INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

ACUTE RENAL FAILURE is commonly associated with defects in the renal tubular epithelial cell surface (25-28, 34). These epithelial defects correlate with the degree of renal insufficiency, and repair of these denuded areas occurs in parallel with recovery of normal renal function (25, 27, 34). Despite the obvious importance of the nephrogenic repair process, there is limited information on the basic mechanisms that promote recovery from renal tubular epithelial cell damage (3, 23, 37).

Growth factors (GFs) promote renal mitogenesis and motogenesis, thereby potentially accelerating the process of renal tubular regeneration (3, 7-11). Several GFs, including epidermal growth factor (EGF), hepatocyte growth factor (HGF), fibroblastic growth factor (FGF), insulin-like growth factor I (IGF-I), endothelin (ET), and transforming growth factor- (TGF-), are normally synthesized in the kidney (5, 8, 9, 13, 14, 24, 29-32, 39). This renal site of synthesis suggests the potential for an autocrine pathway, whereby injury to renal tubular epithelial cells results in enhanced release and/or synthesis of GFs, which, in turn, promote recovery. To date, evidence for the existence of an autocrine mechanism of renal tubular epithelial repair has been conflicting, since renal injury has both enhanced and diminished renal growth factor gene expression (5, 8, 9, 29-32).

The present studies were undertaken in cultured renal tubular epithelial cells to test the hypothesis that mechanical injury of these cells results in an autocrine mechanism that promotes reconstitution of the injured renal tubular epithelium. Our results demonstrate that mechanical injury of renal tubular epithelial cells results in release of selected GFs, including acidic and basic FGF (aFGF and bFGF, respectively) and platelet-derived growth factor BB (PDGF-BB), which are capable of promoting recovery from injury.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Materials. [³H]thymidine (70-85 Ci/mmol) was obtained from Amersham. Culture media RPMI 1640 and Eagle's minimum essential medium (MEM) were obtained from Sigma. Calf bovine serum was obtained from Hyclone. Penicillin, streptomycin, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), suramin, phosphoramidon, glutamine, and nonessential amino acids were obtained from Sigma. Fisher was the source of Scintisafe Econ 2 scintillation fluid. Human recombinant GFs were obtained from R & D Systems (Minneapolis, MN) (EGF, heparin-bound EGF, TGF-1, and aFGF and bFGF), Calbiochem (endothelin-1, -2, and -3; ET-1, ET-2, and ET-3, respectively), and Sigma (HGF, IGF-I, and PDGF-AA and PDGF-BB), respectively. Antibodies for immunoblotting were obtained from R & D Systems (mouse anti-human EGF, goat anti-human heparin-bound EGF, goat anti-human HGF, and rabbit anti-TGF- pan specific), Genzyme (rabbit anti-human PDGF-AA and PDGF-BB), and Sigma (mouse anti-human aFGF, mouse anti-bovine bFGF, and rabbit anti-human endothelin). Endothelin A (BQ123) and B (BQ788) antagonists were obtained from Calbiochem. Enzyme-linked immunosorbent assay (ELISA) kits for TGF-1 and bFGF were obtained from R & D Systems.

Cell and culture conditions. LLC-PK₁ and OK cells were obtained from the American Type Tissue Collection. LLC-PK₁ cells were used at passages 90-125, and OK cells were used at passages 37-45. LLC-PK₁ cells were grown in RPMI 1640 medium supplemented with 20 mM HEPES, 100 µg/ml streptomycin, 100 µg/ml penicillin, and 6% calf bovine serum. After initial seeding, LLC-PK₁ cells typically reached confluence in 48-72 h.

OK cells were grown to confluence (typically in 72-120 h) in MEM supplemented with nonessential amino acids, 100 µg/ml streptomycin, 100 µg/ml penicillin, and 10% fetal bovine serum. Aortic bovine smooth muscle cells (SMC) were obtained from medial explants dissected from freshly harvested bovine aortic strips and plated in six-well tissue culture plates. Aortic SMC were grown in MEM containing nonessential amino acids, 100 µg/ml streptomycin, 100 µg/ml penicillin, and 10% calf bovine serum. Explant-derived cells were passed after 10-14 days with 0.05% trypsin-0.02% EDTA and used in experiments at passages 2-6. These SMC typically required 5-7 days to become confluent. The authenticity of these cells was verified by immunohistochemical staining using a monoclonal antibody specific for smooth muscle actin (CGA-7). All cells cultured were maintained at 37°C in a humidified atmosphere of 95% air-5% CO₂.

Mechanical injury. Modifications of the scrape-wounding method originally described by McNeil et al. (4, 17, 21, 39) were used. This method mechanically detaches cells from the monolayer and induces transient cell membrane disruption (17, 21, 22). Confluent cultures of LLC-PK₁, OK cells, and SMC in 24-well plates were maintained in their respective media plus HEPES (for LLC-PK₁ cells only) and streptomycin/penicillin for 48 h at 37°C-5% CO₂. Cultures of LLC-PK₁ cells were maintained on serum-free media for 48 h, whereas OK cells and SMC were maintained in media with 0.1% calf serum for 48 h prior to injury. Cultures were injured by systematic scraping with a cell scraper (Fisher). Each culture was scraped with two circumferential and three horizontal and vertical scrapes. By direct visualization, this resulted in disruption of 50-80% of the monolayer. In studies done on detached LLC-PK₁ cells removed 15 min after the scraping procedure, 26 ± 5% took up trypan blue (n = 3). There was very little (<2%) trypan blue uptake in the LLC-PK₁ cells that remained attached within the scraped monolayer, and the few cells that did not exclude trypan blue remaining within the scraped monolayer were at the edge of the wound area. In all experiments, control media was obtained for the 15 min prior to scraping, whereas experimental media was collected for the 15 min after scraping. In these studies, scraped cells were allowed to remain in and to condition the media for the 15 min following the scraping procedure. Subsequently, cell debris was removed from the media by centrifugation (2,000 revolutions/min for 5 min at 40°C).

A second type of experiment using the scrape-wounding method was also performed. In these experiments, control media was removed after 1 h of incubation in sham-scraped monolayers and compared with the effects of media removed after 1 h of incubation from two sets of scrape-wounded monolayers. Cells were scrape wounded as noted above, with special care taken to remove 50-60% of the monolayer. One set of scrape-wounded monolayers underwent immediate washing of the monolayer to remove detached cells and debris and was then incubated for 1 h in the same serum-free media the cells were in prior to scraping. Another set of scrape-wounded monolayers was not washed immediately after the scraping, and the detached cells and debris were allowed to remain in and to condition the media for 1 h after the scraping procedure. Subsequently, cell debris was removed from the media by centrifugation (2,000 revolutions/min for 5 min at 40°C).

Growth activity assay. Confluent monolayers of 75-cm² flasks of LLC-PK₁ and OK cells were treated with 0.05% trypsin-0.01% EDTA for 4 min at 37°C. The cells were rinsed from the flask with serum-containing media, centrifuged at 300-400 g for 5 min, resuspended in 2-3 ml of medium, and counted on a hemocytometer. LLC-PK₁ cells were seeded in 24-well plates at 2 × 10⁴ cells/well in RPMI media supplemented with 6% bovine calf serum, HEPES, and streptomycin/penicillin. The cells were allowed to adhere for 24 h. The media were then changed to serum-free RPMI media with HEPES and streptomycin/penicillin, which was maintained for 48 h prior to study. OK cells were seeded at 5 × 10⁴ cells/well in MEM media, supplemented with 10% fetal bovine serum and streptomycin/penicillin, and allowed to adhere for 24 h. The media were then changed to MEM media with 0.1% fetal bovine serum and streptomycin/penicillin for 48 h prior to study. Aortic SMC at 80-90% confluency were treated as above, except that 24-well plates were seeded at 5 × 10⁴ cells/well in MEM media supplemented with nonessential amino acids, streptomycin/penicillin, and 10% bovine calf serum and allowed to adhere for 24 h. After 24 h, the media was changed to MEM media with 0.1% bovine calf serum, nonessential amino acids, and streptomycin/penicillin for 48 h prior to study. All cells were maintained at 37°C and 5% CO₂. GFs/media were then added, and the cells were incubated for 24 h. Cells were pulsed with 0.5 µCi/well [³H]thymidine for either 3 (LLC-PK₁ and OK) or 6 (SMC) h. The cells were harvested by rinsing the wells once with phosphate-buffered saline, and the cellular material was precipitated with perchloric acid and solubilized with 0.01 N NaOH-1% sodium dodecyl sulfate (SDS). Samples were transferred to vials, scintillation fluid was added, and radioactivity was determined by counting (Beckman Instruments).

In some experiments, growth activity was ascertained by counting the cell number. In these studies, LLC-PK₁ cells were seeded in 24-well plates at 2 × 10⁴ cells/well in RPMI media supplemented with 6% bovine calf serum, HEPES, and streptomycin/penicillin. After 24 h for adherence, the media were changed to serum-free RPMI media with HEPES and streptomycin/penicillin for 48 h. One milliliter of either control media or experimental media removed from scrape-injured cells in the 15 min following scrape injury was added to each well and maintained in place for 24 h. The plates were then washed once with phosphate-buffered saline and treated with 0.1 ml/well of 0.05% trypsin-0.2% EDTA for 4 min at 37°C. The cells were removed to 12 × 75-mm polyethylene tubes and counted on a hemocytometer.

Migration assay. Confluent cultures of LLC-PK₁ cells and aortic SMC were seeded in 35-mm dishes at 7-10 × 10⁵ cells/plate, as described in Cell and culture conditions. For 24 h prior to the study, LLC-PK₁ cells were maintained in serum-free media, whereas SMC were maintained in 0.1% bovine calf serum. Rectangular areas of denuded cells were created near the center of the monolayer, using sterile needles as described previously (16, 35, 36). Two vertical lines intersecting two horizontal lines (tic-tac-toe-like) were made, and the rectangular areas where the horizontal and vertical lines crossed were measured (time 0) and labeled. The denuded areas were measured with an eyepiece reticle in an inverted microscope with phase optics at ×10-20 (Nikon IM 35). Next, either GFs at the indicated concentrations or 1.5 ml of control or experimental media were added to the plates, and the cultures were incubated at 37°C with 5% CO₂. The area of the denuded areas was measured subsequently at 6 and 24 h as before. We have previously determined the intra- and interobserver variability of this technique and confirmed eyepiece measurements with time-lapse photography and continuous video microscopy (35, 36). A single observer (C. J. Ray) measured all denuded areas and was unaware of culture treatment conditions. For each study condition, the percentage of surface area healed at 6 and 24 h was calculated and compared with simultaneously studied controls. An average of six to eight treated and control denuded areas were studied under each condition. We have previously demonstrated that these denuded areas heal when proliferation is prevented with irradiation, suggesting that migration is the major process for healing (35).

Preparation of conditioned media for immunoblot analyses. LLC-PK₁ cells were grown to confluence in 75-cm² flasks and maintained in the absence of serum for 48 h. The media was then changed to media containing 2.5 µg/ml protease inhibitors (leupeptin, aprotinin, antipain, and pepstatin) at 4 ml/flask, and the cells were incubated for 15 min at 37°C. The medium was removed and placed on ice, and it served as control, noninjured medium. For injured conditioned medium, the cell monolayer was scrape injured after addition of medium with protease inhibitors (4 ml/flask), allowed to incubate for 15 min at 37°C, and otherwise handled identically to control, noninjured medium. Both control (noninjured) medium and medium from injured cultures were concentrated using a Centriprep 3 or 10 (Amicon) concentrator. Ten flasks, each containing 4 ml of either control or experimental media (total of 40 ml of media), were used for all studies. The concentrators were washed with phosphate-buffered saline and loaded with 15 ml of sample/concentrator. The concentrators were centrifuged (3,000 g) for 4-6 h at 4°C, and the filtrate was decanted until the remaining volume was 1 ml/sample. This 1 ml of concentrated sample was transferred to 12 × 75-mm polypropylene tubes for immunoprecipitation. All immunoprecipitations included negative controls (medium only), positive controls (medium spiked with known concentrations of the GF being sought), and noninjured and injured conditioned media.

Immunoprecipitation assay. Concentrated noninjured control and injured medium samples were incubated overnight with primary antibodies added at 5- to 20-µg samples at 4°C with rocking. Prewashed protein G-agarose (1:1 slurry, Calbiochem) was added at 75 µl/sample, and the samples were incubated for 2 h at 4°C with rocking. The samples were then pelleted, and the supernatant was carefully aspirated. The antibody-protein G-agarose complex was sequentially washed with three buffers. The first contained 10 mM tris(hydroxymethyl)aminomethane (Tris), pH 7.2, 1.0 M NaCl, and 0.1% Nonidet P-40. The second contained 10 mM Tris, pH 7.2, 150 mM NaCl, 1.0% deoxycholate, 1.0% Triton X-100, and 0.1% SDS, and the third contained 10 mM Tris, pH 7.2, 150 mM NaCl, and 1.0 mM EDTA. Subsequently, SDS-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer was added, and the samples were heated to 100°C for 5 min. The supernatants were then loaded onto a SDS polyacrylamide gel, and samples were electrophoresed on a Bio-Rad minigel system. The gel was transferred to a nitrocellulose membrane and incubated overnight at 4°C with a primary antibody. The membrane was washed and incubated with a horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature, and the proteins were detected with ECL enhance chemiluminescence Western blotting reagents and Hyperfilm (Amersham).

Heparin-Sepharose purification. These studies were performed as described by Sasahara et al. (33). Media was concentrated as above then incubated with 1 ml (1:1 slurry) heparin-Sepharose overnight at 4°C with rocking. The samples were pipetted into a column, the beads were allowed to settle, and the column bed was sequentially eluted with four solutions [solution 1: 0.1 M sodium acetate, 0.1 M NaCl, 10 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), pH 5.0, 5 ml; solution 2: 0.1 M sodium acetate, 0.1 M NaCl, 0.16 mM CHAPS, pH 5.0, 5 ml; solution 3: 0.05 M Tris, pH 7.0, 0.1 M NaCl, 5 ml; and solution 4: 0.05 M Tris, pH 7.0, 2.0 M NaCl, 2 ml]. The final 2-ml eluted sample was then further concentrated in a Centricon 10 concentrator (Amicon), desalted in the last centrifugation with 10 mM acetic acid, and concentrated to a volume of 50-100 µl. The samples were combined with 5× SDS-PAGE sample buffer, loaded onto a gel, electrophoresed, transferred to nitrocellulose, and probed as above.

ELISA assays. Because we could not be sure that immunoblotting techniques were sufficiently sensitive to detect release of small amounts of selected GF, we also utilized two sensitive, commercially available ELISA kits. To assay injured and control media for TGF-1 and bFGF, Quantikine and Quantikine HS immunoassay plates (R & D Systems) were used exactly as specified by the manufacturer. Control and experimental media were calculated based on linear regression of the standards, which exhibited a correlation coefficient of r = 0.984.

Neutralizing antibody studies. Noninjured and injured conditioned media were divided into 2.2-ml aliquots to incubate with the neutralizing antibodies or a goat immunoglobulin G (IgG) control. Injured conditioned medium was incubated with the following antibodies or goat IgG: 15 or 45 µg/ml goat anti-human PDGF-BB (R & D Systems), 20 or 60 µg/ml rabbit anti-bovine aFGF (R & D Systems), and 34 or 102 µg/ml goat anti-human bFGF (R & D Systems). These samples, plus samples of noninjured and injured conditioned medium without additions, were then added to a recipient plate of serum-starved LLC-PK₁ cells and incubated for 24 h at 37°C in 5% CO₂. The cells were pulsed with 5 µCi/well [³H]thymidine (Amersham) during the last 3 h of the incubation and harvested as described previously. These experiments were repeated five times.

Statistical analyses. All values are expressed as the means ± SE. Statistical analyses were done using paired and unpaired Student's t-tests, analysis of variance, and coefficient of correlation where appropriate. P < 0.05 was considered significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Response of LLC-PK₁ cells to conditioned media from injured cells. When conditioned media removed after 15 min from mechanically injured LLC-PK₁ cells were applied to near-quiescent LLC-PK₁ cells, a five- to sixfold increase (P < 0.02) in recipient [³H]thymidine uptake was noted, compared with [³H]thymidine exposed to control media (Fig. 1A). In our studies, we noted that control media exerts a modest (5-25%, mean 15%) effect to enhance [³H]thymidine uptake in serum-starved LLC-PK₁ recipient cells. Thus much of our data is normalized in terms of control media. Conditioned media removed from mechanically injured LLC-PK₁ cells did not promote healing of small denuded areas made within confluent monolayers of LLC-PK₁ recipient cells, (Fig. 1B). To confirm that these [³H]thymidine studies were indicative of a proliferative response to conditioned media, studies (n = 4) on cell number were done. Twenty-four-hour exposure to control media resulted in cell counts of 32,800 ± 950/well, whereas 24-h exposure to undiluted conditioned media resulted in cell counts of 52,780 ± 2,500/well (P < 0.05 vs. control). Dilutions of 1:2 and 1:4 of conditioned media resulted in cell counts of 41,830 ± 2,830 and 40,080 ± 1,050/well, respectively.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Effect of conditioned media (CM) removed from mechanically injured LLC-PK1 cells (open and hatched bars, A and B) and bovine aortic vascular smooth muscle cells (AVSMC, solid bars, A and B) on [3H]thymidine uptake (A) and 6-h healing of denuded areas (B) in quiescent recipient cells. Donor cell of origin of experimental media and recipient cell on which media were tested are shown, bottom. Values are expressed as multifold increase seen relative to simultaneously studied control media and are means ± SE for 6-10 experiments. * P < 0.05.

To attempt to delineate the source of growth-promoting factors present in conditioned media, the effect of conditioned media removed 1 h after scraping, in which the detached cells and debris were removed and retained in the media, were examined. In these studies (n = 3), media removed from monolayers in which detached cells and debris were removed increased [³H]thymidine uptake by 2.0 ± 0.2-fold relative to control (P < 0.05), whereas media removed from monolayers in which detached cells and debris conditioned the media increased [³H]thymidine uptake by 1.8 ± 0.2-fold relative to control (P < 0.05). These results suggest that the source of some of the growth-promoting activity can be attributed, at least in part, to noninjured cells remaining within the scraped monolayer. Together, these results suggest that mechanically injured LLC-PK₁ cells release one or more factors into the media that promote proliferation (increased [³H]thymidine uptake and cell number) but not migration (healing of denuded areas) in near-quiescent, recipient LLC-PK₁ cells.

Specificity of injured conditioned media responses. To delineate whether the growth-promoting response of media removed after mechanical injury of LLC-PK₁ cells occurs in another type of renal tubular epithelial cell line, comparable studies were done in opossum OK cells. In these studies, conditioned media removed from mechanically injured OK cells resulted in a significant (P < 0.05) 2.5 ± 0.3-fold increase in [³H]thymidine uptake relative to control media when applied to quiescent OK cells (n = 3, experiments done in triplicate). These results suggest that an autocrine recovery response to mechanical injury occurs in more than a single type of renal tubular epithelial cell line.

To determine whether the effects of conditioned media removed from mechanically injured LLC-PK₁ cells are tissue-type specific, conditioned media from control and injured LLC-PK₁ cells were applied to quiescent aortic bovine SMC (Fig. 1, hatched bars). Conditioned media removed from LLC-PK₁ cells significantly stimulated [³H]thymidine uptake but not 6 h healing of denuded areas when applied to SMC. The effect of conditioned media from mechanically injured LLC-PK₁ cells to promote [³H]thymidine uptake was greater when applied to low-passage quiescent recipient SMC (~13-fold) than when applied to high-passage quiescent recipient LLC-PK₁ cells (~6-fold). We have found that another growth-promoting stimulus (10% bovine calf serum) produces much greater increments in [³H]thymidine uptake in SMC (10- to 20-fold) than in LLC-PK₁ cells (3- to 10-fold), suggesting relatively greater growth responsitivity of SMC (unpublished observations). Together, these results suggest that the effects of conditioned media from mechanically injured LLC-PK₁ cells to promote proliferation without affecting migration occurs in at least two cell types from dissimilar tissues.

To see whether LLC-PK₁ cells respond to conditioned media removed from another type of cell subjected to mechanical injury, we studied quiescent LLC-PK₁ cell responses to media removed from mechanically injured bovine aortic SMC. Previous studies have demonstrated that mechanical injury of these SMC results in release of numerous GFs (4). Exposure of quiescent LLC-PK₁ cells to conditioned media from SMC results in a modest (3.5-fold) but significant (P < 0.05) increase in [³H]thymidine uptake and also a significant (P < 0.05) increase in healing of denuded areas (Fig. 1, solid bars).

Characterization of response to injured conditioned media. To initially determine whether the mitogenic activity of conditioned media removed from mechanically injured LLC-PK₁ cells is due to one or more peptide GFs, recipient LLC-PK₁ cells were exposed to conditioned media in the presence and absence of suramin, a nonspecific peptide growth factor antagonist (19). At 100 µg/ml, suramin inhibited the effect of conditioned media from injured cells to stimulate [³H]thymidine uptake by 65 ± 7% (P <0.05, n = 3, done in triplicate). This suggests a potential role for one or more peptide GFs in the response to conditioned media removed from injured LLC-PK₁ cells.

Delineation of potential candidate GFs in injured conditioned media. To examine which GFs known to be produced by the kidney are potential candidates to mediate the responses seen with injured conditioned media, we examined [³H]thymidine and 6-h healing of denuded areas in quiescent LLC-PK₁ cells in the absence (control) and presence of PDGF-AA and -BB; bFGF and aFGF; EGF and heparin-bound EGF; HGF; ET-1, -2, and -3; IGF-I; and TGF-1. Each concentration of GF was studied in triplicate on at least three occasions. The maximum effect of each GF on [³H]thymidine uptake is in Fig. 2. Based on these results in which all of the tested GFs, except ET-3, IGF-I, and TGF-, increased [³H]thymidine uptake in near-quiescent LLC-PK₁ cells, there are several GFs that are potential candidates to be present in conditioned media removed from mechanically injured LLC-PK₁ cells. When the effects of selected GFs on 6-h healing of denuded areas made within quiescent confluent LLC-PK₁ cells were examined (Fig. 3), EGF, heparin-bound EGF, and ET-1 significantly promoted healing, an effect not seen with exposure to conditioned media from injured LLC-PK₁ cells. Together, these results demonstrate that PDGF-AA, PDGF-BB, aFGF, bFGF, ET-2, and HGF are GFs known to be produced in renal tissue that mimic the biological response (increased [³H]thymidine uptake without an effect on healing of denuded areas) seen when LLC-PK₁ cells are exposed to injured conditioned media.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effect of selected growth factors on [3H]thymidine uptake in near-quiescent LLC-PK1 cells. All growth factors were tested at 1011-109 M concentrations. Platelet-derived growth factors AA and BB (PDGF-AA and -BB, respectively), basic and acidic fibroblast growth factor (bFGF and aFGF, respectively), epidermal growth factor (EGF), heparin-bound EGF (HB-EGF), endothelins 1, 2, and 3 (ET-1, ET-2, ET-3), and insulin-like growth factor I (IGF-I) were also tested at 108 M. Values represent maximum effect and were observed at 109 M for all tested growth factors. Values are expressed as % simultaneously studied untreated controls and are means ± SE of at least 3 experiments performed in triplicate. * P < 0.05.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Effect of selected growth factors on healing of denuded area made within confluent, near-quiescent LLC-PK1 cell monolayers. Dose-response experiments using concentrations noted in Fig. 2 were done. Values depicted represent response seen with 109 M of indicated growth factor. Values are expressed as % vs. simultaneously studied, untreated controls. Bars are means ± SE of 6-8 experiments. * P < 0.05.

Determination of the presence of candidate GFs in conditioned media from injured cells. To determine whether any of the candidate GFs were present in conditioned media from mechanically injured LLC-PK₁ cells, immunoblotting and ELISA experiments were carried out on concentrated samples of control media and of conditioned media from injured cells. In these experiments, we probed with antibodies directed against PDGF-AA, PDGF-BB, aFGF, bFGF, ET, and HGF. Although our functional studies suggested that EGF, heparin-bound EGF, and TGF- would not be present, we also immunoblotted for these substances in control media as well as in media from mechanically injured LLC-PK₁ cells. In all experiments, control lanes spiked directly with the GF of interest were used as positive controls. Also, as a positive control, media spiked with known concentrations of the GF of interest were concentrated and processed identically to control media removed from uninjured LLC-PK₁ cells and conditioned media removed from mechanically injured LLC-PK₁ cells. These immunoblotting experiments demonstrated proteins interacting with antibodies directed against PDGF-BB and aFGF in experimental but not control media (Figs. 4 and 5). By contrast, immunoblots with antibodies directed against EGF (Fig. 6), ET (Fig. 7), bFGF, HGF, and heparin-bound EGF (data not shown) did not demonstrate any evidence of these GFs in either control media or in media removed from mechanically injured cells. These results suggest that PDGF-BB and aFGF are released into conditioned media from mechanically injured cells. Together with the functional studies on [³H]thymidine uptake and healing of denuded areas described previously, these results suggest the potential for PDGF-BB and aFGF to contribute to the autocrine recovery response that we observed with conditioned media from mechanically injured cells.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Immunoblot of concentrated media from control (NON INJ) and mechanically injured (INJ) LLC-PK1 cells probed with an antibody directed against aFGF. Location of molecular weight (MW) standards of 18.4, 14.0, and 5.5 kDa (K) is indicated on vertical axis. First 3 left lanes represented positive controls with, respectively, 2.0, 1.0, and 0.5 ng of aFGF added directly to lane. Fourth lane represents a negative control [media alone (M)], and 5th lane represents media spiked (S) with 10 ng aFGF. Sixth and seventh lanes represent concentrated media removed from noninjured control LLC-PK1 cells (NON INJ) and injured LLC-PK1 cells (INJ), respectively.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Immunoblot of concentrated media from control (NON INJ, 6th lane, right) and mechanically injured (INJ, 7th lane, right) LLC-PK1 cells probed with an antibody directed against PDGF-BB. Location of molecular weight (MW) standards of 28.7 and 20.5 kDa are indicated on vertical axis. First three left lanes represent positive controls with, respectively, 1.0, 0.3, and 0.1 ng PDGF-BB added directly to the lane. Fourth lane represents media (M) alone, and 5th lane represents media spiked (S) with 30 ng PDGF-BB, as noted in text.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Immunoblot of concentrated media from control (NON INJ, 6th lane, right) and mechanically injured (INJ, 7th lane, right) LLC-PK1 cells probed with an antibody directed against EGF. Location of molecular weight (MW) standards of 14.0, 5.5, and 2.8 kDa are noted on the vertical axis. First three left lanes represent 2.0, 1.0, and 0.5 ng EGF, respectively, added to lane. Fourth and fifth lanes represent media (M) alone and media spiked (S) with 20 ng EGF, as noted in text.

Although our immunoblotting experiments did not suggest the presence of bFGF in experimental media, based on our functional data, bFGF appeared to be a viable candidate to mediate some of the autocrine recovery response. Also, strong evidence that aFGF can be released by renal cells subjected to toxic and mechanical injury is available from our and other studies (13, 39). We therefore used a sensitive ELISA assay to see whether bFGF was released into conditioned media removed in the 15-min period after mechanical injury of LLC-PK₁ cells (Fig. 8). This technique revealed the presence of large amounts of bFGF in conditioned media after LLC-PK₁ mechanical injury (125 pg/ml in conditioned medium from mechanically injured cells and 3 pg/ml in control medium from noninjured cells P < 0.05). In additional studies (n = 3), release of bFGF into media removed following 60 min of incubation in control LLC-PK₁ cells averaged 6.1 ± 1.9 pg/ml. Media removed 60 min after scrape injury when scraped cells remained in and conditioned the media and when scraped cells were removed from the media prior to 60-min incubation were 44.8 ± 5.9 and 36.3 ± 8.9 pg/ml, respectively. Both experimental groups were significantly (P < 0.05) greater than control. These results, when combined with the functional data described previously, also suggest a potential role for bFGF in LLC-PK₁ autocrine response to mechanical injury.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Immunoblot of concentrated media from control (NON), mechanically injured (INJ), and media alone (M) from LLC-PK1 cells probed with an antibody directed against endothelin. Location of molecular weight (MW) standards 14.7 and 2.9 kDa are denoted on vertical axis. First two left lanes represent 200 and 100 ng, respectively, of ET-1 added directly to lanes, and 3rd and 4th lanes are 100 and 50 ng, respectively, of ET-2 added directly to the lanes.

View larger version (10K): [in this window] [in a new window]   Fig. 8.   Measurement of bFGF by enzyme-linked immunosorbent assay (ELISA) in control medium (medium, left), in media from noninjured LLC-PK1 cells (NON INJ CM, middle), and in media from mechanically injured LLC-PK1 cells (INJ CM, right). Bars represent means ± SE of 3 experiments performed in duplicate. Mean value from media removed from mechanically injured media is significantly higher than in control samples (P < 0.025).

To determine whether another type of renal insult increases release of bFGF into the media, we examined the effect of nonlethal heat shock (2-h exposure of LLC-PK₁ monolayers to 42°C; Ref. 38). In these paired studies (n = 3), bFGF averaged 4.5 ± 2.0 pg/ml in media removed from control monolayers incubated at 37°C for 2 h and was significantly increased at 13.8 ± 4.1 pg/ml (P < 0.05 vs. control) in media removed from monolayer incubation at 42°C over 2 h (Fig. 9). Release of bFGF after 2 h of 42°C heat shock was less than that observed after scrape injury (Fig. 8). No detachment of cells exposed to 42°C was seen, and no uptake of trypan blue occurred in these heat-shocked LLC-PK₁ cells.

View larger version (10K): [in this window] [in a new window]   Fig. 9.   Measurement of bFGF by ELISA in control media (2 h, 37°C, left) and in media from heat shocked (2 h, 42°C, right) LLC-PK1 cells. Bars represent means ± SE of 3 experiments performed in triplicate. Media concentration from heat shocked cells are significantly higher than in control media (P < 0.05).

Because our studies suggested that aFGF, bFGF, and PDGF-BB were candidates to mediate the autocrine growth response produced by injured conditioned media, neutralizing antibody studies were done. To ensure that the neutralizing antibodies studied were effective, we examined the effect of antibodies directed against aFGF, bFGF, and PDGF-BB in amounts recommended by the manufacturer to be effective (20, 34, and 15 µg/ml, respectively) on aFGF-, bFGF-, and PDGF-BB-stimulated (all 10⁹ M) [³H]thymidine uptake in serum-starved LLC-PK₁ cells. In these studies (n = 3), the above-noted concentrations of neutralizing antibodies inhibited the thymidine uptake response to aFGF, bFGF, and PDGF-BB by 29.0 ± 4.8, 28.9 ± 7.9, and 30.6 ± 6.3%, respectively (all P < 0.05). These concentrations of the antibodies directed against aFGF and PDGF-BB did not significantly attenuate the effect of injured conditioned media to promote [³H]thymidine uptake in serum-starved LLC-PK₁ cells (7.7 ± 8.4% enhancement and 10.0 ± 8.8% inhibition, respectively; both not significant). Moreover, higher concentrations of neutralizing antibodies directed against aFGF and PDGF-BB (60 and 45 µg/ml, respectively) also did not significantly attenuate the effect of injured conditioned media to stimulate [³H]thymidine uptake in recipient LLC-PK₁ cells (3.3 ± 3.4 and 8.7 ± 3.8% inhibition, respectively; both not significant). By contrast, the neutralizing antibody directed against bFGF exerted a modest inhibition of the effect of injured conditioned media to promote [³H]thymidine uptake in each experiment (mean 17.4 ± 4.3% at 34 µg/ml and 26.5 ± 12.8% at 102 µg/ml; both P < 0.05) relative to IgG-treated injured conditioned control media.

Recent studies suggest that TGF- expression is increased in rat renal tubular cells after ischemic injury and that TGF- may be an autocrine regulator of renal tubular function (1, 3, 24). Our functional studies with exogenous TGF- revealed neither promotion of proliferation nor migration. However, TGF- can exert variable effects on these parameters, depending on the presence or absence of other GFs (20). Moreover, TGF- could potentially exert multiple other effects that promote tissue remodeling (20). We therefore used antibodies directed against TGF- in an attempt to determine whether TGF- was present in conditioned media removed from mechanically injured LLC-PK₁ cells. These Western blotting results revealed positive controls but no evidence of TGF-1 in injured media (Fig. 10). Moreover, using a sensitive ELISA kit, we were unable to detect TGF- in conditioned media from mechanically injured LLC-PK₁ cells. In these studies, TGF-1 averaged 3.9 ± 1.2 pg/ml in control media and 2.8 ± 1.4 pg/ml in conditioned media (n = 3).

View larger version (34K): [in this window] [in a new window]   Fig. 10.   Immunoblot of concentrated media from control (NON) and mechanically injured (INJ) LLC-PK1 cells probed with an antibody directed against transforming growth factor- (TGF-). Location of molecular weight (MW) standards 28.3 and 18.4 kDa are noted on vertical axis. First 3 lanes represent addition of, respectively, 5.0, 2.0, and 1.0 ng TGF-1 added directly to lane. Fourth and fifth lanes represent media (M) alone and media spiked (S) with 15 ng TGF-1.

Additional studies on endothelin. Our functional data suggest that ET-2 is a potential mediator of the response to conditioned media removed from mechanically injured LLC-PK₁ cells. Also, recent studies by Ong et al. (29) support a role for ET in the autocrine response to hypoxia in human renal tubular epithelial cells. Although our immunoblots did not demonstrate ET in conditioned media from injured LLC-PK₁ cells (Fig. 7), we further studied a potential role for ET in conditioned media removed from injured LLC-PK₁ cells. Initially, we examined the effect of pretreatment with phosphoramidon, an agent which inhibits ET-1-converting enzyme, on injured conditioned media response to [³H]thymidine uptake. In these studies, we utilized 10⁴ M phosphoramidon, a concentration found by Ong et al. (29) to inhibit ET-1 release by 75% in human renal tubular epithelial cells. [³H]thymidine uptake produced by conditioned media removed from injured LLC-PK₁ cells increased [³H]thymidine uptake nearly identically in the presence and absence of phosphoramidon (2.6 ± 0.2- and 2.7 ± 0.2-fold, respectively, vs. control media). Also, neither BQ123 (10⁶ M), an ET_A receptor antagonist (2.1-fold increase in [³H]thymidine uptake over control in the absence and 1.8-fold increase in the presence), nor BQ788 (10⁶M), an ET_B receptor antagonist (1.4-fold increase in [³H]thymidine uptake over control in its absence and 1.6-fold increase in its uptake in the presence), affected quiescent LLC-PK₁ [³H]thymidine uptake responses when exposed to conditioned media removed from mechanically injured LLC-PK₁ cells. In three studies, BQ788 inhibited the effect of ET-2 (10⁹ M) to stimulate [³H]thymidine uptake in serum-starved LLC-PK₁ cells by 17 ± 4% (P < 0.05). Although our observations do not support a role for endothelins in the scrape injury model we studied, in view of our and others' findings of endothelin effects on LLC-PK₁ cell migration and proliferation, further studies are needed to ascertain the role of endothelins in nephrogenic repair.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The present studies demonstrate that mechanical injury of LLC-PK₁ cells results in release of one or more polypeptide GFs into the culture media that produce a growth response in near-quiescent LLC-PK₁ cells. In addition, exposure of near-quiescent LLC-PK₁ cells to media removed from mechanically injured bovine vascular SMC induces both a growth and a motility response. Together, these observations suggest the presence of potential autocrine and paracrine pathways, in which mechanical injury of renal epithelial and vascular SMCs release GFs that potentially promote nephrogenic repair. Our studies, taken in the context of previous studies with mechanical, toxic, and hypoxic damage to cultured rat and human proximal tubular cells, clearly establish the presence and potential significant role for autocrine and paracrine mechanisms in promoting recovering of the renal tubular epithelium from a variety of insults (3, 12, 13, 23, 24, 29, 39).

Studies of Stevens and collaborators (12, 13, 39) have established a potential role for members of the FGF family, acting in an autocrine and paracrine fashion, to promote nephrogenic repair. These investigators found enhanced expression of aFGF in rat kidney tubular cells after nephrotoxin injury and found that media removed from scrape-injured rat proximal tubular cells stimulated [³H]thymidine uptake in human umbilical vein endothelial cells, also suggesting a role for aFGF in paracrine nephrogenic repair (13, 39). More recently, these workers described nephrotoxin-stimulated increased expression of FGF-7 (keratinocyte growth factor) in rat renal interstitial cells and enhanced expression of FGF-7 receptor in renal tubular epithelial cells (12). Together, these observations support the claim that autocrine (aFGF) and novel paracrine (FGF-7) mechanisms involving the FGF family contribute to renal repair.

The present studies demonstrate that addition of physiologically relevant concentrations of aFGF and bFGF exogenously stimulate [³H]thymidine uptake in LLC-PK₁ cells while not affecting LLC-PK₁ cell motility, thus reproducing the cellular effects of conditioned medium removed from mechanically injured LLC-PK₁ cells. We also found release of aFGF and bFGF into the media of scrape-injured LLC-PK₁ cells. This finding, taken in isolation, must be interpreted with caution in view of the fact that, in some studies, aFGF can be released as an inactive homodimer into conditioned media (15). However, suramin, a nonspecific inhibitor of polypeptide GF receptors (19), significantly inhibits the effect of conditioned medium to promote growth. Moreover, at least some of the effect of conditioned medium removed from mechanically injured LLC-PK₁ cells is inhibited by a neutralizing antibody directed against bFGF. Although aFGF and bFGF lack signal sequences that are typical of secreted proteins, atypical pathways can result in FGF cellular release (18). Also, an effect of mechanical injury to stimulate release of bFGF from cytosolic storage sites has also been demonstrated in endothelial cells (21). Collectively, the present and previous studies strongly suggest that several members of the FGF family, including aFGF (FGF-1), bFGF (FGF-2), and FGF-7 are potential autocrine and paracrine factors important in promotion of nephrogenic repair (12, 13, 39).

We also found that PDGF-BB was released onto culture media after mechanical injury of LLC-PK₁ cells. In addition, small concentrations of exogenous PDGF-BB modestly stimulated LLC-PK₁ [³H]thymidine uptake while not affecting motility. Although PDGFs are not commonly thought of as either being synthesized in the kidney or as exerting activity on renal tubular epithelial cells, human renal tubular epithelium expresses mRNA for PDGF-BB and secretes PDGF-AB into culture media (2, 6). Although our experiments are of interest, further studies are necessary to more completely evaluate the potential role of PDGF-BB in nephrogenic repair.

Several investigators have found a potential role for TGF-1 in renal tubular epithelial responses. For example, Basile et al. (1) found increased expression of TGF-1 in regenerating rat tubules following ischemic injury, and Nowak and Schnellmann (24) found that endogenous production TGF-1 increases with time in culture and regulates cellular metabolism, phenotype, and apoptosis in rabbit proximal tubular cells. However, exogeneously applied TGF-1 and other members of the TGF- family appear to inhibit nephrogenic repair in vitro (3, 16, 35). The results of the present studies also do not find release of TGF-1 into either scrape-injured conditioned media by ELISA and Western blotting or into media removed from heat shocked LLC-PK₁ cells (unpublished observations). Further studies are thus required to establish the role of TGF-1 in nephrogenic repair.

We should note several caveats regarding our studies. First, our approach to determining the substance(s) present in media removed from mechanically injured LLC-PK₁ cells was designed to detect the presence of one or more GFs known to be synthesized by the kidney. It is of interest that the effects of media removed from injured cells to promote [³H]thymidine uptake was several times greater than what we observed for any single exogenously applied GF and that the neutralizing antibody directed against bFGF only partially inhibited the thymidine uptake response to injured conditioned media. This suggests that perhaps as yet unidentified GFs contribute to the responses we described. Second, as is demonstrated by our experience in which bFGF was not detected by immunoblotting but was found by ELISA, it is possible that immunoblotting of severalfold concentrated samples of media removed from mechanically injured cells may not be sufficiently sensitive to detect low concentrations of other GFs that our biological studies indicated were candidates to produce the effects seen with conditioned media removed from injured LLC-PK₁ cells (e.g., PDGF-AA, ET-2, and EGF). Third, our studies do not clarify the exact stimulus that results in LLC-PK₁ cell release of GFs into the culture media. Although our scraping method of mechanical injury results in several factors (e.g., cell detachment, induction of cell membrane defects, and cell damage), our 60 min experiments, in which cells and cellular debris were removed from the experimental media and yet a growth promoting response still occurred, suggests that simply a reduction in cell mass and or loss of continguous cells can trigger release of GFs, one of which is bFGF, into the media from cells remaining within the monolayer. Also, our heat shock studies suggest that release of bFGF into the media can occur independent of cell detachment and death. Clearly, further studies will be required to delineate the individual contributions of membrane disruption, cell death, and detachment of cells from substrate in the release of growth promoting substances into media after mechanical injury. Fourth, we have not ascertained whether nonquiescent and sublethally injured cells can respond to the GFs that we found released into conditioned media, thereby completing an autocrine loop. Our studies do, however, suggest that remaining viable cells can respond to the released GFs. Finally, we have not done precise mass-balance studies to accurately ascertain how much GF was released by our experimental injury procedure and whether this is sufficient to induce a growth response.

In summary, the results of the present studies demonstrate that mechanical injury of cultured LLC-PK₁ and OK renal tubular epithelial cells results in a potential autocrine pathway that promotes recovery. Our experiments suggest that aFGF, bFGF, and PDGF-BB are potential candidates to mediate this autocrine response seen in LLC-PK₁ cells, and our neutralizing antibody studies confirm a role for bFGF in this autocrine response. We found no evidence that several GFs that are synthesized by some kidney cell types (EGF, heparin-bound EGF, ET-1, HGF, IGF-I, and TGF-) contribute to this LLC-PK₁ cell autocrine nephrogenic repair response.