点击显示 收起
【摘要】 Inner medullary collecting ducts (IMCD) are terminally differentiated structures derived from the ureteric bud (UB). UB development is mediated by changes in the temporal and spatial expression of integrins and their respective ligands. We demonstrate both in vivo and in vitro that the UB expresses predominantly laminin receptors ( 3 1 -, 6 1 -, and 6 4 -integrins), whereas the IMCD expresses both collagen ( 1 1 - and 2 1 -integrins) and laminin receptors. Cells derived from the IMCD, but not the UB, undergo tubulogenesis in collagen-I (CI) gels in an 1 1 - and 2 1 -dependent manner. UB cells transfected with the 2 -integrin subunit undergo tubulogenesis in CI, suggesting that collagen receptors are required for branching morphogenesis in CI. In contrast, both UB and IMCD cells undergo tubulogenesis in CI/Matrigel gels. UB cells primarily utilize 3 1 - and 6 -integrins, whereas IMCD cells mainly employ 1 1 for this process. These results demonstrate a switch in integrin expression from primarily laminin receptors in the early UB to both collagen and laminin receptors in the mature IMCD, which has functional consequences for branching morphogenesis in three-dimensional cell culture models. This suggests that temporal and spatial changes in integrin expression could help organize the pattern of branching morphogenesis of the developing collecting system in vivo.
【关键词】 kidney development branching morphogenesis threedimensional culture basement membranes
THE MULTIBRANCHED INNER MEDULLARY collecting ducts (IMCD) of the kidney are highly ordered, terminally differentiated structures consisting of polarized epithelial cells derived embryologically from the ureteric bud (UB). The UB, which originates as an outgrowth from the Wolffian duct, also gives rise to the ureters and the epithelial lining of the trigone of the bladder. For such diverse urogenital segments to develop from the UB, cells must undergo well-defined differentiation processes ( 15 ).
The collecting system of the kidney is formed by iterative branching morphogenesis of the UB. Although this process has been studied in whole animal, organ, and cell culture models, the molecular cues for its control are still poorly understood. Cell culture models, which predominantly utilize Madin-Darby canine kidney (MDCK) cells grown in three-dimensional (3D) collagen-I (3D-CI) gels, recapitulate branching morphogenesis in its simplest form. These models suggest that branching morphogenesis is a multistep process that requires sequential cell adhesion to extracellular matrix (ECM), cell spreading, cell proliferation, and cell migration to ultimately form multicellular tubelike structures ( 14 ). A balance between proliferation and apoptosis is critical for the final formation of the lumen of the tubule ( 5 ).
Integrins, the predominant cell-surface receptors that mediate interactions between cells and ECM, play a crucial role in cellular functions such as cell adhesion, migration, proliferation, and apoptosis ( 7, 21 ). Both in vivo and in vitro data demonstrate the importance of integrins that interact with the laminin basement membrane components in normal UB development; however, there is no information on the role of the collagen receptors, 1 1 - and 2 1 -integrins. Mice deficient in the laminin receptor 3 1 -integrin have fewer collecting ducts in the papilla, resulting in decreased branching morphogenesis ( 11 ). Surprisingly, mice lacking the 6 -integrin subunit, which dimerizes to form the laminin receptors 6 1 and 6 4, do not show any anomalies in the developing UB ( 6 ). However, when 6 - and 3 -null mice are crossed, ureters fail to develop ( 4 ). 3 1 - And 6 -integrin subunits are required for UB branching morphogenesis in organ and cell culture model systems ( 24 ).
The spatiotemporal expression of laminin- and collagen-binding integrins in the development of the UB is poorly described and controversial. Studies performed in human tissues describe 3 - and 6 -integrin subunit ( 9, 10 ) expression in the developing collecting system, whereas 2 -, 3 -, and 6 - containing integrin subunits are present in the adult collecting ducts ( 9, 10 ). Although inconclusive, these studies suggest that the expression pattern of integrins in the UB changes as it undergoes terminal differentiation to the collecting duct.
To further understand how integrin expression modulates UB development, we investigated the functional role of murine laminin- and collagen-binding integrins in cell culture models utilizing cells isolated from the UB at embryonic (E) day 10.5 and the adult IMCD. We show both in vivo and in vitro that the undifferentiated UB and the well-differentiated IMCD express the laminin receptors 3 1, 6 1, and 6 4, whereas expression of the collagen-binding integrins, 1 1 and 2 1, is restricted to the differentiated IMCD. Utilizing 3D tubulogenesis assays, we demonstrate that cells derived from the IMCD, but not the UB, can undergo tubulogenesis in CI gels in an 1 1 - and 2 1 -dependent manner. In contrast, both UB and IMCD cells undergo tubulogenesis in CI/Matrigel (MG) gels; however, UB cells primarily utilize 3 1 - and 6 -integrins, whereas IMCD cells mainly employ 1 1 for this process. These findings suggest that temporal and spatial changes in integrin expression could help organize the pattern of branching morphogenesis of the developing collecting system in vivo.
MATERIALS AND METHODS
Materials. Growth factor-reduced MG and rat CI were obtained from Becton Dickinson (Franklin Lakes, NJ). Blocking antibodies against 1 (Ha31/8)-, 2 (Ha1/29)-, 6 (GoH3)-, 1 (Ha2/5 and 9EG7 )-, and 4 (346-11A)-integrin subunits were obtained from Pharmingen (San Diego, CA). An antibody to the cytoplasmic domain of human 3 -integrin (Ab1920) that cross-reacts with mouse, a monoclonal antibody to the extracellular domain of mouse 1 -integrin (for immunohistochemistry; clone MB1.2), and an antibody specifically directed against the extracellular domain of human 2 (clone 10G11) were obtained from Chemicon (Temecula, CA). An Apoptag Apoptosis Detection Kit was obtained from Serologicals (Norcross, GA), and Vectashield was from Vector Labs (Burlingame, CA). Fast 3,3'-diaminobenzidine (DAB) tablets were from Sigma (St. Louis, MO), and Permount was obtained from Fisher Scientific (Loughborough, UK).
Constructs and cell culture. Immortalized UB (a gift from J. Barasch, Columbia University, New York, NY) and IMCD (a gift from E. Delpire, Vanderbilt University, Nashville, TN) cells were cultured and maintained as previously described ( 18, 19 ). Renal papilla cells derived from E18 kidneys of 3 -integrin-null mice and 3 -integrin-null cells reconstituted with the human 3 -integrin subunit were cultured and maintained as previously described ( 23 ). The human 2 -integrin expression construct was prepared as described previously ( 16 ). A pure UB cell population expressing 2 1 -integrin ( 2 -UB) was derived by transfecting UB cells with the 2 -integrin expression construct, after which cells were sorted via fluorescence-activated cell sorting (FACS). An antibody directed to the extracellular domain of human 2 -integrin was utilized to define 2 -integrin expression. Control cells were transfected with the empty vector.
Inhibition of 1 - and 2 -integrin subunit expression by small interfering RNA. The murine 1 -integrin sequence 5'-GGTCACTGTAGCCTGCATT-3' and the murine 2 sequence 5'-GGAGCGAAAATATTTTCCG-3' were targeted for RNA interference. Double-strand small interfering (si)RNA were obtained from Ambion (Austin, TX). Subconfluent populations of IMCD cells were transfected using siPORT Amine Transfection Agent (Ambion) according to the manufacturer's instructions. Control cells were transfected with Silencer Negative Control No. 1 siRNA obtained from Ambion. Tubulogenesis assays were performed 48 h after transfection, and flow cytometry and migration experiments 7 days after transfection.
3D cell culture. Tubulogenesis of UB- and IMCD-derived cells was performed in 3D ECM gels as previously described ( 1, 18 ). The CI gels were composed of 0.1 mg/ml CI in DMEM containing 20 mM HEPES (pH 7.2). For the MG/CI gels, a 1:1 mixture of the collagen solution described above was mixed with growth factor-reduced MG, giving a final concentration of 0.5 mg/ml of CI and 0.5 mg/ml of MG ( 18 ). One hundred microliters of medium supplemented with 10% FCS were added to the gels after they had solidified. For quantification of the branches, cells that formed branching structures (defined as more than 1 branch) were counted in five randomly picked high-power fields. The image analysis was performed using the Metamorph cell-imaging program (Universal Imaging, Downingtown, PA). The degree of branching morphogenesis was quantified by using the number of end points as a correlative measure of the number of branching events. Assays were performed at least in triplicate, and error bars represent SE. P values were calculated with Student's t -test.
Flow cytometry. A suspension of UB or IMCD cells was incubated with monoclonal antibodies to the appropriate integrin (1:100 dilution), followed by incubation with the appropriate secondary antibodies (FITC-coupled rabbit anti-rat or hamster immunoglobulin; 1:100 dilution). Flow cytometry was performed with a FACScan instrument (Becton Dickinson). Cell suspensions incubated with secondary antibody only were used as a negative control for integrin expression.
Immunohistochemistry. Frozen sections of embryonic or adult mouse kidneys were utilized in these studies. Sections were incubated for 20 min with 2% H 2 O 2 in methanol to quench the endogenous peroxidase. After an additional wash in PBS, sections were incubated for 1 h at room temperature with 3% bovine serum albumin, 3% normal goat serum in PBS (blocking solution), followed by incubation with the appropriate primary antibodies diluted in blocking solution overnight at 4°C. After being washed with PBS, the slides were incubated with horseradish peroxidase-conjugated secondary antibodies (1:100 dilution in blocking buffer) for 2 h at room temperature, and staining was evaluated by incubating the slides with 0.05 M Tris solution containing 0.05% DAB tetrahydrocloride Sigma Fast DAB tablets. Finally, the slides were counterstained in Harris hematoxylin and mounted using Permount.
Cell proliferation and apoptosis. UB or IMCD cells (5 x 10 3 cells/ml) were embedded in the 3D gels (100-µl final volume) in 96-well plates as described above and incubated in DMEM/F-12 containing 2% FCS in the presence or absence of an anti-integrin antibody (10 µg/ml final concentration). After 2 days in culture, cells were pulsed for an additional 48 h with [ 3 H]thymidine (1 µCi/well). The gels were then removed from the plates and dialyzed against PBS for 24 h to remove free [ 3 H]thymidine. The gels were then lysed in 10% SDS (100-µl final volume), and the lysates were measured with a beta counter.
For apoptosis studies, the cells were incubated in 3D gels as described above. After 4 days in culture, the gels were fixed in 4% paraformaldehyde for 30 min, followed by DMSO-methanol in a ratio of 1:1 for 30 min. Apoptosis was detected using the Apoptag Apoptosis Detection Kit, as described by the manufacturer. The gels were also stained with 4-6-diamidino-2-phenylindole to identify all the cell nuclei. Apoptotic cells were counted in five different microscopic fields of a fluorescence microscope, and the apoptotic index was expressed as follows: (no. of positive apoptotic cells/total no. of cells/microscopic field) x 100.
Cell adhesion. Microtiter plates (96-well) were coated with matrix proteins at the indicated concentrations in PBS for 1 h at 37°C. Plates were then washed with PBS and incubated with PBS containing 0.1% BSA for 60 min to block nonspecific adhesion. One hundred microliters of single-cell suspensions (10 6 cells/ml) in serum-free DMEM containing 0.1% BSA were added in triplicate to 96-well plates and incubated for 60 min at 37°C. In some experiments, cell suspensions were preincubated with anti-integrin antibodies (10 µg/ml final concentration) on ice for 30 min before the assay. Nonadherent cells were removed by washing the wells with PBS. Cells were then fixed with 1% formaldehyde, stained with 1% crystal violet, solublized in 2% SDS, and the cell lysates were then read at 570 nm. Cells bound to fetal calf serum were used as a positive control to indicate maximalcell adhesion, whereas cells bound to 1% BSA-coated wells were used as background, and this opitcal density was subtracted from that obtained in serum or extracellular matrix proteins.
Cell migration. Cell migration was assayed in Transwells consisting of polyvinylpyrolidone-free polycarbonate filters with 8-µm pores. The underside of each Transwell was precoated with ligand overnight at 4°C, and the filter was subsequently blocked with 1% BSA for 1 h at 37°C to inhibit nonspecific migration. One hundred microliters of a cell suspension (1 x 10 6 cells/ml) in serum-free medium containing 0.1% BSA were added to the wells, and the cells were allowed to migrate into the matrix coated on the underside of the Transwell for 4 h. Cells on the top of the filter were removed by wiping, and the filter was then fixed in 1% formaldehyde in PBS. Migrating cells were stained with 1% crystal violet, and nine randomly chosen fields from triplicate wells were counted at x 200 magnification.
RESULTS
1 1 - And 2 1 -integrins are sufficient to mediate IMCD tubulogenesis in CI gels. As most renal cell tubulogenesis assays utilizing MDCK cells are performed in 3D-CI gels ( 13 ), we determined whether both IMCD and UB cells undergo branching morphogenesis under these conditions. IMCD formed tubular structures within 12 days of culture ( Fig. 1 A ), whereas the majority of UB cells failed to form organized structures ( Fig. 1 A ) and underwent apoptosis (data not shown).
Fig. 1. Ureteric bud (UB) and inner medullary collecting duct (IMCD) express different integrins. A : phase-contrast image of UB and IMCD cells cultured for 12 days in 3-dimensional collagen-1 (3D-CI) gels. IMCD cells, but not UB cells, form distinct tubelike structures. B : flow cytometry was performed on UB and IMCD cells with monoclonal antibodies against different integrin subunits as described in MATERIALS AND METHODS. Expression of the various integrin subunits is shown as mean fluorescence intensity relative to control cells incubated with secondary antibody only. Expression of the 3 -integrin subunit was evaluated by Western blot analysis of total cell lysates from UB and IMCD cells (50 µg/lane) using a rabbit polyclonal antibody directed against the cytoplasmic tail. Equal amounts of protein were loaded as verified by immunoblotting against actin (data not shown). C - D : immunohistochemical localization of 1 -, 2 -, 3 -, 6 -, 1 -, and 4 -subunits was assessed in frozen sections of embryonic day (E) 15 ( C ) or adult kidneys ( D ) as described in MATERIALS AND METHODS. II only represents UB or IMCD stained with secondary antibody alone; arrows indicate the UB in embryonic kidneys. Scale bar =100 µm.
The discrepancy in growth between UB and IMCD cells in 3D matrices suggested that these cell types express different ECM receptors, especially with respect to CI. We therefore analyzed the integrin profile of UB and ICMD cells by flow cytometry or immunoblotting. UB cells express low levels of 1 1 -integrin, and almost no 2 1 -integrin, whereas IMCD cells express significant amounts of these two major collagen-binding receptors ( Fig. 1 B ). UB cells express more of the laminin receptor 3 1 -integrin as well as the 6 -integrin subunit, which is present predominantly as 6 4 rather than the 6 1 heterodimer ( 24 ) (not shown), than do IMCD cells ( Fig. 1 B ). The discrepancy in 1 1 - and 2 1 -integrin expression between the two cell lines could explain why IMCD, but not UB cells, were able to form tubules in 3D-CI gels.
To determine whether the expression pattern of integrins in the UB and IMCD cells was representative of the anatomic structures from which they were originally derived, we determined the integrin expression patterns in embryonic and adult mouse kidneys. As expected, 1 1 - and 2 1 -integrins were poorly expressed in the UB of E15 embryonic kidneys as determined by immunohistochemistry ( Fig. 1 C ). In contrast, expression of these two collagen-binding receptors was detectable in the collecting ducts of adult kidneys ( Fig. 1 D ). 3 1 -, 6 1 -, And 6 4 -integrins were expressed in the UB and adult collecting ducts. Although these are not quantitative assays, these results suggest that the integrin expression patterns found in the cells are similar to those for the organelles from which they were originally derived. Therefore, these cell populations are potentially valid model systems with which to study the role of integrin-ECM interactions in the formation of tubules at different stages of collecting system development.
To determine which integrins mediated IMCD cell tubulogenesis in CI, cells were grown in 3D-CI gels in the presence of blocking antibodies or isotype-matched immunoglobulins to different integrin subunits. Only antibodies directed against the 1 - and 2 -subunits, added alone or in combination, resulted in a dramatic reduction and/or complete inhibition of tubulogenesis ( Fig. 2 ). The role of 3 1 -integrin in tubulogenesis was determined by comparing 3 -null cells (isolated from mouse E18 renal papilla) with 3 -null cells reconstituted with the 3 -subunit, as there are no blocking antibodies to the mouse 3 -integrin subunit available. These cells have an integrin profile similar to UB cells (with little or no expression of 1 1 or 2 1 ) ( 23 ) and behave like UB cells when grown in 3D gels ( 24 ). Both the 3 -null and the reconstituted cells were unable to form tubules in CI gels, suggesting that 3 1 -integrin does not play a significant role in tubulogenesis in CI gels (data not shown). Blocking antibodies to 1 -integrin impeded tubulogenesis to an extent similar to that observed with a combination of anti- 1 - and - 2 -integrin antibodies, suggesting that only 1 1 - and 2 1 -integrins are required for IMCD tubulogenesis in 3D-CI gels ( Fig. 2 ).
Fig. 2. IMCD cell branching morphogenesis in 3D-CI gels is dependent on 1 1 -and 2 1 -integrins. A : IMCD cells were grown in 3D-CI gels in the presence of either an isotype immunoglobulin control or blocking antibodies to the different integrin subunits indicated. After 12 days, the cells were fixed and the number of branches was evaluated. Values are means ± SE of 3 experiments performed in triplicate and are expressed as the percentage of branching relative to control. *Significant difference between isotype immunoglobulin and antibody-treated cells ( P < 0.001). B : phase-contrast images of IMCD cells as described above in the presence of isotype immunoglobulin (control) and blocking antibodies to the integrins indicated.
The specific role of each of these collagen receptors on growth and apoptosis of IMCD cells cultured in 3D-CI gels was determined. Blocking the 2 -integrin subunit resulted in a significant decrease in proliferation, which was more profound when both 1 and 2 were inhibited ( Fig. 3 A ). Similar results were obtained with anti- 1 -integrin antibodies, whereas no effects were observed after incubation with other anti- -subunit antibodies (data not shown). Both anti- 1 - and - 2 -antibodies induced significant apoptosis; however, 1 had a significantly more potent effect than 2 alone ( Fig. 3 B ). The combined effect of 1 - and 2 -antibodies was equivalent to 1, and no effects were observed with antibodies directed against other -subunits (data not shown).
Fig. 3. Blocking 1 1 - and 2 1 -integrin function alters proliferation and apoptosis of IMCD cells in 3D-CI gels. A : IMCD cells were grown in 3D-CI gels in the presence of either an isotype immunoglobulin control or blocking antibodies to the different integrin subunits indicated. After 2 days, cells were incubated for a further 2 days with [ 3 H]thymidine, and [ 3 H]thymidine incorporation was determined as described in MATERIALS AND METHODS. Values are means ± SE of 3 experiments performed in quadruplicate and are expressed as the percentage of proliferation relative to control. B : IMCD cells were grown in 3D-CI gels as described in A. After 4 days, cells were fixed and apoptosis was determined by terminal transferase-mediated dUTP nick end-labeling (TUNEL) assays as described in MATERIALS AND METHODS. Values are means ± SE expressed as the percentage of apoptotic cells relative to the total number of cells in the gels. Significant difference between untreated and antibody-treated cells (* P < 0.05) or between cells treated with 1 vs. 2 antibodies ( # P < 0.05).
IMCD cells adhered to CI in an 1 1 - and 2 1 -integrin-dependent manner, and these integrins compensated for each other as cell adhesion was only inhibited in the presence of a mixture of anti- 1 - and - 2 -antibodies ( Fig. 4 A ). As expected, blocking 1 -integrin completely abolished cell adhesion, similar to the results observed with the anti- 1 - and - 2 -antibodies ( Fig. 4 A ). IMCD cells also migrate on CI. Blocking either 1 1 - or 2 1 -integrin markedly inhibited migration, and a synergistic result was also observed when both of the blocking antibodies were added to the cells ( Fig. 4 B ). Once again, the 1 -blocking antibody inhibited migration to the same level observed with anti- 1 - and - 2 -antibodies ( Fig. 4 B ). No other -integrin antibodies affected IMCD cell adhesion or migration on CI (data not shown).
Fig. 4. IMCD cell adhesion and migration on CI is mediated by 1 1 - and 1 1 -integrins. A : adhesion assays were performed on IMCD cells on CI (0.5 µg/ml) in the presence of either an isotype immunoglobulin control (control) or blocking antibodies to the integrin subunits indicated. Values are means ± SE expressed as the opitcal density reading of the cells that adhered to the plate. Shown is a representative of 3 separate assays performed in triplicate under the same conditions. Serum refers to the total number of cells that adhered. B : migration assays were performed on CI (2 µg/ml) in the presence of either an isotype immunoglobulin control or blocking antibodies to the integrin subunits indicated. Values are means ± SE of 3 separate assays performed under the same conditions and are expressed as the number of migrating cells. *Significant difference between untreated and integrin antibody-treated cells ( P < 0.01).
To verify the results obtained with the blocking antibodies against 1 - and 2 -subunits, we used double-strand siRNA to specifically reduce 1 - and 2 -integrin subunit expression in IMCD cells. Utilizing this approach, we were able to inhibit 50% of endogenous 1 1 - and 2 1 -integrin expression in IMCD cells relative to IMCD cells transfected with an irrelevant siRNA ( Fig. 5 A ). IMCD cells transfected with siRNA to either 1 -, 2 -, or both 1 - and 2 -integrin subunits showed significantly decreased branching morphogenesis in 3D-CI gels ( Fig. 5 B ) as well as decreased migration on CI ( Fig. 5 C ) compared with cells transfected with the irrelevant siRNA. These results were similar to the results obtained with blocking antibodies ( Figs. 2 B and 4 B ).
Fig. 5. Inhibition of 1 - and 2 -integrin expression reduces IMCD cell branching morphogenesis and migration on CI. A : IMCD cells were transfected with double-strand irrelevant small interfering (si)RNA oligonucleotides (IMCD+irr. siRNA) or siRNA oligonucleotides against murine 1 -integrin (IMCD+ 1 siRNA) or 2 -integrin (IMCD+ 2 siRNA ) subunits. Flow cytometry was performed 7 days later utilizing a primary antibody against either murine 1 ( left )- or 2 ( right )-subunits as described in MATERIALS AND METHODS. Integrin expression is displayed by a shift in mean fluorescent intensity compared with no primary antibody (II Abs) incubation. B : tubulogenesis assays were performed on the IMCD cells transfected with irrelevant siRNA oligonucleotides (control) or against 1 -, 2 -, or both ( 1 + 2 ) subunits. Left : representative images of IMCD cells transfected as indicated above and incubated for 12 days in 3D-CI gels. Right : graph showing percentage of branching relative to cells transfected with irrelevant siRNA oligonucleotides evaluated after 12 days in culture. Values are means ± SE of 3 experiments performed in triplicate. *Significant difference between irrelevant siRNA and integrin siRNA-transfected cells ( P < 0.001). C : migration assay was performed on CI (2 µg/ml) with IMCD cells transfected with irrelevant or integrin siRNA oligonucleotides. Values are means ± SE of 3 independent assays performed under the same conditions. *Significant difference between irrelevant siRNA- and integrin siRNA-transfected cells ( P < 0.001).
Together, these results confirm that 1 1 - and 2 1 -integrin are the principal integrins required for IMCD cell tubulogenesis, proliferation, adhesion, and migration on CI.
Expression of 2 1 -integrin reconstitutes the ability of UB cells to form tubules in 3D-CI gels. Our data suggest that differentiation of IMCD structures from the undifferentiated UB is associated with increased expression of the collagen receptors, 1 1 - and 2 1 -integrin. Antibodies to 2 1 -integrin resulted in more inhibition of IMCD cell tubulogenesis (85% inhibition) in 3D-CI gels than anti- 1 -integrin antibodies (40% inhibition) ( Fig. 2 ), suggesting that 2 -integrin plays a more critical role than 1 -integrin in regulating branching morphogenesis in 3D-CI gels. To determine whether transduction of 2 1 -integrin into UB cells would facilitate their ability to undergo tubulogenesis in 3D-CI gels, we generated UB cells that expressed a full-length human 2 -integrin ( 2 -UB). A pure 2 -UB cell population was derived by sorting, via FACS, cells incubated with an antibody directed to the extracellular domain of the human 2 -integrin (data not shown). 2 -UB cells cultured in 3D-CI gels formed tubelike structures, similar to IMCD cells grown in the same matrix in the presence of antibodies to 1 1 -integrin ( Fig. 2 B vs. Fig. 6 A ). 2 -UB cells proliferated 3 times more ( Fig. 6 B ), adhered 50 times more ( Fig. 6 C ), and migrated 30 times more ( Fig. 6 D ) than UB cells when cultured within or on CI. These results strongly suggest that 2 1 -integrin is required for UB cells to undergo tubulogenesis in 3D-CI gels by mediating cell adhesion, migration, and proliferation on a CI substratum.
Fig. 6. 2 -Integrin partially rescues the branching of UB cells in 3D-CI gels. A : phase-contrast images of UB cells transfected with empty vector (UB cells) or with a human integrin 2 cDNA ( 2 -UB) grown in 3D-CI gels for 12 days as described in MATERIALS AND METHODS. B - D : proliferation ( B ), adhesion ( C ), and migration ( D ) of UB and 2 -UB cells in 3D-CI gels or CI substrata (2 µg/ml) were performed as described in Figs. 3 and 4. Values are means ± SE expressed as fold-change over UB cells set as 100% ( B ) and the optical density of cells that adhered to the plate ( C ) and the number of migrating cells ( D ). Shown is a representative experiment of 3 independent assays performed in triplicate under the same conditions. *Significant difference between UB and 2 -UB cells ( P < 0.001).
UB and IMCD cells require different ECM and integrins to undergo branching morphogenesis in 3D-CI/MG. In contrast to the results in 3D-CI culture, both UB and IMCD cells plated into 3D-CI/MG gels proliferated and formed branched 3D tubular structures ( Fig. 7 A ). We determined whether these cells utilize the same or different integrins to undergo branching morphogenesis in 3D-CI/MG gels. We previously demonstrated that UB cells require 3 1 - and the 6 -integrin subunits to undergo branching morphogenesis in 3D-CI/MG gels ( 24 ) (see also Fig. 7 B ). It appears that 1 1 -integrin plays a minor role in UB branching morphogenesis; however, there is no role for 2 1 -integrin. This is likely related to the fact that 1 1 -integrin is expressed at low levels by UB cells; however, there is almost no 2 1 expression ( Fig. 1 B ). In contrast, IMCD cells pretreated with blocking antibodies to the 1 - and 1 -integrin subunits ( Fig. 7, B and C ), but not to 2 ( Fig. 7, B and C )- or 6 -subunits ( Fig. 7 B ), had decreased branching morphogenesis in 3D-CI/MG gels. Similar results were seen with IMCD cells transfected with siRNA to 1 - or 2 -integrin subunits (data not shown). As already mentioned, the effects of inhibiting 3 1 -integrin in IMCD cells could not be tested. These results suggest that in IMCD cells, 1 1 - and 2 1 -integrin compensate for 6 - and perhaps 3 -integrin function in 3D-CI/MG gels. Furthermore, the decreased tubulogenesis of IMCD cells in 3D-CI/MG gels after inhibition of 1 1 -integrin suggests that 1 1 exerts its effects by interacting not only with CI but also with other ECM components in MG.
Fig. 7. UB and IMCD cell morphogenesis in 3D gels is dependent on different integrins. A : UB and IMCD cells were cultured in 3D-CI/MG gels, and images were recorded 12 days after culture. B, top : UB and IMCD cells were grown in 3D-CI/MG gels in the presence of either an isotype immunoglobulin control or blocking antibodies to the different integrins indicated. The total number of branches was counted and is expressed as a percentage relative to the control group. Values are means ± SE of 3 experiments performed in triplicate. Bottom : branching of 3 -integrin-null cells grown as described above is shown relative to their reconstituted control cells. Values are means ± SE of 3 experiments performed in triplicate. Significant difference between isotype control and integrin antibody-treated cells (* P < 0.05) or between UB and IMCD cells treated with the same blocking antibody ( # P < 0.05). C : phase-contrast images of IMCD cells grown in 3D-CI/MG gels for 12 days in the absence or presence of different integrin antibodies. Note the effect of blocking 1 1 on the number of branches.
To understand how different integrin subunits in UB and IMCD cells control tubulogenesis in 3D-CI/MG gels, we determined their role in cell proliferation and apoptosis. UB, IMCD, and reconstituted 3 -null cells proliferated to a similar degree in 3D-CI/MG gels. Only the anti- 1 -anitibody led to a significant decrease in cell proliferation in both UB (55%) and IMCD (35%) cells ( Fig. 8 A ). 3 -Null cells showed a 50% decrease in proliferation compared with reconstituted cells ( Fig. 8 A ), suggesting that the major proliferative signals for UB cells in 3D-CI/MG gels are mediated by 3 1 -integrin. None of the antibodies against the different -integrin subunits tested significantly inhibited IMCD cell proliferation ( Fig. 8 A ), suggesting integrin redundancy for IMCD cell growth in 3D-CI/MG gels.
Fig. 8. Blocking integrin function of UB and IMCD cells in 3D-CI/MG gels affects growth and apoptosis. A : UB or IMCD cells were grown in 3D-CI/MG gels in the presence of either an isotype immunoglobulin control or blocking antibodies to the different integrin subunits indicated. 3 -Null cells and their reconstituted control cells were grown under similar conditions. Proliferation was evaluated by [ 3 H]thymidine incorporation as described in MATERIALS AND METHODS. Values are means ± SE of quadruplicate samples. Shown is a representative experiment of 3 independent assays performed under the same conditions. B : UB or IMCD cells were grown on 3D-CI/MG gels, and after 4 days, they were fixed and apoptosis was determined utilizing a TUNEL assay as described in MATERIALS AND METHODS. Values are means ± SE of 3 experiments performed in triplicate and are expressed as the percentage of apoptotic cells relative to the total number of cells in the gels. Significant difference between isotype and integrin antibody-treated cells (* P < 0.001) or between UB and IMCD cells treated with the same blocking antibody ( # P < 0.001).
Antibodies directed against 6 -, 1 -, and 4 -integrin subunits led to significant apoptosis in UB cells ( Fig. 8 B ). Similarly, 3 -null cells showed a fivefold increase in cell apoptosis compared with their reconstituted counterparts. In IMCD cells, the only two anti-integrin antibodies able to increase apoptosis more than in control cells were those directed against 1 - and 1 -subunits, suggesting that, under these experimental conditions, 1 1 -integrin is the primary mediator of anti-apoptotic signals for IMCD cells.
Taken together, these results suggest that the major integrin that mediates cell proliferation is 3 1 for UB cell tubulogenesis, whereas anti-apoptotic properties are exerted by 3 1 - and 6 4 -integrins. In contrast, IMCD cell proliferation does not appear to be dependent on any specific -integrin subunit, whereas inhibiting 1 1 -integrin function induces apoptosis.
UB and IMCD cells utilize different integrins for migration. As cell migration is dependent on the ability of cells to adhere to a substrate, we investigated the role of different integrin subunits in mediating adhesion and migration of UB and IMCD cells on the different components of 3D-CI/MG gels. UB cells neither adhered nor migrated on CI (data not shown), corroborating the finding that UB cells express few or none of the two major collagen-binding receptors ( Fig. 1 B ). We therefore restricted our comparison between UB and IMCD cells to MG alone. UB cells adhere and migrate significantly better than IMCD cells on this ECM ( Fig. 9, A and B ). Adhesion by both cell types to MG is dependent on multiple -integrin subunits as only antibodies against 1 inhibited UB or IMCD cell adhesion ( Fig. 9 C ). UB cell migration appeared to be dependent on 6 -integrins, 3 1, and, to a lesser extent, on 1 1 ( Fig. 9 D ). In contrast, IMCD cell migration was primarily dependent on 1 1 -integrin and, to a lesser extent, on the 6 -integrin subunits ( Fig. 9 D ). Due to the lack of blocking antibodies, the importance of 3 1 in IMCD cell migration on MG could not be tested. Together, these results suggest that specific integrins do play a role in UB and IMCD cell migration on MG.
Fig. 9. UB and IMCD cell adhesion and migration on MG is dependent on different integrins. Adhesion ( A ) and migration ( B ) assays were performed on UB and IMCD cells using different concentrations of MG as described in MATERIALS AND METHODS. Values are means ± SE of quadruplicate samples ( A ) expressed as optical density of the cells that adhered to the plate or means ± SE of 3 triplicates expressed as the number of cells that transmigrated per microscopic field ( B ). In B, 5 randomly selected fields/treatment were counted under a microscope and averaged. C : adhesion of UB and IMCD cells was performed on MG (2.5 µg/ml) in the presence of either an isotype immunoglobulin control or blocking antibodies to the integrin subunits indicated. Values are means ± SE of quadruplicate samples expressed as the optical density reading of the cells that adhered to the plate incubated in the presence of either isotype controls or blocking antibodies. 3 -Null cell adhesion was compared with its reconstituted counterpart. D : migration assays were performed on MG (2.5 µg/ml) under the same conditions indicated in ( C ). Values are means ± SE expressed as the number of migrating cells. Shown are representative experiments of 3 independent assays that were performed under the same conditions. Significant difference between isotype and integrin antibody-treated cells (* P < 0.05) or between UB and IMCD cells treated with the same blocking antibody ( # P < 0.05).
DISCUSSION
3D tubulogenesis assays are a well-described model system used to recapitulate the events of branching morphogenesis. We utilized this model system to examine how cell-ECM interactions modulate tubulogenesis in early progenitor UB cells and terminally differentiated IMCD cells. We demonstrated that 1 ) IMCD cells, unlike UB cells, express 1 1 - and 2 1 -integrins, the major collagen-binding receptors; 2 ) both IMCD and UB cells express 3 1 -, 6 1 -, and 6 4 -integrins, the major laminin receptors; 3 ) the expression pattern of integrins in UB and IMCD cells correlates with the in vivo expression pattern of structures from which the cells were derived; 4 ) the collagen receptors 1 1 and 2 1 mediate IMCD cell tubulogenesis in CI gels; 5 ) transfection of the 2 -integrin subunit into UB cells allows them to undergo branching morphogenesis in CI-3D gels; and 6 ) different integrin subunits regulate UB and IMCD branching morphogenesis in CI/MG gels. Taken together, these results demonstrate a switch of integrin expression from primarily laminin receptors in the early UB to both collagen and laminin receptors in the mature IMCD. The alterations in integrin expression in UB and IMCD cells modulate branching morphogenesis in 3D cell culture models, which suggests that temporal and spatial changes in integrin expression could help organize the pattern of branching morphogenesis of the developing collecting system in vivo.
The integrin expression pattern in UB and IMCD cells correlated with that observed in vivo. The developing UB expresses the laminin receptors 3 1, 6 1, and 6 4, and the IMCD expresses these integrins as well as the collagen receptors 1 1 and 2 1. The lack of collagen receptor expression in UB cells explains why these cells are unable to undergo branching morphogenesis in 3D-CI gels. Our data not only clarify the expression pattern of the laminin- and collagen-binding integrins in the developing UB and the IMCD system of the mouse but also parallel the finding described in humans ( 9, 10 ). The relatively late expression of both 1 1 and 2 1 in the collecting system and/or the ability of these collagen receptors to functionally compensate for each other may explain the benign renal phenotypes seen in the 1 - and 2 -integrin-null mice. Although neither of these mice types has been studied in detail, there is no evidence of severe branching morphogenesis abnormalities in the developing UB, and only subtle renal abnormalities are present (Zent R and Pozzi A, unpublished observations). These changes may be similar to the minor differentiation defects found in breast tissue in the 2 -null mice ( 2 ).
IMCD cell branching morphogenesis in CI gels was only dependent on the collagen receptors 1 1 and 2 1, with 2 1 playing the predominant role. Inhibiting or downregulating the expression of either of these receptors inhibited migration and induced apoptosis. The increased effect of decreasing 2 -function on blocking tubulogenesis was likely related to its added effect of inhibiting proliferation. The results with decreased 2 -integrin function are similar to those found in MDCK cells grown in CI gels, where 2 -integrin was shown to be critical for branching morphogenesis ( 17 ). The role of 1 1 -integrin in MDCK cell tubulogenesis in CI has never been fully explored; however, it is likely that, like IMCD cells, both 1 1 - and 2 1 -integrins play a role. The branching studies in CI alone are in contrast to those performed in CI/MG, where 2 plays no role in tubulogenesis. This raises the concern of the physiological relevance of model systems that only utilize CI, which is not a constituent of epithelial basement membranes.
Transduction of 2 -integrin into UB cells confirmed the importance of 2 1 in renal cell branching morphogenesis in CI gels. The 2 -UB cell tubulogenesis was similar to IMCD cells pretreated with 1 -blocking antibodies or transfected with siRNA against the 1 -integrin subunit, and these cells were able to adhere, proliferate, and migrate on a CI substrate. This result demonstrates that the morphological changes that occur when UB cells become terminally differentiated IMCD cells are, at least in part, due to expression of 2 1 -integrin. Although it is unclear what determines the spatial and temporal alterations of integrin expression in UB development, it may be related to expression of collagens and laminins that are ligands for 1 1 - and 2 1 -integrins in tubular basement membranes ( 12 ). The physiological relevance of the temporal changes in integrin expression is also unknown. It is interesting to speculate that the increased expression of collagen receptors may render IMCD cells less migratory on basement membranes rich in collagens and laminins. This increase in the adhesive strength of cells to the basement membrane may result in increased epithelial cell polarization and the formation of tight junctions necessary for tubules to withstand increased luminal pressure and transport fluids and solutes in a unidirectional manner.
In CI gels, IMCD cell tubulogenesis is predominantly dependent on the collagen receptor 2 1 -integrin, and, to a less extent, 1 1. In contrast, the inhibition of 1 1 -integrin function on IMCD cells in CI was similar to that seen in CI/MG gels. These differences are likely related to the fact that 1 1 -integrin is a better receptor than 2 1 for laminin-1 and collagen IV, which are the primary constituents of MG. 1 1 -Integrin preferentially binds to collagen IV and to a lesser extent to collagen I, whereas 2 1 is an excellent receptor for fibrilla-forming type I collagen gels and a relatively poor receptor for type IV collagen ( 8, 22 ). In addition, although both 1 1 - and 2 1 -integrins can also be laminin-1 receptors ( 3 ), IMCD cell migration and proliferation on laminin-1 are primarily dependent on 1 1 -integrin, to a lesser extent on 3 1 - and 6 -integrin, but not on 2 1 (Zent R, unpublished observations). In contrast to IMCD cells, UB cells do not migrate or proliferate on collagen IV, and migration and proliferation on laminin-1 are predominantly 3 1 - and 6 -integrin dependent (Zent R, unpublished observations). Therefore, UB and IMCD cells use distinct integrins for their interactions with MG and laminin-1 specifically. Interestingly, unlike UB and IMCD cells, MDCK cells are unable to form tubules in CI/MG gels ( 20 ). This difference between these cell types may be related to disparities in integrin expression.
Little is known about which integrins mediate cellular functions, such as cell proliferation, apoptosis, adhesion, and migration in the context of tubule formation. We demonstrated that inhibiting many specific integrin-ligand interactions induced cell apoptosis, whereas cellular proliferation was affected in a much less restricted manner in CI/MG gels. Blocking antibodies against 6, 1, and 4 markedly induced apoptosis in UB cells, 3 -null cells underwent apoptosis, and inhibition of 1 function decreased tubulogenesis of IMCD cells in CI/MG gels by inducing apoptosis. In contrast, only inhibiting 3 1 -integrin appeared to affect proliferation of UB cells. Together, these results suggest that inhibiting a specific integrin function of cells grown in 3D-CI/MG gels produces a phenotype of decreased branching morphogenesis predominantly by inducing apoptosis rather than by decreasing proliferation. It is only when all integrin-dependent signals are blocked, as seen with the 1 antibody, that cell proliferation is affected to a significant extent.
Interestingly, the results of cell migration on MG correlate with the phenotype observed in 3D tubulogenesis. This suggests that inhibiting migration may be a cue for the induction of apoptosis in the tubulogenesis assays. The mechanism for this is not known; however, it has been postulated that cells in 3D tubular structures require three different plasma membrane surfaces to survive: a free apical surface that borders the lumen, a lateral surface that adheres to neighboring cells, and a basal surface that adheres to the ECM ( 14 ). As tubules grow, cells that lack one of these surfaces are forced to migrate into a position where they can attain these three surface types, and if the cells are unable to achieve this architecture, they undergo apoptosis. Migration of cells in 3D gels is dependent on integrins; therefore, it is possible that inhibiting this integrin function accounts for increased apoptosis and decreased tubulogenesis.
In conclusion, we have characterized UB and IMCD cell branching morphogenesis with respect to integrin expression and demonstrated that they are representative of the developing UB and IMCD in vivo. These cell systems provide an alternative murine model to MDCK cells with which to study kidney tubule formation. With the generation of many genetically mutated mice, as well as the availability of numerous reagents for mouse cells, we propose that the UB and IMCD cells are excellent models for the study of branching morphogenesis that will be useful adjuncts to determine the molecular mechanisms of different stages of UB development.
GRANTS
This work was supported by Veterans Affairs Advanced Career Development and Merit Awards, an American Heart Association Grant in Aid, an American Cancer Society Institutional Review Grant, and a Clinician Scientist award from the National Kidney Foundation (R. Zent); by National Institutes of Health Grants NCI RO1-CA-94849-01 (A. Pozzi) and RO1-DK-51265 (to R. C. Harris); and by Vanderbilt George M. O'Brien Center for Kidney and Urologic Research Grant DK-39261 (R. C. Harris, A. Pozzi, and R. Zent).
ACKNOWLEDGMENTS
We thank Catherine Allen at the Veterans Affairs flow cytometry core and Ellen Donnert for technical help.
【参考文献】
Cantley LG, Barros EJ, Gandhi M, Rauchman M, and Nigam SK. Regulation of mitogenesis, motogenesis, and tubulogenesis by hepatocyte growth factor in renal collecting duct cells. Am J Physiol Renal Fluid Electrolyte Physiol 267: F271-F280, 1994.
Chen J, Diacovo TG, Grenache DG, Santoro SA, and Zutter MM. The 2 integrin subunit-deficient mouse: a multifaceted phenotype including defects of branching morphogenesis and hemostasis. Am J Pathol 161: 337-344, 2002.
Colognato H and Yurchenco PD. Form and function: the laminin family of heterotrimers. Dev Dyn 218: 213-234, 2000. <a href="/cgi/external_ref?access_num=10.1002/(SICI)1097-0177(200006)218:2
De Arcangelis A, Mark M, Kreidberg J, Sorokin L, and Georges-Labouesse E. Synergistic activities of 3 and 6 integrins are required during apical ectodermal ridge formation and organogenesis in the mouse. Development 126: 3957-3968, 1999.
Debnath J, Mills KR, Collins NL, Reginato MJ, Muthuswamy SK, and Brugge JS. The role of apoptosis in creating and maintaining luminal space within normal and oncogene-expressing mammary acini. Cell 111: 29-40, 2002.
Georges-Labouesse E, Messaddeq N, Yehia G, Cadalbert L, Dierich A, and Le Meur M. Absence of integrin 6 leads to epidermolysis bullosa and neonatal death in mice. Nat Genet 13: 370-373, 1996.
Hynes R. Integrins: bidirectional, allosteric signaling machines. Cell 110: 673-687, 2002.
Kern A, Eble J, Golbik R, and Kuhn K. Interaction of type IV collagen with the isolated integrins 1 1 and 2 1. Eur J Biochem 215: 151-159, 1993.
Korhonen M, Ylanne J, Laitinen L, and Virtanen I. The 1 - 6 subunits of integrins are characteristically expressed in distinct segments of developing and adult human nephron. J Cell Biol 111: 1245-1254, 1990.
Korhonen M, Ylanne J, Laitinen L, and Virtanen I. Distribution of 1 and 3 integrins in human fetal and adult kidney. Lab Invest 62: 616-625, 1990.
Kreidberg JA, Donovan MJ, Goldstein SL, Rennke H, Shepherd K, Jones RC, and Jaenisch R. 3 1 Integrin has a crucial role in kidney and lung organogenesis. Development 122: 3537-3547, 1996.
Miner JH. Renal basement membrane components. Kidney Int 56: 2016-2024, 1999.
Montesano R, Schaller G, and Orci L. Induction of epithelial tubular morphogenesis in vitro by fibroblast-derived soluble factors. Cell 66: 697-711, 1991.
O'Brien LE, Zegers MM, and Mostov KE. Opinion: building epithelial architecture: insights from three-dimensional culture models. Nat Rev Mol Cell Biol 3: 531-537, 2002.
Pohl M, Stuart RO, Sakurai H, and Nigam SK. Branching morphogenesis during kidney development. Annu Rev Physiol 62: 595-620, 2000.
Riikonen T, Koivisto L, Vihinen P, and Heino J. Transforming growth factor- regulates collagen gel contraction by increasing 2 1 integrin expression in osteogenic cells. J Biol Chem 270: 376-382, 1995.
Saelman EU, Keely PJ, and Santoro SA. Loss of MDCK cell 2 1 integrin expression results in reduced cyst formation, failure of hepatocyte growth factor/scatter factor-induced branching morphogenesis, and increased apoptosis. J Cell Sci 108: 3531-3540, 1995.
Sakurai H, Barros EJ, Tsukamoto T, Barasch J, and Nigam SK. An in vitro tubulogenesis system using cell lines derived from the embryonic kidney shows dependence on multiple soluble growth factors. Proc Natl Acad Sci USA 94: 6279-6284, 1997.
Sakurai H and Nigam SK. Transforming growth factor- selectively inhibits branching morphogenesis but not tubulogenesis. Am J Physiol Renal Physiol 272: F139-F146, 1997.
Santos OF and Nigam SK. HGF-induced tubulogenesis and branching of epithelial cells is modulated by extracellular matrix and TGF-. Dev Biol 160: 293-302, 1993.
Schwartz MA and Ginsberg MH. Networks and crosstalk: integrin signalling spreads. Nat Cell Biol 4: E65-E68, 2002.
Tulla M, Pentikainen OT, Viitasalo T, Kapyla J, Impola U, Nykvist P, Nissinen L, Johnson MS, and Heino J. Selective binding of collagen subtypes by integrin 1I, 2I, and 10I domains. J Biol Chem 276: 48206-48212, 2001.
Wang Z, Symons JM, Goldstein SL, McDonald A, Miner JH, and Kreidberg JA. 3 1 Integrin regulates epithelial cytoskeletal organization. J Cell Sci 112: 2925-2935, 1999.
Zent R, Bush KT, Pohl ML, Quaranta V, Koshikawa N, Wang Z, Kreidberg JA, Sakurai H, Stuart RO, and Nigam SK. Involvement of laminin binding integrins and laminin-5 in branching morphogenesis of the ureteric bud during kidney development. Dev Biol 238: 289-302, 2001.
作者单位：1 Division of Nephrology, Department of Medicine, 2 Department of Pathology, 7 Department of Cancer Biology, Vanderbilt University Medical Center, and 6 Veterans Affairs Hospital, Nashville, Tennessee 37232; 3 Departments of Medicine and Pediatrics, University of California, La Jolla, California 920