Kidney International, Vol. 63 (2003), pp. 1736–1748

Increased expression of MMP-2, MMP-9 (type IVcollagenases/gelatinases), and MT1-MMP in canineX-linked Alport syndrome (XLAS)

VELIDI H. RAO, GEORGE E. LEES, CLIFFORD E. KASHTAN, RYOCHI NEMORI, RAKESH K. SINGH,DANIEL T. MEEHAN, KATHYRN RODGERS, BRIAN R. BERRIDGE, GAUTAM BHATTACHARYA,and DOMINIC COSGROVE

Boys Town National Research Hospital, Omaha, Nebraska; Texas A & M University, College Station, Texas; University ofMinnesota Medical School, Minneapolis, Minnesota; Fuji Photo Film, Co., Ltd., Kanagawa, Japan; andUniversity of Nebraska Medical Center, Omaha, Nebraska

Increased expression of MMP-2, MMP-9 (type IV collagenases/ Alport syndrome (AS) is a group of genetic disordersgelatinases), and MT1-MMP in canine X-linked Alport syndrome resulting from mutations in collagen IV genes. The patho-(XLAS). logic hallmark of the disease is irregular thickening, thin-Background. Alport syndrome is a group of genetic disor-

ning, and splitting of the glomerular basement membraneders resulting from mutations in either the �3(IV), �4(IV)(GBM) revealed by electron microscopic examinationor �5(IV) collagen chains. The disease is characterized by a

progressive glomerulonephritis, usually associated with a high- of renal tissue. Patients with AS frequently have progres-frequency specific sensorineural hearing loss, dot and fleck sive high-frequency sensorineural hearing loss, retinalretinopathy, and lens abnormalities. Dogs with naturally oc- flecks, and lens abnormalities in addition to the renalcurring genetic disorders of basement membrane collagen (type

phenotype [1]. In humans, most cases of AS are causedIV) may serve as animal models of Alport syndrome. In thisby mutations in the collagen �5(IV) gene, which residesstudy, a well-characterized naturally occurring canine model

was employed to demonstrate a potential role for matrix metal- on the X chromosome, and thus results in a severe domi-loproteinases (MMPs) in Alport renal disease pathogenesis. nant form of the disease in males, and a less severe pheno-

Methods. Adolescent male dogs that developed renal failure type in females due to random X chromosome inactiva-were euthanized and necropsied. Clinicopathologic features oftion resulting in phenotypic mosaicism [2, 3]. Autosomal-the disease were characterized, and kidneys from normal andrecessive forms of AS are less common (about 15% ofAlport dogs were analyzed by gelatin zymography, Western blot-

ting, in situ zymography, immunohistology, and by reverse tran- cases), and result from mutations in the collagen �3(IV)scription polymerase chain reaction (RT-PCR) for expression and �4(IV) genes [4, 5]. In human AS, the mutations inof MMP-2, MMP-9, and membrane type 1-MMP (MT1-MMP).

�3(IV), �4(IV), or �5(IV) chains of type IV collagenResults. Affected dogs developed proteinuria and rapidlyare often associated with a loss of these chains from theprogressive juvenile-onset chronic renal failure. The activitiesGBM, leaving only collagen �1(IV) and �2(IV) chains,of MMP-2 and MMP-9 were significantly induced in Alport

kidney. In situ zymography confirmed elevated active metallo- although exceptions have been noted [6]. Two gene knock-proteinases in kidney cryosections of affected dogs. The mRNAs out mouse models and a random transgene insertion mouseencoding MMP-2, MMP-9 and MT1-MMP were also increased model for autosomal Alport syndrome have been de-in Alport dogs suggesting that elevated expression of MMPs

scribed [7–9]. Three different dog models, including tworeflects events in the progression of Alport syndrome in dogs.with X-linked inheritance and one with autosomal inher-Conclusion. Elevated expression of MMP-2, MMP-9, and

MT1-MMP is observed in fibrotic renal cortex from X-linked itance, have been described [10–13]. All of these modelsAlport syndrome dogs. These findings suggest that MMPs may are deficient in collagen �3(IV), �4(IV), and �5(IV)play an important role in matrix accumulation associated with chain deposition in the GBMs, and all culminate in glo-progressive renal scarring in this model.

merular and interstitial fibrosis.Clinicopathologic studies of various glomerular dis-

Key words: Alport syndrome, fibrosis, matrix metalloproteinase. eases have demonstrated that tubulointerstitial pathol-ogy correlates with the degree and progression of renalReceived for publication July 30, 2002impairment, regardless of the type and anatomic originand in revised form December 5, 2002

Accepted for publication January 3, 2003 of the inciting injury [14, 15]. The interstitial matrix de-posited in the scarred kidney is produced by resident 2003 by the International Society of Nephrology

1736

Rao et al: MMPs in canine Alport syndrome 1737

cells, including activated fibroblasts and tubular epithe- scores were based on a semiquantitative assessment ofthe extent of cortical interstitial fibrosis observed in 3 �mlial cells with a fibrinogenic phenotype [16]. A net accu-

mulation of matrix may result from changes in synthesis, thick, hematoxylin and eosin–stained sections of kidneyfrom each dog. Sections were assigned scores on a scaledegradation, or both. Changes in synthesis of matrix

components in fibrosis have been addressed; however, from 0 to 3�, with 0 representing normal kidney withscant to no interstitial connective tissue, and 1� to 3�the contributing role of matrix turnover has not received

similar attention. representing progressively increasing severity of intersti-tial fibrosis. Fibrosis scores of 3� were representativeDevelopmental and homeostatic remodeling of the

extracellular matrix (ECM) is a highly regulated process of the upper limit of severity of renal cortical fibrosisgenerally observed in male Alport dogs necropsied atorchestrated by a family of zinc-containing, calcium-depen-

dent, secreted neutral proteases known as the matrix metal- the standardized end point used in this study. For ourstudies, necropsy specimens of kidney from five Alportloproteinases (MMPs). This family of enzymes, which now

contains over 25 members, can collectively degrade all dogs and five normal dogs were analyzed.structural proteins of the ECM, including interstitial col-

Tissue processinglagens (I, II, III and V), basement membrane collagens(IV), fibronectin, laminin, proteoglycans and elastin. The Necropsies were performed after euthanasia and a

portion of renal cortex was formalin-fixed and later par-MMPs include collagenases, gelatinases, stromelysins,macrophage metalloelastase, matrilysin, and membrane- affin-embedded, sectioned, and stained for light micro-

scopic examination. Another portion was immediatelytype MMPs. They are regulated at the level of genetranscription by latent proenzyme activation and are in- immersed in liquid nitrogen and kept at �80�C until use.

For further processing, tissues were thawed, weighed, andhibited by a group of inhibitor proteins known as tissueinhibitors of MMP, which consist of four family members homogenized in Tris buffer (0.5 mol/L Tris-HCl, pH 7.5,

containing 200 mmol/L NaCl and 0.2% Triton X-100),(TIMP-1 to TIMP-4) [17, 18].The relative balance of MMPs and TIMPs is thought centrifuged, and the supernatants were aliquoted and

stored at –80�C. The protein content in the supernatantsto determine the rate of ECM turnover. An imbalance inthese ratios can contribute to the progression of some was determined using the micro BCA assay kit according

to the manufacturer’s instructions (Pierce Biochemicals,pathologic disorders, including tumor invasion and meta-stasis, arthritis, fibrotic diseases, atherosclerosis, and em- Rockford, IL, USA) and stored at –20�C until used.physema [19]. A highly regulated balance of active MMPs

Gelatin zymographyand TIMPs is maintained during normal tissue metabo-lism. To date, there have been no in vivo studies of the Substrate gel electrophoresis (zymography) was per-

formed to identify whether the extracts contained MMPcontribution of MMPs and TIMPs to the developmentof fibrosis in AS. Using dogs with X-linked Alport syn- activity and to identify the enzymes involved. The gela-

tin-degrading activity was examined by electrophoresisdrome, we studied the expression of MMP genes and histo-logic evidence of kidney fibrosis. The results of our study on an 8% sodium dodecyl sulfate polymerase acrylamide

gel electrophoresis (SDS-PAGE) containing gelatin (1.0revealed a strong correlation between MMP gene expres-sion, MMP gelatinase activity, and histopathologic evi- mg/mL) without prior heating or reduction as described

[20, 21]. Prestained SDS-PAGE protein standards (Bio-dence of fibrosis, suggesting that deregulated matrix re-modeling may contribute to the pathology of the disease. Rad Laboratories, Richmond, VA, USA) and condi-

tioned media from human HT1080 cells, which containMMP-2 and MMP-9, were run in parallel to determine

METHODSthe molecular weights of gelatinases present in kidney

Dogs tissue extracts. After electrophoresis, the gels were washedin 2.5% Triton-X 100 and incubated in 50 mmol/L Tris-Animals were from a colony of dogs originally referred

to as the Navasota kindred, which is on a mixed back- HCl buffer, pH 7.5, containing 0.15 mol/L NaCl, 10mmol/LCaCl2, and 0.02% NaN3. Gels were stained withground [11]. Recently, a 10 base pair deletion has been

identified in exon 9 of the type IV collagen �5 gene, Coomassie Brilliant Blue R250 and destained. Gelatino-lytic activity of each gelatinase is evident as a clear bandconfirming these dogs serve as a genetic model for XLAS

(Cox et al, unpublished). Progression of renal disease in against the blue background of stained gelatin. The addi-tion of molecular weight markers and MMP-2 standardAlport dogs was monitored to a standardized end point,

serum creatinine �5.0 mg/dL (normal reference range, to the gels facilitated identification of the enzymes. Theaccuracy and sensitivity of the zymographic technique0.5 to 1.5 mg/dL) or onset of uremic symptoms (anorexia

and vomiting on 2 consecutive days), whereupon each for determining protease levels was analyzed by runningdifferent amounts of purified MMP-2 and MMP-9. Thisdog was necropsied. Normal dogs used as controls were

unaffected littermates of Alport dogs. Renal fibrosis confirmed the linearity of zymography (not shown).

Rao et al: MMPs in canine Alport syndrome1738

In order to determine whether clarified zones were lenses configured with a SpotRT digital camera and Im-age Pro-plus software (Tokyo, Japan). Tissue sectionsdue to MMPs or to serine or aspartic proteases, duplicate

gels were incubated in proteolysis buffer with addition were reacted with preimmune serum and then with sec-ondary antibody as a control for background immuno-of appropriate inhibitors for the enzyme class. The inhib-

itors used were 10 mmol/L ethylenediaminetetraacetic staining.acid (EDTA) for MMPs; 1.0 mmol/L phenylmethylsulfo-

Reverse transcription polymerase chain reactionnyl fluoride (PMSF) for serine proteases, and 5.0 �mol/L(RT-PCR) analysispepstatin A for acid and aspartic proteases.

Semiquantitative analysis of MMP-2, MMP-9, andDetection of gelatinolytic activity by film in situ MT1-MMP transcripts was performed by RT-PCR as de-zymography (FIZ) scribed previously [21]. Briefly, total RNA was isolated

using Trizol (GIBCO BRL, Gaithersberg, MD, USA)The gelatinolytic activity was also detected by FIZusing cross-linked gelatin film (gelatinolytic-type film) ac- according to the manufacturer’s instructions. Total RNA

(5 �g) was reverse transcribed by using Superscript IIcording to manufacturer’s instructions (Fuji Photo FilmCo., Ltd., Tokyo, Japan) with modifications. Briefly, sam- with oligo dT primers (GIBCO BRL, Gaithersberg, MD,

USA)ples of kidney were embedded without fixation in Tissue-Tek 22-oxacalcitriol (OTC) compound (Sonkura Fine- The reverse transcribed cDNA was used in each PCR,

in a total volume of 50 uL, with specific primers for targettek, Torrance, CA, USA). These frozen blocks were thensliced sequentially using a cryostat microtome to prepare molecules [MMP-2, MMP-9, MT1-MMP, and glyceralde-

hydes-3-phosphate dehydrogenase (GAPDH)]. Semiquan-serial frozen-thin sections (6 um thickness). These thinsections were placed on a polyethylene terephthalate titation of MMP-2, MMP-9, and MT1-MMP cDNAs was

performed using densitometry and comparison to GAPDHbase film (Fuji Photo Film Co., Ltd.), coated with cross-linked gelatin of 7 �m in thickness. The films with sec- (internal standard) from the same sample after using an

image analysis system (Imagequant Software, Moleculartions were incubated for 12 hours at 37�C in a humidifiedchamber and stained with 1% Amido Black 10B in 70% Dynamics, Sunnyvale, CA, USA). The intensity of each

band was expressed in arbitrary units.methyl alcohol, 10% acetic acid, and 10% glycerol for10 minutes, destained in 70% methyl alcohol, 10% acetic The oligonucleotide primer pairs used were as follows:

canine MMP-2 (GeneBank accession no. AF177217), 5�-acid, and 10% glycerol for 10 minutes and kept in 20%glycerol for 20 minutes. The sections were cover-slipped, GCC CCA AAG AGA GAT GCA ACC TAT-3� (sense)

and 5�-CGG CCA AAG TTG ATC ATG ATG TC-3�sealed, and imaged. The gelatin in contact with the proteo-lytic areas of the sections was digested, and thus negative (antisense); human MMP-9 (GeneBank accession no.

J05070), 5�-GGT CCC CCC ACT GCT GGC CCT TCTstaining indicated zones of enzymatic activity. To confirmthe specificity of in situ zymography, frozen sections were ACG GCC-3� (sense) and 5�-GTC CTC AGG GCA CTG

CAG GAT GTC ATA GGT-3�(antisense); canine MT1-also placed on gelatin-coated films containing 1,10-phen-anthroline (gelatinase inhibitor [GI] films), which is known MMP (GeneBank accession no. AF097638), 5�-TGT TTG

ATG AAG CTT CTC TGG A-3� (sense) and 5�-TCCto inhibit the activity of MMPs.AGT ATT TGT TCC CCT TGT A-3� (antisense); and

Immunofluorescence evaluations human GAPDH (Genebank accession no. M33197), 5�-GGT GAA GGT CGG AGT CAA CGG ATT TGGKidneys were embedded in Tissue-Tek OTC aqueous

compound. They were frozen and sectioned at 4 �m on TCG-3� (sense) and 5�-GGA TCT CGC TCC TGG AAGATG GTG ATG GG-3� (antisense). Each PCR was car-a cryostat. Tissue sections were fixed for 10 minutes in

acetone. After washing the slides in phosphate-buffered ried out at various numbers of cycles under identicalconditions to optimize the reactions. The amplified prod-saline (PBS), they were incubated in a moist chamber

with the appropriate dilution of primary antibodies. The ucts were identified by electrophoresis of an aliquot on2.0% agarose gel stained with ethidium bromide, visual-antibodies used were anti-human MMP-2, MMP-9 (a gift

from Dr. Z. Gunza-Smith, Miami, FL), and rabbit poly- ized by ultraviolet transillumination, and photographed.DNA bands were analyzed by computer-assisted densi-clonal anti-MT1-MMP antibodies (Calbiochem, La Jolla,

CA, USA). After washing, the slides were incubated with tometric scanning of these images using an image analysissystem (Imagequant Software, Molecular Dynamics).affinity-purified fluorescein isothiocyanate (FITC)-con-

jugated goat antirabbit immunoglobulin G (IgG) (Vector The intensity of each amplified product increased withthe increasing number of PCR cycles. The number of PCRLaboratories, Burlingame, MA) and mounted with Vec-

tashield mounting media for fluorescence (Vector Labo- cycles for each product was determined to lie in thelinear range with respect to specific band intensity. Theratories). The sections were cover-slipped, sealed, and

imaged. Images were collected using an Olympus (Tokyo, PCR cycling parameters for MMP-2 were as follows:initial denaturation temperature at 94�C for 10 minutesJapan) BH-2 fluorescence microscope with PlanAPO

Rao et al: MMPs in canine Alport syndrome 1739

Fig. 1. General morphology of renal cortex from X-linked Alport syndrome (XLAS) dogs typical of those used in this study. Tissue was embeddedin paraffin, cut at 3 �m, and stained using hematoxylin and eosin. Fibrosis scores, determined as described in the text, are indicated. C is normal.

Table 1. Animals used in the study specific cycle numbers were as follows: for MMP-9, 30cycles; for MT1-MMP, 34 cycles; and GAPDH, 25 cycles.Age at Serum Renal

Sample necropsy creatinine fibrosis In each instance, the constitutively expressed GAPDHnumber Status Gender months mg/dL score mRNA was used as an internal standard for RT-PCR.

1 Normal Male 12 1.1 0 The PCR procedure was repeated at least three times2 Normal Male 7 1.1 0

for each sample. No PCR product was observed in sam-3 Normal Male 7 1.0 04 Normal Male 11 1.1 0 ples in which either cDNA or reverse transcriptase was5 Normal Male 11 1.2 0 omitted. To control for sample variation due to handling,6 Affected Male 11 5.0 2�

band intensities for each unknown were normalized to7 Affected Male 13 5.5 2�8 Affected Male 8 5.0 3� band intensities for GAPDH, which was amplified in9 Affected Male 7 4.9 3� parallel.10 Affected Male 11 5.1 3

DNA sequencing

The MMP-2, MMP-9, MT1-MMP, and GAPDH frag-ments amplified from XLAS and control kidneys by PCR(one cycle), annealing at 94�C for 60 seconds, 60�C forwere purified using Multiscreen PCR. The resultant60 seconds, and 72�C for 90 seconds (32 cycles), andDNA was subjected to a cycle sequencing reaction usingfinal extension at 72�C for 10.0 minutes (final cycle).the ABI Prism dye-labeled terminator reagent (PE Bio-For amplification of the other mRNA transcripts, thesystems, Foster City, CA, USA), according to the proto-parameters were kept constant except for the cycle num-

ber, which were optimized for each pair of primers. The col provided by the manufacturer. The reaction product

Rao et al: MMPs in canine Alport syndrome1740

Fig. 2. Immunofluorescence staining for matrix metalloproteinases (MMP-2 and MMP-9) and membrane type 1 MMP (MT1-MMP) in normaland X-linked Alport syndrome (XLAS) renal cortex. Cryosections from normal (A, B, and C) and XLAS (D, E, and F ) dogs were subjected toimmunofluorescence immunostaining using antibodies specific for the indicated MMP. Cryosections (G, H, and I ) were reacted with preimmuneserum as a control for background immunostaining. Bar is 100 �m.

was analyzed with an ABI automated DNA sequencer, TBST and were incubated with peroxidase-conjugatedsecondary rabbit antibody. Finally, the color was devel-model 377 (Applied Biosystems, Palo Alto, CA, USA).oped using diaminobenzidine and hydrogen peroxide as

Western blot analysis described [22].The specificity of MMP-2, MMP-9, and MT1-MMP

Data presentation and statistical analysisantibodies used in the immunohistochemical analyseswas characterized by Western blot analysis. The tissue Data are expressed as mean � SD. Differences be-

tween means were tested for significance using Studentextracts were subjected to standard 8% SDS-PAGE un-t test. Differences were considered significant at the levelder reducing conditions as described earlier [21]. Pro-of P 0.05.teins were then electrophoretically transferred onto a

nitrocellulose membrane, blocked with 5% nonfat drymilk in Tris-buffered saline plus 1% Tween-20 (TBST)

RESULTS(Bio-Rad, Hercules, CA, USA), followed by incubation

Clinicopathologic evaluationswith diluted rabbit polyclonal antibodies (MMP-2, 1:300;MMP-9, 1:300; and MT1-MMP, 1:500) overnight at 4�C. The Alport dogs manifested progressive renal disease

typical of that previously described in males from thisThe membranes were washed again three times with

Rao et al: MMPs in canine Alport syndrome 1741

kindred [11]. Affected dogs used in this study developedpersistent proteinuria [urine protein to creatinine (UPC)ratio 0.5 then and thereafter] at 3 to 5 months of age.Onset of azotemia (serum creatinine �1.6 mg/dL) occurredat 4 to 8 months of age, and their renal failure progressedto the standardized end point at 7 to 13 months of age.Normal dogs did not exhibit proteinuria (UPC �0.5)and were not azotemic. Interstitial fibrosis was not ob-served in the kidneys of normal dogs. However, the tubu-lointerstitium of the renal cortex of all Alport dogs wasexpanded by fibrous connective tissue admixed withmixed mononuclear inflammatory cells typical of thechronic inflammatory renal changes associated with thissyndrome. Characteristic histology of renal cortex fromnormal dogs and from Alport dogs that were scored at1�, 2�, and 3� are provided in Figure 1. Ages at nec-ropsy, serum creatinine levels, and renal fibrosis scoresof dogs used in this study are summarized in Table 1.

Immunofluorescence microscopy for MMP-2, MMP-9,Fig. 3. Western blot analysis of matrix metalloproteinases (MMP-2and MT1-MMPand MMP-9) and membrane type 1 MMP (MT1-MMP) in normal and

We examined MMP-2, MMP-9, and MT1-MMP ex- X-linked Alport syndrome (XLAS) tissue extracts. C is tissue extractsfrom unaffected; and A is from an XLAS dog with a fibrotic score ofpression in normal and XLAS renal cortex. Representa-3�. Results demonstrated that the antibodies used in the immunofluo-tive fields from normal and XLAS dogs with a fibrosisrescence studies are specific for the respective MMPs in dog tissues.

score of 2 are shown in Figure 2. No obvious qualitativedifferences were observed when comparing sectionsfrom animals with a fibrosis score of 2 or 3. Expression dog kidney extracts from necropsy samples was analyzedof MMP-2 and MMP-9 in controls is weakly evident in by gelatin zymography. Figure 4 is provided to validatethe tubular epithelial cells (Fig. 2A and B). In Alport the specificity of the assay to detect latent and active formsdogs, however, brightly stained cells are evident in the of the MMPs. Extracts from XLAS kidney contained sev-interstitial space. MMP-9 also appears to accumulate in eral protease activities at molecular weights of 92, 84,the basement membranes and the interstitial matrix near 72, and 64 kD. The upper gelatinolytic bands representclusters of brightly stained cells (Fig. 2E). MT1-MMP the 92 kD (latent or proform) and 88 kD (active form)expression is again very low in the tubular epithelial

MMP-9 species, whereas the lower gelatinolytic bandscells in normal dog kidneys; however, in XLAS kidneys,represent the 72 kD (latent or proform) and 64 kD (ac-elevated immunostaining is observed at the interfacetive form) MMP-2 species. Standards, which are superna-between the tubular epithelial cells and the tubular base-tants from human fibrosarcoma (HT1080) cells, are pre-ment membranes. This cell surface localization is consis-dominantly the latent forms of MMP-2 and MMP-9.tent with the fact that MT1-MMP is a membrane bound

The gelatinase activity was completely inhibited byproteolytic activator for MMP-2 [23]. No immunostain-EDTA, a chelating agent, confirming that the observeding was observed when tissue sections were incubatedbands were the result of zinc-dependent metalloprotei-with preimmune serum (Fig. 2 G to I), indicating that thenase activity. The gelatinase bands on the zymogramobserved immunostaining is not due to nonspecific binding.were not inhibited by PMSF, an inhibitor of serine prote-Since none of these antibody preparations have pre-ases, or pepstatin A, an inhibitor of aspartic proteases.viously been used to detect MMPs in dog tissues, it neces-

sary to evaluate their specificity. To test for this, tissue Analysis of gelatinase activity in the renal cortex ofextracts from a control and Alport dogs were subjected XLAS dogsto Western blot analysis. The blots developed with the

Gelatin zymography was performed using extracts ofantibodies demonstrated a single band for each antibodyrenal cortex from control and XLAS dogs. Extracts werewith the predicted molecular size (Fig. 3). This confirmsderived from necropsy specimens taken from individualsthat the antibodies are indeed specific, and cross-reactivewith either moderate fibrosis score (2�) or severe fibrosiswith the respective dog MMPs.score (3�). The results in Figure 5 illustrate that significant

Detection of MMP activity by gelatin zymography differences in the levels of MMP-2 and MMP-9 were evi-dent in the XLAS necropsy kidney extracts when com-Proteolytic activity of two forms of type IV collagenase

(72 kD, MMP-2; 92 kD, MMP-9) in control and XLAS pared to normal dogs. The active forms of both MMP-2

Rao et al: MMPs in canine Alport syndrome1742

Fig. 4. Specificity of gelatin zymography fordetecting matrix metalloproteinase (MMP)activity in tissue extracts. Lane 1, media fromhuman fibrosarcoma cells (positive control);lane 2, tissue extract from unaffected; and lane3 from an X-linked Alport syndrome (XLAS)dog with a fibrosis score of 3�. Both latentand active forms of MMP-2 and MMP-9 aredenoted. Labels below the lanes indicate whatwas added to the gel incubation buffer, illus-trating that the gelatinase activity is inhibitedby ethylenediaminetetraacetic acid (EDTA),which is an inhibitor of MMPs, but not by phe-nylmethylsulfonyl fluoride (PMSF) or pepsta-tin A, demonstrating that activity is not dueto serine or aspartate proteases.

Fig. 5. Interstitial disease in the X-linked Alport syndrome (XLAS) dogs is associated with markedly elevated matrix metalloproteinases (MMP-2and MMP-9) activity. Sample numbers are reflecting animals described in Table 1. (A ) Gelatin zymography for MMP-2 and MMP-9 proteolyticactivity in necropsy tissue extracts from normal versus XLAS dogs. Quantitative analysis of MMPs. Gelatin zymograms were analyzed by scanningdensitometry and expressed as arbitrary units to produce histograms. (B ) The activity of total MMP-2 and MMP-9 are elevated in XLAS dogscompared to unaffected dogs. Total gelatinolytic activity represents the sum of both MMP-2 and MMP-9 activities. (C ) Latent and active MMP-9for individual samples are shown. (D ) The values for total, MMP-2, and MMP-9 (both latent and active MMP-9 are shown) as mean and standarddeviation for control and Alport dogs. The zymogram shown is representative of three independent experiments. *Statistical significance comparedwith control kidney extracts at P 0.001.

Rao et al: MMPs in canine Alport syndrome 1743

Table 2. Summary of changes in matrix metalloproteinase (MMP) activity and mRNA levels relative to that observed in normal dogs

MMP activity mRNA expressionSample Serum creatinine Fibrosisnumber mg/dL score MMP-2 MMP-9 MMP-2 MMP-9 MT1-MMP

6 5.4 2� 6.8 �0.5 75.3�8.2 1.7�0.2 17.9�1.9 2.3�0.37 5.0 2� 6.5 �0.7 68.7�7.5 2.0�0.3 17.5�1.8 2.5�0.28 4.9 3� 7.4 �0.8 87.8�8.9 2.1�0.3 21.1�2.2 2.5�0.39 5.3 3� 7.4 �0.8 82.5�9.1 2.2�0.3 19.6�2.0 2.7�0.3

10 5.2 3� 7.6 �0.8 98.6�10.1 1.9�0.2 19.2�1.9 2.8�0.4

Data are expressed as the mean of three independent measurements with standard deviations. MTI-MMP is membrane type 1 matrix metalloproteinase.

and MMP-9 were observed only in XLAS tissue extracts predicted size for selected cDNA fragments encoding(Fig. 5A). Zymographic bands were quantified by scan- the MMPs and GAPDH, a housekeeping gene that wasning densitometry and the data analyzed using Image used to control for cDNA loading in the PCR reactions.Quant software (see Methods section) and expressed as Sequencing of amplified cDNAs from control and XLASarbitrary units of lysis. There was a significant increase indicated 100% homology with the published sequencesin the levels of both MMP-2 and MMP-9 in XLAS dogs validating that the amplified products were indeed spe-in comparison to controls (P 0.001) (Fig. 5B). Fold cific (not shown). The most dramatic induction of MMPinduction was remarkably high, especially for MMP-9, mRNA expression in XLAS was that for MMP-9 (Fig.and did not vary substantially in samples derived from 7A, middle panel), consistent with the data obtainedanimals with fibrosis scores of 3� compared to 2� (an through gelatin zymography. RT-PCR amplification ofaverage of 89-fold versus 72-fold, respectively, for MMP-9, MMP-9 produced a cDNA of the expected size, whichand 7.5-fold versus 6.6-fold, respectively, for MMP-2, see was barely detectable in controls but was significantlyTable 2). Active MMP-9 is seen only in affected animals increased (19-fold on average, see Table 2) in XLASand not in controls (Fig. 5C). (P 0.001). The relatively low levels of MMP-9 amplifi-

cation product in controls was not the result of the qualityIn situ zymography for the detection of gelatinase

or quantity of mRNA because the presence of intactactivity by FIZmRNA, as well as equivalent loading of reverse tran-

In situ zymography using gelatin films (gelatinolytic scribed message in the PCR reactions was demonstratedfilms) on cryosections revealed that strong gelatinolytic by amplification of GAPDH mRNA (Fig. 7A, loweractivity is present in specimens from XLAS dogs, but panel). Amplification of MMP-2 (Fig. 7A, upper panel)very little or no activity was recognized in the normal dogs confirms expression in both control and XLAS dogs and(Fig. 6). The region where gelatinolysis occurred was elevated expression (twofold on average, see Table 2)observed as white to pale blue. The gelatinolytic activity in XLAS dogs as compared to control (P 0.001). Thein the cryosections was almost completely inhibited on relative expression of MMP-2 and MMP-9 specificsections that have been incubated on the film containing mRNA transcripts in XLAS as determined by densito-1,10-phenanthroline (GI film), a broad MMP inhibitor.

metric analysis of RT-PCR is given in Figure 7C. TheseThe FIZ gelatin films employed in this assay have beenresults were qualitatively in agreement with zymographyshown to detect only the active isoforms of the MMPs(gelatin and in situ) and immunofluorescence analysis.[24], confirming that a marked increase in overall metal-Fold induction of mRNA versus gelatinolytic activity,loproteinase activity is evident in the Alport renal cortex.however, revealed some interesting similarities. Fold in-Furthermore, this activity appears predominantly in theduction of MMP-2 and MMP-9 gelatinolytic activity wasinterstitial space, and not in the glomeruli (denoted asgreater than that for MMP-2 and MMP-9 mRNA. TheG in Fig. 6).ratio for fold induction (in XLAS versus control samples)of enzyme activity to mRNA, on average, is between 3Detection of mRNA for MMP-2, MMP-9,

and MT1-MMP and 4 to 1 for both MMP-2 and MMP-9. Gelatinolyticactivity and specific mRNA levels for the MMPs wereRT-PCR analysis was used to determine whether thenot significantly different for samples taken from animalsinduction of MMP-2 and MMP-9 protein was secondarywith a fibrosis score of 2� compared to 3�.to the increased expression of MMP-2 and MMP-9 mRNA

Transcripts for MT1-MMP mRNA were also analyzedtranscripts in XLAS kidneys. RNA was isolated from aby RT-PCR and found to be consistently and significantlyportion of the same necropsy specimen employed in theelevated (2.6-fold on average, Figure 8 and Table 1) ingelatin zymography experiments. Corresponding sampleXLAS-derived samples relative to controls (P 0.001).numbers were derived from the same animals employedThese data reflect elevated MT1-MMP protein observedfor gelatin zymograms (Fig. 5). The primers selected

for amplification produced only a single product of the by immunofluorescence (Fig. 2 C and F).

Rao et al: MMPs in canine Alport syndrome1744

nents, including interstitial and basement membrane col-lagen and glycoproteins and play an important role invarious physiologic and pathologic conditions [25, 26].Here we show that the MMP-2, MMP-9, and the mem-brane-bound proteolytic activator for the MMPs, MT1-MMP, are all up-regulated in the renal cortex of XLASdogs. Induction is apparent at both the level of mRNAand enzyme activities. Further, in situ zymography re-veals the presence of active metalloproteinase in tissuesections of XLAS renal cortex, demonstrating that theelevated expression translates to elevated metalloprotein-ase activity in the organ. Combined, these data providecompelling evidence suggesting a role for MMPs in aber-rant matrix remodeling associated with interstitial fibro-sis in the canine XLAS model. An earlier study examinedthe role of MMP-9 in glomerular pathogenesis in a mu-rine Alport model and concluded that it did not influencedisease progression. This study, however, did not addresstubulointerstitial disease [27].

The regulation of kidney ECM remodeling during nor-mal tissue metabolism and its dysregulation in kidneydiseases may result from a complex interplay of MMPtranscriptional regulation, proenzyme activation, and in-hibition by TIMPs. In interstitial fibrosis, the primaryinsult initiates a sequence of events starting with theinflux of a variety of inflammatory cells, including mono-cytes. These cells produce a variety of cytokines, growthfactors and other mediators that presumably function torepair the damaged tissue and modulate expression ofthe MMPs [28]. Some MMPs, including MMP-2, MMP-3,MMP-9, and metalloelastase, are potentially capable ofdisrupting the kidney architecture by virtue of their spec-ificity for various components of basement membrane,thus enabling further infiltration of inflammatory cellsinto the kidney and thereby setting into motion the dys-regulated matrix remodeling that follows. Recently, wehave demonstrated significantly elevated levels of MMP-2and MMP-9 in a mouse model for AS, which progres-sively develops interstitial fibrosis between 4 and 8 weeksafter birth. Further studies also demonstrated that thetissue monocytes, independent of myofibroblasts, drive

Fig. 6. In situ gelatin zymography using gelatin films demonstrated tubular atrophy in the Alport mice [29]. The data pre-strong gelatinolytic activity in the X-linked Alport syndrome (XLAS) sented here are the first to demonstrate differential ex-dogs (Alport) as compared to control (Control). In this particular field,

pression of gelatinases and MT1-MMP genes in canineactivity was in the interstitial space surrounding the glomerulus (G).The glomerulus itself, however, was negative for active metalloproteinase. XLAS, an animal model for human XLAS. Our resultsThe gelatinolytic activity is almost completely blocked in samples incu- suggest that differential MMP expression may contributebated on film containing 1,10-phenanthroline (Alport GI) indicating

to the molecular pathogenesis of the disease.that the gelatinolytic activity is attributable to MMPs.

We show that XLAS extracts contained activated formsof MMP-2 and MMP-9 but the mechanism for the activa-tion is not clear. MMPs are secreted as inactive latent

DISCUSSION precursors (zymogens) and their activation is a prerequi-Remodeling of ECM accompanies many physiologic site to their functioning in vivo [25, 26]. Activation of

and pathologic processes. The MMPs constitute a multi- MMP-2 is not mediated by serine proteases, such asgene family of over 25 secreted and cell surface enzymes plasmin, as is the case for other MMPs. Instead, protease

inhibitor studies suggest that MMP-2 is activated by met-that are capable of degrading a variety of ECM compo-

Rao et al: MMPs in canine Alport syndrome 1745

Fig. 7. Interstitial disease in X-linked Alport syndrome (XLAS) is associated with elevated matrix metalloproteinases (MMP-2 and MMP-9)mRNA transcripts. (A ) Reverse transcription polymerase chain reaction (RT-PCR) from RNA isolated from normal and XLAS dog kidneysusing primers specific for MMP-2, MMP-9, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Lanes 1 to 3, mRNA expression fromnormal unaffected dogs. Lanes 4 to 8, mRNA expression from XLAS dogs. GAPDH mRNA levels were also analyzed as a control for loading.(B ) Semiquantification of mRNA transcripts for MMPs was performed using densitometry and normalized to GAPDH from the same sample.The results are presented as an index of mRNA expression (arbitrary units). Data are presented as the mean and standard deviation for threeindependent experiments. (C ) The mean and standard deviation across groups of normal and Alport samples for three independent experiments.*Statistical significance compared with control samples (P 0.001).

alloproteinases themselves and requires interaction with that clustering of cell surface molecules can lead to in-creased expression of active MMP-2. In the presentcell surface proteins [30, 31]. This unique property of

MMP-2 activation likely directs proteolytic activity to study, we analyzed the expression of mRNA transcriptsfor MT1-MMP and found them to be significantly ele-either migrating cells or cells involved in ECM turnover.

Recently, the membrane-bound MT1-MMP has been vated in XLAS dogs compared to controls. The balancebetween TIMP-2 and MT1-MMP may be critical for de-identified as a potential physiologic activator of MMP-2

on the cell surface. Studies have shown that a complex termining the activation status of MMP-2. Our studiesshow that the active form of MMP-2, while abundant inof MT1-MMP mediates activation of proMMP-2 with

TIMP-2, acting as a surface receptor for proMMP-2 XLAS renal cortex, is absent in controls. The absence ofactive MMP-2 in controls may be explained by relatively[32–34]. In cell culture, MT1-MMP expression can be

stimulated by the lectin, concanavalin A [23], suggesting decreased activation of MT1-MMP, due to incomplete

Rao et al: MMPs in canine Alport syndrome1746

Fig. 8. Expression of membrane type 1 matrix metalloproteinase (MT1-MMP) mRNA transcripts in X-linked Alport syndrome (XLAS). (A )Shows reverse transcription polymerase chain reaction (RT-PCR) from RNA isolated from normal and XLAS dog kidneys using primers specificfor MT1-MMP and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Lanes 1 to 3, mRNA expression from normal unaffected dogs. Lanes4 to 8, mRNA expression from XLAS dogs. (B ) Semiquantification of mRNA transcripts for MT1-MMP was performed using densitometry andcomparison to GAPDH as done in Figure 6. Data is presented as the mean and standard deviation for three independent experiments. (C ) Themean and standard deviation across groups of normal and Alport samples for three independent experiments. *Statistical significance comparedwith control samples (P 0.001).

processing of pro-MT1-MMP to its active form, instabil- In the present study we show that MMP-9 activity wassignificantly increased in XLAS. Of the total detectedity of the MT1-MMP protein, or lack of translation of

MT1-MMP mRNA into protein. It has been shown that MMP-9, about 35% represented active MMP-9, sug-gesting that the activation of MMP-9 in XLAS is signifi-proteolytic cleavage of pro-MT1-MMP is a prerequisite

for pro-MMP-2 activation [35, 36]. cantly higher than active MMP-2. Recent studies haveshown that the mRNA transcripts for MMP-3 and MMP-7A large number of proteases and organomercurials

have been shown to activate other MMPs in vitro; how- were significantly increased in XLAS kidneys (Rao andCosgrove, unpublished). Therefore, it appears that MT1-ever, the mechanisms of activation are not well docu-

mented in vivo. A potentially important mechanism for MMP, MMP-7, and MMP-3 likely participates in MMP-2and MMP-9 activation in this model. The coexpressionactivation of latent MMPs is intermolecular activation.

Both MMP-3 and MMP-10 are good activators of MMP-7 of MMP-2, MMP-3, MMP-7, and MMP-9 may functionas powerful machinery for ECM degradation in XLAS[37, 38]. MMP-3 can fully cleave pro-MMP-7 to generate

an active species [37]. MMP-7 has been shown to activate kidneys.The common pathway of progressive tubulointerstitialhuman MMP-1, MMP-2, and MMP-9 [37, 39, 40]. It has

been reported that MT1-MMP can activate collagenase 3 fibrosis is the end result for most acute and progressiverenal disease. The process of interstitial scarring and(MMP-13) and MMP-9 [41]. In addition, pro-MMP-9 can

be activated by gelatinase A (MMP-2) [42], MMP-3 [43], cellular destruction associated with interstitial fibrosisshares many characteristics from one organ to another.MMP-13 [44], plasmin [45], and tissue kallikrein [46].

Rao et al: MMPs in canine Alport syndrome 1747

3. Tryggvasan K, Heiskari N: Alport syndrome, in Molecular Ne-Fibrosis is usually associated with a general inflammatoryphrology: Kidney Function in Health and Disease, edited by

response that involves T cells, B cells, and monocytes/ Schlondorff D, Bonventre JV, New York, Marcel Dekker, Inc.,1995, pp 795–808macrophages. In a mouse model for AS, we have demon-

4. Lemmink HH, Mochizuki T, van den Heuvel LP, et al: Mutationsstrated that tissue monocytes, independent of myofi-in the type IV collagen alpha 3 (COL4A3) gene in autosomal

broblasts, drive renal tubular atrophy [29]. Monocytes/ recessive Alport syndrome. Hum Mol Genet 3:1269–1273, 19945. Ding J, Stitzel J, Berry P, et al: Autosomal recessive Alportmacrophages have been implicated in connective tissue

syndrome: Mutation in the COL4A3 gene in a woman with Alportdestruction mediated through MMPs and secrete a num-syndrome and posttransplant antiglomerular basement membrane

ber of MMPs in response to various stimulants [47–49]. nephritis. J Am Soc Nephrol 5:1714–1717, 19956. Mazucco G, Barsotti P, Muda AO, et al: Ultrastructural andTherefore, it appears that monocytes may be involved

immunohistochemical findings in Alport’s syndrome: A study ofin the upregulation of MMPs in XLAS kidneys. It is108 patients from 97 Italian families with particular emphasis on

known that cytokines play a crucial role in the induction COL4A5 gene mutation correlations. J Am Soc Nephrol 9:1023–1031, 1998and perpetuation of inflammatory response [50], while

7. Cosgrove D, Meehan DT, Grunkenmeyer JA, et al: CollagenMMPs are considered to be key enzymes in ECM turn-COL4A3 knockout: a mouse model for autosomal Alport syn-

over and physiologic mediators of cell migration through drome. Genes Dev 10:2981–2992, 19968. Miner JH, Saner JR: Molecular and functional defects in kidneysvarious barriers [51, 52].

of mice lacking collagen alpha 3(IV): Implications for Alport syn-MMPs were notably absent in the glomeruli of Alportdrome. J Cell Biol 135:1403–1413, 1996

dogs based on our immunostaining results (Fig. 2). This 9. Lu W, Phillips CL, Killen PD, et al: Insertional mutation of thecollagen genes Col4a3 and Col4a4 in a mouse model of Alportmay reflect the differences in the fibrogenic processes ofsyndrome. Genomics 61:113–124, 1999the interstitium and the glomerulus. In the interstitium,

10. Lees G, Helman RG, Kashtan CF, et al: A model of autosomalthe interactive roles of monocytes and myofibroblasts are recessive Alport syndrome in English cocker spaniel dogs. Kidney

Int 54:706–719, 1998thought to collectively drive fibrosis [29]. In the Alport11. Lees G, Helman RG, Kashtan CE, et al: A new form of X-linkedglomerulus, however, mesangial expansion associated

dominant hereditary nephritis in dogs. Am J Vet Res 60:373–383,with matrix deposition in the GBMs precedes overt glo- 1999

12. Thorner P, Jansen B, Bauman R, et al: Samoyed hereditary glom-merular sclerosis [7, 53].erulopathy: Immunohistochemical staining of basement mem-branes of kidney for laminin, collagen type IV, fibronectin, andGoodpasture antigen, and correlation with electron microscopy ofCONCLUSIONglomerular capillary basement membranes. Lab Invest 56:435–443,1987This is the first report showing differences in MMP-2,

13. Zheng K, Thorner P, Marrano P, et al: Canine X-chromosomeMMP-9, and MT1-MMP expression in canine XLAS kid-linked hereditary nephritis: A genetic model for human X-linked

neys. Our study demonstrated significant alterations of hereditary nephritis resulting from a single base mutation in thegene encoding the �5 chain of collagen type IV. Proc Natl AcadMMP expression in canine XLAS kidneys. ElevatedSci USA 91:3989–3993, 1994MMPs may serve as a compensatory mechanism aimed

14. Bohle A, Muller GA, Wehrmann M, et al: Pathogenesis ofat limiting the rate of fibrogenesis. Fibrosis proceeds, chronic renal failure in the primary glomerulopathies, renal vascu-however, despite the elevated MMP activity. The diverse lopathies, and chronic interstitial nephritides. Kidney Int 49 (Suppl

54):S2–S9, 1996functional properties of the MMPs warn that this view15. Fine LG, Ong AC, Norman JT: Mechanisms of tubulo-interstitialmay be oversimplified. Thus, the specific role of elevated injury in progressive renal diseases. Eur J Clin Invest 23:259–265,

MMP activity in dysregulated matrix deposition associ- 199316. Kunicio GS, Nielson EG, Haverty T: Mechanisms of tubulointer-ated with renal fibrosis remains unclear and will require

stitial fibrosis. Kidney Int 39:151–160, 1991further study. 17. Birkedal-Hansen H, Moore WGI, Bodden MK, et al: Matrixmetalloproteinases: A review. Crit Rev Oral Biol Med 4:197–250,1993ACKNOWLEDGMENTS

18. Leco KJ, Apte SS, Taniguchi GT, et al: Murine inhibitor of metal-We thank Ms. Nicole Pearsall and Q. Kong for their help in prepar- loproteinases-4 (TIMP-4): cDNA isolation, and expression in adult

ing the histograms and statistical analysis, and Rene Rodgers for se- mouse tissues. Pharmacol Ther 59:329–341, 1993quence analysis. We also thank John (Skip) Kennedy for help in figure 19. Woessner JF, Jr: The family of matrix metalloproteinases. Annpreparation. This work is supported in part by NIH R01 DK55000 to N Y Acad Sci 732:11–22, 1994D.C., NIH R01 DK57676 to C.E.K., and the tobacco settlement fund 20. Rao VH, Singh R, Bridge JA, et al: Regulation of MMP-9 (92from the State of Nebraska. kDa type IV collagenase/gelatinase B) expression in stromal cells

of human giant cell tumor of bone. Clin Exp Metastasis 15:400–409,Reprint requests to Dominic Cosgrove, Ph.D., Boys Town National 1997

Research Hospital, 555 No. 30th St., Omaha, NE 68131. 21. Rao VH, Singh R, Delimont DC, et al: Transcriptional regulationE-mail: cosgrove@boystown.org of MMP-9 expression in stromal cells of human giant cell tumor

of bone by tumor necrosis factor-�. Int J Oncol 14:291–300, 199922. Kalluri R, Cosgrove D: Assembly of type IV collagen: InsightsREFERENCES

from �3(IV) collagen-deficient mice. J Biol Chem 275:12719–12724,20001. Kashtan CE: Alport syndrome: An inherited disorder of renal,

23. Sato H, Takino T, Okada Y, et al: A matrix metalloproteinasesocular and cochlear basement membranes. Medicine (Baltimore)expressed on the surface of invasive tumor cells. Nature 370:61–65,78:338–360, 199919942. Kashtan CE, Michael AF: Perspectives in clinical nephrology:

Alport syndrome. Kidney Int 50:1445–1463, 1996 24. Ikeda M, Maekawa R, Tanaka H, et al: Inhibition of gelatinolytic

Rao et al: MMPs in canine Alport syndrome1748

activity in tumor tissue by synthetic matrix metalloproteinases in- of human stromelysin 2 and its interactions with other matrixmetalloproteinases. Eur J Biochem 253:67–75, 1998hibitor: Application of film in situ zymography. Clin Cancer Res

39. Barille S, Bataille R, Rapp MJ, et al: Production of metallopro-6:3290–3296, 2000teinases-7 (matrilysin) by human myeloma cells and its potential25. Nagase H, Woessner JF: Matrix metalloproteinases. J Biol Cheminvolvement in metalloproteinases-2 activation. J Immunol 163:274:21491–21494, 19995723–5728, 199926. Sternlicht MD, Werb Z: How matrix metalloproteinases regulate

40. Crabbe T, Smith B, O’Connell J, Docherty A: Human progela-cell behavior. Annu Rev Cell Dev Biol 17:463–516, 2001tinase A can be activated by matrilysin. FEBS Lett 345:14–16, 199427. Andrews KL, Betsuyaku T, Rogers S, et al: Gelatinase B (MMP-9)

41. Fridman R, Toth M, Pena D, Mobashery S: Activatioin of progela-is not essential in the normal kidney and does not influence progres-tinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res 55:2548–sion of renal disease in a mouse model of Alport syndrome. Am2855, 1995J Pathol 157:303–311, 2000

42. Ogata Y, Enghild JJ, Nagase H: Matrix metalloproteinases 328. Ries C, Petrides PD: Cytokine regulation of matrix metalloprotein-(stromelysin) activates the precursor for the human matrix metallo-ases activity and its regulatory dysfunction in disease. Biol Chemproteinases 9. J Biol Chem 267:3581–3584, 1992Hoppe Seyler 376:345–355, 1995

43. Knauper V, Smith B, Lopez-Otin C, Murphy G: Activation of29. Rodgers KD, Rao V, Meehan DT, et al: Monocytes promoteprogelatinase B (proMMP-9) by active collagenase 3 (MMP-13).myofibroblast accumulation and apoptosis in Alport renal fibrosis. Eur J Biochem 248:369–373, 1997

Kidney Int 63:71–88, 2003 44. Okada Y, Gonoji Y, Naka K, et al: Matrix metalloproteinases 930. Stettler-Stevenson WG, Krutzsch HC, Wacher MP, et al: The (92 kDa gelatinase/type IV collagenase) from HT 1080 human

activation of human type IV collagenase proenzyme. J Biol Chem fibrosarcoma cells. J Biol Chem 267:21712–21719, 1992264:1353–1356, 1989 45. Menashi S, Fridman R, Desrevieres S, et al: Regulation of 92

31. Brown PD, Kleiner DE, Unsworth EJ, Stettler-Stevenson kDa gelatinase B activity in the extracellular matrix by tissue kalli-WG: Cellular activation of the 72 kDa type IV collagenase/TIMP-2 krein. Ann N Y Acad Sci 732:466–468, 1994complex. Kidney Int 43:163–170, 1993 46. Welgus HG, Campbell EJ, Cury JD, et al: Neutral metalloprotein-

32. Strongin AY, Collier I, Bannikov G, et al: Mechanism of cell ases produced by human mononuclear phagocytes. Enzyme profile,surface activation of 72-kDa type IV collagenase. Isolation of the regulation, and expression during cellular development. J Clinactivated form of the membrane metalloprotease. J Biol Chem Invest 86:1496–1502, 1990

47. Shankavaram UT, DeWitt DL, Funk SE, et al: Regulation of270:5331–5338, 1995human monocyte matrix metalloproteinases by SPARC. J Cell33. Zucker S, Drews M, Conner C, et al: Tissue inhibitor of metallo-Physiol 173:327–334, 1997proteinases-2 (TIMP-2) binds to the catalytic domain of the cell

48. Shankavaram UT, Lai WC, Netzel-Arnett S, et al: Monocytesurface receptor, membrane type 1-matrix metalloproteinases 1membrane type 1-matrix metalloproteinases. Prostaglandin-depen-(MT1-MMP). J Biol Chem 273:1216–1222, 1998dent regulation and role in metalloproteinases-2 activation. J Biol34. Butler JS, Butler MJ, Atkinson SJ, et al: The TIMP2 membraneChem 276:19027–19032, 2001type 1 metalloproteinase “receptor” regulates the concentration

49. Zhang Y, McCluskey K, Fujii K, Wahl LM: Differential regula-and efficient activation of progelatinase A. A kinetic study. J Bioltion of monocyte matrix metalloproteinases and TIMP-1 produc-Chem 273:871–880, 1998tion by TNF-alpha, granulocyte-macrophage CSF, and IL-1 beta35. Lehti K, Lohi J, Valtanen H, Keski OJ: Proteolytic processingthrough prostaglandin-dependent and independent mechanisms. Jof membrane-type-1 matrix metalloproteinases is associated withImmunol 161:3071–3076, 1998gelatinase A activation at the cell surface. Biochem J 334:345–353, 50. Kochanek PM, Hallenbeck JM: Polymorphonuclear leukocytes

1998 and monocytes/macrophages in the pathogenesis of cerebral isch-36. Kurchat P, Zigrino P, Nischt R, et al: Tissue inhibitor of matrix emia and stroke. Stroke 23:1367–1579, 1992

metalloproteinases-2 regulates matrix metalloproteinases-2 activa- 51. Matrisian LM: The matrix degrading metalloproteinases. Bio-tion by modulation of membrane-type 1 matrix metalloproteinases assays 14:455–463, 1992activity in high and low invasive melanoma cell lines. J Biol Chem 52. Chandler S, Miller KM, Lury J, et al: Matrix metalloproteinases,274:21056–21062, 1999 tumor necrosis factor and multiple sclerosis: An overview. J Immu-

37. Imai K, Yokohama Y, Nakanishi I, et al: Matrix metalloproteinases nol 72:155–161, 19977 (matrilysin) from human rectal carcinoma cells. Activation of 53. Cosgrove DE, Rodgers KD, Meehan DT, et al: Integrin �1�1 andthe precursor interaction with other matrix metalloproteinases and TGF-�1 play distinct roles in Alport glomerular pathogenesis andenzyme properties. J Biol Chem 270:6691–6697, 1995 serve as dual targets for metabolic therapy. Am J Pathol 157:1649–

1659, 200038. Nakamura H, Fujii Y, Ohuchi E, et al: Activation of the precursor