Ischemia-induced cleavage of cadherins in NRK cells: evidence for a role of metalloproteinases
http://www.100md.com
《美国生理学杂志》
Departments of Pharmacology and Toxicology and Pathology and Laboratory Medicine, College of Medicine, Texas A&M University System Health Science Center, and Department of Veterinary Integrated Biosciences, College of Veterinary Medicine, Texas A&M University, College Station, Texas
ABSTRACT
Although ischemia has been shown to disrupt cell adhesion, the underlying molecular mechanism is unknown. In these studies, we adapted a model of ischemia-reperfusion to normal rat kidney (NRK) cells, examined disruption of the cadherin/catenin complex, and identified a role for matrix metalloproteinases (MMPs) in ischemia-induced cleavage of cadherins. In NRK cells, ischemia was induced by applying a thin layer of PBS solution supplemented with calcium and magnesium and a layer of mineral oil, which restricts exposure to oxygen. NRK cells exhibited extracellular 80-kDa and intracellular 40-kDa E-cadherin fragments after 4 h of ischemia, and at 6 h the expression of full-length E-cadherin decreased. While no fragments of N-cadherin, -catenin, and -catenin were observed at any time point, the detectable levels of these proteins decreased during ischemia. Ischemia was detected by an increase in pimonidazole adducts, as well as an increase in glucose transporter-1 protein expression. Ischemia did not decrease cell number, but there was a decrease in ATP levels. In addition, there was no evidence of cleaved caspase 3 or 9 during 6 h of ischemia. The MMP inhibitors GM-6001 and TAPI-O inhibited cleavage and/or loss of E- and N-cadherin protein expression. Tissue inhibitors of metalloproteinases (TIMP)-3 and to a lesser extent TIMP-2, but not TIMP-1, inhibit ischemic cleavage and/or loss of E- and N-cadherin. These results demonstrate that ischemia induces a selective metalloproteinase-dependent cleavage of E-cadherin and decrease in N-cadherin that are associated with a disruption of junctional contacts.
normal rat kidney cells; E-cadherin; N-cadherin; matrix metalloproteinase
THE SURVIVAL RATE FOR ACUTE renal failure (ARF) has not improved since the advent of dialysis and, in fact, mortality rates approach 30–50% (42). Although ischemia is a leading cause of ARF, the molecular targets leading to renal injury and failure are not completely understood. In ischemia-induced ARF, a loss of epithelial integrity and shedding of epithelial cells occur in the tubules. After injury, both viable and nonviable cells are shed, leaving the basement membrane as the only barrier between filtrate and interstitium (4a), which allows for backleak of the filtrate and intratubular obstruction from cellular casts (32). The loss of cell-cell attachments may occur via disruption of adherens junctions, which are composed of cadherin and catenin proteins. While evidence suggests that disruption of the cadherin/catenin complex occurs during ischemia (24), the mechanism underlying the loss of cadherin/catenin complex integrity remains unclear.
Adherens junctions are composed of cadherins and catenins and are essential for the formation and maintenance of functional tight junctions (7). Thus the loss of adherens junction integrity may lead to disruption of the normal epithelial barrier due to deficits in the functioning of multiple adhesive complexes (11). The cadherin gene superfamily encodes for transmembrane proteins that regulate Ca2+-dependent cell-cell adhesion (44), which most often requires intracellular association with catenins for adhesive activity. -Catenin is linked to the cytoplasmic domain of cadherins via - or -catenin; -catenin does not directly bind to cadherins, but it is suggested that -catenin links the cadherin/catenin complex to the cytoskeleton (28, 39). The renal expression of cadherin molecules is complex, with reports of at least eight cadherins present in the kidney (E-, K-, Ksp-, N-, OB-, P-, R-, VE-). OB-cadherin is a mesenchymal cadherin (35), whereas P-cadherin is expressed in glomeruli (6). The expression of K- and R-cadherin is mainly confined to the tubules in developing kidneys (6, 8, 10, 34), suggesting that E-, Ksp-, and N-cadherin are the primary cadherins expressed in the tubules of adult kidneys.
Several recent studies have examined the impact of simulated ischemia on the cadherin/catenin complex. Internalization of E-cadherin was seen in Madin-Darby canine kidney (MDCK) cells challenged with antimycin A (10 μM) and 2-deoxyglucose (10 mM) to deplete ATP (18). In addition, ATP depletion using a similar protocol was associated with a loss of full-length E-cadherin, and the generation of an 80-kDa fragment of E-cadherin in MDCK cells, which was not blocked by inhibitors of the proteasomal, lysosomal, or calpain proteolytic pathways (5). Importantly, ischemic rat kidneys demonstrated a reduction in E-cadherin protein levels (5). Taken together, the evidence indicates that ischemia disrupts adherens junctions; however, the mechanism of disruption is not known.
Matrix metalloproteinases (MMPs) are zinc- and calcium-dependent enzymes that are synthesized as zymogens, and once activated, are proteinases that regulate extracellular matrix (ECM) degradation (27, 29). Traditionally MMP substrates were thought to be limited to ECM components, but this view has changed with the discovery of non-ECM substrates, including cadherins. MMP-3 and MMP-7 can cleave E-cadherin, generating an 80-kDa extracellular fragment and a 40-kDa intracellular fragment (30). In addition, MMP-7 has been shown to mediate E-cadherin ectodomain shedding in injured lung epithelium (19). Induced expression of MMP-3 resulted in E-cadherin cleavage, which was blocked by GM-6001, an MMP inhibitor (17).
The objectives of this study were to adapt an in vitro model of ischemic injury to examine ischemia-induced disruption of the cadherin/catenin complex in normal rat kidney (NRK) cells and to determine a potential role for MMPs in this process.
MATERIALS AND METHODS
Cell culture. NRK-52E cells (ATTC, Gaithersburg, MD) were cultured on plastic dishes in Dulbecco’s modified Eagle’s medium containing 1.5 g/l sodium bicarbonate and 5% bovine serum in an atmosphere of 5% CO2-95% air at 37°C. At confluence (4–5 days), subcultures were prepared by treatment with 0.02% EDTA, 0.05% trypsin solution, and cells were seeded at a density of 4 x 104 cells/cm2. Cells were used between passages 3 and 20.
Simulated ischemia. Using a protocol adapted from Meldrum et al. (22), confluent cells were washed twice with PBS before the addition of PBS supplemented with 1.5 mM CaCl2 and 2 mM MgCl2. A layer of mineral oil (Sigma 400–5, St. Louis, MO) was added to the cell culture dish. For a 10-cm2 dish, 2 ml of PBS with Ca2+ and Mg2+ and 10 ml of mineral oil were used. After 6 h of ischemia, cells were washed 5x with PBS and normal growth media was added to simulate reperfusion. To isolate conditioned PBS, the mineral oil and PBS with Ca2+ and Mg2+ were collected, centrifuged, and mineral oil was aspirated. Unless otherwise noted, control cells were incubated for 6 h in PBS with Ca2+ and Mg2+.
Western blot analysis. Cell culture plates were washed twice with PBS, scraped, and centrifuged. The supernatant was removed, and cell pellet was washed with cold PBS and then lysed in buffer (10 mM Tris·HCl, pH 7.6, 1% SDS, 1 mM PMSF, 1 mM leupeptin, 1 mM orthovanadate) and boiled for 10 min. The homogenates were spun at 18,000 g for 10 min, and the supernatant was collected. Proteins were quantified by the Bradford method and diluted to 1 μg/μl in 2x sample buffer (250 mM Tris·HCl, pH 6.8, 4% SDS, 10% glycerol, 2% -mercaptoethanol, 0.006% bromophenol blue). Samples were boiled for 5 min before electrophoresis, and 20 μg of protein were separated by 8, 12, or 15% SDS-PAGE. Separated proteins were transferred onto a Hybond-ECL nitrocellulose membrane (Amersham, Piscataway, NJ) in transfer buffer (25 mM Tris, 200 mM glycine, 20% methanol, and 1% SDS). Nonspecific binding was blocked by incubation with Tris-buffered saline plus Tween 20 (TBST) blocking buffer (0.1% Tween 20, 10 mM Tris, pH 7.5, 100 mM NaCl) supplemented with 5% nonfat dry milk for 1 h at room temperature. Primary antibodies (Table 1) were diluted in the same buffer and incubated at 4°C overnight. After subsequent washes with TBST, membranes were incubated with secondary antibody (1:20,000 in TBST, 5% nonfat dry milk) against the appropriate species for 1 h at room temperature. The blots were washed 3x in TBST, and proteins were detected with the Amersham ECL system and exposure to X-ray film (Kodak, Rochester, NY).
View this table:
Immunofluorescence. Cells grown on Lab Tek Chamber Slides (Nunc, Rochester, NY) were washed twice with PBS, fixed in 2% paraformaldehyde for 10 min, washed two times in 0.02 M PBS for 10 min, and permeabilized in 1% Triton X-100 in 0.02 M PBS for 10 min. Slides were treated with 1:20 blocking solutions of serum related to the species in which the secondary antibody was generated at room temperature for 1 h. Primary antibodies were added at appropriate dilutions overnight. After washing (0.3% Tween in 0.02 M PBS; PBST), FITC-conjugated secondary antibodies (1:200) were added and sections were incubated in the dark at room temperature for 1 h. Slides were mounted with antifade media (Molecular Probes, Eugene, OR) following several washes. Immunostained slides were visualized with a Zeiss Axioplan 2 microscope (Zeiss, Thornwood, NY) fitted with an Axiocam HR digital camera and Axiovision 3.0 software. Negative controls involved substituting IgG for primary antibodies and appropriate species serum for secondary antibodies. The actin cytoskeleton was detected using Alexa Fluor 488-phalloidin (Molecular Probes) staining. A Hypoxyprobe Kit-1 (Chemicon, Temecula, CA) was used to detect ischemia using the manufacturer’s protocol.
ATP levels. ATP levels were detected using an ATP assay kit based on luminescence (Calbiochem, La Jolla, CA) according to the manufacturer’s instructions.
Aggregation assay. The aggregation assay developed for cadherins, where calcium is present during trypsin treatment to preserve cadherin function, was utilized (9). Normal and ischemic NRK cells were washed with PBS containing 1.5 mM CaCl2 and 2 mM MgCl2, before the addition of 0.025% trypsin with Ca2+ and Mg2+ for 10–15 min at 37°C in the presence of 5% CO2. Cells were collected and resuspended at 1.0 x 105 cells/ml in either PBS with 1 mM EDTA or PBS with Ca2+ and Mg2+. From this suspension, 1.5 ml were placed into a 12.5-mm-diameter well coated with 1.5% agarose in PBS. Cells were incubated in a rotary shaker at 80 rpm and 37°C for 20 and 60 min. Aggregated cells were counted manually in a hemacytometer.
Inhibitors. All inhibitors were added to PBS with Ca2+ and Mg2+ 30 min before ischemia. Chemical MMP inhibitors GM-6001 (Calbiochem, San Diego, CA) and TAPI-O (Peptides International, Louisville, KY) were dissolved in DMSO, and the desired concentration was added. Recombinant tissue inhibitors of metalloproteinase (TIMP)-1, TIMP-2, and TIMP-3 (Chemicon) were solubilized at 100 μg/ml in 1 mg/ml of BSA in PBS with Ca2+ and Mg2+. BSA (1 ml/ml) was added to PBS with Ca2+ and Mg2+ in control cells for recombinant TIMP experiments. NRK cells were transfected with 50–100 multiplicity of infection (MOI) of green fluorescent protein or TIMP-3 adenoviruses, constructed as previously described (4, 12).
Statistics. Data are expressed as means ± SE. Groups were compared using ANOVA followed by the Bonferroni post hoc test. P < 0.05 was defined as significant.
RESULTS
Protein expression of cadherins/catenins in ischemic NRK cells. We used a model developed by Meldrum et al. (22) that closely parallels in vivo renal ischemia, in which mineral oil is used to simulate ischemia by restricting exposure to oxygen and metabolite washout. Based on the requirement for Ca2+ in cadherin function, we modified the model by using PBS supplemented with Ca2+ and Mg2+ overlayered with mineral oil to induce ischemia as opposed to a direct overlayering of the confluent cells with mineral oil. Protein expression of cadherins and catenins in ischemic NRK cells was evaluated by Western blot analysis. Using an antibody targeting the cytoplasmic domain of E-cadherin (Transduction Laboratories), cleavage of E-cadherin to a 40-kDa fragment was seen after 3 h of ischemia, and full-length (120 kDa) E-cadherin was lost following 6 h of ischemia (Fig. 1A). Although no fragmentation of N-cadherin was detected, N-cadherin expression began to decrease after 3 h and was virtually lost following 5 h of ischemia (Fig. 1A). Cadherin expression was stable in NRK cells incubated for 6 h in PBS supplemented with Ca2+ and Mg2+ (Fig. 1A), suggesting that the loss of protein expression was due to the ischemic insult, i.e., the mineral oil overlay. However, the simulated mineral oil overlay model includes substrate deprivation as well as oxygen restriction, both of which may be important. As such, a similar pattern of E-cadherin fragmentation (80- and 40-kDa fragments) was seen in ischemic NRK cells in which normal growth media was substituted for PBS with Ca2+ and Mg2+; however, the extent of loss of full-length E-cadherin was not as great (data not shown). Similar results were also seen in NRK cells placed in an Anaeropack system designed for anaerobic cell culture to simulate ischemia (14), suggesting that the fragmentation of E-cadherin and loss of N-cadherin occur in multiple models and is not an artifact of the mineral oil overlay model.
While there was no decrease in -catenin over the ischemia time course, a slight decrease in expression of p120 and -catenin was detected, as well as a dramatic loss of -catenin protein expression at 6 h (Fig. 1A). To examine whether the 40-kDa E-cadherin fragment was generated by extracellular cleavage, the conditioned PBS was isolated from ischemic cells. Using an antibody directed against the extracellular domain of E-cadherin (Santa Cruz Biotechnology), an 80-kDa E-cadherin fragment that is released from the cell surface in ischemic NRK cells was identified (Fig. 1B). A complete recovery of full-length E- and N-cadherin was seen after 24 and 48 h of reperfusion, demonstrating the reversible nature of the ischemic insult (Fig. 2).
Ischemic model. To confirm that the mineral oil overlay model induced ischemia and to compare the biochemical impact of the model with other in vitro systems used to simulate ischemia, the expression of two ischemia-inducible markers was assessed. First, utilizing the Hypoxyprobe Kit-1 (Chemicon), immunofluorescence was performed on ischemic NRK cells. Because pimonidazole forms long-lived adducts with thiol groups in proteins, peptides, and amino acids in ischemic cells, a monoclonal antibody was used to detect the pimonidazole adducts. In control NRK cells, there was no detection of pimonidazole adducts, but after 4–6 h of ischemia, specific staining with the probe was detected (Fig. 3A). To further verify the induction of ischemia, protein expression of an ischemia-inducible protein, the glucose transporter-1 (GLUT-1) (46), was measured. A time-dependent increase in GLUT-1 protein expression was seen in ischemic NRK cells (Fig. 3B).
After 2 h of ischemia, cellular levels of ATP decreased to 46.5 ± 2% of control and rapidly declined to 9.3 ± 0.8% after 6 h of ischemia (Fig. 3C). ATP levels were restored to 51.2 ± 1% of control after 2 h of reperfusion and returned to control levels following 24 h of reperfusion. The decrease in ATP levels is similar to the 90% decrease seen in cells treated with sodium cyanide (15). In addition, caspase activation was determined by Western blot analysis using antibodies for cleaved (i.e., activated) caspase 3 and 9. There was no detection of cleaved caspase 3 or 9 at 2–6 h of ischemia, and total protein expression of caspase 3 was not affected (Fig. 3D). In addition, cleavage of caspases was observed at 12 h of ischemia, suggesting induction of apoptosis following prolonged ischemia (Fig. 3D). As expected, cleavage of both caspases was seen in NRK cells treated with staurosporine. Taken together, these results demonstrate that this in vitro model of ischemia is associated with hypoxia, a transient decrease in cellular ATP levels that is reversible following reperfusion, and is not associated with activation of caspases during the 6-h ischemic insult.
Cadherin/catenin function. Cadherin/catenin protein localization was evaluated by immunofluorescence microscopy. While normal NRK cells expressed E-cadherin at the plasma membrane, 4 h of ischemia resulted in a decrease in E-cadherin at the cell membrane, which was further decreased after 6 h (Fig. 4). N-cadherin expression in normal NRK cells, although patchy, is located at the cell membrane, but after 4 and 6 h of ischemia N-cadherin expression was lost. After 6 h of ischemia, there was a loss of p120, -, and -catenin expression, whereas no change in the expression of -catenin was observed (Fig. 4). Changes in the actin cytoskeleton were examined using Alexa Fluor 488-phalloidin staining. In normal NRK cells, actin filaments displayed staining that outlines the cell periphery as well as actin stress fibers. Ischemia resulted in disruption of the cytoskeleton, primarily observed as a decrease in the actin stress fibers (Fig. 4). However, there was no disruption in the actin at the cell periphery, as was seen when cells were treated with 10 μM cytochalasin D (data not shown).
NRK cells viewed by phase-contrast microscopy revealed characteristically closely packed polygon-shaped cells with little light transmitted between them (Fig. 5A). Following 6 h of ischemia, cells exhibited a loss of cell-cell adhesion and increased transmission of light at cell-cell boundaries, although cells remained attached to the growing surface (Fig. 5A). After 3 h of reperfusion, cells looked similar to control, which is indicative of cell recovery from ischemia. Using the classic aggregation assay developed for cadherins (9), the function of cadherins in ischemic NRK cells was examined (Fig. 5B). Control and ischemic NRK cells were allowed to aggregate for 20 and 60 min at 37°C while being shaken at 80 rpm. A significant decrease in cell aggregation induced by 6 h of ischemia was seen at both 20 (30 ± 3% of control) and 60 min (52 ± 4% of control) of cell aggregation, and this functional deficit was still present after 3 h of reperfusion. However, after 24 h of reperfusion cell aggregation returned to control values (Fig. 5B), which corresponds to the reexpression of full-length cadherins (Fig. 2).
Ischemia-induced cleavage of cadherins: evidence for a role of MMPs. To determine whether MMPs were involved in ischemia-induced disruption of cadherins, the MMP inhibitors GM-6001 and TAPI-O were added to the PBS with Ca2+ and Mg 2+ before ischemia. Both 10 μM GM6001 and 50 μM of TAPI-O blocked E-cadherin fragmentation and loss of N-cadherin (Fig. 6), suggesting that MMPs play a role in ischemia-induced cleavage and loss of cadherins. Interestingly, an increase in N-cadherin expression was seen in NRK cells treated with GM-6001. In addition, recombinant TIMPs were used to inhibit MMP activity. TIMP-1 did not prevent ischemic cleavage and/or loss of E- or N-cadherin (Fig. 7A). Although TIMP-2 protected full-length E-cadherin, there was a 40-kDa fragment of E-cadherin present and a loss of N-cadherin expression. In contrast, TIMP-3 inhibited both cleavage and/or loss of E- and N-cadherin expression (Fig. 7A). Similar inhibitory activity was also seen in NRK cells transfected with a TIMP-3 adenovirus but not the green fluorescent protein control (Fig. 7B). These data collectively indicate the involvement of a membrane-associated metalloproteinase in E- and N-cadherin cleavage following ischemic insult in NRK cells.
DISCUSSION
Because ischemia is a leading cause of ARF, investigating the targets of ischemic insult will provide insight into the mechanisms underlying ARF. The cadherin/catenin complex, which is critical to renal proximal epithelial cell function, is disrupted by ischemic insult. In this study, an ischemic model, which resembles ischemic renal injury in vivo, resulted in selective fragmentation/loss of E-cadherin and loss of N-cadherin expression that could be blocked by MMP inhibitors.
Most in vitro ischemic models use chemicals to induce ischemia, such as ATP depletion with deoxy-D-glucose and antimycin A or sodium cyanide (5, 15). To avoid using chemicals that may have other nonspecific molecular targets and to mimic ischemic insult in vivo (oxygen/nutrient deprivation and metabolite accumulation), an alternative model for ischemia was employed in the present study using PBS supplemented with Ca2+ and Mg2+ overlayered with mineral oil. We have adapted this model for a renal cell line from Meldrum et al. (22); however, the mineral oil model was first used in isolated myocytes (13, 41). In the original and subsequent studies by Meldrum et al. (20–23), mineral oil is directly overlaid onto confluent LLC-PK1 cultures. For our studies of cadherin/catenin expression and function, we removed the confounding factor of divalent cation removal by first using a thin layer of PBS supplemented with Ca2+ and Mg2+. The Meldrum group uses 1 h of simulated ischemia, which is much shorter than the time period used in our studies. However, when mineral oil was directly added to NRK cultures, the fragmentation and loss of E-cadherin were seen at earlier time points (data not shown), suggesting that the supplemented PBS layer may be partially attenuating the ischemic insult. Consistent with this conclusion is the fact that replacement of the supplemented PBS with complete culture media was associated with E-cadherin fragmentation, but to a lesser extent than with PBS plus Ca2+ and Mg2+ (data not shown). In agreement with previous results using TNF- as a marker of ischemia (22), the induction of ischemic markers pimonidazole adducts and GLUT-1 was seen NRK cells. In addition, the decrease in ATP levels following 6 h of ischemia in this study is similar to the 90% decrease seen in cells treated with sodium cyanide (15).
Caspase 9 is activated on cytochrome c release from the mitochondria and further activates other caspases, such as caspase 3, to initiate a cascade leading to apoptosis (36). During the 6-h time course of ischemia, no evidence of caspase activation was detected; however, prolonged ischemia (12 h) resulted in caspase activation. While caspase 3 has been shown to cleave E-cadherin to generate a 24-kDa fragment (38), the lack of temporal correlation between caspase activation and cadherin loss suggests that this protease is not responsible for cadherin/catenin disruption in this system. Cleavage of E-cadherin and loss of N-cadherin occur before caspase activation, suggesting that disruption of the cadherin/catenin complex precedes ischemia-induced apoptosis. However, the possibility remains that significant apoptosis occurs during the reperfusion phase, as reported by Meldrum et al. (20–23).
Disruption of the cadherin/catenin complex at the cell membrane is concurrent with altered cell morphology, as evidenced by the loss of cell-cell contact after 6 h of ischemia. The integrity of intracellular junctions and polarity in epithelial cells depends on cadherin-mediated cell-cell contacts (2). Therefore, cadherin disruption and loss correlate with an altered phenotype in ischemic epithelial cells that includes, but is not limited to, the loss of cell polarity and shedding of tubular epithelial cells. Given the relationship between the cadherin/catenin complex and the actin cytoskeleton, we also examined the pattern of actin staining during ischemia. In normal NRK cells, actin filaments displayed a staining pattern that outlines the cell periphery as well as actin stress fibers, similar to that seen in previous studies (1). Six hours of ischemia resulted in disruption of actin stress fibers, but the junctional staining of actin was preserved. Similar results were seen in mouse proximal tubule cells treated with cyanide to induce chemical anoxia (16).
Disruption of cadherin/catenin integrity may also be associated with activation of the transcriptional activity of -catenin (43). However, during ischemia or reperfusion -catenin was not seen in the nucleus. Thus far, we have not been able to provide evidence for activation of -catenin cell signaling during ischemia in our in vitro model. In contrast, ATP depletion with sodium cyanide induces translocation of -catenin into the nucleus in opossum kidney cells (32). The discrepancy in experiments could be due to the different methods utilized to induce ischemia or the different cell lines used.
Fragmentation of N-cadherin is seen in apoptotic hepatic cells and proliferating vascular smooth muscle cells (26, 40); however, in our model ischemia was associated with a loss of N-cadherin protein expression without detectable fragments. It is possible that the antibody used in our studies cannot recognize the cleaved fragments. However, a panel of four commercially available antibodies did not detect fragments in ischemic NRK cells. The fact that MMP inhibition blocks the loss of N-cadherin suggests that this loss may be due to proteolytic cleavage. In addition, GM-6001 was associated with a significant increase in N-cadherin levels in control NRK cells, further supporting a role for MMPs in regulating N-cadherin expression.
Although rapid internalization of E-cadherin is seen in ATP-depleted cultured renal epithelial cells (18), a longer insult leads to proteolytic clipping of E-cadherin and disruption of the complex (5). Previous studies have shown a loss of E-cadherin in ischemia both in vivo and in vitro (5). ATP-depleted MDCK cells expressed an 80-kDa fragment and loss of full-length E-cadherin that could not be prevented with inhibitors of the proteasomal, lysosomal, or calpain-mediated proteolytic pathways (5). The 40-kDa E-cadherin fragment is believed to correspond to the cytoplasmic and transmembrane domain of the protein. Detection of the extracellular 80-kDa fragment of E-cadherin in conditioned PBS of ischemic NRK cells suggests that the cleavage event occurs extracellularly.
Our results suggest that MMPs play a key role in the cleavage and/or loss of E- and N-cadherin by inhibition with GM-6001 and TAPI-O. Recent data indicate that MMPs play a role in ischemic ARF, as MMPs were increased in rat postischemic kidney tissue and were localized to the renal tubules (3). In addition, MMPs are thought to be involved in ischemia-induced damage to other organs, such as the brain, lung, and heart (25, 33, 37, 45). Furthermore, this study has shown that TIMP-3 completely blocks both cleavage and/or loss of E- and N-cadherin, whereas TIMP-2 protects full-length E-cadherin protein expression but not loss of N-cadherin. These differences imply that different MMP(s) may play a role in E- and N-cadherin regulation. This hypothesis is further supported by the finding that lower concentrations (10 μM) of TAPI-O blocked the loss of N-cadherin but did not attenuate E-cadherin fragmentation in ischemic NRK cells (data not shown). Because ADAMs (for a disintegrin and metalloproteinase domain) are also inhibited by TIMP-3, it is possible that this family of metalloproteinases may play a role in ischemia-induced cleavage and/or loss of E- and N-cadherin.
In summary, simulated in vitro ischemia was associated with fragmentation of E-cadherin and loss of N-cadherin before loss of -, -, and p120 catenin. A loss of cell-cell contacts corresponds to the disruption of cell adhesion. Ischemia-induced cleavage and/or loss of E- and N-cadherin can be blocked by GM-6001 and TAPI-O, suggesting a potential role for MMPs. In addition, the present study shows that the MMP(s) responsible for ischemia-induced E-cadherin cleavage is inhibited by TIMP-2 and TIMP-3, and loss of N-cadherin is inhibited by TIMP-3, indicating a novel role for membrane-associated metalloproteases in these events. Future experiments will examine whether specific MMP inhibitors can decrease the pathological sequelae following ischemia-induced ARF.
GRANTS
These studies were supported by the Department of Medical Pharmacology and Toxicology, the Center for Environmental and Rural Health (P30-ES09106), and National Heart, Lung, and Blood Institute Grant HL-59373 (G. E. Davis).
ACKNOWLEDGMENTS
We thank Bert Vogelstein for kindly providing the AdEasy adenoviral system.
FOOTNOTES
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.
REFERENCES
Alvarez C and Sztul ES. Brefeldin A (BFA) disrupts the organization of the microtubule and the actin cytoskeletons. Eur J Cell Biol 78: 1–14, 1999.
Bannerman DD, Sathyamoorthy M, and Goldblum SE. Bacterial lipopolysaccharide disrupts endothelial monolayer integrity and survival signaling events through caspase cleavage of adherens junction proteins. J Biol Chem 273: 35371–35380, 1998.
Basile DP, Fredrich K, Weihrauch D, Hattan N, and Chilian WM. Angiostatin and matrix metalloprotease expression following ischemic acute renal failure. Am J Physiol Renal Physiol 286: F893–F902, 2004.
Bayless KJ and Davis GE. The Cdc42 and Rac1 GTPases are required for capillary lumen formation in three-dimensional extracellular matrices. J Cell Sci 115: 1123–1136, 2002.
Bonventre JV and Weinberg JM. Recent advances in the pathohysiology of ischemic acute renal failure. J Am Soc Nephrol 14: 2199–2210, 2003.
Bush KT, Tsukamoto T, and Nigam SK. Selective degradation of E-cadherin and dissolution of E-cadherin-catenin complexes in epithelial ischemia. Am J Physiol Renal Physiol 278: F847–F852, 2000.
Cho EA, Patterson LT, Brookhiser WT, Mah S, Kintner C, and Dressler GR. Differential expression and function of cadherin-6 during renal epithelium development. Development 125: 803–812, 1998.
Citi S. The molecular organization of tight junctions. J Cell Biol 121: 485–489, 1993.
Dahl U, Sjodin A, Larue L, Radice GL, Cajander S, Takeichi M, Kemler R, and Semb H. Genetic dissection of cadherin function during nephrogenesis. Mol Cell Biol 22: 1474–1487, 2002.
Goldberg M, Peshkovsky C, Shifteh A, and Al-Awqati Q. μ-Protocadherin, a novel developmentally regulated protocadherin with mucin-like domains. J Biol Chem 275: 24622–24629, 2000.
Goto S, Yaoita E, Matsunami H, Kondo D, Yamamoto T, Kawasaki K, Arakawa M, and Kihara I. Involvement of R-cadherin in the early stage of glomerulogenesis. J Am Soc Nephrol 9: 1234–1241, 1998.
Gumbiner BM. Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell 84: 345–357, 1996.
He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, and Vogelstein B. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA 95: 2509–2514, 1998.
Henry P, Popescu A, Puceat M, Hinescu M, and Escande D. Acute simulated ischemia produces both inhibition and activation of K+ currents in isolated ventricular myocytes. Cardiovasc Res 32: 930–939, 1996.
Kaibori M, Inoue T, Tu W, Oda M, Kwon AH, Kamiyama Y, and Okumura T. FK506, but not cyclosporin A, prevents mitochondrial dysfunction during hypoxia in rat hepatocytes. Life Sci 69: 17–26, 2001.
Keller SH and Nigam SK. Biochemical processing of E-cadherin under cellular stress. Biochem Biophys Res Commun 307: 215–223, 2003.
Kroshian VM, Sheridan AM, and Lieberthal W. Functional and cytoskeletal changes induced by sublethal injury in proximal tubular epithelial cells. Am J Physiol Renal Fluid Electrolyte Physiol 266: F21–F30, 1994.
Lochter A, Galosy S, Muschler J, Freedman N, Werb Z, and Bissell MJ. Matrix metalloproteinase stromelysin-1 triggers a cascade of molecular alterations that leads to stable epithelial-to-mesenchymal conversion and a premalignant phenotype in mammary epithelial cells. J Cell Biol 139: 1861–1872, 1997.
Mandel LJ, Doctor RB, and Bacallao R. ATP depletion: a novel method to study junctional properties in epithelial tissues. II. Internalization of Na+,K+-ATPase and E-cadherin. J Cell Sci 107: 3315–3324, 1994.
McGuire JK, Li Q, and Parks WC. Matrilysin (matrix metalloproteinase-7) mediates E-cadherin ectodomain shedding in injured lung epithelium. Am J Pathol 162: 1831–1843, 2003.
Meldrum KK, Burnett AL, Meng X, Misseri R, Shaw MBK, Gearhart JP, and Meldrum DR. Liposomal delivery of heat shock protein 72 into renal tubular cells blocks nuclear factor-B activation, tumor necrosis factor- production, and subsequent ischemia-induced apoptosis. Circ Res 92: 293–299, 2003.
Meldrum KK, Hile K, Meldrum DR, Crone JA, Gearhart JP, and Burnett AL. Simulated ischemia induces renal tubular cell apoptosis through a nuclear factor-B dependent mechanism. J Urol 168: 248–252, 2002.
Meldrum KK, Meldrum DR, Hile KL, Burnett AL, and Harken AH. A novel model of ischemia in renal tubular cells which closely parallels in vivo injury. J Surg Res 99: 288–293, 2001.
Meldrum KK, Meldrum DR, Hile KL, Yerkes EB, Ayala A, Cain MP, Rink RC, Casale AJ, and Kaefer MA. P38 MAPK mediates renal tubular cell TNF- production and TNF--dependent apoptosis during simulated ischemia. Am J Physiol Cell Physiol 281: C563–C570, 2001.
Molitoris BA and Marrs J. The role of cell adhesion molecules in ischemic acute renal failure. Am J Med 106: 583–592, 1999.
Muhs BE, Plitas G, Delgado Y, Ianus I, Shaw JP, Adelman MA, Lamparello P, Shamamian P, and Gagne P. Temporal expression and activation of matrix metalloproteinases-2, -9, and membrane type 1-matrix metalloproteinase following acute hindlimb ischemia. J Surg Res 111: 8–15, 2003.
Murphy F, Waung J, Collins J, Arthur MJ, Nagase H, Mann D, Benyon RC, and Iredale JP. N-cadherin cleavage during activated hepatic stellate cell apoptosis is inhibited by tissue inhibitor of metalloproteinase-1 (Abstract). Comp Hepatol 3, Suppl 1: S8, 2004.
Murphy GJ, Murphy G, and Reynolds JJ. The origin of matrix metalloproteinases and their familial relationships. FEBS Lett 289: 4–7, 1991.
Nagafuchi A, Takeichi M, and Tsukita S. The 102 kd cadherin-associated protein: similarity to vinculin and posttranscriptional regulation of expression. Cell 65: 849–857, 1991.
Nagase H and Woessner JF Jr. Matrix metalloproteinases. J Biol Chem 274: 21491–21494, 1999.
Noe V, Fingleton B, Jacobs K, Crawford HC, Vermeulen S, Steelant W, Bruyneel E, Matrisian LM, and Mareel M. Release of an invasion promoter E-cadherin fragment by matrilysin and stromelysin-1. J Cell Sci 114: 111–118, 2001.
Price VR, Reed CA, Lieberthal W, and Schwartz JH. ATP depletion of tubular cells causes dissociation of the zonula adherens and nuclear translocation of beta-catenin and LEF-1. J Am Soc Nephrol 13: 1152–1161, 2002.
Schulze CJ, Wang W, Suarez-Pinzon WL, Sawicka J, Sawicki G, and Schulz R. Imbalance between tissue inhibitor of metalloproteinase-4 and matrix metalloproteinases during acute myocardial ischemia-reperfusion injury. Circulation 107: 2487–2492, 2003.
Shimazui T, Oosterwijk-Wakka J, Akaza H, Bringuier PP, Ruijter E, Debruyne FM, Schalken JA, and Oosterwijk E. Alterations in expression of cadherin-6 and E-cadherin during kidney development and in renal cell carcinoma. Eur Urol 38: 331–338, 2000.
Simonneau L, Kitagawa M, Suzuki S, and Thiery JP. Cadherin 11 expression marks the mesenchymal phenotype: towards new functions for cadherins Cell Adhes Commun 3: 115–130, 1995.
Slee EA, Harte MT, Kluck RM, Wolf BB, Casiano CA, Newmeyer DD, Wang HG, Reed JC, Nicholson DW, Alnemri ES, Green DR, and Martin SJ. Ordering the cytochrome c-initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and -10 in a caspase-9-dependent manner. J Cell Biol 144: 281–292, 1999.
Sole S, Petegnief V, Gorina R, Chamorro A, and Planas AM. Activation of matrix metalloproteinase-3 and agrin cleavage in cerebral ischemia/reperfusion. J Neuropathol Exp Neurol 63: 338–349, 2004.
Steinhusen U, Weiske J, Badock V, Tauber R, Bommert K, and Huber O. Cleavage and shedding of E-cadherin after induction of apoptosis. J Biol Chem 276: 4972–4980, 2001.
Tsukita S, Furuse M, and Itoh M. Structural and signalling molecules come together at tight junctions. Curr Opin Cell Biol 11: 628–633, 1999.
Uglow EB, Slater S, Sala-Newby GB, Aguilera-Garcia CM, Angelini GD, Newby AC, and George SJ. Dismantling of cadherin-mediated cell-cell contacts modulates smooth muscle cell proliferation. Circ Res 92: 1314–1321, 2003.
Vanheel B, Leybaert L, de Hemptinne A, and Leusen I. Simulated ischemia and intracellular pH in isolated ventricular muscle. Am J Physiol Cell Physiol 257: C365–C376, 1989.
Venkataraman R and Kellum JA. Novel approaches to the treatment of acute renal failure. Expert Opin Investig Drugs 12: 1353–1366, 2003.
Willert K and Nusse R. Beta-catenin: a key mediator of Wnt signaling. Curr Opin Genet Dev 8: 95–102, 1998.
Wu Q and Maniatis T. A striking organization of a large family of human neural cadherin-like cell adhesion genes. Cell 97: 779–790, 1999.
Yano M, Omoto Y, Yamakawa Y, Nakashima Y, Kiriyama M, Saito Y, and Fujii Y. Increased matrix metalloproteinase 9 activity and mRNA expression in lung ischemia-reperfusion injury. J Heart Lung Transplant 20: 679–686, 2001.
Zhang JZ, Behrooz A, and Ismail-Beigi F. Regulation of glucose transport by hypoxia. Am J Kidney Dis 34: 189–202, 1999.(Marisa D. Covington, Kayl)
ABSTRACT
Although ischemia has been shown to disrupt cell adhesion, the underlying molecular mechanism is unknown. In these studies, we adapted a model of ischemia-reperfusion to normal rat kidney (NRK) cells, examined disruption of the cadherin/catenin complex, and identified a role for matrix metalloproteinases (MMPs) in ischemia-induced cleavage of cadherins. In NRK cells, ischemia was induced by applying a thin layer of PBS solution supplemented with calcium and magnesium and a layer of mineral oil, which restricts exposure to oxygen. NRK cells exhibited extracellular 80-kDa and intracellular 40-kDa E-cadherin fragments after 4 h of ischemia, and at 6 h the expression of full-length E-cadherin decreased. While no fragments of N-cadherin, -catenin, and -catenin were observed at any time point, the detectable levels of these proteins decreased during ischemia. Ischemia was detected by an increase in pimonidazole adducts, as well as an increase in glucose transporter-1 protein expression. Ischemia did not decrease cell number, but there was a decrease in ATP levels. In addition, there was no evidence of cleaved caspase 3 or 9 during 6 h of ischemia. The MMP inhibitors GM-6001 and TAPI-O inhibited cleavage and/or loss of E- and N-cadherin protein expression. Tissue inhibitors of metalloproteinases (TIMP)-3 and to a lesser extent TIMP-2, but not TIMP-1, inhibit ischemic cleavage and/or loss of E- and N-cadherin. These results demonstrate that ischemia induces a selective metalloproteinase-dependent cleavage of E-cadherin and decrease in N-cadherin that are associated with a disruption of junctional contacts.
normal rat kidney cells; E-cadherin; N-cadherin; matrix metalloproteinase
THE SURVIVAL RATE FOR ACUTE renal failure (ARF) has not improved since the advent of dialysis and, in fact, mortality rates approach 30–50% (42). Although ischemia is a leading cause of ARF, the molecular targets leading to renal injury and failure are not completely understood. In ischemia-induced ARF, a loss of epithelial integrity and shedding of epithelial cells occur in the tubules. After injury, both viable and nonviable cells are shed, leaving the basement membrane as the only barrier between filtrate and interstitium (4a), which allows for backleak of the filtrate and intratubular obstruction from cellular casts (32). The loss of cell-cell attachments may occur via disruption of adherens junctions, which are composed of cadherin and catenin proteins. While evidence suggests that disruption of the cadherin/catenin complex occurs during ischemia (24), the mechanism underlying the loss of cadherin/catenin complex integrity remains unclear.
Adherens junctions are composed of cadherins and catenins and are essential for the formation and maintenance of functional tight junctions (7). Thus the loss of adherens junction integrity may lead to disruption of the normal epithelial barrier due to deficits in the functioning of multiple adhesive complexes (11). The cadherin gene superfamily encodes for transmembrane proteins that regulate Ca2+-dependent cell-cell adhesion (44), which most often requires intracellular association with catenins for adhesive activity. -Catenin is linked to the cytoplasmic domain of cadherins via - or -catenin; -catenin does not directly bind to cadherins, but it is suggested that -catenin links the cadherin/catenin complex to the cytoskeleton (28, 39). The renal expression of cadherin molecules is complex, with reports of at least eight cadherins present in the kidney (E-, K-, Ksp-, N-, OB-, P-, R-, VE-). OB-cadherin is a mesenchymal cadherin (35), whereas P-cadherin is expressed in glomeruli (6). The expression of K- and R-cadherin is mainly confined to the tubules in developing kidneys (6, 8, 10, 34), suggesting that E-, Ksp-, and N-cadherin are the primary cadherins expressed in the tubules of adult kidneys.
Several recent studies have examined the impact of simulated ischemia on the cadherin/catenin complex. Internalization of E-cadherin was seen in Madin-Darby canine kidney (MDCK) cells challenged with antimycin A (10 μM) and 2-deoxyglucose (10 mM) to deplete ATP (18). In addition, ATP depletion using a similar protocol was associated with a loss of full-length E-cadherin, and the generation of an 80-kDa fragment of E-cadherin in MDCK cells, which was not blocked by inhibitors of the proteasomal, lysosomal, or calpain proteolytic pathways (5). Importantly, ischemic rat kidneys demonstrated a reduction in E-cadherin protein levels (5). Taken together, the evidence indicates that ischemia disrupts adherens junctions; however, the mechanism of disruption is not known.
Matrix metalloproteinases (MMPs) are zinc- and calcium-dependent enzymes that are synthesized as zymogens, and once activated, are proteinases that regulate extracellular matrix (ECM) degradation (27, 29). Traditionally MMP substrates were thought to be limited to ECM components, but this view has changed with the discovery of non-ECM substrates, including cadherins. MMP-3 and MMP-7 can cleave E-cadherin, generating an 80-kDa extracellular fragment and a 40-kDa intracellular fragment (30). In addition, MMP-7 has been shown to mediate E-cadherin ectodomain shedding in injured lung epithelium (19). Induced expression of MMP-3 resulted in E-cadherin cleavage, which was blocked by GM-6001, an MMP inhibitor (17).
The objectives of this study were to adapt an in vitro model of ischemic injury to examine ischemia-induced disruption of the cadherin/catenin complex in normal rat kidney (NRK) cells and to determine a potential role for MMPs in this process.
MATERIALS AND METHODS
Cell culture. NRK-52E cells (ATTC, Gaithersburg, MD) were cultured on plastic dishes in Dulbecco’s modified Eagle’s medium containing 1.5 g/l sodium bicarbonate and 5% bovine serum in an atmosphere of 5% CO2-95% air at 37°C. At confluence (4–5 days), subcultures were prepared by treatment with 0.02% EDTA, 0.05% trypsin solution, and cells were seeded at a density of 4 x 104 cells/cm2. Cells were used between passages 3 and 20.
Simulated ischemia. Using a protocol adapted from Meldrum et al. (22), confluent cells were washed twice with PBS before the addition of PBS supplemented with 1.5 mM CaCl2 and 2 mM MgCl2. A layer of mineral oil (Sigma 400–5, St. Louis, MO) was added to the cell culture dish. For a 10-cm2 dish, 2 ml of PBS with Ca2+ and Mg2+ and 10 ml of mineral oil were used. After 6 h of ischemia, cells were washed 5x with PBS and normal growth media was added to simulate reperfusion. To isolate conditioned PBS, the mineral oil and PBS with Ca2+ and Mg2+ were collected, centrifuged, and mineral oil was aspirated. Unless otherwise noted, control cells were incubated for 6 h in PBS with Ca2+ and Mg2+.
Western blot analysis. Cell culture plates were washed twice with PBS, scraped, and centrifuged. The supernatant was removed, and cell pellet was washed with cold PBS and then lysed in buffer (10 mM Tris·HCl, pH 7.6, 1% SDS, 1 mM PMSF, 1 mM leupeptin, 1 mM orthovanadate) and boiled for 10 min. The homogenates were spun at 18,000 g for 10 min, and the supernatant was collected. Proteins were quantified by the Bradford method and diluted to 1 μg/μl in 2x sample buffer (250 mM Tris·HCl, pH 6.8, 4% SDS, 10% glycerol, 2% -mercaptoethanol, 0.006% bromophenol blue). Samples were boiled for 5 min before electrophoresis, and 20 μg of protein were separated by 8, 12, or 15% SDS-PAGE. Separated proteins were transferred onto a Hybond-ECL nitrocellulose membrane (Amersham, Piscataway, NJ) in transfer buffer (25 mM Tris, 200 mM glycine, 20% methanol, and 1% SDS). Nonspecific binding was blocked by incubation with Tris-buffered saline plus Tween 20 (TBST) blocking buffer (0.1% Tween 20, 10 mM Tris, pH 7.5, 100 mM NaCl) supplemented with 5% nonfat dry milk for 1 h at room temperature. Primary antibodies (Table 1) were diluted in the same buffer and incubated at 4°C overnight. After subsequent washes with TBST, membranes were incubated with secondary antibody (1:20,000 in TBST, 5% nonfat dry milk) against the appropriate species for 1 h at room temperature. The blots were washed 3x in TBST, and proteins were detected with the Amersham ECL system and exposure to X-ray film (Kodak, Rochester, NY).
View this table:
Immunofluorescence. Cells grown on Lab Tek Chamber Slides (Nunc, Rochester, NY) were washed twice with PBS, fixed in 2% paraformaldehyde for 10 min, washed two times in 0.02 M PBS for 10 min, and permeabilized in 1% Triton X-100 in 0.02 M PBS for 10 min. Slides were treated with 1:20 blocking solutions of serum related to the species in which the secondary antibody was generated at room temperature for 1 h. Primary antibodies were added at appropriate dilutions overnight. After washing (0.3% Tween in 0.02 M PBS; PBST), FITC-conjugated secondary antibodies (1:200) were added and sections were incubated in the dark at room temperature for 1 h. Slides were mounted with antifade media (Molecular Probes, Eugene, OR) following several washes. Immunostained slides were visualized with a Zeiss Axioplan 2 microscope (Zeiss, Thornwood, NY) fitted with an Axiocam HR digital camera and Axiovision 3.0 software. Negative controls involved substituting IgG for primary antibodies and appropriate species serum for secondary antibodies. The actin cytoskeleton was detected using Alexa Fluor 488-phalloidin (Molecular Probes) staining. A Hypoxyprobe Kit-1 (Chemicon, Temecula, CA) was used to detect ischemia using the manufacturer’s protocol.
ATP levels. ATP levels were detected using an ATP assay kit based on luminescence (Calbiochem, La Jolla, CA) according to the manufacturer’s instructions.
Aggregation assay. The aggregation assay developed for cadherins, where calcium is present during trypsin treatment to preserve cadherin function, was utilized (9). Normal and ischemic NRK cells were washed with PBS containing 1.5 mM CaCl2 and 2 mM MgCl2, before the addition of 0.025% trypsin with Ca2+ and Mg2+ for 10–15 min at 37°C in the presence of 5% CO2. Cells were collected and resuspended at 1.0 x 105 cells/ml in either PBS with 1 mM EDTA or PBS with Ca2+ and Mg2+. From this suspension, 1.5 ml were placed into a 12.5-mm-diameter well coated with 1.5% agarose in PBS. Cells were incubated in a rotary shaker at 80 rpm and 37°C for 20 and 60 min. Aggregated cells were counted manually in a hemacytometer.
Inhibitors. All inhibitors were added to PBS with Ca2+ and Mg2+ 30 min before ischemia. Chemical MMP inhibitors GM-6001 (Calbiochem, San Diego, CA) and TAPI-O (Peptides International, Louisville, KY) were dissolved in DMSO, and the desired concentration was added. Recombinant tissue inhibitors of metalloproteinase (TIMP)-1, TIMP-2, and TIMP-3 (Chemicon) were solubilized at 100 μg/ml in 1 mg/ml of BSA in PBS with Ca2+ and Mg2+. BSA (1 ml/ml) was added to PBS with Ca2+ and Mg2+ in control cells for recombinant TIMP experiments. NRK cells were transfected with 50–100 multiplicity of infection (MOI) of green fluorescent protein or TIMP-3 adenoviruses, constructed as previously described (4, 12).
Statistics. Data are expressed as means ± SE. Groups were compared using ANOVA followed by the Bonferroni post hoc test. P < 0.05 was defined as significant.
RESULTS
Protein expression of cadherins/catenins in ischemic NRK cells. We used a model developed by Meldrum et al. (22) that closely parallels in vivo renal ischemia, in which mineral oil is used to simulate ischemia by restricting exposure to oxygen and metabolite washout. Based on the requirement for Ca2+ in cadherin function, we modified the model by using PBS supplemented with Ca2+ and Mg2+ overlayered with mineral oil to induce ischemia as opposed to a direct overlayering of the confluent cells with mineral oil. Protein expression of cadherins and catenins in ischemic NRK cells was evaluated by Western blot analysis. Using an antibody targeting the cytoplasmic domain of E-cadherin (Transduction Laboratories), cleavage of E-cadherin to a 40-kDa fragment was seen after 3 h of ischemia, and full-length (120 kDa) E-cadherin was lost following 6 h of ischemia (Fig. 1A). Although no fragmentation of N-cadherin was detected, N-cadherin expression began to decrease after 3 h and was virtually lost following 5 h of ischemia (Fig. 1A). Cadherin expression was stable in NRK cells incubated for 6 h in PBS supplemented with Ca2+ and Mg2+ (Fig. 1A), suggesting that the loss of protein expression was due to the ischemic insult, i.e., the mineral oil overlay. However, the simulated mineral oil overlay model includes substrate deprivation as well as oxygen restriction, both of which may be important. As such, a similar pattern of E-cadherin fragmentation (80- and 40-kDa fragments) was seen in ischemic NRK cells in which normal growth media was substituted for PBS with Ca2+ and Mg2+; however, the extent of loss of full-length E-cadherin was not as great (data not shown). Similar results were also seen in NRK cells placed in an Anaeropack system designed for anaerobic cell culture to simulate ischemia (14), suggesting that the fragmentation of E-cadherin and loss of N-cadherin occur in multiple models and is not an artifact of the mineral oil overlay model.
While there was no decrease in -catenin over the ischemia time course, a slight decrease in expression of p120 and -catenin was detected, as well as a dramatic loss of -catenin protein expression at 6 h (Fig. 1A). To examine whether the 40-kDa E-cadherin fragment was generated by extracellular cleavage, the conditioned PBS was isolated from ischemic cells. Using an antibody directed against the extracellular domain of E-cadherin (Santa Cruz Biotechnology), an 80-kDa E-cadherin fragment that is released from the cell surface in ischemic NRK cells was identified (Fig. 1B). A complete recovery of full-length E- and N-cadherin was seen after 24 and 48 h of reperfusion, demonstrating the reversible nature of the ischemic insult (Fig. 2).
Ischemic model. To confirm that the mineral oil overlay model induced ischemia and to compare the biochemical impact of the model with other in vitro systems used to simulate ischemia, the expression of two ischemia-inducible markers was assessed. First, utilizing the Hypoxyprobe Kit-1 (Chemicon), immunofluorescence was performed on ischemic NRK cells. Because pimonidazole forms long-lived adducts with thiol groups in proteins, peptides, and amino acids in ischemic cells, a monoclonal antibody was used to detect the pimonidazole adducts. In control NRK cells, there was no detection of pimonidazole adducts, but after 4–6 h of ischemia, specific staining with the probe was detected (Fig. 3A). To further verify the induction of ischemia, protein expression of an ischemia-inducible protein, the glucose transporter-1 (GLUT-1) (46), was measured. A time-dependent increase in GLUT-1 protein expression was seen in ischemic NRK cells (Fig. 3B).
After 2 h of ischemia, cellular levels of ATP decreased to 46.5 ± 2% of control and rapidly declined to 9.3 ± 0.8% after 6 h of ischemia (Fig. 3C). ATP levels were restored to 51.2 ± 1% of control after 2 h of reperfusion and returned to control levels following 24 h of reperfusion. The decrease in ATP levels is similar to the 90% decrease seen in cells treated with sodium cyanide (15). In addition, caspase activation was determined by Western blot analysis using antibodies for cleaved (i.e., activated) caspase 3 and 9. There was no detection of cleaved caspase 3 or 9 at 2–6 h of ischemia, and total protein expression of caspase 3 was not affected (Fig. 3D). In addition, cleavage of caspases was observed at 12 h of ischemia, suggesting induction of apoptosis following prolonged ischemia (Fig. 3D). As expected, cleavage of both caspases was seen in NRK cells treated with staurosporine. Taken together, these results demonstrate that this in vitro model of ischemia is associated with hypoxia, a transient decrease in cellular ATP levels that is reversible following reperfusion, and is not associated with activation of caspases during the 6-h ischemic insult.
Cadherin/catenin function. Cadherin/catenin protein localization was evaluated by immunofluorescence microscopy. While normal NRK cells expressed E-cadherin at the plasma membrane, 4 h of ischemia resulted in a decrease in E-cadherin at the cell membrane, which was further decreased after 6 h (Fig. 4). N-cadherin expression in normal NRK cells, although patchy, is located at the cell membrane, but after 4 and 6 h of ischemia N-cadherin expression was lost. After 6 h of ischemia, there was a loss of p120, -, and -catenin expression, whereas no change in the expression of -catenin was observed (Fig. 4). Changes in the actin cytoskeleton were examined using Alexa Fluor 488-phalloidin staining. In normal NRK cells, actin filaments displayed staining that outlines the cell periphery as well as actin stress fibers. Ischemia resulted in disruption of the cytoskeleton, primarily observed as a decrease in the actin stress fibers (Fig. 4). However, there was no disruption in the actin at the cell periphery, as was seen when cells were treated with 10 μM cytochalasin D (data not shown).
NRK cells viewed by phase-contrast microscopy revealed characteristically closely packed polygon-shaped cells with little light transmitted between them (Fig. 5A). Following 6 h of ischemia, cells exhibited a loss of cell-cell adhesion and increased transmission of light at cell-cell boundaries, although cells remained attached to the growing surface (Fig. 5A). After 3 h of reperfusion, cells looked similar to control, which is indicative of cell recovery from ischemia. Using the classic aggregation assay developed for cadherins (9), the function of cadherins in ischemic NRK cells was examined (Fig. 5B). Control and ischemic NRK cells were allowed to aggregate for 20 and 60 min at 37°C while being shaken at 80 rpm. A significant decrease in cell aggregation induced by 6 h of ischemia was seen at both 20 (30 ± 3% of control) and 60 min (52 ± 4% of control) of cell aggregation, and this functional deficit was still present after 3 h of reperfusion. However, after 24 h of reperfusion cell aggregation returned to control values (Fig. 5B), which corresponds to the reexpression of full-length cadherins (Fig. 2).
Ischemia-induced cleavage of cadherins: evidence for a role of MMPs. To determine whether MMPs were involved in ischemia-induced disruption of cadherins, the MMP inhibitors GM-6001 and TAPI-O were added to the PBS with Ca2+ and Mg 2+ before ischemia. Both 10 μM GM6001 and 50 μM of TAPI-O blocked E-cadherin fragmentation and loss of N-cadherin (Fig. 6), suggesting that MMPs play a role in ischemia-induced cleavage and loss of cadherins. Interestingly, an increase in N-cadherin expression was seen in NRK cells treated with GM-6001. In addition, recombinant TIMPs were used to inhibit MMP activity. TIMP-1 did not prevent ischemic cleavage and/or loss of E- or N-cadherin (Fig. 7A). Although TIMP-2 protected full-length E-cadherin, there was a 40-kDa fragment of E-cadherin present and a loss of N-cadherin expression. In contrast, TIMP-3 inhibited both cleavage and/or loss of E- and N-cadherin expression (Fig. 7A). Similar inhibitory activity was also seen in NRK cells transfected with a TIMP-3 adenovirus but not the green fluorescent protein control (Fig. 7B). These data collectively indicate the involvement of a membrane-associated metalloproteinase in E- and N-cadherin cleavage following ischemic insult in NRK cells.
DISCUSSION
Because ischemia is a leading cause of ARF, investigating the targets of ischemic insult will provide insight into the mechanisms underlying ARF. The cadherin/catenin complex, which is critical to renal proximal epithelial cell function, is disrupted by ischemic insult. In this study, an ischemic model, which resembles ischemic renal injury in vivo, resulted in selective fragmentation/loss of E-cadherin and loss of N-cadherin expression that could be blocked by MMP inhibitors.
Most in vitro ischemic models use chemicals to induce ischemia, such as ATP depletion with deoxy-D-glucose and antimycin A or sodium cyanide (5, 15). To avoid using chemicals that may have other nonspecific molecular targets and to mimic ischemic insult in vivo (oxygen/nutrient deprivation and metabolite accumulation), an alternative model for ischemia was employed in the present study using PBS supplemented with Ca2+ and Mg2+ overlayered with mineral oil. We have adapted this model for a renal cell line from Meldrum et al. (22); however, the mineral oil model was first used in isolated myocytes (13, 41). In the original and subsequent studies by Meldrum et al. (20–23), mineral oil is directly overlaid onto confluent LLC-PK1 cultures. For our studies of cadherin/catenin expression and function, we removed the confounding factor of divalent cation removal by first using a thin layer of PBS supplemented with Ca2+ and Mg2+. The Meldrum group uses 1 h of simulated ischemia, which is much shorter than the time period used in our studies. However, when mineral oil was directly added to NRK cultures, the fragmentation and loss of E-cadherin were seen at earlier time points (data not shown), suggesting that the supplemented PBS layer may be partially attenuating the ischemic insult. Consistent with this conclusion is the fact that replacement of the supplemented PBS with complete culture media was associated with E-cadherin fragmentation, but to a lesser extent than with PBS plus Ca2+ and Mg2+ (data not shown). In agreement with previous results using TNF- as a marker of ischemia (22), the induction of ischemic markers pimonidazole adducts and GLUT-1 was seen NRK cells. In addition, the decrease in ATP levels following 6 h of ischemia in this study is similar to the 90% decrease seen in cells treated with sodium cyanide (15).
Caspase 9 is activated on cytochrome c release from the mitochondria and further activates other caspases, such as caspase 3, to initiate a cascade leading to apoptosis (36). During the 6-h time course of ischemia, no evidence of caspase activation was detected; however, prolonged ischemia (12 h) resulted in caspase activation. While caspase 3 has been shown to cleave E-cadherin to generate a 24-kDa fragment (38), the lack of temporal correlation between caspase activation and cadherin loss suggests that this protease is not responsible for cadherin/catenin disruption in this system. Cleavage of E-cadherin and loss of N-cadherin occur before caspase activation, suggesting that disruption of the cadherin/catenin complex precedes ischemia-induced apoptosis. However, the possibility remains that significant apoptosis occurs during the reperfusion phase, as reported by Meldrum et al. (20–23).
Disruption of the cadherin/catenin complex at the cell membrane is concurrent with altered cell morphology, as evidenced by the loss of cell-cell contact after 6 h of ischemia. The integrity of intracellular junctions and polarity in epithelial cells depends on cadherin-mediated cell-cell contacts (2). Therefore, cadherin disruption and loss correlate with an altered phenotype in ischemic epithelial cells that includes, but is not limited to, the loss of cell polarity and shedding of tubular epithelial cells. Given the relationship between the cadherin/catenin complex and the actin cytoskeleton, we also examined the pattern of actin staining during ischemia. In normal NRK cells, actin filaments displayed a staining pattern that outlines the cell periphery as well as actin stress fibers, similar to that seen in previous studies (1). Six hours of ischemia resulted in disruption of actin stress fibers, but the junctional staining of actin was preserved. Similar results were seen in mouse proximal tubule cells treated with cyanide to induce chemical anoxia (16).
Disruption of cadherin/catenin integrity may also be associated with activation of the transcriptional activity of -catenin (43). However, during ischemia or reperfusion -catenin was not seen in the nucleus. Thus far, we have not been able to provide evidence for activation of -catenin cell signaling during ischemia in our in vitro model. In contrast, ATP depletion with sodium cyanide induces translocation of -catenin into the nucleus in opossum kidney cells (32). The discrepancy in experiments could be due to the different methods utilized to induce ischemia or the different cell lines used.
Fragmentation of N-cadherin is seen in apoptotic hepatic cells and proliferating vascular smooth muscle cells (26, 40); however, in our model ischemia was associated with a loss of N-cadherin protein expression without detectable fragments. It is possible that the antibody used in our studies cannot recognize the cleaved fragments. However, a panel of four commercially available antibodies did not detect fragments in ischemic NRK cells. The fact that MMP inhibition blocks the loss of N-cadherin suggests that this loss may be due to proteolytic cleavage. In addition, GM-6001 was associated with a significant increase in N-cadherin levels in control NRK cells, further supporting a role for MMPs in regulating N-cadherin expression.
Although rapid internalization of E-cadherin is seen in ATP-depleted cultured renal epithelial cells (18), a longer insult leads to proteolytic clipping of E-cadherin and disruption of the complex (5). Previous studies have shown a loss of E-cadherin in ischemia both in vivo and in vitro (5). ATP-depleted MDCK cells expressed an 80-kDa fragment and loss of full-length E-cadherin that could not be prevented with inhibitors of the proteasomal, lysosomal, or calpain-mediated proteolytic pathways (5). The 40-kDa E-cadherin fragment is believed to correspond to the cytoplasmic and transmembrane domain of the protein. Detection of the extracellular 80-kDa fragment of E-cadherin in conditioned PBS of ischemic NRK cells suggests that the cleavage event occurs extracellularly.
Our results suggest that MMPs play a key role in the cleavage and/or loss of E- and N-cadherin by inhibition with GM-6001 and TAPI-O. Recent data indicate that MMPs play a role in ischemic ARF, as MMPs were increased in rat postischemic kidney tissue and were localized to the renal tubules (3). In addition, MMPs are thought to be involved in ischemia-induced damage to other organs, such as the brain, lung, and heart (25, 33, 37, 45). Furthermore, this study has shown that TIMP-3 completely blocks both cleavage and/or loss of E- and N-cadherin, whereas TIMP-2 protects full-length E-cadherin protein expression but not loss of N-cadherin. These differences imply that different MMP(s) may play a role in E- and N-cadherin regulation. This hypothesis is further supported by the finding that lower concentrations (10 μM) of TAPI-O blocked the loss of N-cadherin but did not attenuate E-cadherin fragmentation in ischemic NRK cells (data not shown). Because ADAMs (for a disintegrin and metalloproteinase domain) are also inhibited by TIMP-3, it is possible that this family of metalloproteinases may play a role in ischemia-induced cleavage and/or loss of E- and N-cadherin.
In summary, simulated in vitro ischemia was associated with fragmentation of E-cadherin and loss of N-cadherin before loss of -, -, and p120 catenin. A loss of cell-cell contacts corresponds to the disruption of cell adhesion. Ischemia-induced cleavage and/or loss of E- and N-cadherin can be blocked by GM-6001 and TAPI-O, suggesting a potential role for MMPs. In addition, the present study shows that the MMP(s) responsible for ischemia-induced E-cadherin cleavage is inhibited by TIMP-2 and TIMP-3, and loss of N-cadherin is inhibited by TIMP-3, indicating a novel role for membrane-associated metalloproteases in these events. Future experiments will examine whether specific MMP inhibitors can decrease the pathological sequelae following ischemia-induced ARF.
GRANTS
These studies were supported by the Department of Medical Pharmacology and Toxicology, the Center for Environmental and Rural Health (P30-ES09106), and National Heart, Lung, and Blood Institute Grant HL-59373 (G. E. Davis).
ACKNOWLEDGMENTS
We thank Bert Vogelstein for kindly providing the AdEasy adenoviral system.
FOOTNOTES
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.
REFERENCES
Alvarez C and Sztul ES. Brefeldin A (BFA) disrupts the organization of the microtubule and the actin cytoskeletons. Eur J Cell Biol 78: 1–14, 1999.
Bannerman DD, Sathyamoorthy M, and Goldblum SE. Bacterial lipopolysaccharide disrupts endothelial monolayer integrity and survival signaling events through caspase cleavage of adherens junction proteins. J Biol Chem 273: 35371–35380, 1998.
Basile DP, Fredrich K, Weihrauch D, Hattan N, and Chilian WM. Angiostatin and matrix metalloprotease expression following ischemic acute renal failure. Am J Physiol Renal Physiol 286: F893–F902, 2004.
Bayless KJ and Davis GE. The Cdc42 and Rac1 GTPases are required for capillary lumen formation in three-dimensional extracellular matrices. J Cell Sci 115: 1123–1136, 2002.
Bonventre JV and Weinberg JM. Recent advances in the pathohysiology of ischemic acute renal failure. J Am Soc Nephrol 14: 2199–2210, 2003.
Bush KT, Tsukamoto T, and Nigam SK. Selective degradation of E-cadherin and dissolution of E-cadherin-catenin complexes in epithelial ischemia. Am J Physiol Renal Physiol 278: F847–F852, 2000.
Cho EA, Patterson LT, Brookhiser WT, Mah S, Kintner C, and Dressler GR. Differential expression and function of cadherin-6 during renal epithelium development. Development 125: 803–812, 1998.
Citi S. The molecular organization of tight junctions. J Cell Biol 121: 485–489, 1993.
Dahl U, Sjodin A, Larue L, Radice GL, Cajander S, Takeichi M, Kemler R, and Semb H. Genetic dissection of cadherin function during nephrogenesis. Mol Cell Biol 22: 1474–1487, 2002.
Goldberg M, Peshkovsky C, Shifteh A, and Al-Awqati Q. μ-Protocadherin, a novel developmentally regulated protocadherin with mucin-like domains. J Biol Chem 275: 24622–24629, 2000.
Goto S, Yaoita E, Matsunami H, Kondo D, Yamamoto T, Kawasaki K, Arakawa M, and Kihara I. Involvement of R-cadherin in the early stage of glomerulogenesis. J Am Soc Nephrol 9: 1234–1241, 1998.
Gumbiner BM. Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell 84: 345–357, 1996.
He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, and Vogelstein B. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA 95: 2509–2514, 1998.
Henry P, Popescu A, Puceat M, Hinescu M, and Escande D. Acute simulated ischemia produces both inhibition and activation of K+ currents in isolated ventricular myocytes. Cardiovasc Res 32: 930–939, 1996.
Kaibori M, Inoue T, Tu W, Oda M, Kwon AH, Kamiyama Y, and Okumura T. FK506, but not cyclosporin A, prevents mitochondrial dysfunction during hypoxia in rat hepatocytes. Life Sci 69: 17–26, 2001.
Keller SH and Nigam SK. Biochemical processing of E-cadherin under cellular stress. Biochem Biophys Res Commun 307: 215–223, 2003.
Kroshian VM, Sheridan AM, and Lieberthal W. Functional and cytoskeletal changes induced by sublethal injury in proximal tubular epithelial cells. Am J Physiol Renal Fluid Electrolyte Physiol 266: F21–F30, 1994.
Lochter A, Galosy S, Muschler J, Freedman N, Werb Z, and Bissell MJ. Matrix metalloproteinase stromelysin-1 triggers a cascade of molecular alterations that leads to stable epithelial-to-mesenchymal conversion and a premalignant phenotype in mammary epithelial cells. J Cell Biol 139: 1861–1872, 1997.
Mandel LJ, Doctor RB, and Bacallao R. ATP depletion: a novel method to study junctional properties in epithelial tissues. II. Internalization of Na+,K+-ATPase and E-cadherin. J Cell Sci 107: 3315–3324, 1994.
McGuire JK, Li Q, and Parks WC. Matrilysin (matrix metalloproteinase-7) mediates E-cadherin ectodomain shedding in injured lung epithelium. Am J Pathol 162: 1831–1843, 2003.
Meldrum KK, Burnett AL, Meng X, Misseri R, Shaw MBK, Gearhart JP, and Meldrum DR. Liposomal delivery of heat shock protein 72 into renal tubular cells blocks nuclear factor-B activation, tumor necrosis factor- production, and subsequent ischemia-induced apoptosis. Circ Res 92: 293–299, 2003.
Meldrum KK, Hile K, Meldrum DR, Crone JA, Gearhart JP, and Burnett AL. Simulated ischemia induces renal tubular cell apoptosis through a nuclear factor-B dependent mechanism. J Urol 168: 248–252, 2002.
Meldrum KK, Meldrum DR, Hile KL, Burnett AL, and Harken AH. A novel model of ischemia in renal tubular cells which closely parallels in vivo injury. J Surg Res 99: 288–293, 2001.
Meldrum KK, Meldrum DR, Hile KL, Yerkes EB, Ayala A, Cain MP, Rink RC, Casale AJ, and Kaefer MA. P38 MAPK mediates renal tubular cell TNF- production and TNF--dependent apoptosis during simulated ischemia. Am J Physiol Cell Physiol 281: C563–C570, 2001.
Molitoris BA and Marrs J. The role of cell adhesion molecules in ischemic acute renal failure. Am J Med 106: 583–592, 1999.
Muhs BE, Plitas G, Delgado Y, Ianus I, Shaw JP, Adelman MA, Lamparello P, Shamamian P, and Gagne P. Temporal expression and activation of matrix metalloproteinases-2, -9, and membrane type 1-matrix metalloproteinase following acute hindlimb ischemia. J Surg Res 111: 8–15, 2003.
Murphy F, Waung J, Collins J, Arthur MJ, Nagase H, Mann D, Benyon RC, and Iredale JP. N-cadherin cleavage during activated hepatic stellate cell apoptosis is inhibited by tissue inhibitor of metalloproteinase-1 (Abstract). Comp Hepatol 3, Suppl 1: S8, 2004.
Murphy GJ, Murphy G, and Reynolds JJ. The origin of matrix metalloproteinases and their familial relationships. FEBS Lett 289: 4–7, 1991.
Nagafuchi A, Takeichi M, and Tsukita S. The 102 kd cadherin-associated protein: similarity to vinculin and posttranscriptional regulation of expression. Cell 65: 849–857, 1991.
Nagase H and Woessner JF Jr. Matrix metalloproteinases. J Biol Chem 274: 21491–21494, 1999.
Noe V, Fingleton B, Jacobs K, Crawford HC, Vermeulen S, Steelant W, Bruyneel E, Matrisian LM, and Mareel M. Release of an invasion promoter E-cadherin fragment by matrilysin and stromelysin-1. J Cell Sci 114: 111–118, 2001.
Price VR, Reed CA, Lieberthal W, and Schwartz JH. ATP depletion of tubular cells causes dissociation of the zonula adherens and nuclear translocation of beta-catenin and LEF-1. J Am Soc Nephrol 13: 1152–1161, 2002.
Schulze CJ, Wang W, Suarez-Pinzon WL, Sawicka J, Sawicki G, and Schulz R. Imbalance between tissue inhibitor of metalloproteinase-4 and matrix metalloproteinases during acute myocardial ischemia-reperfusion injury. Circulation 107: 2487–2492, 2003.
Shimazui T, Oosterwijk-Wakka J, Akaza H, Bringuier PP, Ruijter E, Debruyne FM, Schalken JA, and Oosterwijk E. Alterations in expression of cadherin-6 and E-cadherin during kidney development and in renal cell carcinoma. Eur Urol 38: 331–338, 2000.
Simonneau L, Kitagawa M, Suzuki S, and Thiery JP. Cadherin 11 expression marks the mesenchymal phenotype: towards new functions for cadherins Cell Adhes Commun 3: 115–130, 1995.
Slee EA, Harte MT, Kluck RM, Wolf BB, Casiano CA, Newmeyer DD, Wang HG, Reed JC, Nicholson DW, Alnemri ES, Green DR, and Martin SJ. Ordering the cytochrome c-initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and -10 in a caspase-9-dependent manner. J Cell Biol 144: 281–292, 1999.
Sole S, Petegnief V, Gorina R, Chamorro A, and Planas AM. Activation of matrix metalloproteinase-3 and agrin cleavage in cerebral ischemia/reperfusion. J Neuropathol Exp Neurol 63: 338–349, 2004.
Steinhusen U, Weiske J, Badock V, Tauber R, Bommert K, and Huber O. Cleavage and shedding of E-cadherin after induction of apoptosis. J Biol Chem 276: 4972–4980, 2001.
Tsukita S, Furuse M, and Itoh M. Structural and signalling molecules come together at tight junctions. Curr Opin Cell Biol 11: 628–633, 1999.
Uglow EB, Slater S, Sala-Newby GB, Aguilera-Garcia CM, Angelini GD, Newby AC, and George SJ. Dismantling of cadherin-mediated cell-cell contacts modulates smooth muscle cell proliferation. Circ Res 92: 1314–1321, 2003.
Vanheel B, Leybaert L, de Hemptinne A, and Leusen I. Simulated ischemia and intracellular pH in isolated ventricular muscle. Am J Physiol Cell Physiol 257: C365–C376, 1989.
Venkataraman R and Kellum JA. Novel approaches to the treatment of acute renal failure. Expert Opin Investig Drugs 12: 1353–1366, 2003.
Willert K and Nusse R. Beta-catenin: a key mediator of Wnt signaling. Curr Opin Genet Dev 8: 95–102, 1998.
Wu Q and Maniatis T. A striking organization of a large family of human neural cadherin-like cell adhesion genes. Cell 97: 779–790, 1999.
Yano M, Omoto Y, Yamakawa Y, Nakashima Y, Kiriyama M, Saito Y, and Fujii Y. Increased matrix metalloproteinase 9 activity and mRNA expression in lung ischemia-reperfusion injury. J Heart Lung Transplant 20: 679–686, 2001.
Zhang JZ, Behrooz A, and Ismail-Beigi F. Regulation of glucose transport by hypoxia. Am J Kidney Dis 34: 189–202, 1999.(Marisa D. Covington, Kayl)