INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE EXTRACELLULAR MATRIX (ECM) is an important factor in the regulation of cell growth and differentiation (1, 2, 12). During the developmental processes of various tissues and organs, ECM plays important roles in the organization and morphogenesis of epithelia. ECM also confers a survival signal, particularly for endothelial and epithelial cells (17). It has been established that the disruption of epithelial cell-ECM interactions resulted in apoptosis, a phenomenon defined as anoikis, so-called homeless cell death (7). Several conditions were found to circumvent anoikis, such as low cell density, addition of hepatocyte growth factor, overexpression of Bcl-2, and transformation with v-Ha-ras or v-src (7). A recent report indicated that the presence of fibronectin or vitronectin alone was not sufficient for endothelial cell survival. Limiting cell motility or spreading also induced apoptosis of endothelial cells despite the presence of ECM (20).

Proper interactions with the ECM may result in stabilization and facilitate differentiation of the polarized cell. On the other hand, an improper contact or interaction, for instance, the application of ECM from the apical side to epithelial cells, could elicit instability or even cell death. To test this hypothesis, we employed collagen gel overlay. In previously published studies, application of collagen gel over a Madin-Darby canine kidney (MDCK) monolayer induced rapid membrane remodeling of cells and elicited lumen formation within 24 h (8, 24). In our experiments, we observed a similar response and found that apoptosis occurred within 24 h following collagen overlay. We therefore decided to characterize in detail the response of MDCK cells to collagen gel overlay. At the subconfluent phase, collagen gel overlay initiated the formation of mesenchyme-like structures of MDCK cells early and induced apoptosis later; the monolayer eventually completely disintegrated because of widespread apoptosis within 3 days. Because agarose gel overlay did not elicit any morphological alterations, collagen gel overlay-induced apoptosis of MDCK cells is therefore not a result of physical stress. The collagen overlay-induced apoptosis is specific to the polarized cell type. We thus defined it as "disoriented cell death," in contrast to the previously described homeless cell death (7).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell lines and cultures. MDCK and LLC-PK₁ cells were purchased from American Type Culture Collection and regularly maintained in DMEM supplemented with 10% FCS. Bovine aortic endothelial cells (BAEC) were from Dr. Ikuo Morita of Tokyo Medical and Dental University, Tokyo, Japan; NIH/3T3 cells were from Dr. Hsiao-Shen Liu of National Cheng Kung University Medical College, Tainan, Taiwan; and ESK-4 cells were from the cell bank in National Cheng Kung University.

Preparation of hydrated collagen and agarose gel. Type I collagen was prepared from rat tail tendons according to the established procedure (16). The final concentration of type I collagen stock was 1% dissolved in 0.025 N acetic acid. For the preparation of agarose gel, a stock of 1% agarose solution in 0.025 N acetic acid was made. In preparing collagen gel, 3 vol of collagen stock were mixed with 5.7× DMEM (1 vol), 2.5% NaHCO₃ (0.5 vol), 0.1 M HEPES (1 vol), 0.17 M CaCl₂ (0.1 vol), 1 N NaOH (0.1 vol), and 4.3 vol of 1× culture medium (DMEM + 20% FCS) under chilled conditions. The mixtures were dispensed on the cultured cells at 1 ml/35-mm dish and placed in an incubator (5% CO₂ in air, 37°C) to allow for gelation. The agarose gel was prepared using the same protocol. After gelation, each culture was overlaid with 1.5 ml of culture medium that was replaced every other day.

DNA extraction and electrophoresis. The method of extraction of low-molecular-weight genomic DNA has been described previously (7, 26). Briefly, cultured cells were extracted with 0.5% Triton X-100, 10 mM EDTA, 10 mM Tris, pH 7.4, and phenol-chloroform for three times. The DNA was precipitated in propanol and electrophoresed in a 1.5% agarose gel. Finally the DNA was visualized with ethidium bromide staining under ultraviolet light.

Hoechst 33258 staining. MDCK cells cultured under agarose or collagen gel overlay for different incubation times were washed twice with PBS and then fixed with 2% paraformaldehyde in the presence of the gel. After the fixation buffer was washed out, cells were permeabilized with buffer containing 0.1% Triton X-100 and then stained with Hoechst 33258 (5 mg/ml) for 1 h in the dark. Finally, the stained nuclei were visualized under a fluorescence microscope (Nikon).

Histological study. To observe the cell morphology under collagen gel overlay, cells were cultured on a Transwell filter to the subconfluent phase and then overlaid with collagen gel. At intervals, the whole cultures were fixed with Formalin, dissected, and then embedded in paraffin. The specimens were sectioned, deparaffinated, and stained with hematoxylin and eosin.

Cell cycle analysis. The apoptosis was quantitated by flow cytometry with propidium iodide as described by Nicoletti et al. (19). After agarose or collagen gel overlay for 24, 48, and 72 h, the gel was removed and treated with collagenase to release attached cells or apoptotic bodies. MDCK cells remaining on culture dishes were treated with dispase (0.6 units, 0.5 mg/ml PBS) and trypsin to obtain a better cell suspension. The collagenase-treated suspension and trypsin-dispase-digested cells were combined, washed with PBS, and fixed in 70% alcohol. After fixation, cells were treated with RNase (100 mg/ml PBS) and stained with propidium iodide (40 mg/ml PBS). The mixed cells were incubated in the dark at room temperature for 30 min and analyzed by flow cytometry with a FACScan (Becton Dickinson, Mountain View, CA) with excitation set at 488 nm. Data were analyzed by Cell FIT software and represented as either histographs or numbers. The hypodiploid DNA peak of apoptotic cells can be distinguished from the normal diploid DNA peak on the fluorescence profiles of propidium iodide-stained cells.

Transfection. The DNA construct used for the transfection was a gift from Dr. Y. Tsujimoto (27), and the transfection process has been successfully done in our lab (14). The vector pCDj contained the G418 resistance gene and the Epstein-Barr virus-derived replication origin. The bcl-2 cDNA sequences were expressed by simian virus 40 enhancer/promoter regulatory elements (pCDj-bcl-2). A DNA construct without bcl-2 cDNA sequences (pCDj-SV2) was used as a control. MDCK cells were transfected by the method of lipofection. In brief, DNA-liposome complexes were applied to MDCK cells that had been cultured to 80-90% confluence for 24 h in a CO₂ (5%) incubator. The ratio of DNA to liposome complex at 1 mg/20 ml was used to obtain the optimum result for transfection. After the transfection, culture medium was replaced with 10% FCS-DMEM. At 72 h after transfection, cells were passaged at 1:10 dilution into G418-selective medium (effective concentration of 0.5 mg/ml).

Western blot analysis. The expression of Bcl-2 in wild-type MDCK and bcl-2 transfectants was determined by immunoblotting, as described previously (14). Briefly, 100 mg of cell homogenate protein from specific samples were resolved by 10% SDS-PAGE and electrophoretically blotted onto nitrocellulose paper and then incubated with mouse anti-human Bcl-2 polyclonal antibody (purchased from Dako); immunocomplexes were detected with horseradish peroxidase-conjugated goat anti-mouse IgG antibody (1:1,000 dilution) and envisioned by fluorography with an enhanced chemiluminescence detection kit (Amersham International).

Assessment of cell proliferation. Estimation of proliferation rate of MDCK cells and bcl-2 transfectants was assessed by measuring 1) cell number and 2) thymidine incorporation in the culture. Cells were plated at a density of 1 × 10⁵/dish on 60-mm dishes for 1 day, and the medium was changed every 2 days. To determine the cell number, cells were harvested by trypsinization and enumerated using a hemocytometer. To assess thymidine incorporation rate, 2 ml of [³H]thymidine (1 mCi/ml) were added to each dish and cells were incubated at 37°C for 4 h. The incorporation was terminated by adding 1 ml of cold thymidine (100 mg/ml). DNA was precipitated with cold 30% TCA and then washed once with 5% TCA and then 95% ethanol. [³H]DNA was dissolved in 2 ml of 0.5 N sodium hydroxide solution and transferred to vials containing 2.5 ml of Ecolume scintillation fluid. [³H]DNA was measured by a Beckman model LS 6800 scintillation counter.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Collagen gel overlay induced MDCK cell remodeling and apoptosis. MDCK cells were cultured to 80-85% confluence on culture plates before being overlaid with collagen gel (0.3%). To exclude the possibility that nonspecific effects of physical stress due to the gels might be involved, MDCK cells were overlaid by agarose gel (0.3%) in lieu of collagen gel as the control. Under normal culture conditions, MDCK cells form domes during the confluent phase at 48 h. Cell shedding from the monolayer was gradually enhanced due to overconfluence from 48 to 72 h. Agarose gel overlay did not affect the morphology and characteristics of the epithelial monolayer whatsoever (Fig. 1A). In contrast, within 10 h, collagen gel overlay caused MDCK cells to undergo a transformation from epithelial into mesenchymal cell-like morphology, which was characterized by cell elongation, reduction in cell-cell association, and augmented cell migration. Consistent with previous reports, the morphological changes of cells finally culminated in the formation of lumenlike structures at ~24 h (8, 24) (Fig. 1A). Simultaneously, MDCK cells started to disintegrate into tiny vesicles that resembled apoptotic bodies microscopically (Fig. 1A). As cell disintegration continued, widespread cell death was observed by 72 h.

View larger version (159K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Fig. 1.   Growth of Madin-Darby canine kidney (MDCK) cells after collagen gel overlay. MDCK cells cultured to 80-85% confluence were employed for the study. A: phase-contrast pictures show the morphology of MDCK cell monolayer after 24, 48, and 72 h of agarose or collagen gel overlay. Under agarose overlay, MDCK cells formed domes within 48 h and exhibited overconfluence-induced cell shedding within 72 h. In contrast, collagen gel overlay induced cell remodeling to form lumen within 24 h and marked cell degeneration within 48 h. B: growth curve of MDCK cells cultured under normal medium (control), agarose, or collagen gel overlay. Each point represents mean ± SE from 3 experiments in duplicate or triplicate. Only viable cells were counted.

View larger version (142K): [in this window] [in a new window]   View larger version (116K): [in this window] [in a new window]   Fig. 2.   Fluorescence microscope (A) and biochemical (B) examinations of apoptosis in MDCK cells after collagen gel overlay. MDCK cells cultured to 80-85% confluence were employed for the study. A: at 24, 48, and 72 h after agarose or collagen gel overlay, cultures were fixed, stained with Hoechst 33258, and examined under fluorescence microscope. Evidence of nuclear fragmentation could not be observed until 72 h of agarose overlay. In contrast, evidence of nuclear fragmentation could be observed within 24 h of collagen gel overlay. Collagen gel-induced apoptosis was enhanced with the time of the treatment. B: analysis of low-molecular weight (MW) DNA extracted from MDCK cells cultured under normal medium (M), agarose (A), or collagen gel overlay (C) at 24, 48, and 72 h. Low-MW DNA was extracted from whole cultures containing floated and attached cells in one 60-mm dish. Although apoptotic bodies could be observed under a microscope within 24 h, DNA ladder was not visible until 48 h of collagen gel overlay.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Quantitation of collagen gel overlay-induced apoptosis in MDCK cells. A: representative results of cell cycle analysis in cultures under normal medium (control), agarose, and collagen gel for 72 h. Floated and attached cells were combined, and cell cycle was assessed by FACScan. The sub-G0/G1 population (Ao) of the cell cycle indicates apoptosis. B: summarized results of the sub-G0/G1 phase in cultures receiving respective treatment for 24, 48, and 72 h. Each point represents mean ± SE from 3 experiments in duplicate or triplicate.

View larger version (99K): [in this window] [in a new window]   Fig. 4.   Histological (A) and fluorescence microscope (B) examinations of MDCK cells overlaid by collagen gel for 72 h. MDCK cells were cultured on a Transwell filter to 80-85% confluence and then overlaid with collagen gel. At 72 h of collagen gel overlay, whole culture was dissected, fixed, embedded in paraffin wax, and finally sliced by a microtome perpendicularly to the gel plane with 5-mm thickness. Specimens were stained with hematoxylin and eosin (A) or Hoechst 33258 (B) and examined under light or fluorescence microscopy, respectively (magnification = ×400). There are apparently 2 layers of cells under collagen gel, indicating MDCK cells remodel to form lumen. Fragmented nuclei could be seen mostly under collagen gel or between the gel and attached cells. Occasionally, some apoptotic bodies could be found in the lumen. Arrow indicates Transwell filter.

Effect of collagen gel overlay on other polarized cells. Proximal tubule-derived LLC-PK₁ epithelial cells and BAEC were employed to determine whether or not collagen overlay-induced apoptosis described above was a polarized cell-specific phenomenon. Under collagen gel, both LLC-PK₁ and BAEC monolayer underwent cell-cell dissociation and became irregular in shape within 12 h but showed no evidence of lumen formation even at 96 h (Fig. 5). Mild apoptosis of BAEC was observed by 36 h as assessed by DNA ladder analysis, whereas marked apoptosis of LLC-PK₁ could be detected within 24 h of collagen overlay (Fig. 6, A and B). Furthermore, FACScan analysis revealed that collagen gel overlay induced extensive apoptosis in LLC-PK₁ cells within 24 h (56.2 ± 1.2% vs. 10.2 ± 0.4%) (Fig. 6C).

View larger version (139K): [in this window] [in a new window]   Fig. 5.   Morphology of bovine aortic endothelial cells (BAEC) and LLC-PK1 cells before and after collagen gel overlay. Top: phase-contrast pictures show morphological changes of BAEC cell monolayer at 0, 12, and 96 h of collagen gel overlay. Bottom: morphological changes of LLC-PK1 cells at 0, 12, and 48 h of collagen gel overlay. These 2 cell lines did not exhibit any morphological alterations under agarose gel overlay. Note that cell degeneration was markedly higher in LLC-PK1 cells than in BAEC cells when subjected to collagen gel overlay.

View larger version (56K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Fig. 6.   Demonstration of apoptosis in LLC-PK1 and BAEC cells cultured under collagen gel overlay. Low-MW DNA was extracted from LLC-PK1 (A) and BAEC (B) cells cultured under normal medium, agarose, or collagen gel overlay for indicated times and analyzed by electrophoresis. C: quantitative changes of apoptotic ratio in LLC-PK1 cells cultured under normal medium (control), agarose, and collagen gel for 24, 48, and 72 h. Both floated and attached cells were collected, and their cell cycle was analyzed by FACScan. Apoptosis ratios were obtained from the sub-G0/G1 population of the cell cycle. Each point represents mean ± SE from 3 experiments in duplicate.

Collagen gel overlay did not inhibit growth of mesenchymal cells. To test whether collagen gel overlay would affect nonpolarized cells, NIH/3T3 and ESK-4 cells, both mesenchymal in nature, were examined under collagen or agarose gel overlay. We found that the growth of mesenchymal cells was restricted by the physical stress of agarose gel per se. In contrast, collagen gel overlay seemed to provide favorable conditions for cell survival (Fig. 7). The morphology and confluence of the these cells were not altered by collagen gel overlay. Biochemical examination did not exhibit any sign of DNA laddering (data not shown).

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Growth curve of NIH/3T3 (A) and ESK-4 (B) cells cultured under normal medium (control), agarose, or collagen gel overlay. For these experiments, cells were cultured to certain stage and then underwent specific treatment and their cell numbers were assessed at 24, 48, and 72 h. Only viable cells were counted.

Low cell density delayed collagen overlay-induced apoptosis. Previous reports have demonstrated that MDCK cells cultured at a lower density were resistant to homeless cell death (7). Here, we examined whether collagen gel overlay could affect the cell growth of MDCK cells when they were tested at lower density. MDCK cells were plated at 1 × 10⁵ in a 60-mm culture dish for 1 day to allow for cell adherence. On the second day (day 0), cultures were treated with normal culture medium (control), agarose, or collagen gel overlay and the cell numbers were assessed every other day until day 6. As shown in Fig. 8, there was a marked increase in cell number from day 2 to day 4 in both control and agarose overlay groups, manifesting the rapid proliferative phase. At the first 2 days of collagen gel overlay, cultures under agarose gel overlay had as many cells as the control. However, from then on to day 6, there was no further increase in cell number of cultures under collagen overlay, indicating that a balance between cell proliferation and apoptosis was reached. Because collagen gel overlay did not affect the growth of low-density cultures within 2 days and the collagen gel overlay-induced apoptosis appeared later than 48 h, we conclude that lower cell density would delay disoriented cell death.

View larger version (16K): [in this window] [in a new window]   Fig. 8.   Effects of collagen gel overlay on MDCK cells at a lower density. For this experiment, 1 × 105 MDCK cells were plated in a 60-mm culture dish for 1 day and then treated with normal culture media (control), agarose, or collagen gel overlay. Cell numbers were assessed at days 0, 2, 4, and 6 of treatment. There was a rapid increase in cell number between day 2 and day 4 in both control and agarose overlay groups. In contrast, there was no further increase of cell number in cultures under collagen overlay from day 2 to day 6.

Bcl-2 overexpression prolonged collagen overlay-induced apoptosis of MDCK cells. The bcl-2 protooncogene has been shown to be anti-apoptotic in various systems (reviewed in Refs. 21 and 25). To explore whether Bcl-2 overexpression conferred resistance to apoptosis induced by collagen gel overlay, we transfected MDCK cells with bcl-2. Figure 9A shows that Bcl-2 protein levels were expressed in stably transfected MDCK cells. MDCK cells, the control plasmid transfectant (C1), and bcl-2 transfectants (B1, B4, and B9) were cultured to 80-90% confluent and then were overlaid with collagen gel as described above. Like MDCK cells, the bcl-2 transfectants developed lumenlike structures within 24 h of collagen gel overlay. As expected, the control plasmid transfectants behaved very much like the wild-type MDCK cells, exhibiting DNA fragmentation within 1 day under collagen gel, whereas bcl-2 transfectants showed no obvious DNA fragmentation until 72 h (Fig. 9B). Cell cycle analysis by FACScan also indicated that overexpression of Bcl-2 conferred resistance to apoptosis induced by collagen gel overlay to a certain extent (Fig. 9C). Because DNA fragmentation could be observed on bcl-2 transfectants within the 72-h collagen overlay, we further assessed Bcl-2 protein levels during the time course. The results showed that Bcl-2 protein contents gradually decreased following collagen gel overlay (Fig. 10). Interestingly, although certain levels of Bcl-2 were still present in B1 and B6 cells, both bcl-2 transfectants underwent apoptosis within 72 h of collagen gel overlay.

View larger version (41K): [in this window] [in a new window]   Fig. 9.   Bcl-2 overexpression prolongs collagen gel overlay-induced apoptosis. A: examination of Bcl-2 expression in MDCK cells transfected with bcl-2. Western blot analysis was used to demonstrate the expression of Bcl-2 in homogenates of wild-type, plasmid control (C1), and bcl-2-transfected (B1, B4, B6, B9) MDCK clones. B: electrophoretic results of DNA extracted from different MDCK clones (wild type, C1, B1, B4, and B9) cultured under collagen gel overlay for 24, 48 and 72 h. C: quantitation of collagen gel overlay-induced apoptosis in MDCK cells and B4 and B9 clones at 24, 48 and 72 h. Both floated and attached cells were analyzed, and the sub-G0/G1 population of the cell cycle was determined by FACScan analysis. Each point represents mean ± SE from 3 experiments in duplicate or triplicate.

View larger version (85K): [in this window] [in a new window]   Fig. 10.   Bcl-2 levels in bcl-2 transfectants declined during the time course of collagen gel overlay. B1 and B6 clones were cultured to the subconfluent stage and then treated by collagen gel overlay. At indicated time points, cells were harvested. Bcl-2 proteins were assessed by Western blot analysis, and -actin was used as an indicator for the amount of loading. M0 indicates the sample of bcl-2 control transfectants harvested at 0 h.

Apoptosis induced by collagen gel overlay could be alleviated by cycloheximide. Because apoptosis induced by collagen overlay had a 24-h lag period, we wondered whether some de novo gene products were involved in disoriented cell death. When cycloheximide (1 mg/ml) was added to cultures under collagen overlay, it was possible to prevent collagen gel overlay-induced morphological changes until at least 48 h. However, cells rounded up and detached from the dish at 72 h of cyclohexamide treatment in both cultures with or without collagen gel overlay (data not shown). Therefore, the examination of apoptosis was assessed within 48 h. Results showed that cycloheximide conferred resistance to collagen gel overlay-induced apoptosis within 48 h, as confirmed by DNA laddering analysis (Fig. 11A) and FACScan analysis (Fig. 11B). These data indicated that newly synthesized proteins were required for collagen gel overlay-induced apoptosis.

View larger version (31K): [in this window] [in a new window]   Fig. 11.   Cycloheximide (CHX) confers resistance to collagen gel overlay-induced apoptosis. A: result of electrophoresis analysis of low-MW DNA-extracted MDCK cells cultured under agarose or collagen gel overlay for 24 and 48 h with or without the presence of cycloheximide (1 mg/ml). M indicates DNA marker. B: quantitative changes of apoptotic ratio in MDCK cells cultured under agarose and collagen gel in the absence or presence of cycloheximide (1 mg/ml). Data were obtained from the summarized results of the sub-G0/G1 population of the cell cycle assessed by FACScan in cultures receiving respective treatment for 24 and 48 h. Each point represents mean ± SE from 3 experiments in duplicate.

Functional block of ₂-integrin-alleviated apoptosis induced by collagen gel overlay. Epithelial cell-substratum attachment is mediated by integrins, a family of ECM receptors (10). It has been reported that MDCK cells expressed several integrins, including ₂₁, ₃₁, _x1, _v3, and ₆₄ (23), of which ₂₁ is the major receptor for collagens I and IV (22). Although integrins are considered to be localized to the basolateral membrane where they have direct contact with ECM, Schwimmer and Ojakian (24) have demonstrated that ₂₁-integrin on the apical membrane is the one that mediates collagen gel overlay-induced cell remodeling of MDCK cells. To elucidate whether collagen gel overlay-induced apoptosis was also mediated by signals transduced from ₂₁-integrin, MDCK cells were overlaid with collagen gel in the presence of anti-₂-integrin monoclonal antibody (MAb) 5E8, which was previously shown to block ₂₁ signal transduction in these cells (24). As shown in Fig. 12A, collagen gel-mediated cell morphological changes were inhibited by MAb 5E8 within 48 h. In addition, collagen gel-induced DNA fragmentation could be alleviated by MAb 5E8. FACScan analysis demonstrated that collagen gel overlay-induced apoptosis was markedly reduced by MAb 5E8 from 30.42% to 9.43% at 48 h (Fig. 12B). These data indicate that signals transduced from ₂-integrin are required for collagen gel overlay-induced membrane remodeling and apoptosis.

View larger version (108K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Fig. 12.   Anti-2-integrin antibody (Ab) blocks collagen gel overlay-induced cell remodeling and apoptosis. A: morphological changes of MDCK cells cultured under collagen gel overlay for 24 and 48 h with or without the presence of anti-2-integrin monoclonal antibody (MAb) 5E8. Obviously, MAb 5E8 blocks the collagen gel overlay-induced cell remodeling. B: quantitative changes of apoptotic ratio in MDCK cells cultured under control, agarose, or collagen gel in the absence or presence of MAb 5E8. Data were obtained from the summarized results of the sub-G0/G1 population of the cell cycle assessed by FACScan in cultures receiving respective treatment for 48 h. MAb 5E8 markedly reduced collagen gel overlay-induced apoptosis. Each bar represents mean ± SE from 3 experiments in duplicate.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Previous studies have demonstrated that when MDCK cells interact with the ECM from apical sides they would undergo cell membrane remodeling, which resulted in lumen formation (8, 24). In this study, we observed similar responses, except that we also noticed the presence of apoptosis. As the incubation time of collagen overlay was prolonged, the level of apoptosis was enhanced. The coexistence of cell remodeling and apoptosis resembles the conditions in which MDCK cells are cultured in suspension. For instance, Wang et al. (29) found that MDCK cells developed cystlike structures with reversed polarization showing apical membrane facing outside and basolateral membrane facing inside when cultured in suspension, whereas Frisch and Francis (7) observed that MDCK cells developed apoptosis when their interactions with ECM are disrupted by plating them in suspension or pretreating the culture plates with polyhydroxyethylmethacrylate.

When subjected to collagen gel overlay, MDCK cells exhibited morphological changes first and developed apoptosis later. Substances such as cycloheximide and anti-₂-integrin antibody, which prevent collagen overlay-induced morphological changes, could also prevent apoptosis. These data suggest that apoptosis could be targeted to cells that develop cell remodeling under collagen gel overlay. In fact, apoptosis was initially identified at the areas where cell remodeling was present. In addition, the majority of apoptotic bodies lay between the collagen gel and the firmly attached cell layer (Fig. 4). Once the collagen gel was removed, the apoptotic bodies were found stuck onto the gel. These data indicate that cells that develop remodeling are prone to apoptosis. It could be taken to indicate that cells that underwent remodeling but had not positioned well might die of disorientation and that the apoptotic cells might serve to form a barrier to ensure survival of the attached cells. On the other hand, those cells that did not develop cell remodeling when subjected to collagen overlay might not have received ₂-integrin signaling from the apical membrane.

Because disruption of epithelial cell-matrix interaction and inappropriate cell-matrix interaction would both prompt apoptosis, it should be of fundamental importance to compare anoikis with disoriented cell death. Disruption of epithelial cell-matrix interaction triggers apoptosis of MDCK cells within 6-8 h, which cannot be prevented by the protein synthesis inhibitor, cycloheximide. In contrast, collagen gel overlay triggers MDCK cell apoptosis at 24 h, much later than the onset of anoikis. Disoriented cell death can be prevented by cycloheximide, indicating protein synthesis is required for this process. Lower cell density conferred resistance to anoikis, but it could only slightly delay disoriented cell death. Bcl-2 overexpression alleviated anoikis (7, 14) and prolonged disoriented cell death. Obviously, anoikis and disoriented cell death are induced by different stimuli and thereby different signaling mechanisms. We have demonstrated that collagen gel overlay induces a time course-dependent decrease of Bcl-2 protein. However, Bcl-2 proteins are still maintained at certain levels before the death of bcl-2 transfectants (Fig. 10). Recent reports indicated that caspase-3 induced degradation of Bcl-2, the degraded product of which could trigger apoptosis (4). It is therefore possible that collagen gel overlay-induced apoptotic signals could result in degradation of Bcl-2, which may in turn facilitate the disoriented cell death in Bcl-2-overexpressing MDCK cells. Exactly how the apoptosis machinery is turned on by collagen gel overlay is now under investigation.

Epithelial cells in vivo reside on basement membranes consisting of specific types of ECM. The basement membrane not only serves as a survival factor but also facilitates growth and differentiation via signals mediated through various types of integrins. In this study, we demonstrated for the first time that type I collagen fibrils induced apoptosis of polarized cells when applied apically. The functions of disoriented cell death, therefore, are to ensure the polarity of epithelial or endothelial cells both in vitro and in vivo. It is well known that apoptosis plays an important role for the morphogenesis of various organs or tissues during developmental processes. However, what in the physiological environments causes apoptosis remains speculative. Considering that type I collagen is the most abundant substratum providing the framework for tissues and organs and that signals transduced from collagen through ₂-integrin into epithelial cells could result in apoptosis, we can speculate that disoriented cell death plays very important roles in morphogenesis and organogensis under physiological conditions. On the other hand, disoriented cell death may possibly function as a restriction for preneoplastic cells during invasion in pathological conditions.

Evidence is increasing linking cell death to the development of renal cysts; for example, widespread apoptosis was found in the kidneys of autosomal dominant and recessive polycystic kidney diseases (30). In addition, three studies using transgenic mice have shown that knocking out Bcl-2 led to the development of cystic kidney (11, 18, 28). With the use of an in vitro model to study the cell biological mechanism of cyst formation, our laboratory has also shown that MDCK cells, when cultured in collagen gel, develop massive apoptosis in addition to the well-known multicellular cystic structures (13). Moreover, we observed that MDCK cells initially formed cell aggregates in collagen gels and then reorganized into cystic structures containing a central cavity due to apoptosis in the central portion of the aggregates (13), thus evident that apoptosis is involved in pathogenesis of renal cyst formation both in vivo and in vitro. Although the cause of apoptosis is still not understood, collagen gel-induced apoptosis apparently plays an important role in it. During the developmental process of the kidney in embryo, vigorous interactions must occur between cells and their ECM, which ultimately affect cell remodeling and morphogenesis. The cell survival or death therefore could be very critical for the renal organogenesis. In this study, we demonstrated that Bcl-2 overexpression alleviated collagen overlay-induced apoptosis. In addition, overexpression of Bcl-2 in MDCK cells completely averted the development of apoptosis-induced cyst formation (13). Together, prevention of cell death, i.e., through Bcl-2 overexpression, should therefore be a prerequisite not only for cell survival but also for the normal differentiation of the nephrons during renal ontogeny.