The chemokine KC regulates HGF-stimulated epithelial cell morphogenesis

【摘要】  Hepatocyte growth factor (HGF) induces migration, proliferation, and branching in renal epithelial cells from the inner medullary collecting duct (mIMCD-3 cells). Microarray analysis after HGF stimulation of these cells revealed upregulation of the chemokine KC. We found that both the message and protein levels of KC are increased after HGF treatment and that mIMCD-3 cells express the KC receptor CXCR2. Treatment with KC results in stimulation of mIMCD-3 cell proliferation but has no effect on basal rates of cell migration or branching morphogenesis. In contrast to its known stimulatory effect on neutrophil migration, KC markedly inhibits HGF-mediated cell migration and branching morphogenesis, resulting in shorter tubules with fewer branch points. Examination of the mechanism of this effect reveals that KC does not alter phosphorylation of the c-met receptor or the initial activation of the MAPK or phosphoinositide 3-kinase (PI 3-K) signaling pathways. However, sustained activation of the PI 3-K pathway by HGF was inhibited by treatment with KC, and mimicking this effect by treatment with LY-294002 2 h after HGF stimulation reproduced the inhibition of HGF-stimulated branching morphogenesis. These data demonstrate that HGF-mediated KC production can act in an autocrine fashion to downregulate excessive branching and migration of renal epithelial cells in response to HGF, while still supporting cell proliferation. These characteristics may play a role in modulating the response to HGF during developmental tubule formation and/or during the repair of the tubular architecture following injury.

【关键词】  hepatocyte growth factor kidney Akt phosphoinositide kinase

HEPATOCYTE GROWTH FACTOR (HGF) and its receptor c-met are expressed in tubular epithelia such as the kidney, pancreas, and breast, where they are believed to play a role in the morphological events of development and repair following injury. HGF regulates many morphogenic changes of epithelial cells including elongation ( 30 ) and migration ( 10 ). In epithelial cells established from murine collecting duct (mIMCD-3 cells), HGF specifically induces scattering, branching morphogenesis, and migration ( 6 ). The ability of HGF to stimulate branching has been suggested to be important in normal ureteric bud arborization during kidney development ( 44 ), whereas its effects on migration and development of a tubular architecture have been proposed to explain its positive effects on repair of the injured renal tubule ( 27 ). However, the mechanism whereby these morphological effects are regulated is not fully understood.

Binding of HGF to c-met results in phosphorylation of the cytoplasmic domain of c-met and activation of numerous signaling pathways including MAP kinase and phosphoinositide 3-kinase (PI 3-K). After stimulation of these pathways, two classes of signals are activated: local effectors that regulate the machinery of cell morphogenesis [such as the small GTP binding proteins rac and rho ( 40 ) and the focal adhesion proteins FAK and paxillin ( 24 )], as well as downstream signaling intermediates such as STAT3, ERK, and Akt that regulate gene expression ( 33 ).

To determine which genes HGF is regulating during the process of epithelial morphogenesis, we performed a microarray analysis of mIMCD-3 renal epithelial cells following stimulation with HGF. We found that several genes are upregulated by activation of c-met, one of which is the chemokine KC. This protein, also known as cytokine-induced neutrophil chemoattractant (CINC), growth-related protein (Gro- ), and CXCL1, is a secreted factor that acts as a neutrophil chemoattractant ( 53, 54 ). Deletion of KC's receptor, CXCR2 ( 22, 45 ), results in mice deficient in inflammatory responses ( 4, 49 ) and neutrophils that are unable to migrate across epithelial cells ( 11 ). In addition, repair of wounded epithelium in these knockout mice is delayed compared with control mice ( 8, 26 ). In the present study, we report that HGF stimulates the secretion of KC by renal epithelial cells and that KC can act directly on these cells in an autocrine fashion to inhibit HGF-stimulated morphogenesis via downregulation of sustained PI 3-K/Akt activation.

EXPERIMENTAL PROCEDURES

Materials. The mIMCD-3 cells are derived from the mouse inner medullary collecting duct ( 38 ) and were maintained in DMEM/F-12 with 10% fetal calf serum and used between passages 17 and 30. Anti-ERK 1 and 2, anti-phospho-ERK 1 and 2, and anti-Akt antibodies were purchased from Cell Signaling Technologies (Beverly, MA). The anti-phospho-Akt and anti-CXCR2 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-KC antibody was purchased from R & D (Minneapolis, MN). The anti Gab-1 antibody and type I rat tail collagen were purchased from Upstate Biotechnology (Lake Placid, NY). HGF, protein A Sepharose, and tri-reagent were purchased from Sigma (St. Louis, MO). The LY-294002 compound was purchased from Calbiochem (Santa Rosa, CA). PVDF membranes were from Millipore (Immobilon-P, Millipore, Bedford, MA). GeneScreen membranes were purchased from New England Nuclear Research Products (Boston, MA).

Microarray. RNA for microarray analysis was prepared from control mIMCD-3 cells and cells stimulated with HGF for 8 or 24 h via a two-step purification protocol: tri-reagent was used for initial purification, followed by final purification using the RNeasy Midi Kit as per the manufacturer's instructions (Qiagen, Valencia, CA). Purity was followed by visual analysis on formaldehyde-agarose gels and by verification that A 260/280 ratios were greater than 1.8. The resulting purified RNA was used for hybridization in an Affymetrix analysis and in a glass slide microarray analysis of gene expression, both carried out by the Keck Foundation Biotechnology Resource Laboratory at Yale University. The glass slides were prepared by the Keck Foundation Biotechnology Resource Laboratory and contain the mouse NIA15K gene set consisting of 15,000 unique cDNA clones ( 50% from novel genes) obtained from preembryonic, embryonic, and newborn mice with an average size of 1.5 kb ( 47 ). The murine U74 A chips used in the Affymetrix analysis contain sequences from 6,000 est's and 6,000 known genes (Affymetrix, Santa Clara, CA).

Northern analysis. RNA was prepared as above using tri-reagent. Two micrograms of total RNA from each time point were separated on a formaldehyde-agarose gel and transferred to a GeneScreen membrane. To prepare the probes, single-stranded cDNA was prepared from mIMCD-3 total RNA using SuperScript II reverse transcriptase as recommended by the manufacturer (Invitrogen, Carlsbad, CA). With this cDNA as the substrate, KC and GAPDH sequences were synthesized via PCR for use as probes. Primers for GAPDH were 5'-GTGAAGGTCGGTGTGAACGG and 3'-GGCCCCTCCTGTTATTATGG; product size: 1,127 bp. Primers for KC were 5'-GATTCACCTCAAGAACATCC and 3'-CAAGACATACAAACACAGCC; product size: 551 bp. Twenty-five nanograms of each PCR product were labeled with [ - 32 P]dCTP using Exo(-) Klenow (Prime-It II, Stratagene, Cedar Creek, TX). Probes were incubated with membranes in formamide hybridization buffer; membranes were then washed and subjected to autoradiography. Membranes were stripped using an aqueous SDS solution for subsequent reprobing. GAPDH expression was used to normalize loading of total RNA.

RT-PCR. To verify expression of CXCR2 in mIMCD-3 cells, a single-step RT-PCR was performed according to the manufacturer's instructions (Invitrogen). Primers for CXCR2 were 5'-GGTCAAGTTTGTGTGCATAG and 3'-CAGTCTCTGGTAATGATGCC; predicted product size: 674 bp. Primers for -2 microglobulin were 5'-CTCGCGCTACTCTCTCTTTCTGG and 3'-GCTTACATGTCTCGATCCCAC; product size: 337 bp.

Western analysis. Cells were serum-starved for 24 h before stimulation with the appropriate factors for the indicated times. Cells were then lysed in ice-cold RIPA buffer (20 mM Tris, pH 7.4; 160 mM NaCl; 1 mM EDTA; 1 mM EGTA; 1 mM DTT; 1% Triton X-100; 1% sodium deoxycholate) with the following protease and phosphatase inhibitors: 0.5 µg/ml leupeptin, 0.5 µg/ml pepstatin, 10 mM NaF, 1 mM PMSF, and 0.4 mM NaVO 4. After 15 min at 4°C, lysates were centrifuged at 14,000 g for 15 min, and the supernatants were stored at -80°C. Protein concentrations were determined using the Bradford assay, and 40 µg of whole cell lysates were used unless otherwise noted. Samples were boiled 5 min in -ME-containing sample buffer and separated via SDS-PAGE. Proteins were transfered to PVDF membranes, blocked with 5% milk/TBST (100 mM Tris, pH 7.5; 0.9% NaCl; 0.1% Tween 20), and hybridized with the appropriate antibody. Visualization of proteins was accomplished with horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (Pharmacia, Piscataway, NJ).

Cell migration. Migration assays were performed using a modified Boyden chamber assay as previously described ( 7 ). Briefly, a 48-well bottom plate (Neuro Probe, Cabin John, MD) was filled with DMEM/F-12 medium containing the appropriate factors. The well was over-laid with a rat-tail collagen type I (Collaborative Biomedical)-coated polycarbonate filter with 8-µm pores (Nucleopore). The top compartment was connected, and 1.5 x 10 4 cells suspended in DMEM/F-12 containing the appropriate factors were added to the top of each well. After 4 h of incubation at 37°C, the filters were removed, stained with Diff-Quik (Baxter Healthcare), and cells remaining on the top of the membrane were mechanically scraped off. Cells that had passed through the pores were counted to determine the number of cells per millimeter squared of membrane. Each well represents an n of 1, and each experiment was repeated at least three separate times. P values were determined using an unpaired Student's t -test.

Collagen assay of branching morphogenesis. mIMCD-3 cells were harvested by trypsinization, washed in 1% serum/DMEMF-12 medium, and counted. Collagen solution was prepared as follows: type I rat tail collagen was mixed with 10% 1 M HEPES buffer, pH 7.3, and 10% 10 x MEM at 4°C. Cells were added to a final concentration of 1.5 x 10 5 cells/ml, and 100 µl of the mixture were aliquotted into each well of a 96-well plate. The solution was allowed to polymerize for 15 min at 37°C, and 100 µl of DMEMF-12 containing 1% serum and the appropriate growth factor were added to the well. The plates were incubated for 6-24 h at which time 50 cells/well were randomly counted to determine the total number of processes. The number of processes was divided by the total number of cells counted to obtain the average number of processes per cell, with each well representing an n of 1. In the PI 3-K inhibition experiments, the supernatant containing the HGF was removed after 2 h and replaced with the same medium containing HGF and the inhibitor LY-294002 at the indicated concentration. The LY-294002 was dissolved in DMSO to give a stock concentration of 50 mM. In the long-term assay, cells were added to a final concentration of 1.5 x 10 5 cells/ml into a mixture of 70% collagen (prepared as described above) and 30% Matrigel (Collaborative Biochemical, Bedford, MA) prepared as an 80:20 mixture of Matrigel and 150 mM NaCl. After polymerization, DMEM-F-12 medium containing 3% serum and the appropriate growth factors was added to the well, and the cells were maintained in culture for 8 days with a medium change every 3 days.

Analysis of PI 3-K activity. Lysates of mIMCD-3 cells stimulated with 40 ng/ml HGF were precleared for 30 min at 4°C with 30 µl of a 50% protein A-Sepharose slurry in phosphate-buffered saline/0.02% NaN 3 followed by overnight immunoprecipitation at 4°C with 2 µg of anti-Gab1 antibody. The antibody was then precipitated with 40 µl protein A-Sepharose and washed three times with lysis buffer. Immunoprecipitates were collected by centrifugation and washed twice with PBS containing 1% Nonidet P-40 and 100 µM Na 3 VO 4, twice with 100 mM Tris, 500 mM LiCl 2, 100 µM Na 3 VO 4, pH 7.5, and twice with 10 mM Tris, 100 mM NaCl, 1 mM EDTA, and 100 µM Na 3 VO 4, pH 7.5. The pellets were then resuspended in 50 µl of the final wash buffer containing 12 mM MgCl 2 and 20 µg of phosphatidylinositol (Avanti Polar Lipids). To start the PI 3-K reaction, 10 µl of 40 µM ATP containing 30 µCi of [ 32 P]ATP were added to each pellet, which was then incubated at room temperature for 10 min. The reaction was stopped by the addition of 20 µl of 8 M HCl, and the lipids were extracted using 160 µl of CHCl 3 :MeOH (1:1). The phases were separated by centrifugation, and 50 µl of the lower organic phase were spotted onto a glass-backed silicon thin-layer chromatography plate. The lipids were resolved by thin-layer chromatography in MeOHCHCl 3 -H 2 O-NH 4 OH (60:47:11.3:2) and visualized using a Storm PhosphorImager.

Cell proliferation assays. Cells were plated at a density of 5 x 10 4 /well in a 96-well plate in 100 µl of serum-free DMEM/F-12. The supernatant was removed the next day and replaced with medium containing either HGF, KC, or both at the indicated concentrations. To inhibit the CXCR2 receptor, the specific inhibitor SB-225002 (Calbiochem) was added at a concentration of 100 nM in the indicated experiments. This compound has a greater than 150-fold specificity for CXCR2 compared with five other similar receptors ( 56 ). Twenty-four hours later, 10 µl of the WST-1 proliferation reagent (Roche, Indianapolis, IN) were added to each well. The WST-1 reagent is a tetrazolium salt that is cleaved by mitochondrial dehydrogenase to formazan. The amount of formazan directly correlates with the number of metabolically active cells in the culture ( 14, 15 ). After 2 h, this cleavage product was read at 415 nm with the reference wavelength set at 655 nm. This absorbance, corrected for absorbance of the medium, is shown as the proliferation index.

RESULTS

HGF stimulates expression of KC. A glass slide-based microarray analysis of mIMCD-3 cells carried out after 24 h of stimulation with 40 ng/ml HGF showed specific regulation of 28 messages. One of these sequences corresponded to KC, a 6- to 8-kDa ( 58, 59 ) CXC chemokine that has traditionally been studied in the context of immune regulation ( 1 ). We also used Affymetrix technology to analyze this process and again observed upregulation of KC, this time at 8 h. A Northern blot analysis confirmed these observations ( Fig. 1 A ); HGF stimulates a 2.3-fold increase in the KC message after 8 h (normalized to GAPDH), which then decreases to baseline levels by 48 h after stimulation. Because KC is a secreted protein, we analyzed supernatants from HGF-treated cells to determine KC expression. Secretion of KC is apparent after 8 h of HGF stimulation and increases with time, even after the message has decreased to a baseline level ( Fig. 1 B ). In addition, control cells show no detectable secretion of KC until the 48-h time point, despite the presence of basal KC message expression.

Fig. 1. Hepatocyte growth factor (HGF) stimulates KC production and secretion. A : HGF-stimulated inner medullary collecting duct (mIMCD-3) cells were subjected to Northern blot analysis at the indicated times, and levels of KC and GAPDH mRNA were analyzed. B : supernatant from HGF-treated mIMCD-3 cells was collected at the indicated time points and immunoblotted with an antibody to KC. Purified recombinant KC was run in lane 14 as a positive control.

To determine whether mIMCD-3 cells expressed the CXCR2 receptor for KC, we used PCR and Western blot analysis. Reverse-transcriptase PCR of mRNA from mIMCD-3 cells with primers designed to amplify murine CXCR2 demonstrated a single band of the appropriate size that was present only following the addition of reverse transcriptase ( Fig. 2 A ). In addition, Western blot analysis revealed the 45-kDa protein product of the CXCR2 gene in lysates from these cells ( Fig. 2 B ).

Fig. 2. Expression of CXCR2 in mIMCD-3 cells. A : RT-PCR for CXCR2. Primers corresponding to sequences from -2 microglobulin ( 2m) and CXCR2 were used in an RT-PCR analysis of mIMCD-3 message expression. B : protein expression of CXCR2. Whole cell lysate from mIMCD-3 cells was analyzed for CXCR2 expression by Western blotting.

KC inhibits HGF-stimulated epithelial morphogenesis. The expression of both the KC ligand and its receptor by mIMCD-3 cells suggested the possibility of an autocrine effect of KC on these cells. Because KC is a known chemoattractant for leukocytes, we examined the effect of KC on mIMCD-3 cell migration and branching morphogenesis. In contrast to its promigratory effects on leukocytes, addition of KC to renal epithelial cells had no significant effect on basal rates of branching process formation ( Fig. 3 A, quantitated in 3 B ) or cell migration ( Fig. 3 D ), although there was a tendency for both migration and branching morphogenesis to decrease slightly with KC treatment. However, in cells treated with HGF, simultaneous addition of KC caused a dose-dependent decrease in both HGF-stimulated branching morphogenesis ( Fig. 3 A, quantitated in 3 B ) and cell migration ( Fig. 3 D ). At a dose of 50 ng/ml, KC inhibited HGF-stimulated cell migration by 50% and branching process formation by 60%. Preincubation with KC for 30 min before the addition of HGF resulted in a similar degree of inhibition of HGF-stimulated morphogenesis (data not shown). Examination of a time course of the effect of KC revealed that KC inhibited HGF-stimulated branching morphogenesis within 6 h ( Fig. 3 C ).

Fig. 3. Influence of KC on epithelial cell morphogenesis. A : single-cell branching morphogenesis assay. mIMCD-3 cells suspended in a collagen matrix were treated ±40 ng/ml HGF and ±50 ng/ml KC and representative cells were photographed after 24 h. B : process formation by cells treated as in A was quantitated as described in EXPERIMENTAL PROCEDURES; n = 3 wells. * P < 0.008 compared with control + 0 ng/ml KC. ** P < 0.04 compared with HGF + 0 ng/ml KC. This figure is representative of 3 separate experiments. C : time course of branching. The experiment was performed as in A and scored at the indicated times. D : cell migration assay. A modified Boyden chamber analysis was used to measure motility of mIMCD-3 cells exposed to 40 ng/ml HGF ± 50 ng/ml KC. * P < 0.001 compared with control. ** P < 0.001 vs. HGF alone. This figure is representative of 3 separate experiments.

To determine the effect of inhibition of initial branching process formation on long-term tubule formation, we examined the in vitro HGF-stimulated formation of tubules in the presence and absence of added KC. In the absence of HGF, small multicellular cystic structures were the predominate morphology seen in both control and KC-treated cells ( Fig. 4, left ). As previously described, treatment with HGF results in elongated multicellular tubular structures with multiple branch points and nascent lumens, whereas treatment with HGF and KC results in structures that have fewer branches and are consistently shorter than those seen with HGF alone ( Fig. 4, right ). Thus, similar to the effects of transforming growth factor- ( 43 ) and endostatin ( 18 ), KC can negatively regulate HGF-stimulated epithelial branching and additionally appears to regulate tubule length.

Fig. 4. Long-term assay of branching morphogenesis. mIMCD-3 cells suspended in a collagen/Matrigel mixture were cultured ±40 ng/ml HGF and ±50 ng/ml KC for 8 days. Representative multicellular structures are shown.

In contrast to its inhibitory effects on epithelial morphogenesis, KC treatment stimulated basal epithelial cell proliferation ( Fig. 5 A ). This is in keeping with previous results in alveolar epithelial cells ( 9 ) and gastric epithelial cells ( 46 ). Furthermore, addition of KC to HGF-treated cells had neither an additive nor an inhibitory effect on HGF-mediated proliferation. Addition of the CXCR2 inhibitor SB-225002 decreased KC-mediated proliferation to control levels ( Fig. 5 B ) but had no effect on HGF-stimulated proliferation ( Fig. 5 C ), demonstrating that this proliferative effect is mediated by the KC receptor. In cells plated in a collagen matrix, KC only modestly induced proliferation ( Fig. 5 D ). Of note, the experiments shown in Fig. 5 A were performed at a separate time and with separate aliquots of mIMCD-3 cells than those in Fig. 5, B - D, presumably explaining the difference in the basal rates of proliferation. These results suggest that the addition of KC to mIMCD-3 cells results in cross talk between the signaling pathways downstream of the KC receptor and the c-met receptor that specifically interrupts morphogenic signaling events without disrupting other signaling pathways.

Fig. 5. Epithelial proliferation mediated by KC. A : cell proliferation was measured 24 h after culture of ±40 ng/ml HGF and ±50 ng/ml KC as described in EXPERIMENTAL PROCEDURES. This figure is representative of 3 separate experiments. B : the CXCR2 inhibitor SB-225002 was added at a concentration of 100 nM to cells treated with 50 ng/ml KC in the proliferation assay described above. * P < 0.007 compared with 50 ng/ml KC. This figure is representative of 3 separate experiments. C : SB-225002 (100 nM) was added to cells treated ±HGF (40 ng/ml) and the rate of proliferation was determined as above. D : cells were suspended in collagen matrix (described in EXPERIMENTAL PROCEDURES ) and grown ±50 ng/ml KC for 24 h. Cell proliferation was then assayed as described above. * P < 0.04 compared with control.

KC treatment does not prevent HGF-mediated c-met receptor activation. Binding of HGF to c-met results in receptor dimerization and autophosphorylation, initiating the receptor's signaling cascade. Pretreatment with KC does not diminish the HGF-stimulated autophosphorylation of c-met ( Fig. 6 A ), demonstrating that KC is not acting to prevent HGF from binding to its receptor. After receptor phosphorylation, SH2 domain-containing proteins such as Grb2, the p85 subunit of the PI 3-K, and phospholipase C are rapidly recruited to the receptor (or to the associated docking protein Gab1) and activated. Because activation of both Erk (downstream of Grb2) and PI 3-K is necessary for HGF-mediated cell migration and branching morphogenesis ( 7, 19 ), we analyzed the effect of KC addition on the activation of these signaling intermediates. Cells were stimulated for 10 min with HGF in the presence or absence of KC, and cell lysates were immunoblotted with an antibody to the activated form of ERK. These experiments revealed that KC addition did not inhibit HGF-stimulated phosphorylation of either ERK1 or ERK2 at this early time point ( Fig. 6 B ). Similarly, treatment with KC did not alter initial PI 3-K activation 10 min after stimulation with HGF, as judged by either activation of the downstream kinase Akt ( Fig. 6 C ) or PI 3-K enzyme activation ( Fig. 7 A ).

Fig. 6. Analysis of HGF signaling intermediates. A : cells were treated ±40 ng/ml HGF and ±50 ng/ml KC for 10 min followed by immunoprecipitation of c-met and immunoblotting with -phosphotyrosine (pTyr). B : lysates of cells treated for 10 min with HGF (40 ng/ml) and the indicated concentrations of KC were subjected to Western blot analysis to detect ERK phosphorylation (pERK), and reprobed to detect total ERK present in the lysate (tERK). C : lysates of cells treated for 10 min with HGF (40 ng/ml) and the indicated concentrations of KC were subjected to Western blot analysis with an activation state-specific anti-Akt antibody (pAkt) and reprobed to determine total Akt present in the lysate (tAkt).

Fig. 7. KC decreases long-term HGF-mediated activation of the phosphoinositide 3-kinase (PI 3-K) and Akt phosphorylation. A : immunoprecipitation of the Gab-1/PI 3-K complex was performed at the indicated times after HGF stimulation ± KC treatment. Immunoprecipitates were assayed for the in vitro activity of the PI 3-K by phosphorylation of PI to form PI 3 P. B : time course of Akt activation. Lysates of cells treated for the indicated times with 40 ng/ml HGF ± 50 ng/ml KC were analyzed for activated Akt (pAkt) and total Akt (tAkt). C : results of 3 separate experiments carried out as in B were quantitated for Akt phosphorylation, normalized to total Akt levels. * P < 0.03 compared with HGF at 180 min. ** P < 0.04 compared with HGF at 240 min.

KC inhibits sustained activation of the PI 3-K. It has been recently demonstrated that the duration of activation of signaling pathways such as MAPK and PI 3-K can determine whether stimulation of those pathways results in cell division or cell differentiation ( 32, 37 ). HGF treatment resulted in peak stimulation of ERK activation at 10-30 min, with sustained activation at 45% of peak for at least 4 h, an effect that was not inhibited by addition of KC (data not shown). We also examined the long-term activation of the PI 3-K pathway in mIMCD-3 cells. Stimulation with HGF resulted in the expected initial activation of the PI 3-K at 10 min ( Fig. 7 A ), correlating with peak activation of Akt at the same time point ( Fig. 7 B, quantitated in 7 C ). Activation of Akt decreased over the next 90 min but was then sustained at 50% of the initial level of activation for up to 4 h after HGF treatment ( Fig. 7, B and C ). In contrast, when mIMCD-3 cells were stimulated with HGF in the presence of KC, activation of the PI 3-K and Akt declined progressively and returned to baseline levels by 4 h after HGF stimulation ( Fig. 7, A - C ).

Because addition of KC decreased the sustained, but not the initial, PI 3-K activation after HGF treatment, we investigated the possibility that sustained PI 3-K activation was required for HGF-stimulated branching morphogenesis. To study this possibility, we used the synthetic compound LY-294002 that inhibits PI 3-K by binding specifically to the ATP-binding site of the catalytic subunit of PI 3-kinase ( 51 ). Addition of LY-249002 to the mIMCD-3 cells 2 h after the addition of HGF resulted in a decrease in the amount of Akt phosphorylation to control levels by 4 h after HGF treatment ( Fig. 8 A ) and caused marked inhibition of HGF-mediated branching at 24 h, mimicking the effect of KC ( Fig. 8 B ). The delay in loss of Akt activation is presumed to be due to the rate of entry of LY-294002 into the cells and the rate of dephosphorylation of PI 3,4,5 P 3 and Akt itself. A time course of branching morphogenesis revealed that addition of LY-294002 as late as 6-8 h after HGF stimulation resulted in loss of branching ( Fig. 8 C ). These results demonstrate that sustained PI 3-K activation is required for HGF-mediated cell branching.

Fig. 8. Late inhibition of PI 3-K mimics the effects of KC. A : cells were analyzed for Akt phosphorylation at the indicated time points after stimulation with 40 ng/ml HGF. In lanes 5 and 6, the LY-294002 inhibitor was added to the cells 2 h after HGF stimulation. B : branching of mIMCD-3 cells was assayed 24 h after treatment with 40 ng/ml HGF ± simultaneous addition of 50 ng/ml KC or addition of LY-294002 (10 µM) 2 h after HGF treatment ( * P < 0.05 vs. HGF alone). C : HGF-stimulated branching process formation was performed as in Fig. 7 B, with the addition (addn.) of 10 µM LY-294002 (LY) at the indicated times after stimulation with 40 ng/ml HGF. The number of processes/cell was quantitated 24 h after the addition of HGF.

DISCUSSION

In this study, we found that HGF stimulation of a renal epithelial cell line resulted in production of the chemokine KC and that this chemokine can act to influence the normal effects of HGF on epithelial morphogenesis, suggesting that KC can act in an autocrine fashion in these cells. An autocrine role for KC in endothelial cells was noted in a study by Wen and co-workers ( 55 ) in which they demonstrated that KC could upregulate its own expression. However, to date there have been no reports of an autocrine effect of KC in epithelial cells.

In addition to its established role as a neutrophil chemoattractant, KC has also been found to act as a growth factor in some cell types. Driscoll et al. ( 9 ) showed that rat KC can induce proliferation in alveolar epithelial cells and Suzuki et al. ( 46 ) found the same in gastric epithelial cells. KC was separately identified as a necessary growth factor for melanocytes and termed melanocyte growth-stimulatory activity ( 2, 39 ). Melanocyte-derived tumors engineered to produce higher levels of KC give rise to increased infiltration of host CD31+ blood vessels, leading to the proposal that KC is a positive regulator of angiogenesis ( 25 ), possibly by stimulating proliferation and/or migration of endothelial cells. The proliferative effects of KC are believed to be due to activation of the nuclear factor- B transcription factor ( 52 ) in a process that leads from KC-stimulated serine phosphorylation of the COOH terminus of CXCR2 ( 31 ) to activation of heterotrimeric G proteins ( 57 ) and p38 MAP kinase ( 52 ). Consistent with this effect on other epithelial cell types, we found that KC stimulates proliferation in renal epithelial cells ( Fig. 5 ), suggesting that this same pathway downstream of CXCR2 may be activated in response to KC.

In contrast to its ability to induce proliferation in epithelial cells, Unemori et al. ( 50 ) found that KC had no affect on the proliferation of rheumatoid synovial fibroblasts. In addition, KC induced a dose-dependent decrease in the expression of interstitial collagens but had no effect on collagen-degrading metalloproteinases or on tissue inhibitors of metalloproteinases. Thus KC appears to be capable of inducing cell type-specific regulation of cell migration (neutrophils), proliferation (epithelial cells), or matrix expression (fibroblasts).

As opposed to its direct stimulatory effects on mIMCD-3 cell proliferation, we found that KC inhibits HGF-mediated migration and branching, while having little effect on baseline morphogenesis. Examination of branching process formation in real time reveals that HGF stimulates the cells to begin process extension within 4-6 h and that these processes will retract unless contact is made with a process from an adjacent cell (data not shown). Cells that fail to make contact continue to extend and retract processes at a predictable rate such that the total number of processes/cell reaches a steady state at 12 h (see Fig. 3 C ). Based on the constant degree of inhibition of HGF-stimulated processes by KC, it appears that KC inhibits the rate of HGF-stimulated process extension. Examination of the mechanism of this effect reveals that KC does not prevent HGF-mediated phosphorylation of the c-met receptor or the initial activation of downstream signaling pathways such as the PI 3-K and MAPK. However, examination of a time course of PI 3-K activation reveals the sustained activation of the PI 3-K and Akt following HGF treatment is prevented by KC.

We previously demonstrated that single-cell branching morphogenesis correlates with the ultimate development of multicellular branching tubule formation ( 34 ), and it is well known that activation of the PI 3-K is necessary for this process. Constitutively activated PI 3-K increases tubulogenesis ( 21 ), and addition of the PI 3-K inhibitors wortmannin or LY-294002 at the time of HGF stimulation has been shown to inhibit renal epithelial cell branching ( 7 ), epithelial mammary gland cell branching ( 35 ), and ureteric bud outgrowth ( 48 ). Our finding that KC inhibited the sustained phase of PI 3-K activation rather than initial activation led us to investigate the possibility that HGF-stimulated morphogenesis was specifically dependent on sustained PI 3-K activation. Jones et al. ( 17 ) showed that the PI 3-K undergoes two peaks of activation in response to platelet-derived growth factor: an initial, rapid peak and a later, more sustained activation required for cell-cycle progression. This group has also shown that activation of the PI 3-K is responsible for a large portion of the chemotactic response by this receptor ( 41 ). In the present study, the addition of LY-294002 as late as 8 h after HGF stimulation resulted in inhibition of the sustained phase of PI 3-K activation and caused inhibition of branching similar to that seen with the addition of KC, demonstrating that the transient activation of the PI 3-K is not sufficient to mediate branching morphogenesis.

One of the hallmarks of HGF-stimulated tubulogenesis in vitro is the formation of multiple branch points ( 6 ). This branching effect has been proposed to play a role in the development of arborized tubular organs such as the kidney and lung. Indeed, HGF-induced branching morphogenesis of the renal ureteric bud can be greatly reduced by addition of a neutralizing antibody specific for HGF ( 44 ), and exogenous HGF stimulates branching morphogenesis of the fetal lung, whereas neutralization of HGF results in inhibition of branching ( 36 ). However, unregulated branching during development results in abnormalities of the kidney and eventual renal dysplasia ( 5 ). Thus negative regulators of branching morphogenesis such as endostatin, BMP2, and TIMP2 have been shown to play an important role in the normal development of these organs ( 3, 12, 18 ). The present results suggest that HGF-stimulated KC production may serve a similar function to downregulate branching during tubule development.

Another setting where it may be even more important to limit branching morphogenesis is during the repair of existing tubular structures. After renal tubular necrosis, multiple growth factors including HGF, EGF, and IGF-1 are upregulated as part of the repair process ( 13, 16, 20, 23, 28 ). These growth factors are believed to mediate cell migration, proliferation, and reformation of the intact renal epithelium. However, the propensity for HGF to stimulate metalloproteinase activation and branching morphogenesis might be expected to result in attempts by these cells to degrade the tubular basement membrane and form new branch points within the damaged tubule. Therefore, the local secretion of negative regulators of branching morphogenesis would be predicted to play an important role in this setting.

KC is also upregulated during renal injury ( 42 ) where it has been proposed to play a role in the process of neutrophil infiltration and subsequent inflammation. Indeed, inhibition of KC-mediated neutrophil influx using a neutralizing antibody to KC decreases renal injury and improves survivability ( 29 ). However, our results demonstrate that KC may play a beneficial role in the repair process as well. HGF-mediated upregulation of KC by injured tubular epithelial cells with subsequent autocrine stimulation of cell proliferation but inhibition of HGF-mediated branching may act to enhance repair while preventing the futile formation of new tubule branch points.

GRANTS

This work was supported by a National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant R01-DK-54911 to L. G. Cantley and by an NIDDK microarray pilot project grant to J. Ueland.

ACKNOWLEDGMENTS

The authors thank A. Karihaloo for helpful comments and discussion.

【参考文献】
  Baggiolini M. Chemokines and leukocyte traffic. Nature 392: 565-568, 1998.

Balentien E, Mufson BE, Shattuck RL, Derynck R, and Richmond A. Effects of MGSA/GRO on melanocyte transformation. Oncogene 6: 1115-1124, 1991.

Barasch J, Yang J, Qiao J, Tempst P, Erdjument-Bromage H, Leung W, and Oliver JA. Tissue inhibitor of metalloproteinase-2 stimulates mesenchymal growth and regulates epithelial branching during morphogenesis of the rat metanephros. J Clin Invest 103: 1299-1307, 1999.

Boisvert WA, Santiago R, Curtiss LK, and Terkeltaub RA. A leukocyte homologue of the IL-8 receptor CXCR-2 mediates the accumulation of macrophages in atherosclerotic lesions of LDL receptor-deficient mice. J Clin Invest 101: 353-363, 1998.

Cano-Gauci DF, Song HH, Yang H, McKerlie C, Choo B, Shi W, Pullano R, Piscione TD, Grisaru S, Soon S, Sedlackova L, Tanswell AK, Mak TW, Yeger H, Lockwood GA, Rosenblum ND, and Filmus J. Glypican-3-deficient mice exhibit developmental overgrowth and some of the abnormalities typical of Simpson-Golabi-Behmel syndrome. J Cell Biol 146: 255-264, 1999.

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.

Derman MP, Cunha MJ, Barros EJ, Nigam SK, and Cantley LG. HGF-mediated chemotaxis and tubulogenesis require activation of the phosphatidylinositol 3-kinase. Am J Physiol Renal Fluid Electrolyte Physiol 268: F1211-F1217, 1995.

Devalaraja RM, Nanney LB, Du J, Qian Q, Yu Y, Devalaraja MN, and Richmond A. Delayed wound healing in CXCR2 knockout mice. J Invest Dermatol 115: 234-244, 2000.

Driscoll KE, Hassenbein DG, Howard BW, Isfort RJ, Cody D, Tindal MH, Suchanek M, and Carter JM. Cloning, expression, and functional characterization of rat MIP-2: a neutrophil chemoattractant and epithelial cell mitogen. J Leukoc Biol 58: 359-364, 1995.

Gherardi E, Gray J, Stoker M, Perryman M, and Furlong R. Purification of scatter factor, a fibroblast-derived basic protein that modulates epithelial interactions and movement. Proc Natl Acad Sci USA 86: 5844-5848, 1989.

Godaly G, Hang L, Frendeus B, and Svanborg C. Transepithelial neutrophil migration is CXCR1 dependent in vitro and is defective in IL-8 receptor knockout mice. J Immunol 165: 5287-5294, 2000.

Grisaru S, Cano-Gauci D, Tee J, Filmus J, and Rosenblum ND. Glypican-3 modulates BMP- and FGF-mediated effects during renal branching morphogenesis. Dev Biol 231: 31-46, 2001.

Humes HD, Cieslinski DA, Coimbra TM, Messana JM, and Galvao C. Epidermal growth factor enhances renal tubule cell regeneration and repair and accelerates the recovery of renal function in postischemic acute renal failure. J Clin Invest 84: 1757-1761, 1989.

Ishiyama M, Shiga M, Sasamoto K, Mizoguchi M, and He P. A new sulfonated tetrazolium salt that produces a highly water-soluble formazan dye. Chem Pharm Bull (Tokyo) 41: 1118-1121, 1993.

Iwaki T, Iwaki A, Fukumaki Y, and Tateishi J. B-crystallin in C6 glioma cells supports their survival in elevated extracellular K + : the implication of a protective role of B-crystallin accumulation in reactive glia. Brain Res 673: 47-52, 1995.

Joannidis M, Spokes K, Nakamura T, Faletto D, and Cantley LG. Regional expression of hepatocyte growth factor/c-met in experimental renal hypertrophy and hyperplasia. Am J Physiol Renal Fluid Electrolyte Physiol 267: F231-F236, 1994.

Jones SM, Klinghoffer R, Prestwich GD, Toker A, and Kazlauskas A. PDGF induces an early and a late wave of PI 3-kinase activity, and only the late wave is required for progression through G1. Curr Biol 9: 512-521, 1999.

Karihaloo A, Karumanchi SA, Barasch J, Jha V, Nickel CH, Yang J, Grisaru S, Bush KT, Nigam S, Rosenblum ND, Sukhatme VP, and Cantley LG. Endostatin regulates branching morphogenesis of renal epithelial cells and ureteric bud. Proc Natl Acad Sci USA 98: 12509-12514, 2001.

Karihaloo A, O'Rourke DA, Nickel C, Spokes K, and Cantley LG. Differential MAPK pathways utilized for HGF- and EGF-dependent renal epithelial morphogenesis. J Biol Chem 276: 9166-9173, 2001.

Kawaida K, Matsumoto K, Shimazu H, and Nakamura T. Hepatocyte growth factor prevents acute renal failure and accelerates renal regeneration in mice. Proc Natl Acad Sci USA 91: 4357-4361, 1994.

Khwaja A, Lehmann K, Marte BM, and Downward J. Phosphoinositide 3-kinase induces scattering and tubulogenesis in epithelial cells through a novel pathway. J Biol Chem 273: 18793-18801, 1998.

Lee J, Cacalano G, Camerato T, Toy K, Moore MW, and Wood WI. Chemokine binding and activities mediated by the mouse IL-8 receptor. J Immunol 155: 2158-2164, 1995.

Lin JJ, Cybulsky AV, Goodyer PR, Fine RN, and Kaskel FJ. Insulin-like growth factor-1 enhances epidermal growth factor receptor activation and renal tubular cell regeneration in postischemic acute renal failure. J Lab Clin Med 125: 724-733, 1995.

Liu ZX, Yu CF, Nickel C, Thomas S, and Cantley LG. Hepatocyte growth factor induces ERK-dependent paxillin phosphorylation and regulates paxillin-focal adhesion kinase association. J Biol Chem 277: 10452-10458, 2002.

Loukinova E, Dong G, Enamorado-Ayalya I, Thomas GR, Chen Z, Schreiber H, and Van Waes C. Growth regulated oncogene- expression by murine squamous cell carcinoma promotes tumor growth, metastasis, leukocyte infiltration and angiogenesis by a host CXC receptor-2 dependent mechanism. Oncogene 19: 3477-3486, 2000.

Milatovic S, Nanney LB, Yu Y, White JR, and Richmond A. Impaired healing of nitrogen mustard wounds in CXCR2 null mice. Wound Repair Regen 11: 213-219, 2003.

Miller SB, Martin DR, Kissane J, and Hammerman MR. Hepatocyte growth factor accelerates recovery from acute ischemic renal injury in rats. Am J Physiol Renal Fluid Electrolyte Physiol 266: F129-F134, 1994.

Miller SB, Martin DR, Kissane J, and Hammerman MR. Insulin-like growth factor I accelerates recovery from ischemic acute tubular necrosis in the rat. Proc Natl Acad Sci USA 89: 11876-11880, 1992.

Miura M, Fu X, Zhang QW, Remick DG, and Fairchild RL. Neutralization of Gro and macrophage inflammatory protein-2 attenuates renal ischemia/reperfusion injury. Am J Pathol 159: 2137-2145, 2001.

Montesano R, Matsumoto K, Nakamura T, and Orci L. Identification of a fibroblast-derived epithelial morphogen as hepatocyte growth factor. Cell 67: 901-908, 1991.

Mueller SG, White JR, Schraw WP, Lam V, and Richmond A. Ligand-induced desensitization of the human CXC chemokine receptor-2 is modulated by multiple serine residues in the carboxyl-terminal domain of the receptor. J Biol Chem 272: 8207-8214, 1997.

Murillo H, Huang H, Schmidt LJ, Smith DI, and Tindall DJ. Role of PI3K signaling in survival and progression of LNCaP prostate cancer cells to the androgen refractory state. Endocrinology 142: 4795-4805, 2001.

Nakagami H, Morishita R, Yamamoto K, Taniyama Y, Aoki M, Matsumoto K, Nakamura T, Kaneda Y, Horiuchi M, and Ogihara T. Mitogenic and antiapoptotic actions of hepatocyte growth factor through ERK, STAT3, and AKT in endothelial cells. Hypertension 37: 581-586, 2001.

Nickel C, Benzing T, Sellin L, Gerke P, Karihaloo A, Liu ZX, Cantley LG, and Walz G. The polycystin-1 C-terminal fragment triggers branching morphogenesis and migration of tubular kidney epithelial cells. J Clin Invest 109: 481-489, 2002.

Niemann C, Brinkmann V, Spitzer E, Hartmann G, Sachs M, Naundorf H, and Birchmeier W. Reconstitution of mammary gland development in vitro: requirement of c-met and c-erbB2 signaling for branching and alveolar morphogenesis. J Cell Biol 143: 533-545, 1998.

Ohmichi H, Koshimizu U, Matsumoto K, and Nakamura T. Hepatocyte growth factor (HGF) acts as a mesenchyme-derived morphogenic factor during fetal lung development. Development 125: 1315-1324, 1998.

Okuma E, Inazawa Y, Saeki K, and Yuo A. Potential roles of extracellular signal-regulated kinase but not p38 during myeloid differentiation of U937 cells stimulated by cytokines: augmentation of differentiation via prolonged activation of extracellular signal-regulated kinase. Exp Hematol 30: 571-581, 2002.

Rauchman MI, Nigam SK, Delpire E, and Gullans SR. An osmotically tolerant inner medullary collecting duct cell line from an SV40 transgenic mouse. Am J Physiol Renal Fluid Electrolyte Physiol 265: F416-F424, 1993.

Richmond A and Thomas HG. Purification of melanoma growth stimulatory activity. J Cell Physiol 129: 375-384, 1986.

Ridley AJ, Comoglio PM, and Hall A. Regulation of scatter factor/hepatocyte growth factor responses by Ras, Rac, and Rho in MDCK cells. Mol Cell Biol 15: 1110-1122, 1995.

Rosenkranz S, DeMali KA, Gelderloos JA, Bazenet C, and Kazlauskas A. Identification of the receptor-associated signaling enzymes that are required for platelet-derived growth factor-AA-dependent chemotaxis and DNA synthesis. J Biol Chem 274: 28335-28343, 1999.

Safirstein R, Megyesi J, Saggi SJ, Price PM, Poon M, Rollins BJ, and Taubman MB. Expression of cytokine-like genes JE and KC is increased during renal ischemia. Am J Physiol Renal Fluid Electrolyte Physiol 261: F1095-F1101, 1991.

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, Barros EJ, Yang XM, Matsumoto K, Nakamura T, Park M, and Nigam SK. Involvement of hepatocyte growth factor in kidney development. Dev Biol 163: 525-529, 1994.

Shuster DE, Kehrli ME Jr, and Ackermann MR. Neutrophilia in mice that lack the murine IL-8 receptor homolog. Science 269: 1590-1591, 1995.

Suzuki H, Mori M, Seto K, Shibata F, Nagahashi S, Kawaguchi C, Suzuki M, Matsui H, Watanabe K, Miura S, and Ishii H. Rat CXC chemokine GRO/CINC-1 paradoxically stimulates the growth of gastric epithelial cells. Aliment Pharmacol Ther 14, Suppl 1: 94-100, 2000.

Tanaka TS, Jaradat SA, Lim MK, Kargul GJ, Wang X, Grahovac MJ, Pantano S, Sano Y, Piao Y, Nagaraja R, Doi H, Wood WH, 3rd Becker KG, and Ko MS. Genome-wide expression profiling of midgestation placenta and embryo using a 15,000 mouse developmental cDNA microarray. Proc Natl Acad Sci USA 97: 9127-9132, 2000.

Tang MJ, Cai Y, Tsai SJ, Wang YK, and Dressler GR. Ureteric bud outgrowth in response to RET activation is mediated by phosphatidylinositol 3-kinase. Dev Biol 243: 128-136, 2002.

Terkeltaub R, Baird S, Sears P, Santiago R, and Boisvert W. The murine homolog of the interleukin-8 receptor CXCR-2 is essential for the occurrence of neutrophilic inflammation in the air pouch model of acute urate crystal-induced gouty synovitis. Arthritis Rheum 41: 900-909, 1998. <a href="/cgi/external_ref?access_num=10.1002/1529-0131(199805)41:5

Unemori EN, Amento EP, Bauer EA, and Horuk R. Melanoma growth-stimulatory activity/GRO decreases collagen expression by human fibroblasts. Regulation by C-X-C but not C-C cytokines. J Biol Chem 268: 1338-1342, 1993.

Vlahos CJ, Matter WF, Hui KY, and Brown RF. A specific inhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002 ). J Biol Chem 269: 5241-5248, 1994.

Wang D and Richmond A. Nuclear factor- B activation by the CXC chemokine melanoma growth-stimulatory activity/growth-regulated protein involves the MEKK1/p38 mitogen-activated protein kinase pathway. J Biol Chem 276: 3650-3659, 2001.

Watanabe K, Kinoshita S, and Nakagawa H. Purification and characterization of cytokine-induced neutrophil chemoattractant produced by epithelioid cell line of normal rat kidney (NRK-52E cell). Biochem Biophys Res Commun 161: 1093-1099, 1989.

Watanabe K, Koizumi F, Kurashige Y, Tsurufuji S, and Nakagawa H. Rat CINC, a member of the interleukin-8 family, is a neutrophil-specific chemoattractant in vivo. Exp Mol Pathol 55: 30-37, 1991.

Wen DZ, Rowland A, and Derynck R. Expression and secretion of gro/MGSA by stimulated human endothelial cells. EMBO J 8: 1761-1766, 1989.

White JR, Lee JM, Young PR, Hertzberg RP, Jurewicz AJ, Chaikin MA, Widdowson K, Foley JJ, Martin LD, Griswold DE, and Sarau HM. Identification of a potent, selective non-peptide CXCR2 antagonist that inhibits interleukin-8-induced neutrophil migration. J Biol Chem 273: 10095-10098, 1998.

Wu D, LaRosa GJ, and Simon MI. G protein-coupled signal transduction pathways for interleukin-8. Science 261: 101-103, 1993.

Wuyts A, Govaerts C, Struyf S, Lenaerts JP, Put W, Conings R, Proost P, and Van Damme J. Isolation of the CXC chemokines ENA-78, GRO and GRO from tumor cells and leukocytes reveals NH 2 -terminal heterogeneity. Functional comparison of different natural isoforms. Eur J Biochem 260: 421-429, 1999.

Wuyts A, Haelens A, Proost P, Lenaerts JP, Conings R, Opdenakker G, and Van Damme J. Identification of mouse granulocyte chemotactic protein-2 from fibroblasts and epithelial cells. Functional comparison with natural KC and macrophage inflammatory protein-2. J Immunol 157: 1736-1743, 1996.

作者单位：1 Section of Nephrology, and 2 Howard Hughes Medical Institute, Yale University School of Medicine, New Haven, Connecticut 06510