experiment, we added Cd²⁺ (15 µM) to the basolateral compartment of the MDCK monolayer: when the extracellular Ca²⁺ concentration was 2 mM, the monolayer maintained a TER of more than 60% of control values after 24 h. With 0.1 mM Ca²⁺, TER decreased by ~70% of control values after 6 h, and, after 24 h, TER was not measurable. With a concentration of 1 mM extracellular Ca²⁺, the cells showed a decrease to 40% of control values after 24 h (Fig 2).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Influence of different extracellular Ca2+ concentrations on TER of MDCK cells after Cd2+ addition to the basolateral site. Ca2+ concentrations were similar in the apical and basolateral compartment. Cadmium (15 µM) is applied to basolateral site at different concentrations of extracellular Ca2+. Data are means ± SE. *** P < 0.001, * P < 0.05 (against 0 h).

Dose-response curve and Cd²⁺ toxicity: MDCK compared with LLC-PK₁. Different concentrations of Cd²⁺ were added to the basolateral compartment of MDCK and LLC-PK₁ monolayer. The results are shown in Fig. 3. There was a dose-dependent toxicity in MDCK as well as in LLC-PK₁ cells. A 50% decrease in TER in LLC-PK₁ cells occurred with 6.5 mM Cd²⁺. In MDCK cells, a 50% reduction in TER occurred at a concentration of 12 µM Cd²⁺ after 24 h with an extracellular Ca²⁺ of 1 mM.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Influence of Cd2+ concentrations on TER in MDCK compared with LLC-PK1 cells. Different concentrations of Cd2+ are added to the basolateral compartment. Data are means ± SE for a 24-h application time. Concentration of extracellular Ca2+ was 1 mM. # P < 0.005 (MDCK vs. LLC-PK1).

Influence of time and Cd²⁺ toxicity: MDCK compared with LLC-PK₁. Figure 4 shows that cadmium damages LLC-PK₁ cells much faster than MDCK cells. After 6 h with 20 µM Cd²⁺ (basolateral; extracellular Ca²⁺, 1 mM), the MDCK cells had a TER of ~60% of controls, whereas the TER of the LLC-PK₁ monolayer had already decreased to 20% of control values. The TER of LLC-PK₁ cells fell to 60% of control values after 3 h. Both cell lines lost their TER after 24 h under these conditions.

View larger version (8K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Fig. 4.   Comparison of Cd 2+ toxicity in MDCK (A) and LLC-PK1 cells (B). A: MDCK cells for TER and LDH. Cd2+ (20 µM) is added to basolateral compartment. Extracellular concentration of Ca2+ was 1 mM; means ± SE. *** P < 0.001. B: LLC-PK1 cells for TER and LDH. Cd2+ (20 µM) is added to basolateral compartment. Extracellular concentration of Ca2+ was 1 mM; means ± SE. *** P < 0.001.

In these experiments with LLC-PK₁ cells, the LDH concentrations in the medium increased up to 100 U/l after 24 h. In the MDCK experiments, no significant changes were found (Fig. 4A).

Cd²⁺ toxicity: morphology and immunofluorescence findings. Figure 5 compares control MDCK cells with cells exposed to 50 µM Cd²⁺ (basolateral) for 6 h. Below the MDCK monolayer, the collagen-coated filter membrane can be seen as a lucent granular structure. After addition of Cd²⁺ the MDCK cells become rounded, and the integrity of the monolayer is disturbed. The monolayer was partially separated from the collagen matrix of the filter, indicating loss of cell matrix adhesion.

View larger version (113K): [in this window] [in a new window]   Fig. 5.   Influence of Cd2+ on the morphology of MDCK cells. Cd2+ (50 µM) is given from basolateral for 6 h. Concentration of extracellular Ca2+ is 1 mM; magnification, ×200. A: control; B: 50 µM Cd2+. Lucent membrane is the basolateral collagen matrix of the cell culture well. With cadmium, monolayer is disrupted from the collagen matrix (arrowhead). Monolayer increases in height, and cells round up and begin to lose cell contacts, as seen by the increased cell-cell distance.

Because the antibody DECMA-1 specifically binds MDCK E-cadherin and another antibody for LLC-PK₁ E-cadherin was not available, E-cadherin immunofluorescence in LLC-PK₁ cells was not possible. The immunofluorescent distribution of MDCK E-cadherin is given on the left of Fig. 6: under control conditions (DMEM, MOPS), the basolateral cell borders as localizations are pointed out by the fluorescence of E-cadherin molecules. MDCK cells treated with 15 µM Cd²⁺ was only little affected, whereas the immunofluorescence of MDCK E-cadherin at 30 µM Cd²⁺ was no longer restricted to the cell borders. The monolayer was broken. E-cadherin fluorescence was detectable around the whole cells.

View larger version (K): [in this window] [in a new window]   View larger version (79K): [in this window] [in a new window]   Fig. 6.   Influence of Cd2+ on the immunofluorescence pattern of E-cadherin and -catenin distribution in epifluorescence (A) and confocal microscopy (B) (10 mm = 6 µm). Cadmium in concentrations of 15 and 30 µM was added to the incubation medium for 6 h. Concentration of extracellular Ca2+ is 0.1 mM. Under control conditions, staining for E-cadherin and -catenin is restricted basolaterally. Addition of cadmium causes changes in distribution patterns. Immunofluorescence is no longer restricted to basolateral membrane but is seen throughout cytoplasm, indicating a loss of polarity with regard to the zonula adherens and disruption of the cadherin-catenin complex. MDCK cells seem to be more resistant to cadmium compared with LLC-PK1 cells. In epifluorescence, rounding up of cells can be seen as a sign of a loss of cellular contacts. In MDCK cells, pattern for E-cadherin and -catenin are comparable. In LLC-PK1 cells, only -catenin antibodies stained the cells, demonstrating the same behavior after cadmium exposure at lower concentrations.

The immunofluorescent distribution of -catenin is comparable to the immunofluorescent distribution of E-cadherin and shown in the middle and right of Fig. 6. LLC-PK₁ and MDCK cells under control condition both showed an -catenin immunofluorescence pattern that was restricted to the cell-cell borders (basolateral membrane). In cadmium-treated cells, the fluorescence pattern was visible throughout the cells after 6 h in both cell lines.

Figure 6 depicts that cadmium was more toxic to LLC-PK₁ than to MDCK cells. After 6 h under Cd²⁺ (15 µM), the LLC-PK₁ cells have lost almost all cell-cell contacts. A similar effect in MDCK cells was seen only after doubling the Cd²⁺ concentration (30 µM Cd²⁺).

Thus Cd²⁺ changed the distribution of E-cadherin and -catenin; in particular, it caused loss of polarity of the cadherin-catenin complex, which is restricted to the basolateral membrane under control conditions.

Phalloidin staining of cytoskeletal F-actin. Distribution of cytoskeletal F-actin is given on Fig. 7. Under control conditions (DMEM, MOPS), actin filaments are detectable as a network in both cell lines in the whole cell. After treatment with 15 µM Cd²⁺ for 6 h, the intact actin network is still detectable in MDCK cells. In LLC-PK₁ cells, morphological alterations in cell shape and the actin network were visible. After treatment with 30 µM Cd²⁺, this change was found in both cell lines.

View larger version (101K): [in this window] [in a new window]   Fig. 7.   Rhodamine-phalloidin staining of cytoskeletal F-actin filaments (10 mm = 6 µm). Under control conditions, actin filaments pass trough the cells as a network. In Cd2+-treated cells, dose-dependent changes in the cell morphology and actin arrangement are detectable. Stress fibers and pseudopodia can be observed in LLC-PK1 cells after 6 h at 15 µM Cd2+ and at 30 µM Cd2+ in MDCK cells.

Assembly of the cadherin-catenin complex. Results of dot blots from cell extraction and fraction are shown in Table 2. In the detergent-insoluble fraction, E-cadherin and -catenin molecules of the cadherin-catenin complex are detectable in the highest concentration in both cell lines in controls and Cd²⁺-treated cells. In the detergent-soluble fraction, E-cadherin and -catenin molecules can be found in both controls and Cd²⁺-treated cells but in a significantly lower level. Compared with controls, in MDCK cells, a significantly increased content of E-cadherin or -catenin molecules was not detectable in the Triton X-100-soluble fraction after Cd²⁺ treatment, whereas, in LLC-PK₁ cells, E-cadherin as well as -catenin molecules are detectable after treatment with 30 µM Cd²⁺.

View this table: [in this window] [in a new window]   Table 2.   Semiquantitative results of dot blot for -catenin and E-cadherin in LLC-PK1 and MDCK cells

Cd²⁺ toxicity: cell proliferation determined by BrdU incorporation. The proliferation of MDCK cells assessed by BrdU incorporation is proportional to the ELISA reading. Cadmium reduced the proliferation of the cells grown in MOPS buffer compared with controls (Fig. 8). Thus it seems that Cd²⁺ is taken up by the MDCK cells and influences cell proliferation without disruption of the membrane, as shown by the assessed LDH activity in the supernatant (compare Fig. 4).

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Influence of Cd2+ on BrdU incorporation of MDCK cells. Addition of Cd2+ (0.1, 1, 10 µM) to the cells was made for 6 h. Concentration of extracellular Ca2+ is 1 mM. Data are means ± SE. *** P < 0.005 (against control conditions). Cadmium concentrations as low as 0.1 µM caused a decrease in BrdU incorporation, indicating a reduction in DNA synthesis. FCS, fetal calf serum.

ATP levels. In LLC-PK₁ cells grown in DMEM/10% FCS, an intracellular ATP level of 9.4 ± 0.7 nmol ATP/mg total protein (means ± SE) was detectable. In MDCK cells grown under the same conditions, the ATP content was 10.9 ± 0.7 nmol/mg protein. Cd²⁺ was dissolved in MOPS buffer; therefore, MOPS buffer without cadmium was used as reference point. In MOPS buffer, the ATP levels were 7.7 ± 0.4 and 9.5 ± 0.6 nmol ATP/mg protein for LLC-PK₁ and MDCK cells, respectively. For LLC-PK₁ cells, the ATP level in MOPS was slightly lower compared with DMEM/10% FCS (P < 0.05). In Cd²⁺-treated cells, ATP levels decreased significantly concentration dependently, and, at 50 µM Cd²⁺ fell to 1.4 ± 0.1 nmol ATP/mg protein in LLC-PK₁ cells and to 1.6 ± 0.2 nmol ATP/mg protein in MDCK cells (Fig. 9).

View larger version (29K): [in this window] [in a new window]   Fig. 9.   Effect of Cd2+ on ATP levels in LLC-PK1 and MDCK cells. ATP levels in LLC-PK1 controls (con) were 7.7 ± 0.4 nmol ATP/mg protein, and, in MDCK controls, they were 9.5 ± 0.6 nmol ATP/mg protein (MOPS). Effect of Cd2+ was concentration dependent. At 50 µM, Cd2+ in both cell lines ATP levels were decreased to <2 nmol ATP/mg protein. Data are shown as means ± SE of ATP levels (nmol/mg protein); n = 4. *P < 0.05.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture models. In recent years, cell culture models have become more important for physiological and toxicological research. This development is due to new cell lines being introduced and by better characterization of established cell lines. Due to the improved characterization of many cell lines and to special culture techniques (e.g., 2-compartment filter plates), the results of cell culture experiments are more relevant to in vivo conditions (4, 5, 11).

The two-compartment accesses were used to evaluate the effect of apical vs. basolateral application. We used immortalized cell lines because they are well characterized, although their behavior may differ from human cells in vivo.

Choice of Cd²⁺ concentration. Different concentrations of cadmium were chosen for the development of the dose-response curve to compare our system with results of other authors (22). The apical-basolateral experiments (Fig. 1) were undertaken with 10 µM Cd²⁺, a concentration where the effects are more prominent. For the Ca²⁺ experiments (Fig. 2), 15 µM Cd²⁺ was used to get clearer experimental effects on the less-sensitive MDCK cells. For the immunofluorescence and morphological experiments, high Cd²⁺ concentrations were used (15, 30, 50 µM) to get visible effects after 6 h, because the cadmium-treated cells tend to separate from the noncollagen-treated glass surfaces used for these experiments after longer exposure times.

Toxicity of cadmium. The TER is a function of the integrity of the zonula occludens. In MDCK cells, the toxic effect of cadmium was more distinct from the basolateral than from the apical side and does not diffuse from the apical to the basolateral compartment of the monolayer. These characteristics correspond to the usual models of cell-cell contacts: from the apical side, the intercellular space is sealed up by the zonula occludens. From the basolateral side, molecules and ions can diffuse without restriction to the zonula adherens (the cadherin-catenin complex) and to the desmosomes (7, 11, 18).

The damages probably occur directly at the extracellular Ca²⁺ binding domains of the cadherins (11, 18), as high calcium concentrations protect the cells from damage by cadmium. This behavior is similar to previous experiments in LLC-PK₁ cells (22). Cadmium has been shown to be taken up similarly from the basolateral and apical sides of the cells (15). If the damage by cadmium takes place within the intracellular compartment, then it would not matter whether cadmium is taken up from the apical or the basolateral side of the cell.

The surface area may be different on the basolateral compared with the apical side. Assuming that cadmium simply diffuses across the membrane and that the surface area of the basolateral side is considerably higher, cadmium would cause more toxicity from this compartment. Indeed, the surface area of the basolateral membrane is approximately twice compared with the apical membrane (22, 26). However, if simply the surface area would be the major determinant of cadmium toxicity in MDCK cells, the change in extracellular Ca²⁺ concentration would be of no influence. However, that the difference in surface area might contribute to the site-specific Cd²⁺ toxicity cannot be ruled out.

Cadmium damaged the MDCK monolayer at a concentration as low as 10 µM (Ca²⁺ extracellular, 1 mM), and, for LLC-PK₁ cells, the damaging concentration was ~5 µM (Ca²⁺ extracellular, 1 mM), as shown in Fig. 3. This means that cells representing the proximal tubule (LLC-PK₁) decrease their TER under the influence of Cd²⁺ at lower concentrations than cells representing the collecting duct (MDCK). MDCK cells demonstrated a detachment from the basal collagen membrane (Fig. 5). Cell membrane interaction is mainly due to integrin-mediated adhesion. Integrin function is influenced by the calcium concentration in the intra- and extracellular space (23). Thus it seems possible that cadmium interferes with the cell matrix interaction through calcium-dependent mechanisms as well. This has not been studied further, however.

The mechanisms of cadmium toxicity are not identical in both cell lines used. LLC-PK₁ cells are damaged more severely as shown by the LDH release into the supernatant (Fig. 4). This suggests, in agreement with others, that LLC-PK₁ cells are more sensitive to cadmium and accumulate more cadmium than MDCK cells (30).

The initial mechanism of cadmium-induced injury seems to be similar for both cell lines: the heavy metal interferes with the extracellular Ca²⁺ binding sites of cadherins present at the basolateral side of the monolayer cells. However, LLC-PK₁ cells are obviously influenced by additional mechanisms affecting the cell membrane (LDH release into the supernatant), which seem to occur as an advanced process associated with a decrease in TER. Additionally, the cadmium uptake into the MDCK cells (31) is reflected by a decrease in BrdU incorporation.

Cadherin-catenin complex and the actin cytoskeleton. The immunofluorescent pattern of -catenin staining changed under the influence of cadmium in both cell lines. The E-cadherin pattern could be studied in MDCK cells only and changed in the same direction. From the images of immunostaining, it seems likely that the cells first lose their cell-cell contacts, i.e., opening up of the zonula occludens and zonula adherens. This is indicated by the decrease in TER and the sequential loss of cell polarity as demonstrated by the immunofluorescence pattern to -catenin. In analogy to the immunofluorescence studies in ischemia for the pattern of distribution of antibodies against Na⁺-K⁺-ATPase (16) and catenins (33), there are two possible models for the change of the immunofluorescent staining pattern under the influence of cadmium: first, the cadherin-catenin complex is spread out along the apical and basolateral cell-membrane either, i.e., no longer restricted to the basolateral membrane, or, second, the cadherin-catenin complex is destroyed, and more -catenin molecules are solubilized and internalized into the cytosol.

The distribution of the immunofluorescence throughout the cell, as seen on the images, suggests that the loss of polarity is primarily associated with the disruption of the zonula adherens (belt desmosomes, Ref. 18). The appearance of the immunofluorescence in the cytosol, as well as the detection of E-cadherin and -catenin molecules in the detergent-soluble fraction by the immunoblot, supports this notion in LLC-PK₁ cells. The morphological disruption was not detectable in the Triton X-100 solubility experiments for MDCK cells. This supports the notion that LLC-PK₁ cells are more sensitive to cadmium than MDCK cells. Under control conditions, an outnumbered share of E-cadherin and -catenin molecules were detectable in the detergent-soluble fraction, whereas the majority of molecules were detergent insoluble. This supports the assumption that the detergent-soluble fraction in the controls represents the pool of soluble, physiologically not functional molecules. In these cases, the different parts of the cadherin-catenin complex are not associated with each other. As seen for LLC-PK₁ cells, the portion of detergent-soluble molecules was increased after incubation with 30 µM Cd²⁺. This reflects the dose-dependent damage to the cadherin-catenin complex by cadmium in this cell line.

Depolimerization of the F-actin filaments is the main cause for the observed changes in the network structure. It has been described by others before that Cd²⁺ exposure produces depolymerization of the actin cytoskeleton in renal mesangial cells (28). Our results confirm this effect for proximal and distal tubule cells in a dose-dependent fashion.

ATP levels. It could be shown that Cd²⁺ administration to LLC-PK₁ and MDCK cells is connected with an Cd²⁺ concentration-dependent intracellular ATP depletion. Reduced ATP levels could be the trigger for reduced cellular proliferation and may be the initial step, causing morphological changes and alterated cell-cell adherence. However, we have shown recently that chemically induced ATP depletion, using antimycin A and 2-deoxyglucose, leads to a rapid loss of cellular proliferation (29) but different morphologically observed alterations of the cytoskeleton. Furthermore, the cadherin-catenin complex was relatively long preserved during ATP depletion, whereas with Cd²⁺ the cadherin-catenin complex disturbed early. Depolymerization of F-actin could be seen with Cd²⁺ and with ATP depletetion. The pattern was, however, different, and formation of stress fibers could be observed in Cd²⁺ affected but not by antimycin A ATP-depleted cells (chemical anoxia, Ref. 29). In conclusion, Cd²⁺ causes ATP depletion and changes in the cytoskeleton. The difference in the pattern observed leads us to the speculation that ATP alone does not explain the cellular damages seen. In particular, the cadherin-catenin complex pattern is different from the chemical anoxia.

The mechanism of action leading to that damage cannot be further elucidated with the techniques used. In particular, it cannot be clarified whether the protein is degraded via intracellular endonucleases, whether the complex is functionally affected, or whether the catenins and cadherins are falsely sorted by the intracellular sorting machinery. In addition, loss of polarity is also attributed to the dysfunction of the zonula occludens (tight junction, Ref. 18). Actin fibers are also linked to the zonula occludens. Therefore, it seems likely that Cd²⁺ interferes with the zonula occludens, an issue not investigated in the present study.

In conclusion, cadmium induces a calcium-dependent change in cell-cell interaction, intracellular decrease in ATP concentration, and intracellular organization in LLC-PK₁ and MDCK cells. Both cell lines lost their cell-cell contacts, their polarity, and functional properties of renal epithelia. Cd²⁺ toxicity was evident by the effect on the barrier function in proximal and less in distal tubular cells, probably through interaction with the calcium binding motifs of the cadherin molecule, indicated by the dependence on the extracellular Ca²⁺ concentration. In cells representing the proximal tubule, cadmium further caused a destruction of cellular compartments through as yet unknown mechanisms.