| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, California, USA; and
Mayo Clinic College of Medicine, Biochemistry & Molecular Biology, Rochester, Minnesota, USA
1Correspondence: LA BioMed, 1124 West Carson St., Torrance, CA 90502, USA. E-mail: rhirschberg{at}labiomed.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: ECM • renal interstitial fibrosis • unilaterial ureteral obstruction • experimental renal fibrosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
. Despite major variations in the pathogenic causes of diseases of the kidney, the mechanisms causing and propagating interstitial fibrosis appear to be rather uniform and shared. Unilateral ureteral obstruction (UUO) in rodents has long been used for the study of fibrosis. In this animal model, renal interstitial fibrogenesis occurs without initial glomerular abnormalities (4)
. Earlier work has demonstrated that transforming growth factor-ß (TGF-ß) is the single most important fibrogenesis-inducing and propagating cytokine (4)
. Inhibition of endogenous TGF-ß activity in experimental fibrogenic renal disease models (including but not limited to UUO) such as with neutralizing anti-TGF-ß antibodies substantially reduces the development of renal fibrosis (5
6
7)
. In addition, mice lacking functional smad3 are partially protected from fibrosis when challenged by induction of UUO (8)
.
Renal fibroblasts are thought to be the primary target of TGF-ß and are the principal origin of interstitial ECM proteins. Very few fibroblasts and myofibroblasts normally are found in the kidney. However, in fibrogenic renal diseases such as UUO, the number of fibroblasts increases by proliferation and they undergo a maturation and activation process that includes the expression of motor proteins such as myosin and 
-smooth muscle actin (SMA). Additional fibroblasts may be recruited by transdifferentiation of damaged tubular cells and/or from bone marrow stromal cells (9
10
11)
. Tubular epithelial cell injury, such as stretch injury in UUO, results in TGF-ß autoinduction. Together with infiltrating macrophages that secrete TGF-ß, tubular cells are the main source for this profibrogenic cytokine in renal diseases. In mature fibroblasts TGF-ß promotes proliferation and growth (12)
.
The best-studied signaling pathway used by TGF-ß in most cell types involves activation of the cellular substrates smad2 and smad3 by serine/threonine phosphorylation through the receptor kinase upon ligand binding. Activated smad2 and -3 form complexes with the transcriptional coactivator smad4. After nuclear translocation, the smad complex associates with various transcriptional coactivators and/or corepressors (13)
. The complexity of TGF-ß signaling and transcriptional regulation is further increased by the presence of inhibitory smads (smad6 and -7), signal modifiers (smurf, snip), and the need to integrate smad activation with the endocytic machinery (13)
. As smad signaling is observed in essentially all TGF-ß-responsive cells, it does not readily account for the differential (growth) responses to TGF-ß in epithelial cells compared with fibroblasts. Our recent identification of novel TGF-ß targets specifically in fibroblasts may help to resolve this issue (14)
. In (nonrenal) fibroblasts, TGF-ß activates a parallel pathway involving p21-activated kinase-2 (PAK2) (14)
. PAK2 activation is independent of smad2/3 phosphorylation and is regulated by the Rho family GTPases Rac1 and Cdc42. As TGF-ß activation of PAK2 does not occur in various epithelial cell cultures, this finding supports the concept of targeting specific TGF-ß signaling pathways in cell types with distinct biological responses.
We have extended this idea of fibroblast-specific TGF-ß signaling and recently determined that the abelson nonreceptor tyrosine kinase (c-abl) is also stimulated by TGF-ß in nonrenal fibroblasts, but not in epithelial cells (unpublished results). The upstream mechanisms connecting TGF-ß to c-abl activation are not known at this time. Thus, in addition to the smad2/3 pathway, in fibroblasts TGF-ß activates two other cellular signals: PAK2 and c-abl. In the present studies we tested the hypotheses that 1) activation of c-abl by TGF-ß occurs in renal fibroblasts and is a required costimulatory signal for renal fibroblast proliferation; and 2) inhibition of the c-abl kinase with imatinib mesylate reduces renal interstitial fibrosis in experimental obstructive nephropathy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Imatinib was given in a dose-accelerating schedule (days 1 and 2: 50 mg/kg; days 3 and 4: 100 mg/kg; days 5–7: 150 mg/kg). On day 6 one animal in the imatinib group showed signs of peritonitis, either chemical or infectious, and this rat was excluded from further analysis. After 7 days animals were killed and 1 mm coronal sections were obtained on ice from the right (nonobstructed) and the left (obstructed) kidney. Slices were fixed in 4% neutral paraformaldehyde/PBS and embedded in paraffin. Total RNA was extracted from other kidney aliquots with the RNA-Stat 60 reagent using the manufacturer’s instructions (Tel Test, Friendswood, TX, USA). For protein extraction and subsequent Western blots, kidney slices were minced with a razor blade on ice, taken up in buffer (50 M HEPES, 1% triton X-100, 50 mM tetra-Na-pyrophosphate, 100 mM Na-fluoride, 10 mM EDTA, 10 µg/mL leupeptin, 10 µg/mL aprotinin, 2 mM benzamidine, 2 mM PMSF, pH 7.4), homogenized, and stored at –82°C. For some Western blot assays, minced aliquots of tissue were taken up directly in reducing Laemmli buffer, sonicated at 5 W for 5 s, cleared by centrifugation, and stored. Total protein in extracts was measured with the RC DC protein assay kit (Bio-Rad, Hercules, CA, USA).
Cell culture models
Supplemental studies were performed in cultured renal cells. Mouse mesangial cells (ATCC, Manassas, VA, USA) were grown in DMEM/F12 (3:1) containing 14 mM HEPES and 5% FBS. Murine proximal tubular epithelial cells were grown in DMEM/F12 containing 10% FBS. NRK-49F rat renal fibroblasts (ATCC) were grown in DMEM/F12 containing 5% BCS.
Assays and analyses
Activation of PAK2 and c-abl in cultured renal fibroblasts, proximal tubular, and mesangial cells
These three cell types were grown in 6-well plates and serum-starved for 24 h by incubation with serum-free media containing 0.1% BSA. Cells were incubated without (control) or in the presence of TGF-ß1 (1 nM), imatinib (5 µg/mL), or both for 45 min. Cultures receiving imatinib were preincubated with this compound for 30 min. At the end of the period of incubation, cells were washed with PBS and lysed with lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% Na-deoxycholate, 0.1 TIU/mL aprotinin, 50 µg/mL PMSF, 100 µM Na-vanadate, 1 µg/mL leupeptin). PAK2 and c-abl were immunoprecipitated from cleared lysates and their kinase activity was assayed as described below.
3H-Thymidine incorporation in cultured cells
NRK-49F rat renal fibroblasts, mouse mesangial cells and proximal tubular cells were grown to 50–70% confluence and incubated in media containing 3H-thymidine, 1 µCi/mL, without (control) or with TGF-ß, 1 nM, in the presence or absence of imatinib, 5 µg/mL, for 16 h (n=8 each). Cells were washed with ice-cold PBS. Cells were then lysed and DNA was precipitated with TCA. Precipitates were washed with ethanol, dissolved in 1 M NaOH, and radioactivity was measured by liquid scintillation counting.
TGF-ß-induced NRK49F proliferation
Additional experiments were performed to examine the time course of TGF-ß-induced proliferation of NRK49F cells in the absence and presence of imatinib or during inhibition of PDGF activity. Cells were plated at 1800/well in 96-well plates. Upon attachment, cells were incubated with serum-free medium containing 0.1% BSA (control) or TGF-ß1, 0.1 or 1.0 nM (n=8 each). Similar studies were performed in the presence of imatinib, 5 µg/mL (n=8) or of both recombinant proteins PDGF-R/Fc and PDGF-Rß/Fc, each at 2 µg/mL (n=8; R&D Systems, Minneapolis, MN, USA). These two fusion proteins contain the soluble, extracellular ligand binding domains of the PDGF receptor and ß, respectively. In combination, these reagents bind and neutralize all forms of PDGF. Incubations were performed for 0–72 h. Media were then removed, the cells were washed once with PBS at 37°C, and plates were frozen at –83°C. After addition of 200 µL/well of CyQuant GR dye in lysis buffer (Molecular Probes, Eugene, OR, USA) the DNA content in each well was assayed by fluorescence measurement with excitation at 485 nm and emission of 535 nm. The number of cells per well was derived from a standard curve compiled from eight standards between 78 and 10,000 cells/well, each in quadruplicate.
Immunohistochemistry and immunofluorescence
Deparaffinized 4 µm sections were quenched with 3% H2O2/PBS, blocked with 10% normal goat or monkey serum in 3% BSA/PBS, and incubated for 12–16 h at 4°C with primary antibodies at dilutions of 1:50 to 1:400. Primary antibodies used for these studies included anti-collagen type IV (Southern Biotechnology Associates, Birmingham, AL, USA), monoclonal anti-fibronectin (BD Transduction Laboratories, San Diego, CA, USA), anti-SMA (Zymed Laboratories, South San Francisco, CA, USA), monoclonal anti-vimentin (Novacastra Laboratories, Newcastle, UK), monoclonal macrophage-specific antibody (Clone ED-1, Chemicon, Temecula, CA, USA), and anti-phospho-smad2/3 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Negative control slides were incubated with nonimmune IgG instead of the primary antibody. Sections were then incubated sequentially with biotinylated second antibody and streptavidin-HRP, followed by DAB substrate. After color development, sections were lightly counterstained with hematoxylin and coverslips were mounted with Permount (Fisher Scientific, Fair Lawn, NJ, USA). For immunofluorescence visualization of collagen type III, sections were incubated with goat anti-collagen type III (Southern Biotechnology Associates), then with Alexa fluor 488-conjugated second antibody (Molecular Probes). Coverslips were mounted with Vectashield (Vector Laboratories, Burlingame, CA, USA).
Digital histomorphometry analysis was performed to quantitatively assess immunohistological findings. In each specimen, 20 nonoverlapping cortical fields (corresponding up to 90% of the cortex of each coronal section) were microphotographed with a 20x magnification objective in a Nikon Eclipse E400 microscope with a Nikon DXM1200 digital camera. In each digital image, the color-specific pixel number (brown in immunohistochemically stained specimen, green in immunofluorescence slides) was counted with Adobe Photoshop 6.0 software (Adobe Systems Inc., San Jose, CA, USA) and expressed as percent of the total pixel number in each image. Results were averaged over the 20 images of each sample. In SMA-stained sections, positively stained blood vessels were digitally subtracted and thereby excluded from the quantitative analysis. In vimentin-stained slides, glomeruli (which stained positive in mesangial areas in all samples) were digitally subtracted and excluded from the measurements. Thus, quantitative assessment of SMA and vimentin assesses primarily or exclusively fibroblasts.
Fibronectin was found in interstitial spaces but also in tubular cells and glomeruli. For quantitative assessment of fibronectin that specifically accumulates in interstitial spaces, a standard point counting method was used. Twenty nonoverlapping digital images were obtained in each sample with a 40x magnifying objective. These images were digitally overlaid with a 13 x 10 (=130) point grid. Points overlying fibronectin-positive interstitium were counted and expressed as the ratio of total grid points (total of 2600 in 20 images of each specimen).
For evaluation of renal interstitial macrophage infiltrates, slides were immunostained with the rat macrophage-specific monoclonal antibody, clone ED-1 (Chemicon). The number of interstitial ED-1+-macrophages was assessed in each of 20 consecutive, nonoverlapping cortical fields per slide as the absolute number per high-power field. Cells with anti-pSmad2/3-positive nuclei were counted similarly. The grand mean of ED1+ or pSmad2/3+ cells per cortical high-power field in right (unobstructed) kidneys from vehicle-treated rats was arbitrarily set as 100% (control).
PAK2 kinase activity assay
PAK2 was immunoprecipitated from cell or tissue lysates (500–700 µg) with anti-PAK2 antibody (Santa Cruz Biotechnology) overnight; complexes were collected with protein-A agarose beads, washed twice in lysis buffer and twice in kinase buffer (25 mM Tris, pH 7.4, 10 mM MgCl2, 1 mM dithiothreitol), then incubated in 50 µL of kinase buffer containing 5 µM ATP, 5 µg of myelin basic protein (Sigma), and 5 µCi of [
-32P]-ATP/µL. After 10 min at 37°C the reaction was stopped by addition of 2x Laemmli buffer. Proteins were separated electrophoretically in SDS-PAGE gels, transferred, and membranes were autoradiographed and Western blotted with anti-PAK2 (Cell Signaling Technology, Beverly, MA, USA) to assess the amounts of phosphorylated substrate and PAK2, respectively.
c-abl Kinase activity assay
c-abl was immunoprecipitated overnight with anti-abl (antibody K-12, Santa Cruz) and immune complexes were collected with protein A-agarose beads and washed as above. Complexes were then incubated in 40 µL of kinase buffer containing 5 µM ATP, 2 µg GST-Crk, and 0.5 µCi/µL of [-32P]-ATP. After 5 min at 37°C the reaction was stopped by addition of 40 µL of 2x Laemmli buffer. The mixture was electrophoresed, transferred, and membranes were autoradiographed and subsequently blotted with anti-c-abl (BD PharMingen, San Diego, CA, USA) to examine the amounts of phosphorylated GST-Crk substrate and total c-abl, respectively.
PAI-1 in kidney extracts was assessed by immunoblotting of heparin binding proteins after affinity precipitation with 25 µL of a 1:1-slurry of heparin-Sepharose. Precipitates were washed 4x with PBS, taken up in 2x reducing Laemmli buffer, electrophoresed, and transferred to nitrocellulose. Membranes were blotted with anti-PAI-1 antibody (Santa Cruz).
Western blot assays: aliquots of tissue extracts in Laemmli buffer were electrophoresed in SDS-PAGE gels. Resolved proteins were electrotransferred to nitrocellulose. Membranes were blocked with 5% BSA or 5% dried, fat-free milk (DM) in TBS-0.05% Tween-20 (TTBS) for 1–2 h at room temperature, then incubated with primary antibody in blocking buffer for 12–16 h at 4°C. Membranes were then washed 3x with TTBS and incubated with HRP-conjugated second antibody at 1:4000 to 1:10,000 in 5% DM/TTBS for 1 h at room temperature. Membranes were washed 4x with TTBS and 1x with TBS, and bands were visualized by enhanced chemiluminescence and captured on X-ray film. The primary antibodies used in the assays were anti-fibronectin (BD Biosciences, Palo Alto, CA, USA), anti-collagen IV (Southern Biotechnology Associates), anti-pSmad2/3 and anti-PAK2 (Santa Cruz), anti-c-abl (BD PharMingen), and anti-gapdh (Research Diagnostics, Flanders, NJ, USA).
Collagen type IV ELISA
For further quantification of collagen type IV levels in kidney extracts, direct enzyme-linked immunosorbant assays were performed as described (15)
. Briefly, 96-well plates were coated with 100 µL of kidney extracts taken up in 0.1 M NaHCO3, pH 9.8, 2 mg/mL in triplicate for 48 h at 4°C and for a further 24 h after addition of 100 µL of 0.2% BSA in TBS. Plates were washed with TTBS, then incubated with biotinylated anti-col IV (1:1500, Southern Biotechnology Associates) in TTBS containing 0.2% BSA for 2 h at room temperature. Plates were washed again and incubated with Avidin-HRP in 0.2% BSA/TTBS for 1 h. After extensive washes the substrate reaction was induced by addition of 0.2% OPD in 200 mM Tris, pH 6.0, 150 mM NaCl, 0.01% H2O2, and stopped after 45 min by addition of 4 M H2SO4. Absorption was measured at 490 nm in an ELISA plate reader (Molecular Devices, Menlo Park, CA, USA). Results are expressed in % of mean control (right kidney from vehicle-treated rats).
rtPCR assays
PDGF-BB, A1-chain of collagen type III (C3A1), and fibronectin mRNA levels were quantitatively assessed by competitive rtPCR with parallel amplification of 18S rRNA as internal standard in each tube using a commercially available assay kit (Qiagen, Chatsworth, CA, USA). Reactions had been optimized with respect to amounts of template cDNA, temperature cycling for denaturing, annealing and extension, and cycle number. The following primers were used: fibronectin sense 5'-TTT TGA CAA CGG GAA GCA TTA TCA GAT AA-3'; antisense 5'-TGA TCA AAA CAT TTC TCA GCT ATT GG-3'; C3A1 sense 5'-CGA GGT AAC AGA GGT GAA AGA-3'; antisense 5'-AAC CCA GTA TTC TCC GCT CTT-3'; PDGF-BB sense 5'-GTC GAG TCG GAA AGC TCA TC-3'; antisense 5'-ACT GCA CAT TGC GGT TAT TG-3'.
Statistical evaluation of data
Data are presented as means and standard error of means. Group means were compared by ANOVA, followed by Newman-Keuls multicomparison test. A probability of < 5% (P<0.05) was presumed to reflect statistical significant difference between group mean values.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
a). Coincubation of the cells with imatinib does not affect PAK2 kinase but blocks the c-abl kinase activity (Fig. 1a
). In cultured mesangial cells and proximal tubular epithelial cells, neither PAK2 nor abl kinase activity was increased by TGF-ß (data not shown). Thus, both PAK2 and abl are selectively activated in renal fibroblasts but not in renal mesangial or tubular epithelial cells, fulfilling criteria for fibroblast-selective TGF-ß signals. This finding confirms recently published results indicating that TGF-ß activates PAK2 in nonrenal fibroblasts (including AKR-2B, Swiss-3T3, BALB/c-3T3, and human foreskin fibroblasts) and not in epithelial cells (Mv1Lu, HeLa, LLC-PK1, and human prostate epithelium) (14)
. The present results further extend the previous finding to some of the cells that are relevant in fibrogenic renal diseases.
|
To address the role of c-abl activation in TGF-ß-mediated renal cell proliferation, NRK-49F renal fibroblasts and mesangial and tubular cells were treated with TGF-ß (in the presence or absence of imatinib) and 3H-thymidine incorporation into DNA was determined. TGF-ß induced a significant increase in 3H-thymidine uptake in NRK49F cells that was completely blocked by coincubation with imatinib (Fig. 1b
). In contrast, in cultured mesangial and tubular epithelial cells, TGF-ß did not increase 3H-thymidine incorporation (Fig. 1b
). In extended time course studies, TGF-ß (0.1 and 1.0 nM) induces NRK49F cell proliferation within 24 h and causes doubling of the cell number by 72 h (Fig. 1c
). As expected, imatinib quantitatively blocks the TGF-ß-induced proliferation. Since imatinib inhibits the PDGF receptor kinase, we tested whether TGF-ß may induce proliferation of NRK49F cells indirectly through PDGF. Neutralization of PDGF with soluble binding domains of PDGF receptors and ß had no significant effect on NRK49F cell growth (Fig. 1c
, bottom panel). Thus, activation of c-abl is required for TGF-ß-stimulated renal fibroblast proliferation in vitro.
UUO causes PAK2- and c-abl kinase activation and smad2/3 phosphorylation
Since TGF-ß activates PAK2 and c-abl in cultured renal fibroblasts, we examined whether activation of these two kinases occurs in vivo in kidneys with obstructive nephropathy. As shown in Fig. 2
a, in the obstructed (left) kidney from UUO-rats, PAK2 kinase activity is increased in both imatinib and control rats compared with the contralateral, nonobstructed (right) kidney. In addition, ureteral obstruction for 1 wk increases renal c-abl kinase activity 
4-fold (Fig. 2a, b
). Whereas c-abl activation is blocked by imatinib therapy, PAK2 activity is not affected by this treatment (Fig. 2a
). Quantitation of the inhibitory effect of imatinib on activation of renal c-abl is shown in Fig. 2b
.
|
While Fig. 2
clearly documents that UUO activates both PAK2 and c-abl in the obstructed kidney, previous studies have shown that TGF-ß-dependent activation of smad2/3 is necessary for fibrogenesis (8)
. As such, nuclear translocation of phosphorylated smad2/3 was investigated in control and obstructed kidneys. Compared with the unobstructed kidney, unilateral ureteral obstruction substantially increased levels of phosphorylated smad2/3 and its nuclear localization in tubular and interstitial cells (Fig. 3
). This TGF-ß signal is mainly induced in injured cells in more severely congested tubules, but also in interstitial cells (Fig. 3)
. Imatinib does not reduce smad2/3 phosphorylation and nuclear translocation (Fig. 3b
). Thus, whereas obstructive nephropathy increases activation of smad2/3, PAK2, and c-abl in kidney fibroblasts, only c-abl activity is blocked by imatinib therapy.
|
PAI-1 is a TGF-ß-induced enzymatic inhibitor of matrix protein degradation and is known to increase in obstructed nephropathy (16)
. In the present study, PAI-1 levels were increased in the obstructed compared with the contralateral, nonobstructed kidney. Consistent with the smad dependence of PAI-1 induction (17
, 18)
, treatment with imatinib did not substantially affect the rise in the levels of this protein in renal extracts (Fig. 4
).
|
Expression of platelet-derived growth factor-BB (PDGF-BB) increases in obstructive nephropathy (4)
. This growth factor is mainly induced by TGF-ß in injured tubular cells and may contribute to fibroblast activation and interstitial fibrosis (19
, 20)
. PDGF-BB mRNA levels were moderately increased in vehicle-treated rats in obstructed compared with unobstructed kidneys (145±17% vs. 100±4%, P<0.05). In imatinib-treated rats, levels tended to be lower than in vehicle-treated animals, albeit not significantly and were greater in obstructed (114 ± 13%) than unobstructed (89 ± 9%) kidneys (P<0.05). Thus, ureteral obstruction moderately increased PDGF-B expression that was not substantially blocked by imatinib.
Imatinib does not reduce interstitial macrophage infiltration but blocks fibroblast proliferation in obstructive nephropathy
Within 7 days of onset of obstructive nephropathy in rodents, there is a substantial rise in the number of renal interstitial macrophages (21)
. We examined whether imatinib would modulate the distribution of ED-1+ macrophages in the interstitium of nonobstructed and obstructed kidneys. Although few ED-1+ cells were detectable in nonobstructed kidneys from control and imatinib-treated rats, their number per high-power field was substantially increased 15.3 ± 7.7-fold (P<0.05) and 24.1 ± 11.6-fold (P<0.05) in obstructed kidneys from vehicle and imatinib-treated rats, respectively. Thus, imatinib did not block the renal interstitial macrophage infiltrate that occurs after ureteral obstruction.
A pathoetiologic feature of obstruction-driven renal fibrosis is an increase in interstitial fibroblasts (4)
. We assessed the fibroblast cytoskeletal protein vimentin as a measure of fibroblast number. Since glomerular mesangial cells express vimentin, as expected, all glomeruli stained positive for vimentin independent of ureteral obstruction or treatment with imatinib (Fig. 5
a). However, expression of vimentin corresponding specifically to interstitial fibroblasts increased >4-fold in obstructed kidneys (Fig. 5a, b
). In agreement with our in vitro proliferation data (Fig. 1b, c
), the increase in interstitial, vimentin-expressing fibroblasts in obstructed kidneys is substantially reduced by imatinib (Fig. 5a, b
). Compared with the contralateral, unobstructed kidney, ureteral obstruction increased the number of the SMA-expressing subpopulation of fibroblasts (myofibroblasts) (Fig. 5c, d
). The number of SMA-positive myofibroblasts in the interstitium of obstructed kidneys was reduced by 65% after imatinib treatment; we did not observe SMA expression in any tubular cells. TGF-ß is thought to induce SMA expression via smad2/3 in fibroblasts indicating smad dependence of the transition to a myofibroblast phenotype (22)
. As indicated in Fig. 5e
, the ratio of SMA+ myofibroblasts/vimentin+ fibroblasts increases in obstructed kidneys but is not affected by imatinib therapy. Thus, imatinib reduces the total number of interstitial fibroblasts but does not affect the percentage of fibroblasts that are transformed to the myofibroblast (SMA+) phenotype that is thought to be smad3-mediated (22)
. The reduced number of interstitial SMA+ myofibroblasts in obstructed kidneys from imatinib-treated rats simply reflects a reduced total fibroblast number.
|
Imatinib reduces interstitial accumulation of ECM proteins
The foregoing results show that inhibition of c-abl activity prevents renal fibroblast proliferation and reduces the number of myofibroblasts in a rodent model of kidney fibrosis (Figs. 1
and 5)
. We further examined whether this is associated with reduced renal interstitial fibrosis. We assessed the accumulation of collagen type III, collagen type IV, and fibronectin as makers of renal interstitial fibrosis. These three extracellular matrix proteins were chosen because they represent interstitial accumulation of a normal basement membrane collagen component (collagen type IV), an interstitial collagen (collagen type III), and a noncollagenous extracellular matrix protein (fibronectin).
Collagen type IV is substantially increased in obstructed compared with the nonobstructed kidneys in control rats as shown by immunoblotting of kidney lysates and quantitatively by ELISA assay (Fig. 6
b, c). Most excess collagen type IV accumulates in the interstitium (Fig. 6a
). Treatment with imatinib substantially reduces interstitial collagen type IV in the obstructed kidney (Fig. 6a-c
).
|
Collagen type III contributes to interstitial fibrosis in the obstructed compared with the nonobstructed kidney and imatinib therapy substantially lowers collagen type III accumulation (Fig. 7
a, b). Increased levels of collagen III are most likely a result of increased synthesis as suggested by increased collagen III A1 mRNA levels in obstructed kidneys, which are partially lowered by imatinib (Fig. 7c
).
|
In nonobstructed kidneys from control rats, fibronectin is sparsely found in interstitial spaces, the glomerular mesangium, and Bowman’s capsule and is not affected by imatinib therapy (Fig. 8
a). In contrast, fibronectin is substantially increased in obstructed kidneys, especially in interstitial spaces. Some cortical tubular cross sections stain positively with the anti-fibronectin antibody. This increase in fibronectin levels is reflected in results from Western blot and rtPCR analyses of whole kidney lysates (Fig. 8c, d
). In obstructed kidneys from imatinib-treated rats, fibronectin as (assessed by Western blot) and mRNA levels are moderately reduced (Fig. 8c, d
). However, specific quantitative assessments of fibronectin in the interstitium by a point grid analysis method of immunohistological specimen reveal that interstitial fibronectin in obstructed kidneys from imatinib-treated rats is, in fact, substantially reduced on average by 73% compared with obstructed kidneys from vehicle-treated animals (Fig. 8b
). Thus, imatinib therapy reduces primarily interstitial (fibroblast-derived) fibronectin.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This cytokine is the primary mediator of renal interstitial fibrogenesis, in general and specifically in obstructive nephropathy (4)
. Most fibrogenic effects of TGF-ß are thought to be mediated through the smad2/3 pathway, and smad3 knockout mice are partially protected from interstitial fibrosis after unilateral ureteral obstruction (8
, 23
, 24)
. However, the heterogeneity of cellular responses to TGF-ß cannot be explained solely by effects downstream of smad3. Although activation of non-smad signal mediators by TGF-ß (including p42/44 MAPK, p38 MAPK, and JNK) have been described, they do not appear to be unique to fibroblasts but have been demonstrated in other cell types, most notably in epithelial cells (25
, 26)
. Another explanation for differential growth regulation of epithelial cells and fibroblasts by TGF-ß is the presence of alternative smad2/3-independent but parallel signaling pathways that are selectively active in fibroblasts. Indeed, such a parallel TGF-ß pathway was recently identified (14)
. In several lines of nonrenal fibroblasts, but not in epithelial cell cultures, TGF-ß activates PAK2 (14)
and the c-abl tyrosine kinase (unpublished results). The upstream mechanisms mediating activation of c-abl upon TGF-ß are not known.
In the present studies we have extended these earlier observations to cells that are relevant in fibrogenic renal diseases and show that 1) TGF-ß activates PAK2 and c-abl in renal fibroblasts (Fig. 1)
but not in renal tubular or mesangial cells (data not shown); 2) c-abl activity is required for the proliferative response of renal fibroblasts to TGF-ß (Fig. 1b, c
); and 3) blockade of c-abl activation in vivo with imatinib is an effective treatment to prevent the renal fibrogenesis that occurs in obstructive nephropathy (Figs. 6
7
8)
.
Imatinib mesylate (STI-571, Gleevec®) is a 2-phenylaminopyridine derivative with high affinity and specificity for c-abl and BCR- or Tel-abl and blocks their kinase activity (27
, 28)
. This compound blocks the stem cell factor receptor (c-kit) and the PDGF receptor kinase (29
, 30)
, and has been FDA-approved for the treatment of leukemias and gastrointestinal stromal tumors that are driven by BCR-abl, Tel-abl or oncogenic c-kit mutations, respectively.
TGF-ß activates PAK2 as well as c-abl in NRK49F cells (Fig. 1a
) but not in tubular or mesangial cells (not shown). The upstream mechanisms of c-abl activation by TGF-ß and whether PAK2 activation determines or is required for c-abl activity remain unknown and were not the subject of the present experiments. While imatinib blocks c-abl activation in cultured renal fibroblasts and in the kidney in obstructive nephropathy, it has no effect on PAK2 activation (Figs. 1
, 2)
. In rats with obstructive nephropathy, imatinib substantially reduces fibrogenesis, as indicated by reduced accumulation of collagen type III and IV and fibronectin in the renal interstitium (Figs. 6
7
8)
. This is associated with a reduced number of interstitial vimentin-positive fibroblasts and SMA-positive myofibroblasts. In concert, in vitro and in vivo data show that c-abl activation by TGF-ß contributes to fibroblast proliferation in obstructive nephropathy and imatinib reduces renal fibrogenesis by inhibiting this process.
Obstructive nephropathy causes TGF-ß-induced activation of smad2/3 as indicated by increased levels and nuclear accumulation of pSmad2/3 (Fig. 3)
. Moreover, smad2/3-mediated TGF-ß effects are induced in obstructed kidneys and are not opposed by therapy with imatinib. These effects include an increase in PAI-1 levels and induction of SMA in fibroblasts (Fig. 4
and Fig. 5c-e
). The profibrogenic enzyme PAI-1 is up-regulated by TGF-ß in a smad2/3-dependent manner by direct interaction with a smad binding element in the PAI-1 promoter (18)
. Induction of SMA in fibroblasts by TGF-ß is similarly regulated through smad3 (22)
. In obstructed kidneys, smad2/3 activation is not interrupted by imatinib therapy. Consequently, increased PAI-1 levels are not lowered by inhibition of c-abl activation with imatinib (Fig. 4)
and induction of SMA in a subpopulation of fibroblasts is not reduced as indicated by similar ratios of SMA+ myofibroblasts/vimentin+ fibroblasts in obstructed kidneys in vehicle- and imatinib-treated rats (Fig. 5e
). The reduced number of interstitial SMA-positive myofibroblasts in obstructed kidneys in imatinib-treated rats simply reflects a reduction in the total number of fibroblasts. Thus, the present data suggest that TGF-ß 1) stimulates fibroblast proliferation in the renal interstitium via a c-abl-dependent mechanism and 2) induces the myofibroblast phenotype through smad2/3. The differential cellular response mediated by c-abl and smad activation documents the complexity inherent in TGF-ß signaling as well as the potential to modulate specific TGF-ß-dependent biologies.
The current results suggest that the major mechanism underlying the beneficial actions of imatinib in obstructive nephropathy is the inhibition of TGF-ß-stimulated c-abl activation and the ensuing reduction in fibroblast proliferation. These latter cells are targets of a variety of cytokines and growth factors that are released primarily from injured tubular cells and to a lesser extent from infiltrating macrophages. For instance, PDGF-BB is modestly increased in obstructive nephropathy, reflecting release from injured tubular cells (31
, 32)
perhaps through an indirect action of TGF-ß (19
, 33)
. As such, inhibition of the PDGF receptor kinase by imatinib may contribute to its anti-fibrogenic effects in experimental UUO. However, TGF-ß-induced activation of c-abl and fibroblast proliferation do not require PDGF. This is demonstrated by recent unpublished findings indicating that TGF-ß induces proliferation similarly in normal fibroblasts as well as in fibroblasts that lack functional PDGF receptors (unpublished results). Moreover, as shown in Fig. 1c
(bottom panel), the proliferative effects of TGF-ß in NRK49F fibroblasts are not blocked during neutralization of PDGF. Finally, PDGF induces growth in fibroblasts as well as in epithelial and mesangial cells and, hence, is not a candidate mediator for the differential growth effects of TGF-ß on these cell types.
From the current experimental observations, we propose the following model describing the profibrogenic actions of TGF-ß in obstructive nephropathy (and perhaps in other fibrogenic renal diseases). TGF-ß activates the smad2/3 pathway in all renal cell types and activates c-abl specifically in fibroblasts. C-abl contributes to fibrogenesis by increasing fibroblast proliferation and is required for maximal fibrogenesis by TGF-ß. During the process of disease progression, TGF-ß contributes to the smad-dependent myofibroblast transition whereas c-abl contributes indirectly to this process by increasing the number of fibroblasts, the precursors for myofibroblasts. Interference with the smad2/3 or c-abl pathway should then reduce renal fibrogenesis. Indeed, interruption of the smad3 pathway reduces renal fibrosis as recently shown by other investigators (8
, 24)
. The present studies show that prevention of c-abl activation with imatinib reduces renal fibrogenesis. The role or requirement for PAK2 activation by TGF-ß in renal fibrogenesis, if any, remains to be elucidated.
In contrast to other therapeutic approaches to interrupt TGF-ß-driven renal fibrogenesis, inhibition of c-abl with imatinib is specific to a subset of TGF-ß actions and the most practical, as this compound is readily available for use in humans. For instance, interruption of the smad2/3 pathway or inhibition of TGF-ß receptor kinase with an experimental Alk5 kinase antagonist (34)
would affect numerous cell types and may cause lymphoproliferative disorders (35
, 36)
. Although imatinib mesylate can cause side effects that include, perhaps rarely, acute nephrotoxicity (37
, 38)
, new imatinib-derived c-abl inhibitors presently in development may have improved efficacy and tolerability (39)
.
In summary, the proliferative effects of TGF-ß in (renal) fibroblasts require activation of a novel fibroblast-specific pathway involving c-abl. TGF-ß is the single most important profibrogenic mediator in renal fibrosis; the present data show that inhibition of this pathway with imatinib mesylate in rats with unilateral ureteral obstruction substantially reduces renal fibrogenesis in obstructive nephropathy. Imatinib may be a promising treatment to prevent progressive renal interstitial fibrogenesis and thereby help preserve renal function.
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 20, 2004. Accepted for publication August 30, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Akhmetshina, C. Dees, M. Pileckyte, B. Maurer, R. Axmann, A. Jungel, J. Zwerina, S. Gay, G. Schett, O. Distler, et al. Dual inhibition of c-abl and PDGF receptor signaling by dasatinib and nilotinib for the treatment of dermal fibrosis FASEB J, July 1, 2008; 22(7): 2214 - 2222. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Leipner, K. Grun, A. Muller, E. Buchdunger, L. Borsi, H. Kosmehl, A. Berndt, T. Janik, A. Uecker, M. Kiehntopf, et al. Imatinib mesylate attenuates fibrosis in coxsackievirus b3-induced chronic myocarditis Cardiovasc Res, July 1, 2008; 79(1): 118 - 126. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Gioni, T. Karampinas, G. Voutsinas, A. E. Roussidis, S. Papadopoulos, N. K. Karamanos, and D. Kletsas Imatinib Mesylate Inhibits Proliferation and Exerts an Antifibrotic Effect in Human Breast Stroma Fibroblasts Mol. Cancer Res., May 1, 2008; 6(5): 706 - 714. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. P. Sfikakis, V. G. Gorgoulis, C. G. Katsiari, K. Evangelou, C. Kostopoulos, and C. M. Black Imatinib for the treatment of refractory, diffuse systemic sclerosis Rheumatology, May 1, 2008; 47(5): 735 - 737. [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Horan, S. Wood, V. Ona, D. J. Li, M. E. Lukashev, P. H. Weinreb, K. J. Simon, K. Hahm, N. E. Allaire, N. J. Rinaldi, et al. Partial Inhibition of Integrin {alpha}v 6 Prevents Pulmonary Fibrosis without Exacerbating Inflammation Am. J. Respir. Crit. Care Med., January 1, 2008; 177(1): 56 - 65. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Floege, F. Eitner, and C. E. Alpers A New Look at Platelet-Derived Growth Factor in Renal Disease J. Am. Soc. Nephrol., January 1, 2008; 19(1): 12 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
