Stable Knockdown of Polycystin-1 Confers Integrin-21–Mediated Anoikis Resistance
Lorenzo Battini*,
Elena Fedorova*,
Salvador Macip,
Xiaohong Li*,
Patricia D. Wilson* and
G. Luca Gusella*
* Division of Renal Medicine and Department of Oncological Sciences, Mount Sinai School of Medicine, New York, New York
Address correspondence to: Dr. G. Luca Gusella, Mount Sinai School of Medicine, One Gustave Levy Place, Box 1243, New York, NY 10029. Phone: 212-241-9597; Fax: 212-987-0389; E-mail: luca.gusella{at}mssm.edu
Received for publication March 16, 2006.
Accepted for publication August 17, 2006.
The mechanisms of action of polycystin-1 (PC1) have been difficultto dissect because of its interaction with multiple factors,the heterogeneity of the genetic mutations, and the complexityof the experimental animal models. Here, stable knockdown ofPC1 in MDCK epithelial cells was achieved by lentiviral-mediateddelivery of a specific small interfering RNA for PKD1. The reductionof PC1 expression prevented tubulogenesis in three-dimensionalcollagen type I culture in response to hepatocyte growth factorand induced formation of cysts. PC1 knockdown created a conditionof haploinsufficiency that led to hyperproliferation, increasedadhesion to collagen type I, and increased apoptosis. It wasshown that the suppression of PC1 was associated with the increasedexpression of integrin-21 and reduced apoptosis in cells grownon collagen type I. The engagement of integrin-21 seemed tobe essential for the survival because PC1 knockdown cells weresignificantly less susceptible to anoikis by a mechanism thatwas reversible by anti–integrin-21 blocking antibodies.Overall, these data link integrin-21 to some of the biologicfunctions that are ascribed to PC1 and establish the potentialof this approach for the direct study of PC1 functions in agenetically defined background. Furthermore, these findingsindicate that reduction of PC1 expression levels, rather thanthe loss of heterozygosity, may be sufficient to induce cystogenesis.
Alterations in the PKD1 and PKD2 genes account for cases ofautosomal dominant polycystic kidney disease (ADPKD) with arelative frequency of approximately 85 and 15%, respectively(1–3). PKD1 encodes polycystin-1 (PC1), a large 4302–aminoacid membrane protein that participates in multiprotein complexesthat are involved in cell–matrix and cell–cell interactionsand signaling (4,5). Through the heterodimerization with PC2,a nonselective cationic channel that is encoded by the PKD2gene, PC1 can function as a gating receptor to control calciuminflux (6). The interdependent functions of PC1 and PC2 in kidneyare supported by the overlapping ADPKD phenotypic manifestationthat is caused by mutations of either gene, although the severityand the progression of the disease are exacerbated when PKD1is involved. The co-localization of the polycystins in primarycilia suggests their involvement in a cellular mechanosensorysystem (7). These findings are supported by the evidence thatpolycystic renal phenotypes arise from the mutation or deletionof various ciliary and adhesion proteins (8–10). To date,however, it has been difficult to define clearly the functionsof PC1 and the molecular mechanisms of cystogenesis that itcontrols. In part, this difficulty derives from the heterogeneityof the genetic mutations that are identified in patients withADPKD as well as from the lack of functionally characterizedmutants. Furthermore, most of the animal models for polycystickidney share phenotypic manifestations, but some are geneticallydifferent or incompletely characterized.
Data concerning the functions of PKD1 have been acquired fromthe study of knockout mouse models and from cells that havebeen isolated from patients with ADPKD. Although these sourceshave been extremely informative, they derive from the developmentand differentiation of genetically mutated cells, which mayhave evolved mechanisms compensating for PC1 dysregulation.As a result, some of the available small animal models do notparallel the progression of the disease that is observed inhumans (11), and most have not been evaluated beyond the juvenilestage. Similarly, in studies that use PC1 transdominant mutants,it may be difficult to determine whether the ensuing effectsare produced by the constitutive activation or the irreversibleinhibition of PC1 normal functions. The specific inhibitionof PC1 in characterized cell lines can facilitate the studyof protein function by reproducing a condition of haploinsufficiencyin the absence of other genetic variables.
In this study, constitutive knockdown of PC1 in MDCK cells wasachieved by lentiviral-mediated expression of anti-PC1 siRNA.The stable expression of small interfering RNA (siRNA) thatwas specific for PKD1 inhibited the expression of PC1 and producedbiologic features that are characteristic of the ADPKD phenotype,such as increased cellular adherence, cystic growth, and apoptosis.The apoptotic phenotype could be reverted by culturing the PC1knockdown cells on type I collagen. Importantly, we found thatPC1 knockdown cells became resistant to anoikis through a mechanismthat is mediated by integrin-21. These results suggest thatanoikis resistance may play an important role in the expansionof the epithelia cells within the peritubular space and pointat integrin-21 as a possible target to slow the cystogenic process.This novel approach circumvents possible compensatory developmentalmechanisms, which cannot be excluded in cells that are derivedfrom patients or animal models, and provides a unique tool forthe genomic and proteomic comparative analyses in the presenceor absence of functional PC1.
Cell Culture
Swiss 3T3 fibroblasts, MDCK-G (from hereon referred to as MDCK)and lentivirus-transduced cells lines were maintained at 37°Cin a humidified, 5% CO2 atmosphere in DMEM (Life Technologies,Grand Island, NY) supplemented with 10% FBS (Life Technologies)and 2 mM glutamine (Life Technologies). Conditioned medium fromconfluent Swiss 3T3 fibroblast monolayers was collected every48 h, filtered through a 0.2-µm filter, and stored inaliquots at –80°C until used as source of hepatocytegrowth factor (HGF) as previously reported (12,13).
Generation of the Lentiviral Vector for the Expression of siRNA
The vector VIRHD/E is a derivative of the VVCW self-inactivatinglentivector (14). First, a NheI site was created by digestionwith HindIII, fill-in with Klenow polymerase and self-ligationof the pPD31/C2 plasmid (15), which contains the murine phosphoglyceratekinase (PGK) promoter. The PGK promoter was excised from theresulting pPD31/CN plasmid with ClaI and NheI and ligated intothe corresponding sites of VVCW, in place of the cytomegaloviruspromoter, to generate the VVPW lentivector. The enhanced greenfluorescence protein (EGFP) reporter then was amplified fromthe pEGFP1 plasmid (Clontech, Mountain View, CA) using the HighFidelity Taq polymerase (Roche Diagnostics, Indianapolis, IN)using the primers EGFP/SV40NLS/5 5'-caggatccgccaccatgggcgacccaaaaaagaaaagaaaagtgagcaagggcgaggagctg-3',which contains the SV40 nuclear localization signal, and EGFP/35'-gtgaattcttacttgtacagctcgtccatgccgag-3'. The amplified productwas digested with BamHI and EcoRI (italicized in primer sequences),filled in with Klenow polymerase, and ligated into the Klenow-filledin BamHI and NotI sites of the VVPW vector under the controlof the PGK promoter to generate the VVWP/EGFP3 plasmid. A vectorVVPW/E4 was generated from the insertion of unique XbaI andEcoRV restriction sites between the AscI and ClaI sites of VVPW/EGFP3using a linker derived from the annealing of the following oligonucleotidesthat contained the XbaI (boldface) and EcoRV (italic) restrictionsites: 5'-cgcgcctctagatatcg-3' and 5'-cgcgcctctagatatcg. Thesynthetic human polymerase III H1 promoter then was generatedby 10 cycles of PCR using High Fidelity Taq polymerase and theprimers H1/A 5'-ggcgcgccgctagcaatttgcatgtcgctatgtgttctgggaaatcaccataaacgtgaatgtctttggatttgggaat-3'and H1/B 5'-gtgctagcgaattctggatccgagtggtctcatacagaacttataagattcccaaatccaaagacatttcacgtttatgg-3'that contained the NheI and BamHI and EcoRI cloning sites. Theresulting 127-bp H1 product is flanked by NheI restriction sites(in bold in the primer) and provides the BamHI and EcoRI cloningsites (italicized in the primer) at the 3' end of the promoter.After digestion with NheI, the H1 promoter was ligated intothe XbaI site of VVPW/E4 to produce the VIRHC/E and VIRHD/Elentivectors. For the following study, we used the VIRHD/E lentiviralvector that contained the H1 promoter in the 3'-5' orientation(Figure 1A). The following oligonucleotides were annealed andinserted between the BamHI and EcoRI sites of VIRHD/E so thatspecific siRNA were produced as short hairpin RNA under thecontrol of the H1 promoter: PKD1/3211/S 5'-gatccgccacgtgagcaacgtcaccattagatcaggtgacgttgctcacgtggttttttggaag-3';PKD1/3211/AS 5'-aattcttccaaaaaaccacgtgagcaacgtcacctgatctaatggtgacgttgctcacgtggcg-3';Luc/850/S 5'-gatccgtgcgttgctagtaccaacttcaagagagttggtactagcaacgcacttttttgg-3';and Luc/850/AS 5'-aattccaaaaaagtgcgttgctagtaccaactctcttgaagttggtactagcaacgcacg-3'(specific complementary siRNA sequences are italicized).
Figure 1. Knockdown of PKD1 in MDCK cells. (A) Scheme of the VIRHD/E lentiviral vector and its derivatives VIRHD/E/siLuc and VIRHD/E/siPKD, which contain hairpin sequences for the expression of control anti-luciferase siRNA and anti-PKD1 siRNA, respectively. CMV, cytomegalovirus early promoter; RRE, rev responsive element; cPPT, central polypurine tract; H1, human H1 promoter; PGK, murine phosphoglycerate kinase promoter; nEGFP, enhanced green fluorescence protein with a nuclear localization signal; WPRE, Woodchuck hepatitis virus posttranscriptional regulatory element; , deletion in 3'LTR. (B) Polycystin-1 (PC1) mRNA was detected by Northern analysis on total RNA that was isolated from the indicated cell populations using a probe that was specific for PKD1 (see Materials and Methods). The membrane then was stripped and rehybridized to a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe to control for equal sample loading (bottom). (C) Expression of PC1 was detected in total cell lysate from confluent cell cultures using the C20 antibody (Santa Cruz Biotechnology) specific for the PC1 carboxyl terminus. The position of molecular weight markers (Kdal) and PC1 are indicated. The approximately 150-Kdal band that was observed in all of the cell lines represents a nonspecific signal. The bars in the bottom panel represent the densitometric analysis that was performed on the radiographic film using the ImageJ software. The ratios of PC1 to actin expression are shown as percentage of that in parental MDCK cells. The expression of actin (indicated) was determined as control of equal loading after stripping of the membranes above and immunoblotting with an anti-actin antibody (dilution 1:20,000; Chemicon). (D) Expression of PC1 (white arrowhead) in individual clones that were derived from the control MDCK/E/siLuc (3L, 4L, 5L, 6L, and 7L) and PC1 knockdown MDCK/E/siPKD (2P, 5P, 8P, 9P, and 10P) polyclonal cell populations. Actin was used as loading control. The bars below the lanes indicate the densitometric analysis of the expression of PC1 relative to the correspondent actin signal in the control () and PC1 knockdown () clones. MDCK/E/siLuc and MDCK/E/siPKD indicate the control and PC1 polyclonal populations, respectively.
Establishment of Cell Lines
Lentiviral vectors were produced and titered as described previously(14). MDCK cell were transduced at a multiplicity of infectionof 20 in the presence of 5 µg/ml polybrene (Sigma, St.Louis, MO). One week after transduction, polyclonal cell populationsthat expressed the EGFP reporter protein were sorted using atriple-laser MoFlo high-speed cell sorter (Cytomation, FortCollins, CO). Cells that were sorted at sheath pressure of 30PSI and speed of 15,000 events/s then were expanded to preparefrozen batches. Individual clones from the control MDCK/E/siLucand PC1 knockdown MDCK/E/siPKD cell populations were derivedby limiting dilution.
Three-Dimensional Collagen Gels
MDCK and stably transduced cells were trypsinized to a single-cellsuspension and resuspended at a concentration of 1 x 105 cells/mlin an ice-cold collagen solution that was prepared as describedpreviously (13). Briefly, collagen type I solution (2.4 mg/ml;BD Biosciences, San Diego, CA) was mixed with 10x DMEM (LifeTechnologies) and 7.5% (vol/vol) bicarbonate solution (LifeTechnologies) in a ratio of 8:1:1. The cell suspension thenwas dispensed at 2.5 ml/well in six-well plates and allowedto gel for 20 min at 37°C before being overlaid with 2.5ml of fibroblast-conditioned medium. The conditioned mediumwas changed daily, and pictures were taken at day 7.
Northern Blot Analysis
Total RNA was extracted from confluent monolayers using TRIzol(Invitrogen, Carlsbad, CA), and 20 µg from each samplewas resolved by electrophoresis in 1.2% agarose gel. The RNAthen was transferred onto nylon membrane and ultraviolet cross-linked.PC1 mRNA was detected using a 32P-labeled fragment that wasamplified after reverse transcription of MDCK total RNA usingthe PKD1-specific primers PKDC-212/5 5'-gccaccgcgctagacgtc-3'and PKDC-406/3 5'-cagaccgcagttacactcca-3'. The stripped membranethen was rehybridized with a probe specific for the housekeepingglyceraldehyde-3-phosphate dehydrogenase gene.
Immunodetection
Cells were lysed by incubation for 30 min on ice in lysis bufferthat contained 1% Triton X100, 10 mM Pipes, 3 mM MgCl2, 300mM sucrose, and 5 mM EGTA and supplemented with a cocktail ofprotease inhibitors (Roche Diagnostics). For PC1 detection,250 µg of each sample was resolved on a 4% SDS-polyacrylamidegel and electrophoretically blotted onto Immobilon-P membrane(Millipore, Bedford, MA). PC1 was detected using the affinity-purifiedgoat anti-PC1 C-20 polyclonal antibody (Santa Cruz Biotechnology,Santa Cruz, CA) diluted 1:1000 in PBS and 0.1% Tween-20 thatcontained 5% BSA and a secondary rabbit peroxidase-conjugatedanti-goat antibody (1:5000 dilution in 5% nonfat dry milk inPBS and 0.1% Tween-20). Immunodetection of integrin-2 and integrin-1was performed on 100 µg of protein extracts resolved on10% PAGE using a rabbit anti-human integrin-2 (Chemicon International,Temecula, CA) and anti-human integrin-1–specific antibody(Upstate, NY) at a final dilution of 1:2000 and 1:5000, respectively.Actin expression was detected using a monoclonal mouse anti-actinantibody (Chemicon International) diluted 1:20,000. Immunocomplexeswere visualized using the enhanced chemiluminescence detectionkit Lumi-Light Plus (Roche Diagnostics).
FACS Analysis of Integrin Expression
For the detection of integrin-21 surface expression, 106 controlMDCK/E/siLuc or PC1 knockdown MDCK/E/siPKD cells were harvestedafter incubation with dissociation buffer (Invitrogen). Thecells then were washed once with medium that contained 10% FBS,twice in PBS that contained 0.1% BSA (PBS-0.1%BSA), and finallyresuspended in 100 µl of PBS-0.1%BSA. The mAb specificfor the integrin-21 heterodimer (clone BHA2.1; Chemicon International)(16,17) or isotype control antibody then was added to a finalconcentration of 10 µg/ml before the cells were incubatedat 4°C for 30 min. The cells were washed twice in PBS-0.1%BSAand finally resuspended in 50 µl of the same buffer thatcontained 3 µg of CY3-conjugated goat anti-mouse IgG andincubated at 4°C for 20 min. After three washes in PBS-0.1%BSA,red fluorescence was detected and quantified by flow cytometry.
Cell Adhesion Assay
Cells in single-cell suspension were plated in complete mediumin a collagen-coated 96-well plate in sextuplicate at a concentrationof 1 x 104 cells/well. After 1 h of incubation at 37°C,half of the wells were washed thrice with PBS before fresh mediumwas added to all of the wells. After a 24-h incubation, thenumber of remaining cells was determined using the Cell Titer96 AQueous kit (Promega, Madison, WI). The cell count was normalizedby expressing it as percentage of the number of originally platedcells in the unwashed wells.
Cell-Cycle Analysis
Parental and lentivector-transduced cells were trypsinized toa single-cell suspension, washed twice in PBS-0.1%BSA, and resuspendedat the concentration of 1 to 2 x 106 cells/ml. Cells were fixedin 75% ethanol for 1 h at 4°C. After being washed twicein PBS, cells were incubated for 1 h at 4°C in 1 ml of stainingsolution that contained 50 µg/ml propidium iodide, 3.8mM sodium citrate, and 500 ng/ml RNase A. The cells then werewashed and analyzed by flow cytometry.
Apoptosis and Anoikis Assay
Apoptosis was determined by assessment of the expression ofboth annexin V and caspase-3 on cells that were plated as describedin the text and figure legends. Annexin V was detected usingthe Vibrant Apoptosis Assay Kit # 9 (Molecular Probes, Eugene,OR), following the manufacturer’s protocol, and then analyzedby flow cytometry. Caspase-3 was measured on equal amounts ofcell extracts using the EnzChek Caspase-3 Assay Kit #1, (MolecularProbes) according to the manufacturer’s instructions.For measurement of anoikis, cells were grown in nonadherentconditions using 100-mm Petri dishes that were coated with polyhydroxyethylmethacrylate(polyHEMA; Aldrich Chemical Co., Milwaukee, WI) as described(18). A total of 3,000,000 cells/ml were cultured for 12 h inregular tissue culture plates or in nonadherent conditions inpolyHEMA-coated dishes before they were collected in the mediumin which they had been incubated by scraping or pipetting, respectively.Floating cells in the tissue culture controls on uncoated plateswere combined with the attached cells before protein extractpreparation. The level of anoikis then was measured using theEnzChek Caspase-3 Assay Kit #1 (Molecular Probes). Blockingantibodies anti–integrin-21 were purchased from Chemiconand used at a final concentration of 10 µg/ml.
Statistical Analyses
Statistical analyses were performed by one-way ANOVA with Bonferronipost test or unpaired t test with Welch correction, using theInStat 3.0 software (GraphPad Software, San Diego, CA).
Posttranscriptional Silencing of the PKD1 Gene
Silencing of PC1 expression was achieved by RNA interference.We used lentiviral vectors to stably express siRNA in the targetcells to avoid possible variability of exogenously added siRNAand to allow the analysis of the suppression of PC1 during anextended period of time. MDCK cells were transduced with thecontrol parental lentivector VIRHD/E and its derivative VIRHD/E/siLuc,which expresses a control anti-luciferase hairpin siRNA, orthe VIRHD/E/siPKD lentivector, for the expression of a shorthairpin sequence specific for PC1 (Figure 1A; see Materialsand Methods for details). These vectors also carry the EGFPreporter gene whose expression was used to sort stably transducedpolyclonal cell lines MDCK/E, MDCK/E/siLuc, and MDCK/E/siPKD,respectively. To assess the efficacy of the siRNA, the levelsof PC1 mRNA were measured in transduced and nontransduced cells.Expression of PC1 mRNA was comparable in control MDCK, MDCK/E,and MDCK/E/siLuc cells, whereas it was dramatically decreasedin MDCK/E/siPKD cells (Figure 1B). Immunodetection of PC1 usingan anti-PC1–specific antibody indicated that the downregulationof PC1 mRNA in the MDCK/E/siPKD cells corresponded to an approximately75% decrease in the amount of PC1 protein (Figure 1C), confirmingthe efficiency of the anti-PC1 siRNA. The expression of PC1also was determined in five randomly selected individual clonalpopulations that were derived from the control MDCK/E/siLuc(L clones) and PC1 knockdown MDCK/E/siPKD cells (P clones).Whereas it remained similar in the L clones, the expressionof PC1 was reduced in all of the P cell lines, showing a degreeof inhibition that ranged between 70 and 85% (Figure 1D). Theaverage suppression of PC1 in the P clones confirmed the PC1knockdown observed with the polyclonal populations. Therefore,unless differently specified, in the following experiments,the polyclonal populations were chosen because they were representativeof different inhibition levels that may resemble more closelythe in vivo expression distribution. The use of these populationsalso rules out clonal idiosyncratic responses.
Effects of PC1 Knockdown on In Vitro Tubulocystogenesis
We analyzed how the inhibition of PC1 expression would affectthe structural organization of MDCK cells under these cultureconditions. When grown in three-dimensional collagen gel culturein the absence or presence of exogenous HGF, MDCK cells haveproved very useful as a model for cystogenesis or tubulogenesis,respectively. Within 3 d of culture in the presence of HGF,parental MDCK as well as control vector-transduced MDCK/E andMDCK/E/siLuc cells formed elongated cellular structures thatacquired characteristic tubular shapes by day 7 (Figure 2A).In contrast, MDCK/E/siPKD cells failed to develop tubules andgrew in small clusters that developed increasingly irregularcystic appearance with time (Figure 2A). Occasionally, someof the cellular formations in MDCK/E/siPKD were mixed structureswith tubulocystic appearance (Figure 2A, bottom right). Thisobservation suggested that fluctuations in the level of PC1might be important to determine the final outcome, because itis expected that polyclonal cell populations exhibit differentefficiency of PC1 mRNA inhibition and variable PC1 expression.A detailed histologic analysis was performed by sectioning thecollagen-embedded cultures. The hematoxylin and eosin stainingof preparations from MDCK/E/siPKD and control MDCK/E/siLuc culturesconfirmed the cystic and tubular morphology, respectively (Figure 2B).Whereas tubules presented tight lumens, cysts were characterizedby the typical flattened epithelium and larger, distended lumens.The relative numbers of tubular and cystic formations that weregrown in the three-dimensional collagen cultures were scoredunder light microscopy (Figure 2C). A significantly higher numberof tubular versus cystic structures was observed in parentalMDCK and control MDCK/E or MDCK/E/siLuc cell cultures for 7d. In contrast, in the MDCK/E/siPKD cultures, cystic structuressignificantly outnumbered tubules (Figure 2C).
Figure 2. PC1 knockdown present a cystic phenotype of cells in three-dimensional collagen matrix culture. (A) Cells were grown in collagen type I gel in the presence of hepatocyte growth factor (HGF)-containing conditioned medium for 7 d before pictures were taken using a light microscope. Control cell lines MDCK, MDCK/E, and MDCK/E/siLuc produced extended tubule-like processes (top), whereas PC1 knockdown cells MDCK/E/siPKD presented irregular cystic-like structures (bottom). (B) Hematoxylin and eosin staining of the control MDCK/E/siLuc and PC1 knockdown MDCK/E/siPKD cultures grown in the collagen matrix gels. The luminal space () and the flattened epithelial lining the cyst (arrowhead) are indicated. (C) The number of cysts versus tubules was scored in control and PC1 knockdown cells after 1 wk of culture in three-dimensional collagen gels in HGF. The bars indicate the average number of cysts and tubules ± SD for each cell line after evaluation of five random fields. **P < 0.01 cyst or tubule count in PC1 knockdown cells versus controls. Original magnification, x200.
Alteration of Cell Cycle and Apoptosis in PC1 Knockdown Cells
ADPKD is characterized by epithelial hyperproliferation, andoverexpression of PC1 has been shown to inhibit cell growth(19). We assessed the growth rate in subconfluent cultures ofcontrol and PC1 knockdown cells. Suppression of PC1 expressioncorrelated with significant increase of cell proliferation ofMDCK/E/siPKD cells (Figure 3A). This results were supportedfurther by the faster growth rate of PC1 knockdown clones ascompared with the control clones (Figure 3B). Analysis of thecell cycle in nonsynchronized cultures in complete medium indicateda more pronounced decline of the fraction of PC1 knockdown cellsin G1 by 72 h. At the same time, whereas in control MDCK/E/siLuc,the percentage of cells in the sub-G0/G1 fraction stabilizedat approximately 4%, in PC1 knockdown MDCK/E/siPKD, the sub-G0/G1fraction expanded to 16%, suggesting a higher rate of cell death(Figure 3C). The measurement of annexin V and caspase-3 levelsconfirmed that apoptosis was significantly increased in PC1knockdown cells (Figure 4, A and C). It is interesting thatgrowth on collagen type I–coated plates prevented theapoptosis that was associated with the inhibition of PC1 (Figure 4, B and D).
Figure 3. Increased proliferation and alteration of the cell cycle in PC1 knockdown cells. (A) Equal numbers of cells were plated and grown at subconfluent conditions for up to 72 h. At the indicated times, cells were harvested and counted. Cell counts are reported as percentage of the number of cells scored 2 h after plating (time 0). Each value represents the mean ± SD of three wells in one of five representative experiments. A significant increase in the number of the PC1 knockdown cells was detectable within 48 h and maximal at 72 h (*P < 0.05 and **P < 0.01, respectively, PC1 knockdown cells versus controls). (B) The same analysis that was performed on the clonal cell lines further indicated the faster growth rate of the PC1 knockdown cell lines. As expected, the clones display a narrower SD as compared with the polyclonal population and consequently a higher value of significance (***P < 0.001). (C) Control MDCK/E/siLuc and PC1 knockdown MDCK/E/siPKD cells were plated at subconfluent conditions and subjected to cell-cycle analysis at the indicated times of culture. PC1 knockdown MDCK/E/siPKD showed a decrease in the fraction of cells in G0/G1 by 72 h and a progressive increase in cellular death as compared with the control MDCK/E/siLuc culture.
Figure 4. Apoptosis accounts for increased cell death in PC1 knockdown cells. An equal number of cells were cultured in adherent conditions on plastic (A and C) or collagen type I–coated plates (B and D) for 24, 48, or 72 h before being harvested to measure caspase-3 levels or annexin V. For each sample, the readings of caspase-3 adjusted for the relative protein concentration were expressed as percentage increase over the caspase-3 values of the respective cell population harvested after 2 h of culture (point 0). Shown is the mean ± SEM of three independent experiments. (A) Significant differences were observed between PC1 knockdown cells and controls (*P < 0.05; ***P < 0.001). (B) Apoptosis of PC1 knockdown cells was prevented when the cells were cultured on collagen type I as adhesion substrate. (C and D) Annexin V, measured by FACS analysis, confirmed the highly apoptotic phenotype of PC1 knockdown MDCK/E/siPKD cells as compared with the control MDCK/E/siLuc (C) and its reversal after plating on collagen type I (D).
Increased Adherence and Expression of Integrin-2 and Integrin-1 in PC1 Knockdown Cells
The cystic phenotype is associated with multiple structuralchanges of the basement membrane and with abnormalities in thecomposition of matrix and cell–matrix interactions. ADPKDcells display increased adherence to type I collagen and increasedlevels of integrin-21 compared with normal age-matched humanrenal epithelia (20).
When we assessed the ability of PC1 knockdown cells to adhereto collagen type I, we measured a significantly increased attachmentas compared with the parental or transduced control cells (Figure 5A,left). Similarly, PC1 knockdown clones were more adherent thancontrol clones (Figure 5A, right). It is interesting that thereseemed to be a correlation between the faster growth rate andthe increased adherence within the PC1 knockdown clones andthe relative PC1 expression (2P and 8P versus 5P, 9P, and 10P).Cell interactions with the extracellular matrix are mediatedmainly by integrins. In particular, integrin-21, which functionsas a receptor for collagen type I, has been shown to play acritical role in tubulocysts formation and the survival andthe ability of renal tubular epithelia and MDCK cells to maintainpolarization (21–23). To determine whether changes inintegrin-2 and integrin-1 expression may account for the differentialadhesion to collagen type I, we measured the levels of theseintegrins in the various cell lines. Integrin-2 and integrin-1were equally expressed in MDCK, MDCK/E, and MDCK/E/siLuc. However,in the PC1 knockdown MDCK/E/siPKD cells, integrin-2 and integrin-1were increased three- and six-fold, respectively (Figure 5B).The higher relative expression of integrin-2 and integrin-1in the PC1 knockdown cells was maintained during culture inthe presence of collagen type I (Figure 5B). FACS analysis usingan antibody specific for the integrin-21 heterodimer showeda corresponding increase of integrin-21 on the surface of theMDCK/E/siPKD cells as compared with the control cells, indicatingthe proper processing of the receptor (Figure 5C). This evidence,along with the prevention of apoptosis in PC1 knockdown cellsthat were plated on collagen type I (Figure 4, B and D), suggestedthat the engagement of integrin-21 by its ligand could providea survival signal (24).
Figure 5. PC1 knockdown cells display increased adherence to collagen type I and higher expression of integrin-2 and integrin-1. (A, left panel) Single-cell suspensions of control and PC1 knockdown cells were plated on collagen type I–precoated plates for 1 h. Cells then were washed three times with PBS before fresh medium was added. The number of remaining adherent cells, normalized by the number of originally plated cells (unwashed), was determined using the Cell Titer 96 AQueous kit (Promega). Data are means ± SD of sextuplicate counts (***P < 0.001) in one of five independent experiments. (A, right panel) Adhesion experiments on individual clones confirmed the highly significant increase of adherence of PC1 knockdown P (P2, P5, P8, P9, and P10) cell lines as compared with control L (3L, 4L, 5L, 6L, and 7L) clones (P < 0.0001, unpaired t test with Welch correction performed on aggregate values from individual clones). (B) Cells were plated either on uncoated or collagen type I–precoated plates for 24 h, as indicated. Integrin-2 and integrin-1 expression was detected by immunoblot. Actin expression then was determined on the stripped membrane to control for equal sample loading. The expression of integrin-2 and integrin-1 relative to actin was quantified using ImageJ software (NIH) and depicted in the bar graph below the immunoblot. The increased expression of either integrin was not significantly affected by culturing on collagen type I. (C) Surface expression of integrin-21. FACS analysis that was performed using an anti–integrin-21–specific antibody indicated that the heterodimer was more highly expressed on the plasma membrane of PC1 knockdown MDCK/E/siPKD as compared with control MDCK/E/siLuc cells.
PC1 Knockdown Cells Are Resistant to Anoikis
The prevention of integrin-21 binding to collagen in MDCK cellswas shown previously to induce anoikis, the mechanism of apoptosisthat is triggered by the lack of cell adhesion to the matrixsubstrate (25–28). We observed that PC1 knockdown andcontrol cells that were growing on collagen detached within1 h from the addition of a blocking antibody that specificallyinterferes with the binding of the integrin-21 receptor to itssubstrates (Figure 6A). The cell detachment was followed byan increase in caspase-3 by 12 h in all of the cell lines, indicatingthe induction of anoikis. We therefore asked whether the augmentedexpression of integrin-21 in the PC1 knockdown cells may affectsurvival in conditions of anchorage-independent growth. Whenplated in nonadherent conditions, MDCK/E/siPKD seemed significantlymore resistant to anoikis than the parental and control cells(Figure 6B). The direct role of integrin-21 in anoikis resistancein the MDCK/E/siPKD was assessed using the specific anti–integrin-21antibodies. The presence of the anti–integrin-21 antibodiesreturned the level of anoikis in MDCK/E/siPKD to that observedin the parental or control transduced cells (Figure 6B), suggestingthat the increase in integrin-21 is a determinant of the resistanceto anoikis that is mediated by the integrin-21 heterodimer inPC1-deficient cells.
Figure 6. PC1 knockdown cells are resistant to anoikis by a mechanism that is mediated by integrin-21. (A) Control and PC1 knockdown cells were grown on collagen type I–coated plates for 24 h and then incubated in the absence or the presence of the blocking antibody specific for the integrin-21 heterodimer. The presence of the antibody equally induced cell detachment (data not shown) and caspase-3 activation in control and PC1 knockdown cells. (B) Equal numbers of control and PC1 knockdown cells were plated either under adherent conditions on tissue culture dishes or under nonadherent conditions on polyhydroxyethylmethacrylate-coated Petri dishes in the absence or presence of blocking anti–integrin-21 antibodies (Chemicon; see Materials and Methods). Anoikis was determined by measurement of the levels of caspase-3 in nonadherent conditions expressed as a percentage of the relative caspase-3 values in adherent conditions. Data are means ± SD. Significant change in anoikis as a result of the co-incubation with the anti–integrin-21 antibodies (checkered bars) was observed in PC1 knockdown cells (**P < 0.01 versus PC1 knockdown cells under nonadherent condition in the absence of anti–integrin-21 antibodies).
To analyze directly the biologic functions of PC1, we used siRNAinterference to suppress the expression of PC1 in MDCK cells,which provide a useful model for studying the mechanisms oftubulocystogenesis. We showed that lentiviral-mediated deliveryof anti-PC1–specific siRNA efficiently suppresses PC1expression in MDCK cells and induces a state of haploinsufficiencythat recapitulates the major biologic alterations that are pathognomicto ADPKD. Specifically, PC1 knockdown MDCK lost the abilityto form tubules when grown in three-dimensional collagen gelin the presence of HGF. Similar to observations in human renalADPKD cyst epithelial cells, PC1 knockdown cells were more adherentto collagen type I and grew at a faster rate than the controls,while maintaining a higher apoptotic phenotype (4,29). Thesedata support the findings of Boletta et al. (17) showing slowgrowth, inhibition of apoptosis, and induction of tubule formationin MDCK cells that overexpressed PC1.
We used the siRNA knockdown model to show that the specificsuppression of PC1 leads to stronger adherence and increasedexpression of integrin-2 and integrin-1, whose heterodimer recognizescollagen type I as ligand. These observations complement thoseby Wilson et al. (20), who reported enhanced integrin-2 expressionin cell lines that were established from patients with ADPKDand inhibition of adherence to collagen type I by anti–integrin-21antibodies. Increased adherence to collagen type I, possiblymediated by integrin-1, also was observed in cell lines thatwere derived from the C57Bl/6Jcpk mouse model of autosomal recessivepolycystic kidney disease (ARPKD), which carries a truncationin the cystin gene (29,30). Taken together, these findings emphasizethe importance of integrin-1 in polycystic kidney disease independentof the genetic basis underlying the cystogenic process. In addition,we demonstrated that growth in the presence of collagen typeI reverts the apoptotic phenotype of PC1 knockdown cells, suggestingthat a positive feedback may potentiate the integrin-mediatedsurvival signal when PC1 is absent or expressed at low levels.These results are consistent with the finding that suppressionof integrin-21 leads to increased cell death of MDCK and failsto branch in response to HGF or to form cysts in collagen typeI (22). Although integrins are involved mostly in the bindingto extracellular matrix, they also play a role in cell–cellcontact (31). In particular, integrin-1 has been shown to bindE-cadherin in a divalent cation-dependent manner (32). The co-localizationof integrin-21 with PC1 at the focal adherens (20) and the presenceof E-cadherin in the multiprotein complex with PC1 (33) suggestthat interactions among these molecules may modulate cell–celland/or cell–matrix adhesion during epithelial remodeling.Recently, the re-expression of N-cadherin along with a reductionof E-cadherin has been demonstrated on the membranes of cyst-liningepithelial cells from patients with ADPKD (34), an effect thatis consistent with the epithelial-mesenchymal transition andloss of polarization and dedifferentiation of cystic epithelialcells (4). Whether N-cadherin can interact with integrin-21and provide a survival signal remains to be established. However,because integrin-1 has been localized in the cilium, where itparticipates in the flow-induced Ca2+ response (35), it remainspossible that its modulation plays a role in the ciliary mechanosensoryfunction of polycystins.
Our data also showed for the first time that PC1 knockdown cellsseem more resistant to anoikis—the apoptotic pathway thatis triggered by the loss of cell anchorage—through a mechanismthat involves integrin-21. It is possible that integrin-21-–mediatedcytoskeletal rearrangement and other changes at the membraneof PC1 knockdown cells allow different surface molecules tointeract and transmit survival cues and favor the expansionof the cystic epithelium into the renal interstitium. Changesin the extracellular matrix composition are likely to contributeto these interactions and modulate the cellular architecture.In fact, in agreement with previous observations (36,37), theculture of both control and PC1 knockdown cells in basementmembrane matrigel failed to develop tubules and grew as cystsregardless of the presence of HGF (data not shown).
It is interesting that both the increase in integrin expressionand reduced anoikis are characteristic of the metastatic evolutionof tumors (38). Similarly, N-cadherin has been involved in theprogression of the metastatic phenotype (39). Although we didnot assess N-cadherin expression, the overexpression of integrin-1,hyperproliferation, and resistance to anoikis after the inhibitionof PC1 are suggestive of the emergence of a preneoplastic profileand support the proposed role of PC1 as a tumor suppressor gene(19). To date, no correlation between ADPKD and cancer has beenestablished, but the occurrence of renal carcinoma after acquiredpolycystic kidney disease secondary to hemodialysis is welldocumented (40). How these two cystogenic processes differ remainsan intriguing question, and the definition of the variablesand the conditions that lead to transforming events in acquiredpolycystic kidney but not in ADPKD or ARPKD may reveal new functionsof PC1.
PC1 knockdown also inhibited tubulogenesis of murine inner medullarycollecting duct cells in three-dimensional collagen culture(L.B. and G.L.G., data not shown). Despite that similar effectsare observable in various species, the use of culture-adaptedcell lines, which may carry unknown genetic alterations, preventsfirm conclusions on the contribution of other mutational mechanismsto the onset of ADPKD. However, the specificity of PC1 suppressionby siRNA strongly suggests that the levels of PC1 expressionplay a fundamental role in the cystogenic process. Overall,our in vitro findings support the haploinsufficiency model forADPKD development in agreement with the observation of cystformation in PKD1 null heterozygotes (41) and a PKD1 hypomorphicmouse (42). Because the overexpression of PKD1 produces a cysticphenotype (43,44), it seems that PC1 needs to be expressed withina specific range to maintain normal renal tubule dimensions.These data are consistent with the notion that fluctuationsof the protein levels above or below the normal thresholds willunbalance its stoichiometric relationship with other interactingfactors leading to the cystogenic signal. These changes likelyoccur stochastically and may require the contribution of othergenetic and/or environmental events to establish the cystogenicprocess. Such a mechanism may explain the low frequency of cysticnephrons and the different onset and variable course of thedisease within members of a family who carry the same germlinemutation. Our results further indicate that the stable siRNA-mediatedPC1 knockdown represents an important complement to geneticstudies and may be uniquely suited for the comparative genomicand proteomic analyses to identify cystogenic pathways thatare triggered by the fluctuation in the expression of PC1.
Acknowledgments
This work was supported by the National Institutes of Healthgrants RO1DK63611 (G.L.G.) and P01DK 62345 (P.D.W., and X.L.)and from a research grant from the PKD Foundation (G.L.G.).
We are grateful to Debbie Hyinck (Mount Sinai School of Medicine)for kindly reviewing the manuscript and for the helpful discussion.We also are thankful to the personnel in the Mount Sinai FlowCytometry Shared Resource for help with cell sorting.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Fletcher J, Hu M, Berman Y, Collins F, Grigg J, McIver M, Juppner H, Alexander SI: Multicystic dysplastic kidney and variable phenotype in a family with a novel deletion mutation of PAX2.
J Am Soc Nephrol 16
: 2754
–2761, 2005[Abstract/Free Full Text]
Wilson PD: Polycystic kidney disease.
N Engl J Med 350
: 151
–164, 2004[Free Full Text]
Ong AC, Harris PC: Molecular pathogenesis of ADPKD: The polycystin complex gets complex.
Kidney Int 67
: 1234
–1247, 2005[CrossRef][Medline]
Nauli SM, Alenghat FJ, Luo Y, Williams E, Vassilev P, Li X, Elia AE, Lu W, Brown EM, Quinn SJ, Ingber DE, Zhou J: Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells.
Nat Genet 33
: 129
–137, 2003[CrossRef][Medline]
Yoder BK, Hou X, Guay-Woodford LM: The polycystic kidney disease proteins, polycystin-1, polycystin-2, polaris, and cystin, are co-localized in renal cilia.
J Am Soc Nephrol 13
: 2508
–2516, 2002[Abstract/Free Full Text]
Pazour GJ: Intraflagellar transport and cilia-dependent renal disease: The ciliary hypothesis of polycystic kidney disease.
J Am Soc Nephrol 15
: 2528
–2536, 2004[Abstract/Free Full Text]
Ong AC, Wheatley DN: Polycystic kidney disease: The ciliary connection.
Lancet 361
: 774
–776, 2003[CrossRef][Medline]
Kreidberg JA, Symons JM: Integrins in kidney development, function, and disease.
Am J Physiol Renal Physiol 279
: F233
–F242, 2000[Abstract/Free Full Text]
Guay-Woodford LM: Murine models of polycystic kidney disease: Molecular and therapeutic insights.
Am J Physiol Renal Physiol 285
: F1034
–F1049, 2003[Abstract/Free Full Text]
Montesano R, Matsumoto K, Nakamura T, Orci L: Identification of a fibroblast-derived epithelial morphogen as hepatocyte growth factor.
Cell 67
: 901
–908, 1991[CrossRef][Medline]
Montesano R, Schaller G, Orci L: Induction of epithelial tubular morphogenesis in vitro by fibroblast-derived soluble factors.
Cell 66
: 697
–711, 1991[CrossRef][Medline]
Fedorova E, Battini L, Prakash-Cheng A, Marras D, Gusella GL: Lentiviral gene delivery to CNS by spinal intrathecal administration to neonatal mice.
J Gene Med 8
: 414
–424, 2006[CrossRef][Medline]
Gusella GL, Fedorova E, Marras D, Klotman PE, Klotman ME: In vivo gene transfer to kidney by lentiviral vector.
Kidney Int 61
[Suppl 1]: 32
–36, 2002[CrossRef][Medline]
Henriet P, Zhong ZD, Brooks PC, Weinberg KI, DeClerck YA: Contact with fibrillar collagen inhibits melanoma cell proliferation by up-regulating p27KIP1.
Proc Natl Acad Sci U S A 97
: 10026
–10031, 2000[Abstract/Free Full Text]
Wang XQ, Frazier WA: The thrombospondin receptor CD47 (IAP) modulates and associates with alpha2beta1 integrin in vascular smooth muscle cells.
Mol Biol Cell 9
: 865
–874, 1998[Abstract/Free Full Text]
Folkman J, Moscona A: Role of cell shape in growth control.
Nature 273
: 345
–349, 1978[CrossRef][Medline]
Boletta A, Qian F, Onuchic LF, Bhunia AK, Phakdeekitcharoen B, Hanaoka K, Guggino W, Monaco L, Germino GG: Polycystin-1, the gene product of PKD1, induces resistance to apoptosis and spontaneous tubulogenesis in MDCK cells.
Mol Cell 6
: 1267
–1273, 2000[CrossRef][Medline]
Wilson PD, Geng L, Li X, Burrow CR: The PKD1 gene product, "polycystin-1," is a tyrosine-phosphorylated protein that colocalizes with alpha2beta1-integrin in focal clusters in adherent renal epithelia.
Lab Invest 79
: 1311
–1323, 1999[Medline]
Zuk A, Matlin KS: Apical beta 1 integrin in polarized MDCK cells mediates tubulocyst formation in response to type I collagen overlay.
J Cell Sci 109
: 1875
–1889, 1996[Abstract]
Saelman EU, Keely PJ, Santoro SA: Loss of MDCK cell alpha 2 beta 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[Abstract]
van Adelsberg J: Murine polycystic kidney epithelial cell lines have increased integrin-mediated adhesion to collagen.
Am J Physiol 267
: F1082
–F1093, 1994[Medline]
Hou X, Mrug M, Yoder BK, Lefkowitz EJ, Kremmidiotis G, D’Eustachio P, Beier DR, Guay-Woodford LM: Cystin, a novel cilia-associated protein, is disrupted in the cpk mouse model of polycystic kidney disease.
J Clin Invest 109
: 533
–540, 2002[CrossRef][Medline]
Arnaout MA, Mahalingam B, Xiong JP: Integrin structure, allostery, and bidirectional signaling.
Annu Rev Cell Dev Biol 21
: 381
–410, 2005[CrossRef][Medline]
Whittard JD, Craig SE, Mould AP, Koch A, Pertz O, Engel J, Humphries MJ: E-cadherin is a ligand for integrin alpha2beta1.
Matrix Biol 21
: 525
–532, 2002[CrossRef][Medline]
Huan Y, van Adelsberg J: Polycystin-1, the PKD1 gene product, is in a complex containing E-cadherin and the catenins.
J Clin Invest 104
: 1459
–1468, 1999[Medline]
Roitbak T, Ward CJ, Harris PC, Bacallao R, Ness SA, Wandinger-Ness A: A polycystin-1 multiprotein complex is disrupted in polycystic kidney disease cells.
Mol Biol Cell 15
: 1334
–1346, 2004[Abstract/Free Full Text]
Praetorius HA, Praetorius J, Nielsen S, Frokiaer J, Spring KR: Beta1-Integrins in the primary cilium of MDCK cells potentiate fibronectin-induced Ca2+ signaling.
Am J Physiol Renal Physiol 287
: F969
–F978, 2004[Abstract/Free Full Text]
Santos OF, Nigam SK: HGF-induced tubulogenesis and branching of epithelial cells is modulated by extracellular matrix and TGF-beta.
Dev Biol 160
: 293
–302, 1993[CrossRef][Medline]
Bao Q, Hughes RC: Galectin-3 expression and effects on cyst enlargement and tubulogenesis in kidney epithelial MDCK cells cultured in three-dimensional matrices in vitro.
J Cell Sci 108
: 2791
–2800, 1995[Abstract]
Lu W, Peissel B, Babakhanlou H, Pavlova A, Geng L, Fan X, Larson C, Brent G, Zhou J: Perinatal lethality with kidney and pancreas defects in mice with a targeted Pkd1 mutation.
Nat Genet 17
: 179
–181, 1997[CrossRef][Medline]
Lantinga-van Leeuwen IS, Dauwerse JG, Baelde HJ, Leonhard WN, van de Wal A, Ward CJ, Verbeek S, Deruiter MC, Breuning MH, de Heer E, Peters DJ: Lowering of Pkd1 expression is sufficient to cause polycystic kidney disease.
Hum Mol Genet 13
: 3069
–3077, 2004[Abstract/Free Full Text]
Pritchard L, Sloane-Stanley JA, Sharpe JA, Aspinwall R, Lu W, Buckle V, Strmecki L, Walker D, Ward CJ, Alpers CE, Zhou J, Wood WG, Harris PC: A human PKD1 transgene generates functional polycystin-1 in mice and is associated with a cystic phenotype.
Hum Mol Genet 9
: 2617
–2627, 2000[Abstract/Free Full Text]
Thivierge C, Kurbegovic A, Couillard M, Guillaume R, Cote O, Trudel M: Overexpression of PKD1 causes polycystic kidney disease.
Mol Cell Biol 26
: 1538
–1548, 2006[Abstract/Free Full Text]
This article has been cited by other articles:
L. Battini, S. Macip, E. Fedorova, S. Dikman, S. Somlo, C. Montagna, and G. L. Gusella Loss of polycystin-1 causes centrosome amplification and genomic instability
Hum. Mol. Genet.,
September 15, 2008;
17(18):
2819 - 2833.
[Abstract][Full Text][PDF]
2
1–Mediated Anoikis Resistance
Top
Abstract
Introduction
2
1 and reduced apoptosis in cells grown on collagen type I. The engagement of integrin-
, deletion in 3'LTR. (B) Polycystin-1 (PC1) mRNA was detected by Northern analysis on total RNA that was isolated from the indicated cell populations using a probe that was specific for PKD1 (see Materials and Methods). The membrane then was stripped and rehybridized to a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe to control for equal sample loading (bottom). (C) Expression of PC1 was detected in total cell lysate from confluent cell cultures using the C20 antibody (Santa Cruz Biotechnology) specific for the PC1 carboxyl terminus. The position of molecular weight markers (Kdal) and PC1 are indicated. The approximately 150-Kdal band that was observed in all of the cell lines represents a nonspecific signal. The bars in the bottom panel represent the densitometric analysis that was performed on the radiographic film using the ImageJ software. The ratios of PC1 to actin expression are shown as percentage of that in parental MDCK cells. The expression of actin (indicated) was determined as control of equal loading after stripping of the membranes above and immunoblotting with an anti-actin antibody (dilution 1:20,000; Chemicon). (D) Expression of PC1 (white arrowhead) in individual clones that were derived from the control MDCK/E/siLuc (3L, 4L, 5L, 6L, and 7L) and PC1 knockdown MDCK/E/siPKD (2P, 5P, 8P, 9P, and 10P) polyclonal cell populations. Actin was used as loading control. The bars below the lanes indicate the densitometric analysis of the expression of PC1 relative to the correspondent actin signal in the control (
) and PC1 knockdown (
) clones. MDCK/E/siLuc and MDCK/E/siPKD indicate the control and PC1 polyclonal populations, respectively.
) and the flattened epithelial lining the cyst (arrowhead) are indicated. (C) The number of cysts versus tubules was scored in control and PC1 knockdown cells after 1 wk of culture in three-dimensional collagen gels in HGF. The bars indicate the average number of cysts and tubules ± SD for each cell line after evaluation of five random fields. **P < 0.01 cyst or tubule count in PC1 knockdown cells versus controls. Original magnification, x200.
