| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* The Department of Medicine, SFVAMC/University of California, San Francisco; and
the Department of Pathology, University of California, San Francisco, USA
1Correspondence: 111J Medical Service, SFVAMC, 4150 Clement St., San Francisco, CA 94121, USA. E-mail: david.lovett{at}med.va.gov
ABSTRACT
Chronic kidney disease (CKD) and failure are problems of increasing importance. Regardless of the primary etiology, CKD is characterized by tubular atrophy, interstitial fibrosis, and glomerulosclerosis. It has been assumed that diminished matrix metalloproteinase (MMP) activity is responsible for the accumulation of the extracellular matrix (ECM) proteins and collagens that typify the fibrotic kidney. Here we demonstrate that transgenic renal proximal tubular epithelial expression of a specific enzyme, MMP-2, is sufficient to generate the entire spectrum of pathological and functional changes characteristic of human CKD. At the earliest point, MMP-2 leads to structural alterations in the tubular basement membrane, a process that triggers tubular epithelial-mesenchymal transition, with resultant tubular atrophy, fibrosis and renal failure. Inhibition of MMP-2, specifically in the early, prefibrotic stages of disease may offer an additional approach for treatment of these disabling disorders.—Cheng, S., Pollock, A. S., Mahimkar, R., Olson, J. L., Lovett, D. H. Matrix metalloproteinase 2 and basement membrane integrity: a unifying mechanism for progressive renal injury.
Key Words: epithelial-mesenchymal transition • chronic kidney disease • tubular atrophy • interstitial fibrosis
THE INCREASING INCIDENCE of obesity, with attendant hypertension and diabetes mellitus, has led to an epidemic of chronic kidney disease (CKD) in the industrialized world (1
, 2)
. Regardless of the primary etiology, CKD is characterized by glomerulosclerosis, loss of functional nephron units via tubular atrophy, and development of interstitial fibrosis. Recent studies have emphasized the potential role of epithelial mesenchymal transition (EMT) in the development of CKD (3
4
5
6
7)
. According to this concept, tubular epithelial cell injury results in the progressive loss of defined epithelial features, including disassembly of E-cadherin junctional complexes, nuclear translocation of ß-catenin, and diminished cytokeratin expression. The loss of epithelial features is accompanied by the acquisition of a mesenchymal phenotype, including expression of 
-smooth muscle actin and vimentin, followed by matrix metalloproteinase-2 or -9-dependent dissolution of the basement membrane and invasion/migration into the interstitial space. Interstitial fibrosis may also be the consequence of various degrees of activation and expansion of the intrinsic fibroblast population or infiltration by mesenchymal stem cells (3)
.
Transforming growth factor-ß1 (TGF-ß1) has been the most commonly studied cytokine thought to be responsible for the generation of EMT. However, there is an ever-increasing number of factors capable of initiating or promoting EMT in vitro. These include epidermal growth factor (EGF), fibroblast growth factor-2, interleukin (IL) -1, angiotensin (ANG) II, advanced glycation end products, and connective tissue growth factor, among others (reviewed in refs. 8
, 9
). Furthermore, EMT can be induced in vitro by plating epithelial cells on fibrillar collagens or by disrupting type IV collagen basal lamina organization (10
,11)
.
While enhanced expression of matrix metalloproteinases-2 (MMP-2) or -9 has commonly been associated with the later stages of the EMT continuum (invasion/migration into the interstitium), we recently reported that active MMP-2 was absolutely required for the development of EMT in a TGF-ß1-dependent in vitro model (12)
. Furthermore, active MMP-2 alone was sufficient to drive the full spectrum of EMT. This process required enzymatically active MMP-2 and was not replicated with the closely related MMP-9 protein (12)
.
MMP-2 synthesis is enhanced by each of the cytokines and factors detailed above as capable of inducing EMT (12
13
14
15
16)
. Analyses of the MMP-2 promoter indicate that transcriptional regulation occurs through the convergence of multiple signaling cascades generally activated by these factors, including the TGF-ß1 and MAPK/ERK signaling pathways (17
18
19
20)
. These considerations led to the hypothesis that enhanced expression of MMP-2 represents a primary mediator of injury, as opposed to the generally accepted concept that limits the pathogenic role of MMP-2 activity to the phase of epithelial cell migration/invasion into the interstitial space. In addition, we hypothesized that the initial effect of MMP-2 would be the disruption of tubular basement membrane (TBM) integrity, a process sufficient to trigger EMT in vitro (10
, 11)
. To approach this problem, we generated transgenic mice with targeted expression of active MMP-2 in the proximal tubular epithelial cells. Here we show that transgenic expression of active MMP-2 alone induced widespread TBM alterations, followed by the subsequent evolution of the entire continuum of renal EMT. We conclude that transgenic expression of MMP-2 is sufficient to drive the development of glomerulosclerosis, interstitial fibrosis, tubular atrophy, and renal failure in the absence of superimposed injury, findings that have important clinical implications.
MATERIALS AND METHODS
MMP-2 transgenics
A murine MMP-2 cDNA was cloned into the TOPO-TA vector (Invitrogen, Carlsbad, CA, USA) and Val 107 in the prodomain mutated to Gly107, generating a constitutively active MMP-2 enzyme (42)
. The MMP-2 cassette was recovered from the TOPO-TA plasmid using primers encoding a Kozak consensus sequence and the c-myc epitope tag (E410QKLISEEDL) at the C terminus. This was subcloned into plasmid p1.2C, which contains a cytomegalovirus (CMV) promoter, a human growth hormone intron, and a polyadenylation tail. The rat type I 
-GT promoter was a gift from Dr. Michael W. Lieberman (Baylor University College of Medicine) in plasmid pG2.2GGTI. The promoter was cloned into p1.2C with Mlu I and BstBI, removing the CMV promoter. The excised cassette was microinjected into FVB/N strain fertilized pronuclei (Xenogen Biosciences, Cranbury, NJ, USA). Transgenics were identified by polymerase chain reaction (PCR) of tail genomic DNA using a 5' primer overlapping the MMP-2 coding sequence and a 3' primer for the c-myc epitope. Southern blot analysis was performed with a 32P-labeled PCR product. Nine founders were obtained and three lines (753, 755, and 757, with 4–6 insert copies and single integration sites) were carried as homozygotes. Each line manifested a similar phenoptype and data from the 755 line are shown in this report.
Transgene expression
Kidneys were homogenized in lysis buffer (50 mM Tris/HCL, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.5% CHAPS), incubated with gelatin-Sepharose to affinity absorb MMP-2, and eluted into SDS-PAGE sample buffer. Western blots used murine monoclonal anti-cmyc epitope antibody (Ab) (Novus Biologicals, Littleton, CO, USA), followed by HRP-conjugated goat antimouse IgG (Zymed, South San Francisco, CA, USA) and detection with enhanced chemiluminescence (ECL) reagent (Amersham Biosciences, Piscataway, NJ, USA). For immunohistochemistry, paraformaldehyde-fixed cortical sections were incubated with murine monoclonal antic-myc (9E11, Abcam, Cambridge, MA, USA), followed by the M.O.M kit (Vector, Burlingame, CA, USA) and development with VIP purple substrate. In situ zymography was performed according to Mook et al. (43)
using quenched fluorogenic DQ gelatin (Molecular Probes, Eugene, OR, USA) in the presence or absence of the cyclic peptide MMP-2 inhibitor CTTHWGFTLCGG at 25 µmol/l (12)
.
Histology
For electron microscopy (EM), kidneys were perfused with 4°C PBS and fixed with modified Karnovsky’s fixative. Semithin sections were stained with Toluidine blue; ultrathin sections were stained with lead citrate/uranyl acetate. For light microscopic studies, kidneys were perfused with 4°C PBS and fixed in 4% buffered paraformaldehyde. Endogenous peroxidase was blocked with 3% H2O2 for 30 min, followed by sequential block with 5% normal goat serum and avidin/biotin blocking solution (Vector). Rabbit primary antibodies anti- smooth muscle actin, NeoMarkers, Fremont, CA, USA; anticytokeratin (Pan), Zymed, S. San Francisco, CA, USA; anti-HSP-47, Biovision, Mt. View, CA, USA; anti-FSP-1/S100A4, Neomarkers; antivimentin, Abcam; anti-ZO-1, Zymed) were used at 5 µg/ml PBS/0.1% BSA for 30 min, followed with biotinylated F(ab)2 goat anti-rabbit IgG (Zymed). Sections were developed with Vectastain Elite avidin-biotin complex (ABC) (Vector) followed by the peroxidase substrates enhanced 3,3'-diaminobenzidine (DAB), NovaRed, or VIP purple (Vector).
Morphometric analysis
The extent of EMT, defined by TBM disruption and cellular invasion, was assessed in 20 randomly chosen fields of semithin sections of midcortex in wild-type (WT) and MMP-2 transgenics (n=6/group). Renal injury was assessed according to Raij et al. (44)
. Twenty randomly selected cortical fields of H/E-stained sections from each kidney (6/group) were rated. Tubular lesions were graded on a scale of 0 to 3+ on the basis of the extent of tubular dilation/cellular loss (atrophy). The extent of mononuclear cell infiltration was determined using a 0 to 2+ scale, where 0, 1, and 2 correspond to 0%, 1–10%, and 10–20% of the field. Picrosirius red staining was performed according to Junqueira et al. (45)
and sections were examined under polarized light. Ten randomly chosen low-power cortical fields/kidney (6/group) were digitized and the collagen volumes determined using ImageJ (NIH). Glomerular lesions were scored on stained sections and graded from 0 to 2+ according to the extent of mesangial cell proliferation, hyalinosis, and sclerosis.
Quantitative RT-polymerase chain reaction (RT-PCR)
RNA was isolated from kidneys (n=6 for each group) at 4 and 8 months of age. RNA integrity and quantity were determined with the Agilent 2100 Bioanalyzer. cDNA templates were generated by oligo-dT priming (Transcriptor, Roche, Alameda, CA, USA). Quantitative PCR (Agilent 9800) for FSP1, HSP47, vimentin, and 1 chain of type I collagen was performed using SYBR Green incorporation (Applied Biosystems, Foster City, CA, USA) with the following primer pairs:
FSP1: (Tm=58°C) 5'-ATGGCAAGACCCTTGGAGGAGG-3'/5'-TCACTTCTTCCG-GGGTTCCTTATCTG3';
HSP47: (Tm=60°C) 5'-TGACCGAAGCCATCGACAAGAA-3'/5'-CTACAACTCATC-TCGCATCTTGTCTCC-3';
Vimentin: (Tm=57.3°C) 5'-ATGCGTGAGATGGAAGAGAATTTTGC-3'/5'-TTATT-CAAGGTCATCGTGATGCTGAGA-3';
Type I collagen, 1 (Tm=60°C) 5'-CCGATGGATTCCCGTTCGAGT-3'/5'-GCCGTCCACAAGGGTGCTGTA-3';
Results were normalized to GAPDH (Tm=60°C). 5'-TGACATCAAGAA-GGTGGTGAAGCAGGCAT-3'/5'-CACCCTGTTGCTGTAGCCGTAT-TCATTGTCAT-3';
Reactions were performed in quadruplicate; quantitation of mRNA expression was performed by the 
Ct method. Results are normalized to fold change in transgenics compared with WT.
Data analysis and statistics
Data are expressed as the means ± SE. Statistical differences were assessed using an unpaired t test with significance assigned a P value of <0.05.
RESULTS
Generation and validation of renal proximal tubule-specific MMP-2 transgenics
The transgenic construct consisted of an expression cassette for constitutively active MMP-2, achieved through mutation of the prodomain, coupled to a c-myc epitope tag to distinguish transgenic from native MMP-2 protein (Fig. 1
A). Expression was driven by the renal proximal tubule-specific type I GT promoter (21
, 22)
, which is relatively inactive until 4–6 wk of age, when nephrogenesis is complete in the mouse. Nine founders were identified (Fig. 1B
) and lines 753, 755, and 757 were characterized. Each transgenic line displayed a similar phenotype, and data from line 755 are detailed in this report. Western blots of cortical extracts, which are enriched in proximal tubule cells, confirmed MMP-2/c-myc transgenic protein expression in all three transgenic lines (Fig. 1C
). Proximal tubule epithelial-specific transgene expression was confirmed by immunohistochemical (IHC) staining for the c-myc epitope tag (Fig. 1D
), which revealed a strong signal in proximal tubule cells, while more distal structures were not stained. Transgenic MMP-2/c-myc protein was primarily expressed in a basolateral distribution, with absent apical (brush border) staining (Fig. 1E, a
). Fluorescent in situ zymography, using a quenched gelatin substrate, confirmed that transgenic MMP-2 was enzymatically active. As shown in Fig. 1E, b
, WT kidneys did not demonstrate in situ enzymatic activity, while in the transgenics (c) there was clear evidence for enzymatic activity in a basolateral distribution, consistent with the localization obtained with IHC. Inclusion of a specific MMP-2 cyclic peptide inhibitor eliminated in situ enzymatic activity (d), confirming that the observed activity was due to active MMP-2.
|
Transgenic MMP-2 induces renal tubular epithelial mesenchymal transition
Transgenic kidneys were initially examined at 4 months of age, when the tissues were morphologically normal when examined at the level of conventional light microscopy with a hematoxylin/eosin stain. Furthermore, at this time both WT and MMP-2 transgenics had creatinine levels in the normal range (control: 0.2 mg/dl±0.05; transgenic: 0.3±0.04 mg/dl; P>0.05), as well as normal levels of urinary protein excretion (control Ualb/Ucreat 0.16±0.03; transgenic: Ualb/Ucreat 0.19±.04; P>0.05). Toluidine blue-stained semithin sections of controls revealed normal tubular epithelial cell morphology, with small clusters of resident interstitial fibroblasts (Fig. 2
A). One pattern of EMT in the MMP-2 transgenics was characterized by focal areas of tubular basement membrane (TBM) disruption, with columns of proliferating cells with a mesenchymal phenotype invading the interstitial space (Fig. 2B
). Semiquantitative analysis of foci with clear-cut TBM disruption and interstitial invasion indicated that these events were relatively uncommon but significant (MMP-2 transgenics:14.6±2.2 foci/20 fields vs. WT: 0.4±0.2 foci/20 fields; n=6, P<0.05).
|
Transmission electron microscopy (TEM) demonstrated a number of ultrastructural features consistent with an EMT of the proximal tubular epithelial cells. Many of the proximal tubular epithelial cells examined contained filamentous actin bundles (Fig. 3
A), as well as loss of basolateral infolding and centralization of mitochondria (not shown). Most notably, there was widespread evidence for MMP-2-mediated basement membrane proteolytic alteration, as evidenced by patchy lucencies and lytic areas (Fig. 3A
). The foci of EMT at sites of overt TBM disruption and interstitial invasion revealed dense bundles of interstitial collagen in close apposition to cells with a mixed fibroblastic and mesenchymal phenotype, with abundant rough endoplasmic reticulum (RER) and secretory vesicles (Fig. 3B)
. The intrinsic interstitial fibroblasts also demonstrated an activated morphological phenotype (Fig. 3C
), with conspicuous RER and organized pericellular bundles of interstitial collagen. The TBM were also remarkable for frequent areas of reduplication with inclusions, consistent with persistent dysfunctional turnover (Fig. 3D
).
|
Although the transgenic kidneys generally appeared normal when examined with conventional hematoxylin/eosin staining at the age of 4 months (Fig. 4
A, cf. a, b), there was extensive loss of tubular epithelial cell staining for the epithelial marker cytokeratin (Fig. 4C, D
) as well as loss of ZO-1 in the adherens junctions of the tubules (Fig. 4
, cf. e, f). This loss of epithelial marker expression was associated with variable increases in mesenchymal marker expression (Fig. 4B
). There was widespread expression of fibroblast-specific protein-1 (FSP-1; also denoted S100A4) in the tubular epithelial cells of the MMP-2 transgenics, as well as intense staining of the intrinsic interstitial fibroblasts (Fig. 4B
, cf. a, b). Heat shock protein-47 (HSP-47) is a molecular chaperone involved in the processing and secretion of interstitial collagens (23
, 24)
, and there was extensive tubular epithelial cell expression of this mesenchymal marker in the MMP-2 transgenics (Fig. 4
, cf. c, d). In contrast to the widespread expression of FSP-1/S100A4 and HSP-47 at 4 months, expression of -smooth muscle actin (-SMA; Fig. 4
, cf. e, f) and vimentin (Fig. 4
, cf. g, h) was restricted to a more limited number of tubular epithelial and interstitial cells. Thus, there exists a range of cellular phenotypes in the MMP-2 transgenics at 4 months, suggesting that FSP-1/S100A4 and HSP-47 represent early and widely expressed markers of epithelial mesenchymal transition, while -SMA and vimentin reflect acquisition of a fully developed mesenchymal phenotype by a considerably more limited subset of tubular epithelial cells. We note (data not shown) that tissue inhibitor of metalloproteinase (TIMP)-1,-2, and -3 expression by Western blot was not significantly different in the WT and transgenics at 4 months.
|
We used quantitative PCR analysis for FSP-1/S100A4, HSP-47, vimentin, and the 1 chain of type I collagen (COL1A1) to confirm the IHC analysis (Fig. 4C
). Transcript abundance was significantly increased at 4 months for FSP-1/S100A4, HSP-47, and the 1 type I collagen chain. At this time there was no significant increase in vimentin transcript abundance, consistent with the rather limited expression observed with IHC. Transcript abundance for FSP-1/S100A4, HSP-47, and COL1A1 remained elevated at 8 months, although to a lesser degree than that observed at 4 months. Vimentin transcript abundance was increased nearly 3-fold compared with age-matched controls at 8 months, consistent with a progressive evolution to a more fully mesenchymal phenotype.
Assessment of renal injury in MMP-2 transgenic mice
We used a battery of quantitative and semiquantitative measures to gauge the extent of renal injury in transgenic mice compared with controls. Renal cortical sections were stained with Picrosirius red (PSR) and examined by polarizing microscopy. Interstitial collagens were detected by yellow birefringent staining and the collagen volume fraction (CVF) was determined by digital image analysis. Figure 5
shows representative PSR staining of WT controls and MMP-2 transgenics at 4 and 8 months of age. At 4 months, MMP-2 transgenics (control, a; transgenics, b) displayed substantial peritubular staining for interstitial collagen in a pattern consistent with tubular epithelial cell synthesis. By 8 months there was extensive deposition of collagen in the interstitium associated with a loss of tubules (control, Fig. 5c
; transgenics, Fig. 5d
). The quantitative assessment of CVF is summarized in Fig. 5e
). At 4 months, the CVF of the MMP-2 transgenics was significantly elevated (transgenics: 5.6±0.5%; controls: 1.5±0.3%, P<0.05). By 8 months of age there was a modest increase in the control CVF (2.4±0.5%); however, the CVF of the MMP-2 transgenics was greatly increased (28±3%; P<0.01).
|
Tubular atrophy is a common morphological feature of progressive renal disease and is characterized by apoptotic loss of tubular epithelial cells in response to injury (Fig. 6
A). There was no significant tubular atrophy in the MMP-2 transgenics at 4 months, but by 8 months tubular atrophy was readily evident (Fig. 6a
, cf. a, b). At 8 months the semiquantitative tubular atrophy score (Fig. 6c
) for the wild-types was 0.5 ± 0.15, while the MMP-2 transgenics had a tubular atrophy score of 1.9 ± .2 (P<0.05). Mononuclear cell infiltration is an additional measure of progressive renal injury, and the 8-month-old MMP-2 transgenics had significant degrees of interstitial cellular infiltration (Fig. 6B
, cf. a, b; scores given in panel c). Finally, there was a significant increase in the glomerular injury score characterized by mesangial matrix expansion and hypercellularity in the 8-month-old MMP-2 transgenics compared with controls (Fig. 6C
, cf. a, b; score given in panel c).
|
These morphological markers of renal injury correlated with a decrease in renal function measured by plasma creatinine. At 8 months, when established structural injury was present, MMP-2 transgenics demonstrated increases in creatinine levels, consistent with a loss of 50% of renal function (controls: 0.3±0.06 mg/dl; transgenics: 0.7±0.4 mg/dl; n=6, P<0.01).
DISCUSSION
In this report we demonstrate that targeted expression of active MMP-2 in the renal proximal tubule is sufficient to generate the entire spectrum of EMT in the absence of superimposed injury. The most evident initial changes in the transgenic kidneys were widespread lucencies within the tubular basement membrane, consistent with proteolysis by MMP-2. Although MMP-2 has multiple substrates, it is probable that the primary proteolytic target in the TBM is type IV collagen, which forms the major structural scaffolding of the basement membrane. Proteolytic targeting of the basement membrane is consistent with the localization of MMP-2 to the basolateral surfaces of the tubular epithelial cells. TBM alterations identical to those observed in the MMP-2 transgenics have been demonstrated in biopsy material from a wide variety of human glomerular diseases (25
, 26)
. Consistent with the MMP-2 transgenic, only occasional sites of overt TBM disruption were noted in the human renal biopsies (26)
.
EMT develops in two distinctive patterns. In the first, which is predominant, tubular epithelial cells express a range of mesenchymal markers associated with a loss of epithelial cytokeratin and ZO-1. When examined at the level of TEM, many cells contained bundles of actin filaments, a finding attributed to a cellular response to injury (26)
. Most cells manifested a mixed epithelial/mesenchymal phenotype, and only a relatively minor component progresses to a more fully mesenchymal phenotype with -SMA and vimentin expression. Expression of the interstitial collagen chaperone molecule, HSP-47, coupled with the peritubular localization of interstitial collagen with PSR staining, is consistent with tubular epithelial cell synthesis as opposed to a primarily interstitial fibroblast source. For purposes of clarification, we denote this pattern as intratubular EMT. At 4 months, these extensive ultrastructural and immunohistochemical changes were present in the setting of preserved renal function and normal conventional morphology by hematoxylin and eosin staining.
A second and less common pattern of EMT was observed, which we denote as extratubular EMT. This pattern consisted of small foci of overt TBM disruption associated with the migration of cells with a predominantly mesenchymal phenotype into the interstitium. While these cells contained to a varying extent filamentous actin, few, if any, conformed to the ultrastructural definition of myofibroblasts, as most lacked fibronectin fibrils and fibronexus junctions (26)
.
The two patterns of EMT continued to progress and by 8 months resulted in kidneys characterized by tubular atrophy, extensive interstitial collagen deposition, glomerulosclerosis, and impaired function in the absence of superimposed injury.
The potential role of EMT in the development of human renal disease was first analyzed by Jinde et al. (27)
, who examined 127 renal biopsies from a variety of glomerular diseases. This group noted a small number of -SMA-positive tubular epithelial cells (TEC), which also stained for collagen types I and III. More recently, Rastaldi et al. (28)
examined the hypothesis that EMT induces tubular epithelial cells to become collagen-synthesizing cells, i.e., intratubular EMT. This study of a broad spectrum of human renal diseases used a very similar panel of immunohistochemical markers employed in this report and noted that most TEC demonstrated mixed or intermediate epithelial/mesenchymal phenotypes. In addition, EMT associated with interstitial invasion (i.e., extratubular EMT) was rare, and most TEC remained within tubular structures with various degrees of acquired mesenchymal phenotypes. Taken together, these results demonstrate that the renal MMP-2 transgenic model accurately recapitulates the patterns of EMT and renal injury observed in human disease. Furthermore, the MMP-2 transgenic model extends the in vitro observations of Zeisberg et al. (10
, 11)
, which emphasized the importance of basement membrane organization for the development of EMT.
A higher frequency of the extratubular pattern of EMT has been reported using the unilateral ureteral obstruction model of renal fibrosis (3
, 29)
. This rapidly developing model is characterized by high intratubular pressures and extensive inflammatory cell infiltration, a setting conducive to more extensive TBM structural disruption. This would indicate that the frequency of the extratubular pattern of EMT is dependent on the degree of TBM dissolution and that within the setting of lesser degrees of TBM alteration, the intratubular pattern is predominant.
MMP-2 has been associated with EMT in other cellular systems. Seomun et al. (30)
demonstrated that MMP-2 mediated TGF-ß1-dependent EMT of cultured lens epithelial cells, resulting in the generation of a myofibroblast-like phenotype. Song et al. (31)
examined the role of MMP-2 in the epithelial-mesenchymal transition of cardiac endocardial cushions. These studies demonstrated that type IV collagen serves as a permissive substrate for cardiac endocardial cushion EMT only if active MMP-2 is present and type IV collagen degraded, a pattern consistent with the observations of the present study.
MMP-2 expression was recently evaluated in the pathogenesis of chronic renal allograft rejection, a process hypothesized to be mediated by EMT (32)
. Intense expression of MMP-2 mRNA was noted in the TEC and interstitium in an experimental model of renal allograft rejection (33)
. In this model, MMP-2 expression preceded the development of fibrosis and MMP-2 levels ultimately correlated with increased net collagen volumes. Lutz et al. (34)
examined the effect of a relatively selective MMP-2 inhibitor, BAY12–9566, on the development of experimental allograft nephropathy. Intriguingly, early inhibition of MMP-2 activity resulted in an amelioration of allograft nephropathy, while use of the inhibitor in the setting of established fibrosis resulted in more severe allograft nephropathy. The authors speculated that use of the MMP-2 inhibitor reduced basement membrane injury and subsequent alteration of TEC phenotypes, while later use of the inhibitor may possibly reduce proteolytic removal of accumulated ECM material.
While experimental models of type I diabetes mellitus have been associated with decreased MMP activity (35
36)
, Ergul and colleagues recently demonstrated an early up-regulation of TEC MMP-2 expression in the Goto-Kakizaki rat model of type II diabetes mellitus (37)
. Notably, early MMP-2 expression occurred within the setting of normal conventional morphology, as observed in our study with the MMP-2 transgenics. With advancing age, the Goto-Kakizaki rats develop typical lesions of diabetic nephropathy (38)
.
A clinically relevant conclusion from this study concerns the absence of a relation between expression of multiple immunohistochemical (and ultrastructural) markers of EMT and the presence of apparently normal morphology as assessed by the universally employed hematoxylin/eosin stain. The ultrastructural and histochemical methods used in this study are readily available for analysis of human biopsy tissue (27
, 28)
. We would argue that more routine application of these procedures to clinical material could provide important insights into the extent of EMT, and hence ultimate injury, even within the setting of relatively normal morphology and function. Finally, routine application of this method of analysis could provide the basis for clinical trials examining the potential benefits of selective MMP-2 inhibitors.
We believe it is important to carefully consider both the spatial (i.e., cellular sources) and temporal contexts of MMP-2 synthesis. Thus, early MMP-2 expression by TEC results primarily in an intratubular pattern of EMT, suggesting that MMP-2 inhibitors may be helpful in the restoration or maintenance of an epithelial phenotype. In contrast, within the setting of established tubular atrophy and interstitial fibrosis, therapeutic strategies aimed at reducing the activities of protease inhibitory molecules, such as plasminogen activator inhibitor-1 (39
, 40)
, or inducing MMP synthesis (e.g., bone morphogenic protein-7, ref. 41
) may be more appropriate. The MMP-2 transgenic provides an excellent model to test this hypothesis using recently developed specific MMP-2 inhibitors.
In summary, the current report provides proof-of-principle for a central and initiating role for MMP-2 in the pathogenesis of progressive renal injury. MMP-2 induction and targeted proteolysis of the TBM provides a unifying mechanism integrating the multiple pathophysiologic processes that induce epithelial cell injury. Strategies aimed at the inhibition of MMP-2, specifically in the early prefibrotic stages of disease, could have a major impact on the reduction of CKD and its attendant costs and morbidities.
ACKNOWLEDGMENTS
This study was supported by National Institutes of Health grants RO1 DK39776 (D.H.L.) and K08 DK 59383 (S.C.). We thank Leslie Cape and Anita Nguyen for technical assistance and Drs. Kirsten Johansen, Joel Karliner, Anthony Baker, and Malcolm Davies for their thoughtful comments
Received for publication February 13, 2006. Accepted for publication April 17, 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  C. C. Berthier, P. R. Wahl, M. L. Hir, H.-P. Marti, U. Wagner, H. Rehrauer, R. P. Wuthrich, and A. L. Serra Sirolimus ameliorates the enhanced expression of metalloproteinases in a rat model of autosomal dominant polycystic kidney disease Nephrol. Dial. Transplant., March 1, 2008; 23(3): 880 - 889. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Cohen-Mazor, S. Sela, R. Mazor, N. Ilan, I. Vlodavsky, A. L. Rops, J. van der Vlag, H. I. Cohen, and B. Kristal Are primed polymorphonuclear leukocytes contributors to the high heparanase levels in hemodialysis patients? Am J Physiol Heart Circ Physiol, February 1, 2008; 294(2): H651 - H658. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Boor, K. Sebekova, T. Ostendorf, and J. Floege Treatment targets in renal fibrosis Nephrol. Dial. Transplant., December 1, 2007; 22(12): 3391 - 3407. [Full Text] [PDF]
|
|
|
|
 |
|
 P. Kumpers, F. Gueler, S. Rong, M. Mengel, I. Tossidou, I. Peters, H. Haller, and M. Schiffer Leptin is a coactivator of TGF-beta in unilateral ureteral obstructive kidney disease Am J Physiol Renal Physiol, October 1, 2007; 293(4): F1355 - F1362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Krag, J. R. Nyengaard, and L. Wogensen Combined effects of moderately elevated blood glucose and locally produced TGF-{beta}1 on glomerular morphology and renal collagen production Nephrol. Dial. Transplant., September 1, 2007; 22(9): 2485 - 2496. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Daniel, K. Amann, B. Hohenstein, P. Bornstein, and C. Hugo Thrombospondin 2 Functions as an Endogenous Regulator of Angiogenesis and Inflammation in Experimental Glomerulonephritis in Mice J. Am. Soc. Nephrol., March 1, 2007; 18(3): 788 - 798. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Catania, G. Chen, and A. R. Parrish Role of matrix metalloproteinases in renal pathophysiologies Am J Physiol Renal Physiol, March 1, 2007; 292(3): F905 - F911. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A. Eddy and A. B. Fogo Plasminogen Activator Inhibitor-1 in Chronic Kidney Disease: Evidence and Mechanisms of Action J. Am. Soc. Nephrol., November 1, 2006; 17(11): 2999 - 3012. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
