Overexpression of MMP-7 increases collagen 1A2 in the aging kidney

The percentage of the U.S. population over 65 is rapidly increasing, as is the incidence of chronic kidney disease (CKD). The kidney is susceptible to age-dependent alterations in structure, specifically tubulointerstitial fibrosis that leads to CKD. Matrix metalloproteinases (MMPs) were initially characterized as extracellular matrix (ECM) proteinases; however, it is clear that their biological role is much larger. We have observed increased gene expression of several MMPs in the aging kidney, including MMP-7. MMP-7 overexpression was observed starting at 16 months, with over a 500-fold upregulation in 2-year-old animals. Overexpression of MMP-7 is not observed in age-matched, calorically restricted controls that do not develop fibrosis and renal dysfunction, suggesting a role in the pathogenesis. In order to delineate the contributions of MMP-7 to renal dysfunction, we overexpressed MMP-7 in NRK-52E cells. High-throughput sequencing of the cells revealed that two collagen genes, Col1a2 and Col3a1, were elevated in the MMP-7 overexpressing cells. These two collagen genes were also elevated in aging rat kidneys and temporally correlated with increased MMP-7 expression. Addition of exogenous MMP-7, or conditioned media from MMP-7 overexpressing cells also increased Col1A2 expression. Inhibition of protein kinase A (PKA), src, and MAPK signaling at p38 and ERK was able to attenuate the MMP-7 upregulation of Col1a2. Consistent with this finding, increased phosphorylation of PKA, src, and ERK was seen in MMP-7 overexpressing cells and upon exogenous MMP-7 treatment of NRK-52E cells. These data suggest a novel mechanism by which MMP-7 contributes to the development of fibrosis leading to CKD.

matrix, predominantly collagen, which is associated with loss of organ function as normal tissue is replaced by scar tissue (Wynn 2007). CKD is a prototypical example of progressive fibrosis leading to organ failure (Hewitson 2009;Boor et al. 2010;Zeisberg and Neilson 2010). Both glomerulosclerosis and tubulointerstitial fibrosis are involved in CKD, however, the latter is the better histological predictor of progression (Bohle et al. 1987). Increased expression of Col1a2 and Col3a1 have been previously described to correlate with aging, injury, and fibrotic changes in the kidney (Bielesz et al. 2010;Gaikwad et al. 2010;Fragiadaki et al. 2011), as well as in other systems (Wu and Chakravarti 2007;van Almen et al. 2011).
Numerous animal models have been described to study age-related alterations in the kidney (Baylis and Corman 1998). Many of the structural changes in the aged human kidney are observed in rats, such as degenerative changes in the proximal tubules and thickening of the glomerular basement membrane. Other notable functional deficits in the rat include proteinuria and reduced urine concentration (Haley and Bulger 1983;Sands 2003). Of note, the development of renal disease is more severe in males as compared to females (Baylis 1994;Sasser et al. 2012), and that nutrition affects age-related renal dysfunction (Zawada et al. 1997). In male Fischer 344 rats, we observe a progression of kidney deterioration similar to end-stage renal disease including severe glomerulosclerosis and interstitial fibrosis (Corman and Owen 1992). Lifelong caloric restriction will ameliorate this effect (Stern et al. 2001). Rat models present a well-characterized and invaluable tool to investigate age-related changes in the kidney, including consequences of glomerulosclerosis and fibrosis.
Given the development of glomerulosclerosis and tubulointerstitial fibrosis in the aging kidney, both of which are associated with increased ECM deposition, it was suggested that MMP activity would decrease during aging. In aging male Wistar kidneys, proximal tubules have been shown to have lower cysteine and metalloproteinase activity (Schaefer et al. 1994); similar results were seen in brush border-enriched fractions of male Sprague-Dawley rats (Reckelhoff and Baylis 1992). In both studies, however, the activities of specific MMPs were not characterized. However, in a microarray analysis of kidney samples from 74 patients between 27 and 92 years indicated a 2.90-fold increase in MMP-7 expression with increasing age (Rodwell et al. 2004). Interestingly, the fold change was the second largest. This finding has been confirmed in a separate study (Melk et al. 2005). Previous studies from our laboratory have indicated that MMP-7 is overexpressed in the aging rat kidney .
MMP-7 is the smallest member of the metalloproteinase family and has gained attention in the recent years for its role in abnormal tissue remodeling (Nagase and Woessner 1999). The secreted protein is minimally expressed in the adult, with the notable exceptions of the small intestine and bladder. MMP-7 is not detected in normal human renal tubular epithelium, but significant expression was seen in a number of pathologic states including polycystic kidney disease in humans and unilateral ureteral obstruction or acute folic acid nephropathy in mice (Surendran et al. 2004). It has been proposed as a new screening marker for kidney damage (Reich et al. 2011), cardiovascular complications in patients with CKD (Musial and Zwolinska 2012), and possibly for the prediction of kidney transplant rejection (Jovanovic et al. 2008;Rodder et al. 2010). In addition, MMP-7 may be involved in the development of fibrosis in the lung (Zuo et al. 2002;Rosas et al. 2008) and liver (Huang et al. 2005). There have been reports of MMP inhibitors, specifically doxycycline, successfully reducing proteinuria in patients with diabetic nephropathy (Aggarwal et al. 2010) and glomerulonephritis (Ahuja 2003), suggesting that MMPs play a pathogenic role in the development of chronic renal dysfunction. In this study, we investigated the mechanistic link between MMP-7 overexpression and fibrosis in the aging kidney.

Animals
Male Fisher 344 rats were obtained from the National Institute of Aging, Bethesda, MD, and housed in the Animal Facilities at the College of Medicine, Texas A&M Health Science Center or the University of Missouri School of Medicine. All animal protocols were submitted and approved by the Texas A&M and University of Missouri Animal Care and Use Committee in accordance with the NIH.
Animals were purchased at the indicated ages and housed for a week before being placed in metabolic cages (Tecniplast, Exton, PA) 18 h prior to sacrifice. Animals were fed ad libitum (AL) or calorie restricted (CR); CR was initiated at 14 weeks of age at 10% restriction, increased to 25% restriction at 15 weeks, and to 40% restriction at 16 weeks, which was subsequently maintained throughout the remaining life of the animal. The animal room was temperature controlled and maintained on a 12:12 h light:dark cycle. Following anesthesia (ketamine 87 mg/kg and xylazine 13 mg/kg body weight), rats were sacrificed by heart puncture, the abdominal cavity was opened, and the kidneys were removed and weighed. Kidneys were sliced into 1-mm-thick sections and either snap frozen in liquid nitrogen or frozen in liquid nitrogen-cooled optimal cutting temperature compound (Tissue-Tek; Sakura Finetek, Torrance, CA) for cryosectioning or fixed in formalin and paraffin embedded for immunohistochemistry.

MMP-7 clones
The full-length wild-type human MMP7 (NM_002423) clone in pCMV6-Neo was purchased from OriGene (Rockville, MD). The sequence was altered by oligonucleotide-directed mutagenesis exchange reactions as described previously (Geiser et al. 2001) using Quick-Change II Site-Directed Mutagenesis Kit (Stratagene/Agilent Technologies, Santa Clara, CA). The active mutant with a substitution of valine to glycine at amino acid 92 (Witty et al. 1994) was generated using the following oligonucleotides: antisense 5′-CAG ATG TGG AGG GCC AGA TGT TG-3′, and sense 5′-CAA CAT CTG GCC CTC CAC ATC TG-3′. The inactive mutant with a substitution of glutamic acid to glutamine at amino acid 216 was generated using the following oligonucleotides: antisense 5′-ATG GCC AAG TTG ATG AGT TGC-3′ and sense 5′-GCA ACT CAT CAA CTT GGC CAT-3′. Mutations were confirmed by sequencing.

Western blot
Subconfluent cells were washed twice with ice-cold PBS (phosphate buffered saline; Gibco, Life Technologies) and lysed with 10-mmol\L Tris-1% sodiumdodecyl sulphate (SDS) buffer with Halt Protease/Phosphatase inhibitor. Cells were scraped and incubated for 15 min at 4°C on a rocker. Cells were further disrupted by passing through a 20-gauge needle and spun at 12,000g for 15 min at 4°C. Tissue lysates were isolated using a 10-mmol\L Tris-1% SDS buffer supplemented with Halt Protease Inhibitor Cocktail (Thermo Fisher-Pierce, Rockford, IL). Protein concentration was determined by absorbance readings at 280 nm on a Nanodrop 2000c spectrophotometer (Thermo Fisher Scientific).

Immunohistochemistry
Kidneys were sliced with a razor blade into four sagittal sections and placed in 4% paraformaldehyde for 24 h. The sections were subsequently rinsed repeatedly with PBS, and placed in 70% ethanol for embedding. Sections were deparaffinized by xylene incubation for 12 min and rehydrated in a graded series of ethanol (95%, 80%, 70%, and 50% ethanol) for 5 min each, and then washed with PBS for 10 min. Slides were stained for collagen deposition using the NovaUltra Sirius Red Stain Kit, IHC WORLD, Woodstock, MD.
Cells were imaged on an Olympus IX51 microscope with a UC50 digital camera using cellSense software (Olympus, Center Valley, PA) at equal exposure times.

In-cell Western blot
Subconfluent cells grown in 96-well opaque clear bottom cell culture plates were washed with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were permeabilized with 0.1% Triton X-100 and endogenous peroxidase was quenched with H 2 O 2 and NaN 3 for 20 min. Cells were blocked with normal goat serum for 1 h and incubated with primary antibody at a 1:100 dilution overnight followed by washing as above and addition of secondary antibody at 1:1000 for 1 h. Blots were developed using West Femto (Pierce, Thermo Fisher Scientific), and chemiluminescence was read using a Synergy HT microplate reader with Gen5 software (BioTek, Winooski, VT) and imaged with ChemiDoc imaging system (Bio-Rad). Cells were then washed with PBS, stained with Janus Green stain for 1 min, washed and eluted in 100% ethanol. Absorbance was read at 594 nm. Chemiluminescence signal was normalized per cell number, and the negative control (secondary antibody only) signal was subtracted from an average of three wells per antibody. Expression was then reported relative to the b-actin signal.

RNA isolation and cDNA synthesis
RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA) for animal tissue analysis and sequencing samples, and with the Tissue/cell total RNA mini kit (EZ BioResearch, St. Louis, MO) for inhibitor studies. Snap-frozen kidney tissues were lysed with RNeasy lysis (RTL) buffer (Qiagen) supplemented with b-mercaptoethanol and homogenized using a motorized pellet pestle (Kontes, Vineland, NJ) followed by centrifugation in the Qiashredder (Qiagen). Cultured NRK-52E cells were trypsinized, pelleted, and lysed with RTL buffer (Qiagen) supplemented with b-mercaptoethanol and passed 5 times through a 20-gauge needle. On-column DNase digestion was performed for both tissues and cells. RNA concentration and quality was determined by spectrophotometry on a Nanodrop 2000c and confirmed by agarose gel electrophoresis. cDNA was generated using the iScript cDNA Synthesis Kit (Bio-Rad) for initial MMP and TIMP screening, and the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Life Technologies) was used for later experiments.

Real-time polymerase chain reaction
Initial MMP and TIMP screening was performed using the iCycler iQ real-time polymerase chain reaction (PCR) detection system (Version 3.1; Bio-Rad) and iQ SYBR â Green Supermix (Bio-Rad). Genes of interest were targeted using specific RT² Real-Time PCR primer sets (SuperArray; SABiosciences, Qiagen). Relative quantitation was performed using the ΔΔCt method in which the quantity of target gene mRNA in each experimental sample (young, aged-AL or aged-CR) relative to an internal standard (ß-actin mRNA) is normalized to an arbitrary reference sample (Universal Rat Reference RNA; Stratagene) (Akintola et al. 2008). In subsequent experiments, we used custom primer/probe Taqman â Assays (Applied Biosystems, Life Technologies) and the Sso Fast mix (Bio-Rad) with the CFX96 Touch real-time PCR system (Bio-Rad). Analysis was performed using the ΔΔCt method relative to Casc3 and ß-actin.

Illumina sequencing
RNA from the normal rat kidney parent cell-line NRK-52E, as well as cells stably expressing wild-type MMP7, active mutant MMP7, and control vector, was submitted for high-throughput sequencing. A mRNA-focused, barcoded library was generated using the TruSeq kit (Illumina, San Diego, CA) and analyzed using the HiSeq 2000 platform from Illumina at the DNA Core Facility at the University of Missouri. The sequencing reaction yielded~7.5 Gb of data, corresponding to around 30 million 50-base reads per sample across the whole transcriptome. The Informatics Research Core Facility at the University of Missouri aligned the reads against the rat genome (Rattus norvegicus RGSC3.4; Ensemble, Hinxton, UK) and analyzed them using Bowtie (Langmead and Salzberg 2012), TopHat and Cufflinks (Trapnell et al. 2012) software. Differential expression values defined as fragments per kilobase of transcript per million mapped reads with a false discovery-corrected P-value equal or lower than 0.05 were considered significant. The raw data from our Illumina high-throughput sequencing has been deposited in the Sequence Read Archive (SRA) with the National Center for Biotechnology Information (Bethesda, MD) under the project PRJNA213322, accession number SRP02851, experiment MMP7, accession number SRX327868, and will be made available upon publication of this manuscript.

Statistics
For mRNA expression, in-cell Western, and enzymatic assay results, a two-tailed t-test assuming two-sample equal variance was performed with P-values <0.05 considered statistically significant.

Results
Age-related overexpression of MMP-7 Given the importance of MMPs in acute and chronic renal pathophysiologies ), we examined the mRNA expression of all MMPs and TIMPs in young (4 month-old), aged, 24-month-old AL fed, and aged CR rat kidneys by quantitative PCR. Using rat-specific primers, we found expression of many MMPs that have not yet been linked to the kidney, including MMP-1, -16, -17, -20, -21, and -25 (Fig. 1A). In contrast to a previous report investigating human MMP-2 and MMP-24 (Romanic et al. 2001), expression of MMP-15 and -24 was not detected in the rat kidney. Importantly, we identified several MMPs whose gene expression was significantly changed as a function of aging, including MMP-2, -3, -7, -9, -12, -13, -14, -16, -17, -19, -20, -23, and -25, as well as TIMP-1. Of these, the increased expression of MMP-2, -7, -9, -12, -13, -14, -16, -20, -23, and -25 was attenuated by caloric restriction, as was TIMP-1. As MMP-7 exhibited the most dramatic increase in the aged animals and is overexpressed in the aging human kidney (Rodwell et al. 2004;Melk et al. 2005), we examined MMP-7 expression over an extensive time course. At 16 months expression was significantly upregulated, and increased to over 500-fold upregulation in 2-year-old animals (Fig. 1B). Importantly, increased gene expression correlated with increased protein expression as assessed by Western blot (Fig. 1C). The temporal pattern of MMP-7 overexpression, and the finding that it is not overexpressed in caloric restriction controls, suggests that MMP-7 may play a pathogenic role in the development of chronic renal dysfunction.

MMP-7 overexpression: collagen expression
In order to delineate the effects of MMP-7 overexpression in the kidney, we stably overexpressed MMP-7 in NRK-52E cells. As epithelial cells do not activate MMP-7 in vitro (Witty et al. 1994), we overexpressed wild-type MMP-7, an active mutant of MMP-7, and a catalytically inactive mutant. The active mutant has a point mutation resulting in a valine to glycine substitution at position 92 ( Fig. 2). This mutation in the prodomain allows for an autocatalytic cleavage of the zymogen to produce a catalytically active MMP-7. The inactive mutant has a point mutation in the catalytic domain at position 216. Overexpressed MMP-7 was detectable in the NRK-52E cells and was secreted into the medium (Fig. 2). In conditioned medium from wild-type and the inactive mutant overexpressing NRK-52E cells, only the 30 kDa zymogen was visible on the Western blot. Expression of the active form was lower as determined by real-time PCR and Western blot, and bands representing both the 30 kDa pro-and a 18 kDa active form were detected. Each of the MMP-7 overexpressing cells exhibited comparable doubling times, which were shorter than those of the parent NRK-52E cell line, probably due to the strong cytomegalovirus promoter in the vector (data not shown). It is important to note that the relative expression of pro-MMP-7 appears to be higher in the wild-type and inactive mutant constructs than in the active mutant, which still expressed pro-MMP-7. High-throughput sequencing of mRNA libraries generated from MMP-7 overexpressing cells yielded promising target genes, including Col1a2 and Col3a1, interestingly, in both the WT and active mutant MMP-7 overexpressing cells ( Fig. 3A; Table 1). While WT overexpressing cells had the largest increase in collagen expression, the catalytic activity of MMP-7 may be important given the findings that the active mutant cells also were characterized by collagen overexpression and that this effect was significantly decreased in the inactive mutant cells. Increased collagen deposition is characteristic of the aging rat kidney ( Fig. 3B). As expected, expression of both collagens increased with age and paralleled the temporal changes in MMP-7 overexpression (Fig. 3C).

Discussion
Chronic kidney disease is accompanied by excessive accumulation of extracellular matrix resulting in renal fibrosis. Fibrosis is a slow and incremental process resulting from repeated injury events accumulating over time. The process, which takes several decades in the human, is accelerated in the rat. As with the individual variability across the human population, the aging process between rat strains varies in respect to the kidney (Baylis and Corman 1998).
The male F344 rat used in this study represents a population prone to developing CKD; we detected increased collagen deposition in these animals by 18 months. MMP-7 [aka matrilysin (Abramson et al. 1995), matrilysin-1, pumppunctuated metalloproteinase (Woessner and Taplin 1988), pump-1putative metalloproteinase 1 (Muller et al. 1988;Quantin et al. 1989), matrin (Miyazaki et al. 1990)] is the smallest member of the matrix metalloproteinase family. It is structurally different  from other members of the MMP family in that it lacks the C-terminal hemopexin domain, and has instead an atypical sixth exon (Gaire et al. 1994). The protease is synthesized as a 30 kDa (267aa) inactive proform and is then stepwise activated to a final 18 kDa (177aa) form. MMP-7 is fully activated by trypsin and MMP-3, and is partially activated by plasmin, leukocyte elastase (Imai et al. 1995), or aminophenylmercuric acetate (APMA) in vitro. MMP-7 is expressed at very low levels in the adult, and only in a few tissues; however, it has gained attention due to its presence in a variety of disease states including cancer (Ramankulov et al. 2008) and CKD (Musial and Zwolinska 2012). In aging male Fisher 344 rats, MMP-7 was upregulated by over 500-fold in old animals compared to young. MMP-7 activity has been previously reported in association with fibrotic changes in the kidney ) and other fibrotic conditions, such as idiopathic pulmonary fibrosis (Zuo et al. 2002;Rosas et al. 2008) and liver fibrosis (Huang et al. 2005). In these studies, we demonstrate a link between MMP-7 and collagen expression, suggesting a mechanistic link to fibrosis that is counterintuitive given the role of MMP-7 in degradation of the extracellular matrix (Fig. 5).
We found that upregulation of MMP-7 in a normal rat cell-line NRK-52E results in upregulation of two collagen genes, Col1a2 and Col3a1. Both genes are also upregulated in aging Fisher 344 rat kidneys. As Col1a2 was upregulated fourfold and Col3a1 only twofold, we focused our inhibitor experiments on type I collagen. In the MMP-7 overexpressing NRK-52E cells, we were able to inhibit the MMP-7-induced upregulation of Col1a2 by using inhibitors against PKA, PI3K, src, p38, and ERK. When analyzing sequencing data, we were surprised to find no significant changes in expression in any of the major pathway members identified by the inhibitor screen (data not shown). However, it has been reported that inhibiting PI3K and MEK1/2 reversed the proliferative effects of MMP-7 in human gastric myofibroblasts by inhibiting phosphorylation of Akt and ERK1/2 (Varro et al. 2007). Exogenous MMP-7 treatment has also been reported to promote EGFR-activated MEK signaling, as demonstrated by increase in p-EGFR, p-MEK, and p-ERK in pancreatic cancer cells (Tan et al. 2005). We therefore investigated the effect of MMP-7 overexpression on activating phosphorylation status of ERK, src, and PKA. We found increased phosphorylation of each of these proteins in the MMP-7 overexpressing cells compared to vector control cells and we were also able to induce phosphorylation by exogenous MMP-7 treatment of vector control cells.
The human COL1A2 promoter has been described previously (Ramirez et al. 2006). Stimulation of transforming growth factor beta (TGFb) signaling results in upregulation  of Col1A2, via transmembrane serine/threonine kinases and intracellular Smad proteins (Massague et al. 2005). This requires the interactions of Sp1, Smad3/4 (Zhang et al. 2000), and p300/CREB-binding protein (Ghosh et al. 2000) on the COL1A2 promoter. MMP-7 has been implicated in the activation of EGFR and upregulation of TGFb (Mimori et al. 2004). In the MMP-7 overexpressing cells, however, TGFb expression was not altered, nor was that of any of the Smad proteins (data not shown). Thus, MMP-7 may be regulating Col1A2 via a non-TGF pathway. While a paradoxical relationship between expression of MMP-7 and fibrosis has been demonstrated, putatively due to an aberrant wound healing response, (Huang et al. 2005;Wu and Chakravarti 2007;Rodder et al. 2010), a mechanistic link has not been delineated. Our data suggest that MMP-7 increases collagen expression in an autocrine fashion, independent of inflammation. This is consistent with the autocrine activation of ERK1/2 induced by MMP-2 (Xue and Jackson 2008). Our data suggest that the proteolytic activity of MMP-7 may not be required for induction of collagen expression, as the WT MMP-7, which is not processed to an active form in vitro results in elevated Col1a2 and Col3a1 expression. The fact that the collagen expression is higher in the WT than in the active mutant could result from the fact that there is significantly more total MMP-7 in the WT that in the active mutant, both at mRNA and secreted protein level. However, the fact that we do not see similar increases in collagen expression in the inactive mutant cell line does suggest a role for activation. Interestingly, in whole kidney lysates from the aging kidney, we have only observed pro-MMP-7 and not the active form, and we have not detected active MMP-7 by zymography in either kidney lysates or urine (data not shown). We conclude, based on the inability to detect active MMP-7 in the aging kidney, that pro-MMP-7 is upregulating collagen expression and, therefore, has a pathophysiological role in renal fibrosis. In addition, MMP-7 has not been reported to degrade Col1a2 and Col3a1. The only collagens demonstrated to be MMP-7 targets are collagen type 4 (Kraft et al. 2001) and collagen type 18 (Lin et al. 2001). However, MMP-7 activates the gelatinases MMP-2 and -9 (von Bredow et al. 1998), and the collagenases MMP-1 and -8, which in turn degrade collagen, but we have not detected MMP-8 expression in the rat kidneys, and MMP-1 expression decreases with age. We have also observed decreased total collagenase and increased gelatinase activity in the aging kidney (24 month) in whole kidney lysates (data not shown). Interestingly this effect is only observed in the presence of APMA to activate latent MMPs. Recent studies have shown that noncatalytic domains of MMPs have signaling effects (Correia et al. 2013;Mori et al. 2013;Vandooren et al. 2013), suggesting that noncatalytic functions of MMPs may have important implications. Although MMP-7 lacks many domains common to other MMPs, future studies will focus on identifying specific MMP-7 domains that mediate collagen overexpression.
In this study we demonstrate a mechanistic link between MMP-7 and fibrosis. The early upregulation of MMP-7 causes increased transcription of Col1a2 and Col3a1 genes primarily via PIK3, p38, ERK, src, and PKA signaling, leading to subsequent collagen deposition in the kidney.