CD44 Deficiency Increases Tubular Damage But Reduces Renal Fibrosis in Obstructive Nephropathy
Kasper M.A. Rouschop*,
Miguel E. Sewnath,
Nike Claessen*,
Joris J.T.H. Roelofs*,
Inge Hoedemaeker*,
Ronald van der Neut*,
Jan Aten*,
Steven T. Pals*,
Jan J. Weening* and
Sandrine Florquin*
Departments of *Pathology and Surgery, Academic Medical Center, Amsterdam, The Netherlands.
Correspondence to Dr. Kasper Rouschop, Department of Pathology, Academic Medical Center, P.O. Box 22660, 1100 DD, Amsterdam, The Netherlands. Phone: +31-20-5665653; Fax: 31-20-6960389; E-mail: k.m.rouschop{at}amc.uva.nl
ABSTRACT. CD44 is a glycoprotein involved in inflammation andcell-cell/cell-matrix interactions. CD44 is upregulated in thekidney upon injury; however, its role in the pathogenesis ofrenal damage and fibrosis remains largely unknown. The authorsshow that mice lacking CD44 developed more tubular damage, associatedwith decreased proliferation and increased apoptosis of tubularepithelial cells, but less renal fibrosis after unilateral ureteralobstruction. In addition, impaired influx of macrophages anddecreased accumulation of myofibroblasts was observed in theobstructed kidney of CD44-/- mice compared with CD44+/+ mice.Hepatocyte growth factor (HGF) and transforming growth factor-1(TGF-1) exert reciprocal functions in the progression of renaldiseases and interact with CD44 in vitro. For the first time,the authors establish diminished HGF-signaling, via its highaffinity receptor c-Met, in the absence of CD44 in vivo. Inparallel, the signaling of TGF-1 reflected by the relative phosphorylationand nuclear translocation of Smad-2 and Smad-3 was reduced inthe obstructed kidney of CD44-/- mice. In conclusion, CD44 exertsprotective effects on tubuli but contributes to renal fibrogenesisat least in part through enhancement of HGF and TGF-1 signalingpathway in obstructive nephropathy.
Tubulointerstitial injury is a common finding in the chronicallydiseased kidney and is the main predictor for the progressionto end-stage renal disease. Progression of renal diseases ischaracterized by tubular damage, macrophage infiltration, accumulationof myofibroblasts, and renal fibrosis. One of the moleculesthat may orchestrate this cascade is CD44. Under normal conditions,CD44 is hardly expressed in the kidney except for passengerleukocytes (1,2). However, in inflammatory renal diseases, CD44expression is markedly enhanced, particularly in crescents andinjured tubuli as documented in human diseases and in severalanimal models (1,3–5). Altogether, these observationssuggest a central but still unknown role for CD44 in renal injury.
CD44 family glycoproteins are encoded by a single gene consistingof 19 exons. By alternative splicing, different isoforms canbe generated (6,7). These isoforms have been implicated in manyimportant physiologic and pathologic processes, such as cell-celland cell-matrix interaction, lymphocyte extravasation, woundhealing/scarring, cell migration, lymphocyte activation, andbinding/presentation of growth factors (8–11).
Hyaluronic acid (HA) and osteopontin are the major ligands ofCD44 (9,12). HA is a glycosaminoglycan of the extracellularmatrix, which markedly accumulates in the kidney cortex uponinjury and may undergo degradation into low-molecular weightproducts (3,13) that exert proinflammatory effects (14,15).Interestingly, HA fragments accumulate in the absence of CD44at the site of injury (16), suggesting a role for CD44 in theclearance of HA. Osteopontin, the second major ligand of CD44,promotes accumulation of macrophages, decreases renal cell apoptosisand participates in the regeneration of tubular epithelial cells(TEC) upon renal injury (17,18).
The CD44-variant containing variable-exon 3 (CD44v3) is capableof binding growth factors at its attached heparan sulfate-chainand presents these factors to their high-affinity receptors(19,20). Due to binding to heparan sulfate, a local gradientis created that facilitates cross-linking of growth factor receptors(10,21). One of the heparan sulfate-binding growth factors thatexerts potent renoprotective actions is hepatocyte growth factor(HGF) (22–25). In a B cell line it was shown that CD44v3binds HGF and presents it to its high-affinity receptor c-Met(11).
CD44 is also implicated in the activation and signaling of theprofibrotic agent, transforming growth factor-1 (TGF-1). Afterbinding to CD44, the matrix metalloproteinase-9 (MMP-9) is ableto cleave pro-TGF-1 into its active form (26). Furthermore,upon binding with HA, CD44 interacts with TGF- receptor I, therebyenhancing TGF-1 signaling (27).
These results suggest an important, yet unknown role for CD44in renal injury. Our results reveal that CD44 disruption leadsto increased tubular injury but decreased renal fibrosis duringobstructive nephropathy.
Mice and Experimental Protocol
Mice, CD44 knockout on C57Bl/6 background (CD44-/-) (28) andC57Bl/6 wild-type (CD44+/+) origin were bred in our animal facility.Right kidney unilateral ureteral obstruction (UUO) or sham surgerywas performed under general anesthesia (0.07 ml/10 g mouse ofFFM mixture, containing: 1.25 mg/ml midazolam [Roche, Mijdrecht,The Netherlands], 0.08 mg/ml fentanyl citrate, and 2.5 mg/mlfluanisone [Janssen Pharmaceutica, Beerse, Belgium]) on 6- to8-wk-old male mice. The right ureter was ligated with 6-0 silkand all mice received postoperative analgesia (0.15 mg/kg buprenorfine,subcutaneously; Shering-Plough, Brussels, Belgium). Sham-operatedmice underwent the same procedure without the ligation of oneureter. To mark proliferating cells, 5-bromo-2'-deoxyuridine(BrdU; Sigma Chemical Co, St. Louis, MO) was injected intraperitoneally(50 mg/kg body wt) 1 h before sacrifice. Mice (n = 6 per group)were sacrificed 1, 3, 7, and 14 d. All experimental procedureswere approved by the Animal Care and Use Committee of the Universityof Amsterdam, the Netherlands.
Antibodies
Rat IgG2b anti-CD44 was obtained from concentrated supernatantof the hybridoma IM 7.8.1 (ATCC, Livermore, CA). Goat anti-osteopontinwas purchased from R&D Systems (Abingdon, UK), the biotinylatedHA binding protein from Calbiochem (Darmstadt, Germany), anti--actinand anti-BrdU antibodies from Sigma, anti-active caspase 3 fromCell Signaling Technology (Beverly, MA), anti-F4/80 antibodiesfrom Serotec (Oxford, UK), anti-c-Met (B2) from Santa Cruz Biotechnology(Santa Cruz, CA), anti-phospho c-Met (pYpYpY1230/1234/1235)from Biosource International (Nivelles, Belgium), anti--smoothmuscle action (SMA) and all HRP-labeled secondary antibodiesfrom DAKO (Glostrup, Denmark), anti–TIMP-1 and anti–TIMP-2from Oncogene (Cambridge, UK), and anti-Smad-2/3 antibodiesfrom BD Pharmingen (San Diego, CA). The anti-phospho-Smad-2and –3 antibodies used for Western blotting were a kindgift of P. ten Dijke (NKI, Amsterdam, the Netherlands). Theanti-phospho-Smad-2/3 antibody used for immunohistochemistrywas purchased from Santa Cruz Biotechnology.
Histology and Immunohistochemistry
Renal tissues were fixed in 10% formalin for 12 h and embeddedin paraffin in a routine fashion. Four micrometer sections werestained with hematoxylin and eosin (H&E), periodic acidSchiff (PAS-D), or Sirius Red. For detection of CD44, osteopontin,macrophages, and apoptosis, antigen retrieval was performedby microwave treatment. To detect BrdU, DNA was denatured in2 N HCl, and antigen retrieval was performed by 0.4% pepsin(Sigma). Immunostainings were performed in accordance with standardprocedures. The slides were counterstained with methyl green(Sigma). Anti-phospho c-Met and anti-phospho Smad-2 and –3were stained using frozen sections.
Histopathological Scoring
All histopathological scorings were made in the cortex and performedin a blinded fashion. Tubular injury was assessed by gradingtubular dilatation, epithelial simplification and brush borderloss in ten randomly chosen, non-overlapping fields (x200 magnification).Lesions were graded on a scale from 0 to 4: 0 = normal; 1 =mild, involvement of less than 25% of the cortex; 2 = moderate,involvement of 25 to 50% of the cortex; 3 = severe, involvementof 50 to 75% of the cortex; 4 = extensive damage involving morethan 75% of the cortex. To evaluate the number of proliferatingTEC, BrdU-positive TEC were counted for a total of 1000 TEC.The number of active caspase-3–positive apoptotic tubularcells, macrophages, and TEC positive for nuclear p-Smad-2/3were counted in ten non-overlapping fields.
Osteopontin was expressed as the percentage of positive tubuli.An area of 10 mm2 was analyzed for HA and Sirius Red using adigital image analysis program (Image pro-plus; Mediacybernetics,Germany). Results are expressed as a percentage of the analyzedcortex.
Real-Time Quantitative RT-PCR and Conventional RT-PCR
Total RNA was isolated from frozen kidney cortex or microdissectedtubuli sections (performed with a PALM laser-microbeam system;PALM GmbH, Bernried, Germany) using Trizol reagent (Life Technologies,Breda, The Netherlands). cDNA was synthesized using anchored5'(dT)14-d(A/G/C)-d(A/G/C/T)-3' primers. To exclude genomic-DNAamplification, RNA samples were analyzed without RT-procedure.Real-time RT-PCR was performed on a LightCycler system (RocheDiagnostics, Almere, the Netherlands) using FastStart DNA MasterSYBR Green I reagent (Roche). Specific primers (synthesizedby Sigma-Genosys, Cambridgeshire, UK) for CD44-pan, CD44v3,collagen IV, and house-keeping gene TATA-box binding protein(TBP) were designed and are listed in Table 1. To adjust forvariable input, values were corrected for TBP mRNA. Values areexpressed as x-fold upregulation (obstructed kidney versus contralateralkidney).
Total Collagen Assay
Hydroxyproline concentrations in hydrolyzed (6 M HCL, 110°C,12 h) accurately weighed frozen kidney samples were chemicallymeasured according to the method of Kivirikko et al. (29). Totalcollagen was assumed to contain 12.7% hydroxyproline by weight,and final results were expressed as µg of collagen/mgof kidney weight
Gelatin Zymographic Analysis
Frozen tissue was sonicated in extraction buffer (10 mM NaCacodilate,1 M NaCl, 0.1% Triton, 1 µM ZnCl2, and 0.1 mg/ml NaN3).Equal quantities of protein were loaded onto a 10% polyacrylamidegel containing 1% gelatin (Bloom 225, Sigma) next to a proteinmarker. To induce MMP activity, gels were incubated overnightin a buffer containing 50 mM Tris, 5 mM CaCl2, and 1% Triton,pH 7.5. To visualize MMP, activity gels were stained with CoomassieBrilliant Blue and subsequently destained.
HGF and TGF-1 ELISA
Kidney cortex was sonicated in PBS containing 1% Triton, 1 mMEDTA, and 1% protease inhibitor cocktail (P8340, Sigma). Kidneytissue HGF levels were assayed by two-site ELISA using a mouseanti-HGF antibody (R&D Systems) and a goat anti-HGF (R&DSystems) in accordance with standard procedures. Activated TGF-1was determined using a Quantikine TGF-1 ELISA kit in accordancewith the protocol of the manufacturer (R&D systems). Activationof latent TGF-1 was done by incubation kidney lysates with anequal volume of 2.5 N acetic acid/10 M urea
Immunoblotting
Kidney cortex lysates were prepared as described for the ELISAwith addition of 1 mM sodium orthovanadate (Sigma). Samples(20-µg proteins) were separated by SDS-PAGE and transferredonto activated PVDF membranes (Millipore, Etten Leur, The Netherlands).Membranes were blocked with either 5% (wt/vol) nonfat dry milkin Tris-buffered saline containing 0.1% Tween (TBS-T) (-SMA,c-Met, Smad-2/3, phospho Smad-3, -actin) or 5% bovine serumalbumin (Sigma) in TBS-T (phospho c-Met and phospho Smad-2).The blots were probed with primary antibody followed by incubationwith HRP-conjugated secondary antibody. HRP activity was visualizedwith ECL-reagent (Amersham Pharmacia Biotech, Roosendaal, TheNetherlands). Densitometric quantification analysis was conductedon directly scanned images using National Institutes of HealthImage 1.62 for Macintosh software.
Statistical Analyses
All data were analyzed by comparison with unpaired t test, exceptfor tubular injury, which was analyzed by using a nonparametricMann-Whitney U test.
Expression of CD44 Is Induced by UUO
As assessed by real-time RT-PCR, a strong upregulation of CD44mRNA was observed in obstructed kidneys that peaks 7 d afterUUO (Figure 1A). To localize CD44 expression, immunostainingswere performed. One day after obstruction, CD44 protein washardly detectable, except for a few passenger leukocytes (Figure 1B).Three days after UUO, CD44 expression was increased ontubuli and capillary endothelial cells (Figure 1C). At 7 d (Figure 1D)and 14 d (Figure 1E) after UUO, the expression of CD44 waseven more pronounced.
Figure 1. De novo CD44-expression after unilateral ureteral obstruction (UUO). Quantitative real-time PCR for CD44 presented as x-fold increase of CD44 transcripts, corrected for the number of TATA-box binding protein (TBP) transcripts, of the obstructed kidney versus the contralateral kidney. Sham-operated CD44+/+ (hatched bars) and UUO CD44+/+ (white bars), presented as mean ± SEM, n = 6 (A). Immunostaining for pan-CD44 on CD44+/+ kidneys, 1 d (B), 3 d (C), 7 d (D), and 14 d (E) after UUO. Magnification, x100; representative for n = 6.
Increased Tubular Injury in CD44-/- Obstructed Kidneys
To study the physiologic role of de novo expression of CD44in UUO, we compared renal injury in CD44-/- and CD44+/+ mice.Tubular damage was significantly more severe in CD44-/- comparedwith CD44+/+ obstructed kidneys at all time points (Figure 2A).Increased tubular damage in CD44-/- kidneys was associated withdecreased proliferation (Figure 2B) and increased apoptosisof CD44-/- compared with CD44+/+ TEC (Figure 2C).
Figure 2. Tubular damage after UUO. (A) Semiquantitative scoring of tubular injury revealed more damage in CD44-/- compared with CD44+/+ and representative picture of tubular lesions at day 3 after UUO (periodic acid Schiff [PAS-D] staining; magnification, x200). (B) Tubular epithelial cells (TEC) proliferation as assessed by the number of 5-bromo-2'-deoxyuridine (BrdU)–positive nuclei per 1000 TEC showing more proliferation of TEC in CD44+/+ compared with CD44-/- and representative immunostaining for BrdU 3 d after UUO (magnification, x200). (C) More apoptotic TEC in CD44-/- compared with CD44+/+ were counted, and representative immunostaining of anti-active caspase 3 (magnification, x400). White bars represent CD44+/+, and black bars represent CD44-/-. Mean ± SEM, n = 6.
Decreased Macrophage Influx in CD44-/- Obstructed Kidney
At all time points after UUO, the influx of macrophages wassignificantly impaired in the CD44-/- compared with CD44+/+mice (Figure 3, A and B).
Figure 3. Macrophages infiltration after UUO. (A) Immunostaining for macrophages revealed impaired influx of macrophages into CD44-/- compared with CD44+/+ kidneys 3 d after UUO (magnification, x100). (B) Data are presented as number of macrophages/mm2 cortex. White bars represent CD44+/+, and black bars represent CD44-/-. Mean ± SEM, n = 6.
Hyaluronic Acid and Osteopontin Expression in CD44-/- and CD44+/+ Obstructed Kidneys
Because HA and osteopontin are the principal ligands of CD44and promote inflammation, we analyzed HA and osteopontin expressionby immunohistochemistry. Interstitial HA-positive areas expandedin the obstructed kidneys of both genotypes (Figure 4A), reachingup to a tenfold increase relative to the contralateral kidneysby day 7 after UUO (Figure 4B). Interestingly, the increasein HA was significantly higher in the CD44-/- mice comparedwith the CD44+/+ mice. In contrast, osteopontin expression wascomparable in CD44-/- and CD44+/+ obstructed kidneys (Figure 5, A and B).
Figure 4. Hyaluronic acid (HA) accumulation after UUO. (A) Representative microphotographs of HA staining (magnification, x200) revealed accumulation of HA in CD44+/+ mice (left panel) and CD44-/- (right panel) mice 7 d after UUO. (B) Quantification by digital image analysis revealed less HA accumulation in CD44+/+ (white bars) than in CD44-/- (black bars) kidneys. Mean ± SEM, n = 6.
Figure 5. Osteopontin expression after UUO. (A) Immunostaining for osteopontin revealed comparable expression of osteopontin in renal cortex after UUO in CD44+/+ and CD44-/- mice (magnification, x100). (B) The number of positive tubuli were counted and expressed as percentage of the total. White bars represent CD44+/+, and black bars represent CD44-/-. Mean ± SEM, n = 6.
Attenuation of Renal Fibrosis in CD44-/- Obstructed Kidney
As expected, total kidney collagen increased in time in CD44+/+mice in response to UUO. In sharp contrast, the increase incollagen deposition was attenuated in CD44-/- mice (Figure 6A).These data were confirmed by Sirius Red staining (Figure 6, B and C).
Figure 6. CD44 deficiency diminishes collagen accumulation. Hydroxyproline assay (A) and quantitative analysis of picro Sirius Red staining (B) of kidneys from CD44+/+ (white bars) and CD44-/- (black bars) mice. Data are presented as mean ± SEM, n = 6. (C) Representative micrographs of picro Sirius Red staining in CD44+/+ and CD44-/- renal cortex 14 d after UUO (magnification, x200).
Myofibroblasts play an important role in interstitial fibrosis;therefore, we followed the accumulation of -SMA+ cells. Accumulationof myofibroblasts was delayed in CD44-/- mice compared withCD44+/+ mice as quantified by digital image analysis (Figure 7A).This was confirmed by Western blot analysis, showing less-SMA in renal homogenates of CD44-/-versus CD44+/+ obstructedkidneys at day 1 and day 3 (Figure 7B).
Figure 7. CD44 deficiency delays accumulation of myofibroblasts. (A) Representative picture of –smooth muscle action (-SMA) immunostaining in CD44+/+ and CD44-/- renal cortex 3 d after UUO (magnification, x100) and digital analysis of CD44+/+ (white bars) and CD44-/- (black bars) renal cortex. Data are presented as mean ± SEM, n = 6. (B) Western blotting for -SMA of CD44+/+ and CD44-/- renal cortex; blots are representative for n = 6.
To determine whether the absence of fibrosis in CD44-/- obstructedkidneys was caused by differences in synthesis or degradationof collagen, quantitative real-time RT-PCR was performed forcollagen type IV, which transcripts were comparable in CD44+/+and CD44-/- mice (Figure 8A). We further assessed MMP activityin renal homogenates by zymography. A marked but similar inductionof MMP-2 and MMP-9 activity was observed in CD44+/+ and CD44-/-obstructed kidneys (Figure 8B). To further assess the MMP’scapacity to degrade collagens, we determined the level of tissueinhibitors of MMP (TIMP) by immunoblotting. TIMP-1, the inhibitorwith the highest affinity for MMP-9, was less present at days7 and 14 in CD44-/- compared with CD44+/+ obstructed kidney(Figure 8C). The level of TIMP-2, with the highest affinityfor MMP-2, was comparable in both groups (Figure 8C).
Figure 8. Synthesis and degradation of collagen after UUO. (A) Quantitative real-time PCR for collagen type IV in renal homogenates showed no difference between CD44+/+ (white bars) and CD44-/- (black bars). (B) Gelatin zymography for matrix metalloproteinase (MMP) activity showed comparable activities of MMP-2 and MMP-9 in CD44+/+ and CD44-/- after UUO; zymographic activity was quantified by digital analysis. (C) Western blotting for tissue inhibitors of MMP, TIMP-1 and TIMP-2, on renal cortex of CD44+/+ and CD44-/- mice; data are quantified by densitometric analysis and corrected for loading differences (-actin). All data are presented as mean ± SEM, n = 6.
CD44 Facilitates HGF-Signaling In Vivo
Because signaling by HGF is facilitated by the expression ofthe v3 isoform of CD44 (CD44v3) (11), we determined mRNA levelsof CD44v3 in the CD44+/+ mice by quantitative real-time PCR(Figure 9A). Upon obstruction, mRNA levels of CD44v3 increased,starting at day 1 to culminate after 3 d, indicating its potentialrole early after obstruction. To get insight in the cellularlocalization of CD44v3, tubuli were microdissected and conventionalPCR for CD44v3 mRNA performed. As shown in Figure 9B, CD44v3mRNA was clearly present in tubuli 3 d after obstruction.
Figure 9. CD44-variant containing variable-exon 3 (CD44v3) mRNA after UUO. (A) Quantitative real-time PCR data are presented as x-fold increase of CD44 transcripts, corrected for the number of TBP transcripts of the obstructed kidney versus the contralateral kidney; sham-operated CD44+/+ mice (hatched bars) and CD44+/+ after UUO (white bars). Mean ± SEM, n = 6. (B) Conventional RT-PCR was performed on microdissected tubuli obtained 3 d after obstruction. Total kidney cDNA of day 3 after UUO was used as positive control; total kidney cDNA of sham-operated kidneys was used as negative control.
The HGF level of the obstructed CD44+/+ kidneys decreased after7 d and was nearly undetectable at day 14. In obstructed CD44-/-kidneys, a clear increase in HGF was observed after 3 d followedby a rapid decrease (Figure 10A).
Figure 10. Hepatocyte growth factor (HGF) levels and signaling after UUO. (A) Kidney HGF levels were determined by ELISA and corrected for quantity of protein. (B) Western blot analysis of phosphorylated c-Met revealed less activation of the HGF receptor in CD44-/- kidneys compared with CD44+/+. The upper panels are probed with an antibody directed against phospho-c-Met; blots were stripped and reprobed with an antibody directed against c-Met (middle panels). In each panel, two bands are visualized: the upper band represents pre-c-Met, and the lower band represents c-Met. ß-actin was probed as loading control (lower panels). (C) Blots are quantified by densitometric analysis and corrected for expression of c-Met. White bars represent CD44+/+ kidneys, and black bars represent CD44-/- kidneys. Mean ± SEM, n = 6. (D) Representative immunostainings for phospho-c-Met of CD44+/+ and CD44-/- renal cortex 3 d after obstruction (magnification, x100).
To study whether CD44 expression could enhance HGF signalingvia its high-affinity receptor, c-Met, expression and phosphorylationof the receptor were determined. c-Met expression increasedas obstruction continued (data not shown). Initially (data notshown) and 1 d after obstruction, c-Met expression is more pronouncedin the CD44-/- kidney compared with the CD44+/+ kidney (Figure 10B).After 3 d, no difference in expression of c-Met was observedbetween CD44+/+ and CD44-/- obstructed kidneys, but the capacityof HGF to activate c-Met was less efficient in CD44-/- thanin CD44+/+ mice (Figure 10, B and C). After 7 d of obstruction,phosphorylation of c-Met was hardly detectable in both groups(data not shown). To obtain insight into the site of c-Met phosphorylation,immunostainings for phospho-c-Met were performed that showeddiffuse positive tubuli in CD44+/+ obstructed kidney but onlya few positive tubuli in CD44-/- obstructed kidneys (Figure 10D).
CD44 Contributes to TGF-1 Signaling In Vivo
Despite the difference in TGF-1 levels in CD44+/+ and CD44-/-obstructed kidneys (Figure 11, A and B) and the higher expressionof Smad-2 and Smad-3 (the major signaling molecules of TGF-1)in CD44-/- kidneys, comparable phosphorylation of Smad-2 andSmad-3 was observed in CD44+/+ and CD44-/- obstructed kidneys(Figure 11C), suggesting an impaired TGF-1 signaling pathwayin CD44-/- mice. In addition, nuclear translocation of p-Smad-2/3,essential for TGF-1 signaling, was impaired in the absence ofCD44 (Figure 11D). Both observations suggest a crucial rolefor CD44 in TGF-1 signaling.
Figure 11. Transforming growth factor–1 (TGF-ß1) levels and Smad-2 and Smad-3 activation after UUO. Total TGF-ß1 (A) and activated TGF-ß1 (B) levels were quantified by ELISA and corrected for quantity of protein, CD44+/+ (white bars), and CD44-/- (black bars); data are presented as mean ± SEM, n = 6. (C) TGF-ß1 signaling was determined by Western blotting by assessment of the total levels of Smad-2 and Smad-3 and their phosphorylation; ß-actin was used as loading control. Blots were analyzed by densitometric analysis, the ratio of phosphorylated Smad-2 and Smad-3 versus Smad-2 and Smad-3 is depicted in the graphs. Data are presented as mean ± SEM, n = 6. (D) Representative immunostainings for phospho-Smad-2/3 of CD44+/+ and CD44-/- renal cortex 7 d after obstruction (magnification, x500); the percentage of TEC with positive nuclear staining for phospho Smad-2 and Smad-3 are depicted in the graph.
The decline of renal function in a variety of pathologic statesclosely correlates with the degree of tubulointerstitial damage.A cascade of events takes place during the progression of tubulointerstitiallesions, including release of cytokines/chemokines and growthfactors, expression of adhesion molecules, inflammatory infiltrate,renal epithelial cell damage, accumulation of myofibroblasts,and finally fibrosis. In this study, we show that CD44 playsa crucial role in this cascade.
First, we show by quantitative real-time PCR that CD44 mRNApeaked at day 7 after UUO and by immunohistochemistry that CD44is predominantly expressed by damaged tubuli and inflammatorycells. This is in agreement with previous studies reportingthe expression of CD44 in different models of kidney diseases(2–4) and in human nephropathies (1,5).
Second, we demonstrate that CD44 expression decreases tubularinjury as a consequence of increased tubular proliferation anddecreased tubular apoptosis. In vitro, CD44 has been implicatedin cell proliferation (30) and apoptosis (31). CD44 expressionmay promote the maintenance of tubular cell viability in responseto renal injury, because cell-cell and cell-matrix interactionsmight be facilitated by CD44 (4). Moreover, ligand-receptorinteraction of osteopontin with CD44 induces proliferation anddecreases apoptosis of TEC (17,18). The survival signal inducedby CD44-osteopontin interaction involves activation of the phosphatidylinositol3-kinase/Akt signaling pathway (32), which is also used by HGF,a well-known renoprotective molecule (22,33). In vitro studiesrevealed that CD44v3 binds HGF and presents it to its high-affinityreceptor c-Met (11). In this study, we observed for the firsttime a role for CD44 in HGF signaling in vivo because phosphorylationof c-Met was attenuated in obstructed CD44-/- kidneys. Altogetherfacilitation of HGF signaling by CD44 is very likely to contributeto the preservation of tubuli in our model.
Third, we observed that the two major ligands of CD44, HA, andosteopontin are highly expressed in obstructed kidneys. Interactionsof CD44 with HA and osteopontin play an important role in renalinflammation (3,34,35). Interaction of CD44 with HA is a crucialstep in the rolling of leukocytes along endothelial cells (36,37).Lack of CD44 was associated with increased HA accumulation inobstructed kidneys. This is in agreement with the study of Tederet al. (16), in which HA accumulated in a model of pulmonaryfibrosis as a consequence of CD44 deficiency, suggesting a crucialrole for CD44 in the homeostasis of HA. Osteopontin binds tomacrophages and mediates their adhesion and migration (18,38).Absence of functional interaction between CD44 and both HA andosteopontin results in less infiltration of macrophages in theCD44-/- compared with CD44+/+ obstructed kidneys. Although macrophagesare often considered to induce damage to the tubulointerstitialcompartment (39), the influx of macrophages is not correlatedto the severity of tubular damage in our model, which indicatesthat macrophage influx is not the main determinant of the tubularresponse leading to tubular cell loss. Macrophage clearancecapacity of apoptotic bodies is critical in resolution of inflammation.If apoptotic clearance capacity is exceeded, apoptotic cellsmay progress to secondary necrosis, resulting in the releaseof harmful cellular contents and in damage to the surroundingtissue. Lack of CD44 results in decreased clearance capacityof apoptotic bodies (16) and may therefore contribute to increasedtubular damage.
Fourth, our study reveals that renal fibrosis was clearly attenuated,despite the extensive tubular damage observed in CD44-/- miceafter UUO. This was associated with less myofibroblasts accumulationcompared with CD44+/+ mice. Although the differences in accumulationof myofibroblasts are only evident early in obstruction (days1 and 3), the earlier myofibroblast recruitment may contributeto the development of fibrosis in CD44+/+ obstructed kidneysas the obstruction continues. Because accumulation of extracellularmatrix may be caused by either increased synthesis, decreaseddegradation, or both, the levels of collagen type IV mRNA wereassessed by quantitative PCR, which mRNA was induced at thesame levels in CD44-/- and CD44+/+ obstructed kidneys. SinceMMP play a central role in the degradation of matrix proteins,we analyzed MMP activity by zymography. No differences in activitycould be detected between CD44+/+ and CD44-/- mice that couldexplain the striking difference in collagen accumulation. Thedifferences in collagen accumulation between CD44-/- and CD44+/+mice may be related to altered production of other collagens,such as collagen type III, differential expression of otherMMP, or altered expression of tissue inhibitors of MMP (TIMP)(40,41). Accordingly, Western blot analyses revealed enhancedexpression of TIMP-1, but not TIMP-2, in the CD44+/+ comparedwith CD44+/+ obstructed kidneys. Although TIMP-1 deficiencyhas not been shown to attenuate renal fibrosis (42), increasedTIMP-1 expression in the presence of CD44 is one of the possiblegenes that may contribute to the development of fibrosis. Thisdoes not rule out other factors involved in fibrogenesis.
TGF-1 is probably the most important pro-fibrotic agent duringprogression of renal disease. Indeed, TGF-1 induces myofibroblastictransition (43–45), promotes collagen type IV production(46), decreases MMP expression, and increases expression ofTIMP (47). Unexpectedly, regarding the absence of renal fibrosisin CD44-/- mice, higher levels of TGF-1 and activated TGF-1were observed in obstructed CD44-/- compared with CD44+/+ kidney.TGF-1 is secreted in a latent form and needs to be convertedinto an active form to exert its biologic activity. In vitro,TGF-1 is activated by a variety of mechanisms, but in vivo mechanismsare still not fully understood. A novel mechanism has been suggestedthat involves direct interaction of CD44 and MMP-9 to activateTGF-1 in vivo (48). In a model of pulmonary fibrosis, the lackof CD44 was found to lead to decreased levels of active TGF-1(16). Our findings in the obstructed kidney are at variancebecause we found higher levels of active TGF-1 in CD44-/- obstructedkidney compared with CD44+/+ (circa 60% in CD44-/-versus circa35% in CD44+/+ of total TGF-1). Thrombospondin-1 is also consideredto be a major activator of TGF-1 in vivo (49), yet thrombospondin-1expression was comparable in both types of mice (data not shown).Although higher levels of TGF-1 and Smad-2 and Smad-3 (two majorsignaling proteins of TGF-1) were found in CD44-/- renal homogenatescompared with CD44+/+, relatively less phosphorylation and nucleartranslocation (essential for TGF-1 signaling) of Smad-2 andSmad-3 was measured in CD44-/- compared with CD44+/+ obstructedkidneys. An essential role for CD44 in TGF-1 signaling in vivois suggested by the fact that, in the absence of CD44, moreSmad-2 and Smad-3 expression is required to obtain identicallevels of phosphorylated Smad-2 and Smad-3 and that, despiteequal levels of phosphorylated Smad-2 and Smad-3, less nucleartranslocation was observed in CD44-/- mice. Accordingly, Bourguignonet al. (27) recently showed that HA promotes signaling interactionbetween CD44 and TGF-RI receptor in metastatic mammary carcinoma.
In conclusion, de novo expression of CD44 in injured kidneyspromotes macrophage recruitment. In addition, expression ofCD44 protects TEC at least in part by enhancing survival signalsthrough its interaction with osteopontin and HGF. Moreover,expression of CD44 in the injured kidney contributes to thedevelopment of fibrosis at least in part through enhanced TGF-1signaling. Altogether, our data suggest that CD44 contributesto the delicate balance between HGF and TGF-1 in the progressionof renal disease (50).
Acknowledgments
We thank Dr. Jaklien Leemans for critically reviewing the manuscript,Joost Daalhuizen and Ingvild Kop for their excellent technicalhelp, and Wilfried Meun for photographic expertise. This researchwas funded by the Netherlands Organization for Scientific Researchand the Dutch Kidney Foundation.
Roy-Chaudhury P, Khong TF, Williams JH, Haites NE, Wu B, Simpson JG, Power DA: CD44 in glomerulonephritis: Expression in human renal biopsies, the Thy 1.1 model, and by cultured mesangial cells. Kidney Int 50: 272–281, 1996[Medline]
Benz PS, Fan X, Wuthrich RP: Enhanced tubular epithelial CD44 expression in MRL-lpr lupus nephritis. Kidney Int 50: 156–163, 1996[Medline]
Sibalic V, Fan X, Loffing J, Wuthrich RP: Upregulated renal tubular CD44, hyaluronan, and osteopontin in kdkd mice with interstitial nephritis. Nephrol Dial Transplant 12: 1344–1353, 1997[Abstract/Free Full Text]
Takazoe K, Foti R, Tesch GH, Hurst LA, Lan HY, Atkins RC, Nikolic-Paterson DJ: Up-regulation of the tumour-associated marker CD44V6 in experimental kidney disease. Clin Exp Immunol 121: 523–532, 2000[CrossRef][Medline]
Florquin S, Nunziata R, Claessen N, van den Berg FM, Pals ST, Weening JJ: CD44 expression in IgA nephropathy. Am J Kidney Dis 39: 407–414, 2002[CrossRef][Medline]
Screaton GR, Bell MV, Jackson DG, Cornelis FB, Gerth U, Bell JI: Genomic structure of DNA encoding the lymphocyte homing receptor CD44 reveals at least 12 alternatively spliced exons. Proc Natl Acad Sci USA 89: 12160–12164, 1992[Abstract/Free Full Text]
Dougherty GJ, Landorp PM, Cooper DL, Humphries RK: Molecular cloning of CD44R1 and CD44R2, two novel isoforms of the human CD44 lymphocyte "homing" receptor expressed by hemopoietic cells. J Exp Med 174: 1–5, 1991[Abstract/Free Full Text]
Weiss L, Slavin S, Reich S, Cohen P, Shuster S, Stern R, Kaganovsky E, Okon E, Rubinstein AM, Naor D: Induction of resistance to diabetes in non-obese diabetic mice by targeting CD44 with a specific monoclonal antibody. Proc Natl Acad Sci USA 97: 285–290, 2000[Abstract/Free Full Text]
Siegelman MH, DeGrendele HC, Estess P: Activation and interaction of CD44 and hyaluronan in immunological systems. J Leukoc Biol 66: 315–321, 1999[Abstract]
Jones M, Tussey L, Athanasou N, Jackson DG: Heparan sulfate proteoglycan isoforms of the CD44 hyaluronan receptor induced in human inflammatory macrophages can function as paracrine regulators of fibroblast growth factor action. J Biol Chem 275: 7964–7974, 2000[Abstract/Free Full Text]
van der Voort R, Taher TE, Wielenga VJ, Spaargaren M, Prevo R, Smit L, David G, Hartmann G, Gherardi E, Pals ST: Heparan sulfate-modified CD44 promotes hepatocyte growth factor/scatter factor-induced signal transduction through the receptor tyrosine kinase c-Met. J Biol Chem 274: 6499–6506, 1999[Abstract/Free Full Text]
Weber GF, Ashkar S, Glimcher MJ, Cantor H: Receptor-ligand interaction between CD44 and osteopontin (Eta-1). Science 271: 509–512, 1996[Abstract]
Johnsson C, Tufveson G, Wahlberg J, Hallgren R: Experimentally-induced warm renal ischemia induces cortical accumulation of hyaluronan in the kidney. Kidney Int 50: 1224–1229, 1996[Medline]
Sun LK, Beck-Schimmer B, Oertli B, Wuthrich RP: Hyaluronan-induced cyclooxygenase-2 expression promotes thromboxane A2 production by renal cells. Kidney Int 59: 190–196, 2001[Medline]
Teder P, Vandivier RW, Jiang D, Liang J, Cohn L, Pure E, Henson PM, Noble PW: Resolution of lung inflammation by CD44. Science 296: 155–158, 2002[Abstract/Free Full Text]
Noiri E, Dickman K, Miller F, Romanov G, Romanov VI, Shaw R, Chambers AF, Rittling SR, Denhardt DT, Goligorsky MS: Reduced tolerance to acute renal ischemia in mice with a targeted disruption of the osteopontin gene. Kidney Int 56: 74–82, 1999[CrossRef][Medline]
Ophascharoensuk V, Giachelli CM, Gordon K, Hughes J, Pichler R, Brown P, Liaw L, Schmidt R, Shankland SJ, Alpers CE, Couser WG, Johnson RJ: Obstructive uropathy in the mouse: role of osteopontin in interstitial fibrosis and apoptosis. Kidney Int 56: 571–580, 1999[CrossRef][Medline]
Bennett KL JD, Simon JC, Tanczos E, Peach R, Modrell B, Stamenkovic I, Plowman G, Aruffo A: CD44 isoforms containing exon V3 are responsible for the presentation of heparin-binding growth factor. J Cell Biol 128: 687–698, 1995[Abstract/Free Full Text]
Jackson DG BJ, Dickinson R, Timans J, Shields J, Whittle N: Proteoglycan forms of the lymphocyte homing receptor CD44 are alternatively spliced variants containing the v3 exon. J Cell Biol 128: 673–685, 1995[Abstract/Free Full Text]
van der Voort RKR, Beuling EA, Spaargaren M, Pals ST: Regulation of cytokine signaling by B cell antigen receptor and CD40-controlled expression of heparan sulfate proteoglycans. J Exp Med 2000 192: 1115–1124, 2000[Abstract/Free Full Text]
Kawaida K, Matsumoto K, Shimazu H, Nakamura T: Hepatocyte growth factor prevents acute renal failure and accelerates renal regeneration in mice. Proc Natl Acad Sci USA 91: 4357–4361, 1994[Abstract/Free Full Text]
Matsumoto K, Nakamura T: Hepatocyte growth factor: Renotropic role and potential therapeutics for renal diseases. Kidney Int 59: 2023–2038, 2001[Medline]
Mizuno S, Kurosawa T, Matsumoto K, Mizuno-Horikawa Y, Okamoto M, Nakamura T: Hepatocyte growth factor prevents renal fibrosis and dysfunction in a mouse model of chronic renal disease. J Clin Invest 101: 1827–1834, 1998[Medline]
Yu Q, Stamenkovic I: Localization of matrix metalloproteinase 9 to the cell surface provides a mechanism for CD44-mediated tumor invasion. Genes Dev 13: 35–48, 1999[Abstract/Free Full Text]
Bourguignon LY, Singleton PA, Zhu H, Zhou B: Hyaluronon promotes signaling interaction between CD44 and the TGF-phi RI receptor in metastatic breast tumor cells. J Biol Chem 277: 39703–39712, 2002[Abstract/Free Full Text]
Chen DMR, Zeytun A, Do Y, Lombard C, Robertson JL, Mak TW, Nagarkatti PS, Nagarkatti M: CD44-deficient mice exhibit enhanced hepatitis after concanavalin A injection: Evidence for involvement of CD44 in activation-induced cell death. J Immunol 166: 5889–5897, 2001[Abstract/Free Full Text]
Kivirikko KI, Laitinen O, Prockop DJ: Modifications of a specific assay for hydroxyproline in urine. Anal Biochem 19: 249–255, 1967[CrossRef][Medline]
Ahrens TSJ, Schempp CM, Howells N, Hofmann M, Ponta H, Herrlich P, Simon JC: Soluble CD44 inhibits melanoma tumor growth by blocking cell surface CD44 binding to hyaluronic acid. Oncogene 20: 3399–3408, 2001[CrossRef][Medline]
Yu Q, Toole BP, Stamenkovic I: Induction of apoptosis of metastatic mammary carcinoma cells in vivo by disruption of tumor cell surface CD44 function. J Exp Med 186: 1985–1996, 1997[Abstract/Free Full Text]
Lin YH, Yang-Yen HF: The osteopontin-CD44 survival signal involves activation of the phosphatidylinositol 3-kinase/Akt signaling pathway. J Biol Chem 276: 46024–46030, 2001[Abstract/Free Full Text]
Liu Y: Hepatocyte growth factor promotes renal epithelial cell survival by dual mechanisms. Am J Physiol 277: F624–F633, 1999
DeGrendele HC, Estess P, Siegelman MH: Requirement for CD44 in activated T cell extravasation into an inflammatory site. Science 278: 672–675, 1997[Abstract/Free Full Text]
Wuthrich RP: The proinflammatory role of hyaluronan-CD44 interactions in renal injury. Nephrol Dial Transplant 14: 2554–2556, 1999[Free Full Text]
Gal I, Lesley J, Ko W, Gonda A, Stoop R, Hyman R, Mikecz K: Role of the extracellular and cytoplasmic domains of CD44 in the rolling interaction of lymphoid cells with hyaluronan under physiologic flow. J Biol Chem 278: 11150–11158, 2003[Abstract/Free Full Text]
Xu H, Manivannan A, Liversidge J, Sharp PF, Forrester JV, Crane IJ: Involvement of CD44 in leukocyte trafficking at the blood-retinal barrier. J Leukoc Biol 72: 1133–1141, 2002[Abstract/Free Full Text]
Diamond JR, Kees-Folts D, Ricardo SD, Pruznak A, Eufemio M: Early and persistent up-regulated expression of renal cortical osteopontin in experimental hydronephrosis. Am J Pathol 146: 1455–1466, 1995[Abstract]
Main IW, Nikolic-Paterson DJ, Atkins RC: T cells and macrophages and their role in renal injury. Semin Nephrol 12: 395–407, 1992[Medline]
Oda T, Jung YO, Kim HS, Cai X, Lopez-Guisa JM, Ikeda Y, Eddy AA: PAI-1 deficiency attenuates the fibrogenic response to ureteral obstruction. Kidney Int 60: 587–596, 2001[CrossRef][Medline]
Duymelinck C, Dauwe SE, De Greef KE, Ysebaert DK, Verpooten GA, De Broe ME: TIMP-1 gene expression and PAI-1 antigen after unilateral ureteral obstruction in the adult male rat. Kidney Int 58: 1186–1201, 2000[CrossRef][Medline]
Kim H, Oda T, Lopez-Guisa J, Wing D, Edwards DR, Soloway PD, Eddy AA: TIMP-1 deficiency does not attenuate interstitial fibrosis in obstructive nephropathy. J Am Soc Nephrol 12: 736–748, 2001[Abstract/Free Full Text]
Fan JM, Ng YY, Hill PA, Nikolic-Paterson DJ, Mu W, Atkins RC, Lan HY: Transforming growth factor-beta regulates tubular epithelial-myofibroblast transdifferentiation in vitro. Kidney Int 56: 1455–1467, 1999[CrossRef][Medline]
Yang J, Liu Y: Dissection of key events in tubular epithelial to myofibroblast transition and its implications in renal interstitial fibrosis. Am J Pathol 159: 1465–1475, 2001[Abstract/Free Full Text]
Yang J, Liu Y: Blockage of tubular epithelial to myofibroblast transition by hepatocyte growth factor prevents renal interstitial fibrosis. J Am Soc Nephrol 13: 96–107, 2002[Abstract/Free Full Text]
Grande JP, Warner GM, Walker HJ, Yusufi AN, Cheng J, Gray CE, Kopp JB, Nath KA: TGF-beta1 is an autocrine mediator of renal tubular epithelial cell growth and collagen IV production. Exp Biol Med (Maywood) 227: 171–181, 2002[Abstract/Free Full Text]
Edwards DR, Leco KJ, Beaudry PP, Atadja PW, Veillette C, Riabowol KT: Differential effects of transforming growth factor-beta 1 on the expression of matrix metalloproteinases and tissue inhibitors of metalloproteinases in young and old human fibroblasts. Exp Gerontol 31: 207–223, 1996[CrossRef][Medline]
Yu Q, Stamenkovic I: Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis. Genes Dev 14: 163–176, 2000[Abstract/Free Full Text]
Crawford SE, Stellmach V, Murphy-Ullrich JE, Ribeiro SM, Lawler J, Hynes RO, Boivin GP, Bouck N: Thrombospondin-1 is a major activator of TGF-beta1 in vivo. Cell 93: 1159–1170, 1998[CrossRef][Medline]
Florquin S, Rouschop KM: Reciprocal functions of hepatocyte growth factor and transforming growth factor-beta1 in the progression of renal diseases: A role for CD44? Kidney Int Suppl 15–20, 2003
Received for publication July 28, 2003.
Accepted for publication December 12, 2003.
This article has been cited by other articles:
D. Vansthertem, N. Caron, A.-E. Decleves, S. Cludts, A. Gossiaux, D. Nonclercq, B. Flamion, A. Legrand, and G. Toubeau Label-retaining cells and tubular regeneration in postischaemic kidney
Nephrol. Dial. Transplant.,
December 1, 2008;
23(12):
3786 - 3797.
[Abstract][Full Text][PDF]
G. Stokman, J. C. Leemans, I. Stroo, I. Hoedemaeker, N. Claessen, G. J. D. Teske, J. J. Weening, and S. Florquin Enhanced mobilization of bone marrow cells does not ameliorate renal fibrosis
Nephrol. Dial. Transplant.,
February 1, 2008;
23(2):
483 - 491.
[Abstract][Full Text][PDF]
K. M. A. Rouschop, M. Sylva, G. J. D. Teske, I. Hoedemaeker, S. T. Pals, J. J. Weening, T. van der Poll, and S. Florquin Urothelial CD44 Facilitates Escherichia coli Infection of the Murine Urinary Tract
J. Immunol.,
November 15, 2006;
177(10):
7225 - 7232.
[Abstract][Full Text][PDF]
W. Kim, S.-O. Moon, S. Y. Lee, K. Y. Jang, C.-H. Cho, G. Y. Koh, K.-S. Choi, K.-H. Yoon, M. J. Sung, D. H. Kim, et al. COMP-Angiopoietin-1 Ameliorates Renal Fibrosis in a Unilateral Ureteral Obstruction Model
J. Am. Soc. Nephrol.,
September 1, 2006;
17(9):
2474 - 2483.
[Abstract][Full Text][PDF]
K. M.A. Rouschop, N. Claessen, S. T. Pals, J. J. Weening, and S. Florquin CD44 Disruption Prevents Degeneration of the Capillary Network in Obstructive Nephropathy via Reduction of TGF-beta1-Induced Apoptosis
J. Am. Soc. Nephrol.,
March 1, 2006;
17(3):
746 - 753.
[Abstract][Full Text][PDF]
J. J.T.H. Roelofs, K. M.A. Rouschop, J. C. Leemans, N. Claessen, A. M. de Boer, W. M. Frederiks, H.R. Lijnen, J. J. Weening, and S. Florquin Tissue-Type Plasminogen Activator Modulates Inflammatory Responses and Renal Function in Ischemia Reperfusion Injury
J. Am. Soc. Nephrol.,
January 1, 2006;
17(1):
131 - 140.
[Abstract][Full Text][PDF]
S. Weijer, C. W. Wieland, S. Florquin, and T. van der Poll A thrombomodulin mutation that impairs activated protein C generation results in uncontrolled lung inflammation during murine tuberculosis
Blood,
October 15, 2005;
106(8):
2761 - 2768.
[Abstract][Full Text][PDF]
K. M. A. Rouschop, J. J. T. H. Roelofs, A. T. Rowshani, J. C. Leemans, T. van der Poll, I. J. M. ten Berge, J. J. Weening, and S. Florquin Pre-transplant plasma and cellular levels of CD44 correlate with acute renal allograft rejection
Nephrol. Dial. Transplant.,
October 1, 2005;
20(10):
2248 - 2254.
[Abstract][Full Text][PDF]
J. W. A. M. Celie, E. D. Keuning, R. H. J. Beelen, A. M. Drager, S. Zweegman, F. L. Kessler, R. Soininen, and J. van den Born Identification of L-selectin Binding Heparan Sulfates Attached to Collagen Type XVIII
J. Biol. Chem.,
July 22, 2005;
280(29):
26965 - 26973.
[Abstract][Full Text][PDF]
K. M. A. Rouschop, J. J.T.H. Roelofs, N. Claessen, P. d. C. Martins, J.-J. Zwaginga, S. T. Pals, J. J. Weening, and S. Florquin Protection against Renal Ischemia Reperfusion Injury by CD44 Disruption
J. Am. Soc. Nephrol.,
July 1, 2005;
16(7):
2034 - 2043.
[Abstract][Full Text][PDF]
A. Krettek, G. K. Sukhova, U. Schonbeck, and P. Libby Enhanced Expression of CD44 Variants in Human Atheroma and Abdominal Aortic Aneurysm: Possible Role for a Feedback Loop in Endothelial Cells
Am. J. Pathol.,
November 1, 2004;
165(5):
1571 - 1581.
[Abstract][Full Text][PDF]
Top
Abstract
Introduction
1 (TGF-
-smooth muscle action (SMA) and all HRP-labeled secondary antibodies from DAKO (Glostrup, Denmark), anti–TIMP-1 and anti–TIMP-2 from Oncogene (Cambridge, UK), and anti-Smad-2/3 antibodies from BD Pharmingen (San Diego, CA). The anti-phospho-Smad-2 and –3 antibodies used for Western blotting were a kind gift of P. ten Dijke (NKI, Amsterdam, the Netherlands). The anti-phospho-Smad-2/3 antibody used for immunohistochemistry was purchased from Santa Cruz Biotechnology. 
