Matrix metalloproteinase-8 overexpression prevents proper tissue repair

Matrix metalloproteinase-8 overexpression prevents proper tissue repair Patricia L. Danielsen, MD, PhD,a,b Anders V. Holst, MD,c Henrik R. Maltesen, MD,d Maria R. Bassi, MSc,e Peter J. Holst, MD, PhD,e Katja M. Heinemeier, PhD,f Jørgen Olsen, MD, PhD,d Carl C. Danielsen, MD, PhD,g Steen S. Poulsen, MD, PhD,h Lars N. Jorgensen, MD, DrMedSci,a and Magnus S.
Agren, DrMedSci,a,b Copenhagen and
Arhus, Denmark

Background. The collagenolytic matrix metalloproteinase-8 (MMP-8) is essential for normal tissue repair but is often overexpressed in wounds with disrupted healing. Our aim was to study the impact of a local excess of this neutrophil-derived proteinase on wound healing using recombinant adenovirusdriven transduction of full-length Mmp8 (AdMMP-8). Methods. The effect of MMP-8 overexpression was evaluated in dermal fibroblasts and in two wound healing models in male Wistar rats: subcutaneously positioned ePTFE catheters and linear incisional skin wounds. Results. Fibroblasts transduced with AdMMP-8 secreted MMP-8 with type I collagenolytic activity that could be blocked by a selective MMP-8 inhibitor. AdMMP-8 (5 3 1010 viral particles) administered in homologous fibrin increased MMP-8 mRNA (P < .05) levels compared to parallel wounds treated with a control adenovirus expressing lacZ (AdLacZ). Impaired wound healing was demonstrated with AdMMP-8 by decreased collagen deposition and breaking strength of incisional wounds on day 7 compared to AdLacZ-treated wounds (P < .05). We found no significant effect of AdMMP-8 on mRNA levels of MMP-9, COL1A1, or COL3A1, but AdMMP-8 treatment decreased the number of neutrophils. In the incisional wounds, MMP-8 gene transfer was not associated with significant changes in macrophage numbers or amount of granulation tissue but did increase MMP-8 protein by 76% (P < .01) and decrease type I collagen protein by 29% (P < .05) compared with AdLacZ. Conclusion. These results demonstrate that superphysiologic levels of the proteinase MMP-8 can result in decreased collagen and lead to impaired wound healing. This observation makes MMP-8 a potential drug target in compromised human wound healing associated with MMP-8 overexpression. (Surgery 2011;150:897-906.) From the Department of Surgery K,a Department of Dermatology and Copenhagen Wound Healing Center,b Bispebjerg Hospital, University Clinic of Neurosurgery,c Rigshospitalet, Departments of Molecular and Cellular Medicine,d International Health, Immunology and Microbiology,e and Anatomy,h Panum Institute, University of Copenhagen, Copenhagen, Institute of Sports Medicine,f Bispebjerg Hospital and Center for Healthy Aging, Faculty of Health Sciences, University of Copenhagen, and Institute of Anatomy,g Aarhus University,
Arhus C, Denmark

Supported by The Danish Medical Research Council (22-020287 and 271-07-0742), the Pharmacy Foundation of 1991, Denmark, and Bloddonorernes Forskningsfond. This work was awarded Third Prize for Young Investigators at the Joint Meeting of the European Tissue Repair Society and Wound Healing Society, Limoges, France 2009. Accepted for publication June 15, 2011. Reprint requests: Magnus S.
Agren, MD, Department of Surgery K and Copenhagen Wound Healing Center, Bispebjerg Hospital, Bispebjerg Bakke 23, DK-2400 Copenhagen, Denmark. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2011 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.06.016

NONHEALING WOUNDS represent an immense socioeconomic problem and continue to be a therapeutic challenge.1 Patient heterogeneities and multiple etiologies add further to the complexity.2 Several putative local cellular and molecular pathogenic factors have been identified.2,3 Excessive activity of endogenous proteinases may contribute to defective collagen turnover in nonhealing cutaneous wounds.2,4 The zinc-dependent, matrix metalloproteinases (MMPs) comprise 23 protein members in man.5,6 MMPs have been implicated in pathologic processes associated with remodeling of the extracellular matrix.7 MMP-8 is a candidate due to its collagenolytic efficiency with tissue-destructive SURGERY 897

898 Danielsen et al

potential that may delay wound healing.8-12 On a molecular level, MMP-8 initiates degradation of native interstitial collagens by cutting the three a-chains at one specific locus.13 This proteolytic activity generates a 3/4-length piece from the N-end and a 1/4-length piece from the C-end of the collagen molecule. MMP-8 is expressed primarily by neutrophils14,15 but also by macrophages and fibroblasts during wound healing.12,16 Increased tissue levels of the MMP-8 transcript and protein have been observed in chronic wounds caused by vascular disturbances, including local ischemia,8 venous insufficiency,11,12 vasculitis,12 and diabetes.9,10,12 Human diabetic fibroblasts also show abnormally increased expression of MMP-8 mRNA.17 In addition, smoking increases local MMP-8 protein levels in diseased periodontium18 and regenerating skin.19 Taken together, increased MMP-8 level appears to be a universal biochemical abnormality in wounds exhibiting disrupted healing. We have developed an adenovirus vector for efficient transfer of the Mmp8 gene (AdMMP-8) to mimic the clinical situation of localized increase in MMP-8 activity. Local MMP-8 overexpression was evaluated in the paired ePTFE tube20 and the model of surgical incision wound healing in rats with adenovirus expressing lacZ (AdLacZ) serving as control. The primary end-points were collagen deposition in ePTFE tubes and biomechanical strength of the incisional wounds on day 7. MMP8 and MMP-9 mRNA and protein levels were also measured in addition to neutrophil and macrophage infiltration, and collagen synthesis. MATERIALS AND METHODS Cloning of cDNA encoding rat MMP-8 and expression in replication-deficient recombinant adenovirus. A cDNA library was constructed from total RNA obtained from a 5-day-old rat incisional wound. The cDNA encoding full-length rat MMP-8 (1401 bp) was obtained by RT-PCR amplification of total RNA using the forward primer (59-TTAAGGC GCGCCATGCTTCACCTGAAGACACT-39) and reverse primer (39-TTAAGTTTAAACCTATGGACAG TTAAGCCATAA-59). The PCR fragment was then cloned into a pACCMV-based shuttle vector using AscI and PmeI sites that were inserted in the primers and the shuttle vectors. Human recombinant adenovirus serotype 5 vectors encoding MMP-8 under the control of the cytomegalovirus immediate early (CMV IE) promoter (AdMMP-8) were produced from homologous recombination with the shuttle plasmid and pJM17 plasmid in human HEK 293 embryonic

Surgery November 2011 kidney cells.21 Adenovirus (AdLacZ) containing the E. coli b-galactosidase (lacZ) gene under the regulation of the CMV IE promoter was cloned by plaque assay and amplified in HEK 293 cells. The crude AdMMP-8 and AdLacZ lysates were purified by cesium-chloride density centrifugation. Viral particles were enumerated by OD260 nm, where OD260 nm = 1.00 corresponds to 1012 particles/mL. Adenoviral stocks were stored at
808C in 10% glycerol.21 Evaluation of the AdMMP-8 construct in dermal fibroblasts. Rat dermal fibroblasts (CRL-1213; ATCC, Rockville, MD) were cultured (2 3 106) on 10-cm tissue culture dishes coated with polylysine in Eagle’s minimum essential medium supplemented with 10% fetal bovine serum. The fibroblasts were transduced with non-cytotoxic AdMMP-8 or AdLacZ at a multiplicity of infection of 250 for 24 h at 378C in 5% CO2/95% air. The cells were then washed and incubated for 48 h in 5 mL of serum-free medium. Supernatants were concentrated 10 times (Amicon Ultra-15; Millipore, Billerica, MA). Cell lysates were prepared with 250 mL extraction buffer22 for 24 h at 48C followed by sonication and centrifugation. Samples were dialyzed against assay buffer (50 mM Tris-HCl, 10 mM CaCl2, 25 mM ZnCl2, 150 mM NaCl, 0.05% Brij 35, pH 7.5) and subjected to MMP-8 ELISA, Western blot, and collagenolysis analyses (Fig 1). Animals, wound-healing models, and administration of adenoviral treatments. Male Wistar rats (200–300 g) from Charles River Laboratories (Sulzfeld, Germany) were kept in individual cages and maintained under controlled environmental conditions with a constant temperature (218C), relative humidity (45–65%), a 12 h-light/dark cycle, and with free access to standard laboratory rat pellets and tap water. Animals were acclimatized for 7 days before being operated on. The rats were anesthetized with 15 mg/kg midazolam, 3.3 mg/kg fluanizon, and 0.105 mg/kg fentanyl injected subcutaneously. The dorsum was shaved and skin prepped with 70% ethanol. Twenty rats had two parallel, 6.5-cm, high porosity, expanded polytetrafluoroethylene (ePTFE) tubes20 inserted subcutaneously at 2 cm to the right and left of the spine extending from the prominent seventh cervical spine and caudally via cannula with an inner diameter of 3.2 mm. In another group of 10 rats, two parallel, 6-cm linear incisions were made with a scalpel (#11) through the panniculus carnosus paravertebrally at the same position as the ePTFE tubes. Bupivacain (1.5 mg) was applied to each incision for local analgesia.

Surgery Volume 150, Number 5

Danielsen et al 899

Fig 1. MMP-8 protein production by transduced rat dermal fibroblasts demonstrated by ELISA (A), type I collagenolysis (B) and Western blot (C) analyses. A. MMP-8 levels in cell lysates (open bars) or medium (closed) of cells transduced with AdMMP-8 (mean ± SEM, n = 3). B. Native (trypsin-resistant) bovine type I collagen (15 mg) was incubated in assay buffer, 2 mg total proteins of cell lysate, 300 ng of recombinant rat MMP-8 (R&D Systems), or 2 mg total proteins of conditioned medium without or with 10 mM BB-3103,22 with 1, 10 or 100 nM of MMP-8 inhibitor47 (444237; Calbiochem) at 238C for 72 h in the presence of 1 mM 4-aminophenylmercuric acetate in 60 mL. In MMP-8I, A refers to the 3/4 fragment TCA and B to the 1/4 fragment TCB of 2.5 mg digested collagen analyzed on NuPAGE 4–12% Bis-Tris gel under reducing conditions. C. Western blot analysis of 1 mg total proteins of lysate or medium of AdMMP-8-transduced fibroblasts. Proteins were separated on NuPAGE 12% Bis-Tris gel, transferred to polyvinylidene fluoride membrane and probed with anti-rat MMP-8 polyclonal antibody (1:1000 dilution; R&D Systems). Upper solid arrow indicates the position of proMMP-8 and dotted arrow active recombinant rat MMP-8.

The AdLacZ (5 3 1010 viral particles) or AdMMP-8 (5 3 1010 viral particles) treatments were randomized and administered in 0.5 mL of diluted fibrin (Tisseel Duo Quick; Baxter, Vienna, Austria) to each wound.23 To the 1-mL fibrinogen (about 90 mg/mL) component was added 0.5 mL Tris-buffered saline (pH 7.0). One mL of human thrombin (500 IU) was diluted 125 times with 40 mM calcium chloride in Tris-buffered saline. The 1:1 mixture of these diluted Tisseel components are supposed to prolong the release of bioactive adenovirus, increase gene transfer to wounds,24 and have no significant effect on day-7 breaking strength.20 The incisional wounds were closed with 6 interrupted 4.0 polyamide sutures 8 mm apart. Insertion of tubes, creation of incisions, and treatment applications were done by the same investigator under sterile conditions. The study was approved by the Animal Welfare Committee and Danish Working Environment Authority, and conducted in a class II gene laboratory according to the Danish Ministry of Justice law on animal welfare. Tissue procurement. Rats were killed by intracardial injection of 3 mL pentobarbital (200 mg/mL) containing lidocaine hydrochloride (20 mg/mL). The ePTFE tubes were pulled out on postoperative days 2 and 7 and then divided into 1-cm pieces; the

0.75-cm ends were discarded. The two most cranial pieces were stored at
808C, the next piece in 10 mL acetone at 48C for hydroxyproline analysis on day 7, and the caudal 1-cm ePTFE piece was placed in RNAlater (Ambion, Austin, TX) for 2 mo at 48C for mRNA analyses.25 From each incisional wound, four 8-mm wide strips were cut perpendicular to the incision line through the suture holes. The most cranial biopsy was kept in phosphate-buffered paraformaldehyde (4%) for 24 h at 48C for histology and immunohistochemistry, the next two strips fixed in phosphate-buffered (pH 7.4) 10% formalin for 7 days at ambient temperature for breaking strength analyses, and the caudal biopsy (4 mm) was stored in RNAlater at 48C. mRNA quantification. Tissues were ground in liquid nitrogen and homogenized in 25% guanidine hydrochloride lysis buffer solution containing 2% (v/v) beta-mercaptoethanol using the UltraTurrax T8 Homogenizer (Colonial Scientific, Richmond, VA). Copy numbers of MMP-8, MMP-9, type I procollagen a1 (COL1A1), and type III procollagen a1 (COL3A1) were determined with real-time quantitative RT-PCR using the primers specified in Table I.26 The use of glyceraldehyde-3phosphate dehydrogenase (GAPDH) mRNA, a widely accepted standard was used as an internal control and was validated by the measurement of

900 Danielsen et al

Surgery November 2011

Table I. Primers used for quantitative real-time RT-PCR Gene

Forward primer (59
39)

Reverse primer (59
39)

MMP-8y MMP-8z MMP-9y COL1A1y COL3A1y GAPDHy GAPDHz RPLP0y

CCATGGATCCAGGTTACCCCACT CAGACAACCCTGTCCAACCT GGATGTTTTTGATGCCATTGCTG ATCAGCCCAAACCCCAAGGAGA TGATGGGATCCAATGAGGGAGA CCATTCTTCCACCTTTGATGCT AGACAGCCGCATCTTCTTGT AGGGTCCTGGCTTTGTCTGTGG

TGTGGTCCACTGAAGAAGAGGAAGA GGATGCCGTCTCCAGAAGTA CCACGTGCGGGCAATAAGAAAG CGCAGGAAGGTCAGCTGGATAG GAGTCTCATGGCCTTGCGTGTTT TGTTGCTGTAGCCATATTCATTGT TGATGGCAACAATGTCCACT AGCTGCAGGAGCAGCAGTGG

Accession number* NM_022221 NM_022221 NM_031055 NM_053304 NM_032085 NM_017008 NM_017008 NM_022402

*National Center for Biotechnology Information. yPrimers used for ePTFE tube mRNA analyses. zPrimers used for incisional wounds and skin mRNA analyses.

the stability of the GAPDH mRNA expression relative to ribosomal protein, large, P0 (RPLP0) mRNA and relative to total RNA. MMP-8 mRNA levels in incisional wounds and adjacent skin were determined using a slightly different protocol and primers (Table I).20 Analysis of collagen (hydroxyproline). The ePTFE tubes were delipidated, lyophilized to constant weight, hydrolyzed in 6 N HCl at 1008C for 16 h, and hydroxyproline measured colorimetrically.27,28 Results are expressed as mg collagen (7.46 3 mg hydroxyproline)/mg delipidated dry tissue including the ePTFE tube material. Measurement of wound breaking strength. The strips from the incisional wounds were stretched at a constant deformation rate (10 mm/min) in a materials testing machine until rupture, and the maximal load (breaking strength) was derived from the load-strain curve.29 The mean of the two strips in N was used for the statistical analyses. Myeloperoxidase (MPO) was measured using a specific rat ELISA (HK105) according to manufacturer’s instructions (Hycult, Uden, the Netherlands). MMP-8 protein and gelatin zymographic analyses. Two-mm pieces of ePTFE were incubated for 24 h at 48C in extraction buffer.22 The levels of MMP-8 were measured by rat MMP-8 ELISA kit (RayBiotech, Norcross, GA). Gelatin zymography was performed using gels, apparatus, and reagents from Invitrogen (Carlsbad, CA). After electrophoresis, gels were incubated for 18 h at 378C in developing buffer, stained with EZBlueÔ, destained, and assembled. Densitometry was performed using ImageJ (National Institutes of Health, Bethesda, MD).30 Total proteins were measured in cell lysates, supernatants and tissue extracts according to Bradford. Histopathology and quantification of immunohistochemical staining. The fixed tissues were

embedded in paraffin. Tissue sections (5 mm) were cut perpendicular to both the anteriorposterior axis and the surface of the wound from each paraffin block. In hematoxylin-eosin-stained sections, the amount of granulation tissue between the dermal wound edges was measured in mm31 and cellularity estimated on a four-graded scale (0-3).32 For MMP-8 and macrophage immunohistochemical analysis, sections were boiled in 10 mM sodium citrate (pH 6.0), quenched with 3% hydrogen peroxide, and incubated overnight at 48C with rabbit anti-human MMP-8 polyclonal antibody (1:250 dilution; SA-370; Biomol Research Laboratories, Plymouth Meeting, PA) or with mouse antirat CD68 monoclonal antibody (1:200 dilution; MCA341R; Serotec, Oxford, UK).33 For type I collagen, adjacent sections were treated with 3% hydrogen peroxide, blocked with goat serum, and incubated for 60 min at ambient temperature with rabbit anti-rat type I collagen polyclonal antibody (1:500 dilution; AB755P; Millipore).34 MMP-8 sections were then incubated with the peroxidaseconjugated anti-rabbit EnVisionTM system (K4002; Dako, Glostrup, Denmark).33 The type I collagen and CD68 sections were treated with the Vectastain Elite ABC (Vector Laboratories, Burlingame, CA) followed by reaction with 3,39-diaminobenzidine. Sections were counterstained with Mayers hematoxylin. MMP-8 and type I collagen immunoreactivity was quantified in digital images using Image-Pro Plus 4.1 (Media Cybernetics, Silver Spring, MD). We assessed 4 fields in the wounds, 2 in subepidermal region and 2 in deep dermal region just above the level of panniculus carnosus as detailed elsewhere,34 and 4 fields in mid-dermis of adjacent normal skin 2 cm from the incisions. For each animal, the ratio of the sum of all immunopositive areas to the total visual field area (550 mm 3 420 mm) 3100 was calculated, and the mean of the 4 ratios was

Danielsen et al 901

Surgery Volume 150, Number 5

Table II. Analyses of treated ePTFE tubes Day 2

MPO (mg/mg total protein)* MMP-8 mRNA/GAPDH mRNAz MMP-8 (ng/mg total protein)x MMP-9 mRNA/GAPDH mRNAz MMP-9 activityjj MMP-2 activityjj COL1A1 mRNA/GAPDH mRNAz COL3A1 mRNA/GAPDH mRNAz

Day 7

AdLacZ (n = 9)

AdMMP-8 (n = 9)

AdLacZ (n = 10)

AdMMP-8 (n = 10)

389 ± 44 1.00 1760 ± 133 1.00 2070 ± 146 776 ± 174 1.00 1.00

369 ± 40 0.93 2120 ± 528 0.95 1980 ± 126 683 ± 76 0.95 1.07

107 ± 39 0.06 913 ± 410 0.40 992 ± 445 1500 ± 270 322 435

36 ± 9y 0.33y 395 ± 165 0.34 237 ± 61 1420 ± 254 312 480

*MPO, Myeloperoxidase (marker for neutrophils). Mean ± SEM. yP # .05 vs AdLacZ on day 7. zFold changes relative to the mean of AdLacZ day 2 values and presented as geometric means. xMeasured by ELISA. The MMP-8 concentration in normal, uninjured skin was 17.6 ± 1.0 ng/mg total protein (n = 5). Mean ± SEM. jjMeasured by gelatin zymography and expressed in arbitrary densitometric units per 5 mg of total protein. Mean ± SEM.

used for the statistical analyses. The number of CD68+ cells was counted manually just below the epithelium and just above the level of panniculus carnosus of the central granulation tissue and expressed per field of view (330 mm 3 240 mm). Blinding and statistical analyses. Operations and analyses were conducted by investigators blinded to group allocation. The Student paired t test was applied and significance level was set to P # .05. Numerical data are presented as mean ± SEM. RESULTS To increase expression of MMP-8 locally, we constructed a replication-deficient adenovirus vector overexpressing full-length rat MMP-8 cDNA under the control of the human CMV IE promoter (AdMMP-8). Assessments of the adenoviral MMP-8 construct. PmeI digest of purified DNA run on agarose gel produced the anticipated band. Sequence analysis of three clones showed that we had recovered adenovirus containing a complete MMP-8 cDNA insert. Furthermore, the 59 and 39 ends of the insert were intact and positioned correctly in the adenoviral vector. The functionality of the AdMMP-8 construct was evaluated first in rat dermal fibroblasts. The conditioned medium from AdMMP-8-transduced cells contained conspicuous amounts of MMP-8 protein with collagenolytic activity (Fig 1, A and B). Collagenolytic activity was assessed using native type I collagen and indicated by the generation of the characteristic 3/4 and 1/4 fragments. Collagenolysis was blocked by nonselective and selective hydroxamate MMP inhibitors but required the presence of the MMP-activator 4-aminophenylmercuric acetate (Fig 1, B). Western blot analysis indicated that

some of the secreted MMP-8 protein was in an active form (Fig 1, C). MMP-8 was also made by AdLacZtreated fibroblasts but at low levels (440 ng/mg protein in cell lysate and 19 ng/mg protein in media) without detectable collagenolytic activity. Taken together, these findings suggest strongly that MMP-8 with collagenolytic capacity is secreted by fibroblasts treated with our adenoviral MMP-8 construct. Subcutaneous ePTFE tube wound model. The ePTFE tube model was used to study the temporal effects of MMP-8 overexpression. This model allows the isolation of newly deposited collagenous repair tissue that correlates with early biomechanical wound strength.20 One animal died during anesthesia. There were no wound infections, and all animals gained body weight (14 ± 1%) over the 7 postoperative days. Neutrophils are important sources of MMP-8 as well as of MMP-9 and were quantified indirectly by the marker MPO. MPO levels decreased (P < .001) from day 2 to day 7 in AdLacZ-treated sites. There were no significant differences in neutrophil infiltration on day 2, but on day 7, MPO levels were less (P = .05) in AdMMP-8-treated tubes compared to control tubes. Although there was no significant difference on day 2, the level of MMP-8 mRNA was greater (P < .05) in AdMMP-8-treated ePTFE tubes on day 7 compared to the contralateral AdLacZtreated ePTFE tubes (Table II). The early dominance of neutrophils may explain why MMP-8 mRNA levels were not increased on day 2 by AdMMP-8 treatment, because neutrophils are poorly permissive for adenoviral transduction.35 Furthermore, the increased MMP-8 mRNA was not reflected in increased MMP-8 protein levels in AdMMP-8treated ePTFE tubes. Neither MMP-9 mRNA nor protein levels differed significantly between the 2 groups at either time point (Table II).

902 Danielsen et al

Fig 2. Effects of AdMMP-8 treatment on (A) collagen deposition in subcutaneous ePTFE tubes and (B) breaking strength of incisional wounds postoperative day 7. The values for AdLacZ (B) and AdMMP-8 (C) treatment in each animal are connected with a line. The white bars represent the mean ± SEM of the AdLacZ treatment and the black bars the mean ± SEM of the AdMMP-8 treatment in the 10 animals. Collagen deposition and breaking strength were less for the AdMMP-8-treated sites in 7 of the 10 rats. *P < .05 (paired t-test), AdLacZ vs AdMMP-8.

Because MMP-236 and one of its activators MMPhave collagenolytic activity, MMP-2 was 14 analyzed by gelatin zymography. This control is particularly important, because Siller-L opez et al38 reported increased hepatic MMP-2 protein levels with systemic, full-length human AdMMP-8 treatment in cirrhotic rats. While total MMP-9 levels decreased (P < .05) from days 2 to 7, total MMP-2 levels increased (P < .05) in AdLAcZtreated ePTFE tubes. Notable was the almost total lack of active MMP-2 molecular protein species on day 2, while about 50% of the total MMP-2 proteins 37

Surgery November 2011

were in an active form day 7. Importantly, MMP-2 protein levels, pro and active forms, were similar in AdMMP-8 and AdLacZ-treated ePTFE tubes days 2 and 7 (Table II). The latter finding may indicate that the endogenous MMP-2 activator MMP-14 was not altered either. Recombinant rat proMMP-8 (500 ng) did not degrade gelatin, whereas the proMMP-2 activity was detectable at 0.5 ng. Collagen turnover in the ePTFE tubes was determined by measuring procollagen mRNA levels and collagen deposition. Both COL1A1 mRNA and COL3A1 mRNA were detectable by real-time RT-PCR already on day 2, but the expression of both genes had increased several 100-fold by day 7. There were no significant differences in either COL1A1 mRNA or COL3A1 mRNA between AdMMP-8 and AdLacZ-treated tubes on days 2 or 7 (Table II). In contrast, collagen protein deposition measured on day 7 as hydroxyproline was less with AdMMP-8 treatment (P < .05; Fig 2, A). Collectively, these data suggest that MMP-8 overexpression decreases collagen levels during early wound repair without a significant effect on the synthesis of type I or type III collagens or on MMP-9, MMP-2 or MMP-14 protein levels. Incisional wound healing model. Next, we used the paired rat surgical incision model to evaluate the local effect of AdMMP-8 treatment using the clinical outcome of wound repair. Successful transduction of the MMP-8 gene to the wounds was indicated by the increased (P < .05) MMP-8 mRNA level in AdMMP-8-treated wounds compared to contralateral AdLacZ-treated wounds day 7 and also compared with adjacent uninjured skin. In contrast, MMP-8 mRNA levels did not differ significantly between AdLacZ-treated wounds and skin (Fig 3, A). The breaking strength was determined of formalin-fixed wound strips. Formalin maximizes cross-linking of available scar collagen.20 There was a negative effect of AdMMP-8 compared with AdLacZ on wound healing (P < .05) representing a 41% decrease in breaking strength on day 7 (Fig 2, B). The amount of granulation tissue did not differ significantly between AdLacZ-treated and AdMMP-8-treated wounds (899 ± 327 mm vs 735 ± 300 mm). Neutrophils were absent or distributed very sparsely in the granulation tissue at this stage of healing day 7.32 Rather, the inflammatory infiltrate was monocytic, although the cellularity scores were similar in the AdLacZ and AdMMP-8 groups (2.5 ± 0.3 vs 2.4 ± 0.3). Protein expression of MMP-8 was quantified by image analysis of immunostained sections. AdMMP-8-treated wounds displayed increased MMP-8 immunostaining by

Danielsen et al 903

Surgery Volume 150, Number 5

B

10

25

**

Skin AdLacZ AdMMP-8

*

1

20

MMP-8 protein (%)

MMP-8 mRNA/GAPDH mRNA (relative change)

A

15

10

5

0.1

0 Skin

D

200

25

20

150

Type I collagen (%)

CD68+ cells/field

C

AdLacZ AdMMP-8

100

50

15

*

10

5

0

0 Skin

AdLacZ AdMMP-8

Skin

AdLacZ AdMMP-8

Fig 3. Effects of AdMMP-8 treatment of incisional wounds. (A) MMP-8 mRNA in normal skin and incisional wounds on day 7. mRNA were normalized to GAPDH and presented as geometric means ± SEM (n = 10) of fold changes relative to the mean of AdLacZ values. Quantification of MMP-8 protein (B), CD68+ cells (C), and type I collagen (D) of immunostained sections (cf. Fig 4). MMP-8 and type I collagen were expressed as percentage (%) of total stained area. Mean ± SEM, n = 10. *P < .05, **P < .01, AdLacZ vs AdMMP-8.

76% compared to AdLacZ-treated wounds and by 167% compared to adjacent normal skin (P < .01; Fig 3, B, and Fig 4, A–F). Because macrophages are important sources of MMP-8 and regulate collagen deposition during wound healing,33,39 CD68+ cells were counted. As anticipated, the number of macrophages was (P < .001) increased in the wounds compared to adjacent skin, but there was no significant difference in the density of macrophages in the granulation tissue between AdLacZ and AdMMP-8-treated wounds (Fig 3, C, and Fig 4, G–I). This observation suggests that the difference in MMP-8 content was not due to differences in the number of MMP-8 producing macrophages. Type I collagen was 29% less in AdMMP-8-treated compared with AdLacZ-treated wounds (P < .05; Fig 3, D, and Fig 4, J–O). DISCUSSION The primary focus of our research is to identify pathophysiologic factors in nonhealing cutaneous wounds in the search for novel interventions.3 In

the present study, we examined the pathologic role of the neutrophil-derived proteinase MMP-8. This specific topic has not been addressed before. Our strategy was to overexpress this interstitial collagenase locally using an adenoviral vector to ascertain prolonged and localized exposure of the enzyme at injury sites. Accordingly, a single administration of AdMMP-8 decreased the amount of collagen (specifically type I collagen) and the breaking strength of incisional skin wounds through degradative rather than synthetic collagen pathways. These results suggest strongly that excessive levels of MMP-8 are detrimental to tissue repair. Secondarily, we observed fewer neutrophils with MMP-8 overexpression. Consistent with this antiinflammatory role, MMP-8 can decrease the number of neutrophils through apoptosis.15,40 In another rodent study, neutrophil-depletion had no significant effects on either collagen deposition or breaking strength.41 Thus, it is remarkable that the presence of fewer neutrophils in AdMMP-8-treated

904 Danielsen et al

Surgery November 2011

Fig 4. MMP-8 (A–C), CD68 (G–I) and type I collagen (J–L) representative immunostained images of normal skin (A, G, J), and granulation tissue of AdLacZ-treated (B, H, K) and AdMMP-8-treated incisional wounds (C, I, L) day 7. The corresponding converted gray scale images used for quantification of positive MMP-8 and type I collagen immunostaining are shown in D–F and M–O. A–C, Latent and active forms of MMP-8 were recognized by the antibody. Omission of primary MMP-8 and type I collagen antibodies showed no brown staining. Treatment with isotypic control (mouse IgG1, 1:250 dilution; MCA1209; Serotec) produced no positive immunostaining.

Danielsen et al 905

Surgery Volume 150, Number 5

wounds was accompanied with a decreased amount of collagen, because the specific granules of neutrophils contain large amounts of preformed MMP-8. Because most of the latent and active MMP-8 is bound to the membrane of activated neutrophils,42 collagen degradation is perhaps restricted to the very vicinity of the cells with no net effect on the bulk content of tissue collagen. Decreased numbers of neutrophils may have masked increased MMP-8 protein production by the transduced cells in the subcutaneous ePTFE tubes. In contrast, the fibroblast-rich dermis may explain the increased levels of MMP-8 protein observed in incisional wounds.43 Contrary to the weak link between neutrophils and wound healing,41 several studies support the essential regulatory role of macrophages.39 Skin wounds depleted of macrophages contain less collagen.39 In our study, there was no significant difference in the number of macrophages and hence, the detrimental effect of overexpression of MMP-8 activity on collagen deposition and biomechanical strength was unlikely due to an effect of AdMMP-8 on macrophage recruitment or survival. Loss of one MMP can be compensated for by up-regulation of another MMP with a similar biochemical profile.15,16 This concept was studied in detail by Sabeh et al44 who found that fibroblasts isolated from the skin of MMP-8-null mice preserved their collagenolytic phenotype. Although MMP-2 or MMP-14 enzymes both have collagenolytic capacity,36,37,44 we found no evidence for altered protein levels of MMP-2 or MMP-14 with MMP-8 overexpression. Moreover, rat MMP-13 collagenase is expressed only at very low levels in granulation tissue of incisional wounds on day 7.45 The obvious clinical implication of our findings is to block the excessive MMP-8 activity associated with different conditions of pathologic wound healing.7-10,12,18,19,33 Before embarking on this approach, it will be necessary to consider the ample evidence for the physiologic roles of MMPs in general and in wound healing specifically.6,7,15,22,46 Indiscriminate pharmacologic inhibition of MMPs also impairs several mechanisms of wound healing.22,46 Thus, only highly selective MMP-8 inhibitors should be considered for this therapeutic strategy.7 Our initial trial suggests in rat fibroblasts that it is possible to efficiently and dosedependently decrease excessive MMP-8 collagenolytic activity using a selective inhibitor.47 In conclusion, this study strongly suggests that excessive MMP-8 inhibits wound repair through degradation of newly synthesized fibrillar collagens.

Selective targeting of excessive MMP-8 activity may restore the balance in proteinase function and hence wound repair. The authors thank Lotte Laustsen for technical assistance and Peter Schjerling for his advice on the mRNA measurements.

REFERENCES 1. Sen CK, Gordillo GM, Roy S, Kirsner R, Lambert L, Hunt TK, et al. Human skin wounds: a major and snowballing threat to public health and the economy. Wound Repair Regen 2009;17:763-71. 2. Menke NB, Ward KR, Witten TM, Bonchev DG, Diegelmann RF. Impaired wound healing. Clin Dermatol 2007; 25:19-25. 3.
Agren MS, Werthen M. The extracellular matrix in wound healing: a closer look at therapeutics for chronic wounds. Int J Low Extrem Wounds 2007;6:82-97. 4. Meyer FJ, Burnand KG, Abisi S, Tekoppele JM, van Els B, Smith A. Effect of collagen turnover and matrix metalloproteinase activity on healing of venous leg ulcers. Br J Surg 2008;95:319-25. 5. Nagase H, Visse R, Murphy G. Structure and function of matrix metalloproteinases and TIMPs. Cardiovasc Res 2006;69: 562-73. 6. Gill SE, Parks WC. Metalloproteinases and their inhibitors: regulators of wound healing. Int J Biochem Cell Biol 2008; 40:1334-47. 7. Xue M, Le NT, Jackson CJ. Targeting matrix metalloproteases to improve cutaneous wound healing. Expert Opin Ther Targets 2006;10:143-55. 8. Nwomeh BC, Liang HX, Cohen IK, Yager DR. MMP-8 is the predominant collagenase in healing wounds and nonhealing ulcers. J Surg Res 1999;81:189-95. 9. Lobmann R, Ambrosch A, Schultz G, Waldmann K, Schiweck S, Lehnert H. Expression of matrix-metalloproteinases and their inhibitors in the wounds of diabetic and non-diabetic patients. Diabetologia 2002;45:1011-6. 10. Muller M, Trocme C, Lardy B, Morel F, Halimi S, Benhamou PY. Matrix metalloproteinases and diabetic foot ulcers: the ratio of MMP-1 to TIMP-1 is a predictor of wound healing. Diabet Med 2008;25:419-26. 11. Beidler SK, Douillet CD, Berndt DF, Keagy BA, Rich PB, Marston WA. Multiplexed analysis of matrix metalloproteinases in leg ulcer tissue of patients with chronic venous insufficiency before and after compression therapy. Wound Repair Regen 2008;16:642-8. 12. Piril€a E, Korpi JT, Korkiam€aki T, Jahkola T, Gutierrez-Fernandez A, Lopez-Otin C, et al. Collagenase-2 (MMP-8) and matrilysin-2 (MMP-26) expression in human wounds of different etiologies. Wound Repair Regen 2007;15:47-57. 13. Pelman GR, Morrison CJ, Overall CM. Pivotal molecular determinants of peptidic and collagen triple helicase activities reside in the S39 subsite of matrix metalloproteinase 8 (MMP-8): the role of hydrogen bonding potential of ASN188 and TYR189 and the connecting cis bond. J Biol Chem 2005;280:2370-7. 14. Hasty KA, Jeffrey JJ, Hibbs MS, Welgus HG. The collagen substrate specificity of human neutrophil collagenase. J Biol Chem 1987;262:10048-52. 15. Gutierrez-Fernandez A, Inada M, Balbın M, Fueyo A, Pitiot AS, Astudillo A, et al. Increased inflammation delays wound

906 Danielsen et al

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28. 29.

30. 31.

32.

healing in mice deficient in collagenase-2 (MMP-8). FASEB J 2007;21:2580-91. Hartenstein B, Dittrich BT, Stickens D, Heyer B, Vu TH, Teurich S, et al. Epidermal development and wound healing in matrix metalloproteinase 13-deficient mice. J Invest Dermatol 2006;126:486-96. Burrow JW, Koch JA, Chuang HH, Zhong W, Dean DD, Sylvia VL. Nitric oxide donors selectively reduce the expression of matrix metalloproteinases-8 and -9 by human diabetic skin fibroblasts. J Surg Res 2007;140:90-8. Liu KZ, Hynes A, Man A, Alsagheer A, Singer DL, Scott DA. Increased local matrix metalloproteinase-8 expression in the periodontal connective tissues of smokers with periodontal disease. Biochim Biophys Acta 2006;1762:775-80. Sørensen LT, Zillmer R,
Agren M, Ladelund S, Karlsmark T, Gottrup F. Effect of smoking, abstention, and nicotine patch on epidermal healing and collagenase in skin transudate. Wound Repair Regen 2009;17:347-53. Danielsen PL,
Agren MS, Jorgensen LN. Platelet-rich fibrin versus albumin in surgical wound repair: a randomized trial with paired design. Ann Surg 2010;251:825-31. Becker TC, Noel RJ, Coats WS, Gomez-Foix AM, Alam T, Gerard RD, et al. Use of recombinant adenovirus for metabolic engineering of mammalian cells. Methods Cell Biol 1994;43(Pt A):161-89. Mirastschijski U, Impola U, Karsdal MA, Saarialho-Kere U,
Agren MS. Matrix metalloproteinase inhibitor BB-3103 unlike the serine proteinase inhibitor aprotinin abrogates epidermal healing of human skin wounds ex vivo. J Invest Dermatol 2002;118:55-64. Schek RM, Hollister SJ, Krebsbach PH. Delivery and protection of adenoviruses using biocompatible hydrogels for localized gene therapy. Mol Ther 2004;9:130-8. Breen A, Dockery P, O’Brien T, Pandit A. Fibrin scaffold promotes adenoviral gene transfer and controlled vector delivery. J Biomed Mater Res A 2009;89:876-84. Danielsen PL, Koudahl V, Christensen B, Olsen J. Optimization of RNA preservation in granulation tissue in the expanded polytetrafluoroethylene (ePTFE) tube model. Prep Biochem Biotechnol 2009;39:413-8. Heinemeier KM, Olesen JL, Haddad F, Langberg H, Kjaer M, Baldwin KM, et al. Expression of collagen and related growth factors in rat tendon and skeletal muscle in response to specific contraction types. J Physiol 2007;582: 1303-16. Woessner JF Jr. The determination of hydroxyproline in tissue and protein samples containing small proportions of this imino acid. Arch Biochem Biophys 1961;93:440-7. Grant RA. Estimation of hydroxyproline by the AutoAnalyzer. J Clin Pathol 1965;18:686.
Agren MS, Andersen L, Heegaard AM, Jorgensen LN. Effect of parenteral zinc sulfate on colon anastomosis repair in the rat. Int J Colorectal Dis 2008;23:857-61. Hu X, Beeton C. Detection of functional matrix metalloproteinases by zymography. J Vis Exp 2010;45:2445. Akasaka Y, Ono I, Yamashita T, Jimbow K, Ishii T. Basic fibroblast growth factor promotes apoptosis and suppresses granulation tissue formation in acute incisional wounds. J Pathol 2004;203:710-20. Gal P, Toporcer T, Vidinsky B, Mokry M, Novotny M, Kilık R, et al. Early changes in the tensile strength and morphology

Surgery November 2011

33.

34.

35.

36.

37.

38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

of primary sutured skin wounds in rats. Folia Biol (Praha) 2006;52:109-15.
Agren MS, Andersen TL, Mirastschijski U, Syk I, Schiødt CB, Surve V, et al. Action of matrix metalloproteinases at restricted sites in colon anastomosis repair: an immunohistochemical and biochemical study. Surgery 2006;140:72-82. Shah M, Foreman DM, Ferguson MW. Neutralising antibody to TGF-beta 1,2 reduces cutaneous scarring in adult rodents. J Cell Sci 1994;107(Pt 5):1137-57. Lyons M, Onion D, Green NK, Aslan K, Rajaratnam R, Bazan-Peregrino M, et al. Adenovirus type 5 interactions with human blood cells may compromise systemic delivery. Mol Ther 2006;14:118-28. Aimes RT, Quigley JP. Matrix metalloproteinase-2 is an interstitial collagenase. Inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native type I collagen generating the specific 3/4- and 1/4-length fragments. J Biol Chem 1995;270:5872-6. Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M, Okada Y. Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J Biol Chem 1997;272:2446-51. Siller-L opez F, Sandoval A, Salgado S, Salazar A, Bueno M, Garcia J, et al. Treatment with human metalloproteinase-8 gene delivery ameliorates experimental rat liver cirrhosis. Gastroenterology 2004;126:1122-33. Mirza R, DiPietro LA, Koh TJ. Selective and specific macrophage ablation is detrimental to wound healing in mice. Am J Pathol 2009;175:2454-62. Cox JH, Starr AE, Kappelhoff R, Yan R, Roberts CR, Overall CM. Matrix metalloproteinase 8 deficiency in mice exacerbates inflammatory arthritis through delayed neutrophil apoptosis and reduced caspase 11 expression. Arthritis Rheum 2010;62:3645-55. Dovi JV, He LK, DiPietro LA. Accelerated wound closure in neutrophil-depleted mice. J Leukoc Biol 2003;73:448-55. Owen CA, Hu Z, Lopez-Otin C, Shapiro SD. Membrane-bound matrix metalloproteinase-8 on activated polymorphonuclear cells is a potent, tissue inhibitor of metalloproteinaseresistant collagenase and serpinase. J Immunol 2004;172: 7791-803. Bhattacharyya TK, Merz M, Thomas JR. Modulation of cutaneous aging with calorie restriction in Fischer 344 rats: a histological study. Arch Facial Plast Surg 2005;7:12-6. Sabeh F, Li XY, Saunders TL, Rowe RG, Weiss SJ. Secreted versus membrane-anchored collagenases: relative roles in fibroblast-dependent collagenolysis and invasion. J Biol Chem 2009;284:23001-11. Porras-Reyes BH, Blair HC, Jeffrey JJ, Mustoe TA. Collagenase production at the border of granulation tissue in a healing wound: macrophage and mesenchymal collagenase production in vivo. Connect Tissue Res 1991;27:63-71. Mirastschijski U, Haaksma CJ, Tomasek JJ,
Agren MS. Matrix metalloproteinase inhibitor GM 6001 attenuates keratinocyte migration, contraction and myofibroblast formation in skin wounds. Exp Cell Res 2004;299:465-75. Matter H, Schwab W, Barbier D, Billen G, Haase B, Neises B, et al. Quantitative structure-activity relationship of human neutrophil collagenase (MMP-8) inhibitors using comparative molecular field analysis and X-ray structure analysis. J Med Chem 1999;42:1908-20.