Identification of Proteinase 3 as the Major Caseinolytic Activityin Acute Human Wound Fluid

Yajuan He, Patty K. Young, and Frederick GrinnellDepartment of Cell Biology and Neuroscience, University of Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

Wound fluid contains several proteinases that areimportant in the repair process. In this study, weanalyzed caseinolytic activity in wound fluid obtainedfrom acute (burn) wounds. Caseinolytic activity inwound fluid increased markedly 2 d after injury andappeared on casein zymographs as a series of bands ora smear ranging from 30 to 100 kDa. Most of theenzyme activity was inhibited by the synthetic humanneutrophil elastase inhibitor MDL 27,367 but not bythe naturally occurring inhibitor of elastase, humansecretory leukoproteinase inhibitor. Fractionation ofwound fluid indicated that a single enzyme accountedfor µ80% of the caseinolytic activity. This enzyme

Wound healing is a coordinated process involvingtissue degradation and regeneration. Excessivedestruction of extracellular matrix molecules byproteinases has been implicated as a contributingfactor in the failure of wounds to heal (Falanga

et al, 1994). Because control of wound proteinases may provide atherapeutic strategy to promote healing, identifying the proteinasesthat participate in repair has become a major focus of wound healingresearch. One approach to accomplishing this goal is analysis of woundfluid (Wysocki and Grinnell, 1990; Kirsner et al, 1993; Tarnuzzer andSchultz, 1996). Wound fluid contains a mixture of plasma componentssuch as the adhesion proteins fibronectin and vitronectin (Grinnellet al, 1992) and locally secreted substances such as proteinases (seebelow) and growth factors (Alper et al, 1985; Matsuoka and Grotendorst,1989; Cooper et al, 1994).

A wide array of proteinases have been identified in wound fluid.These include matrix metalloproteinases (MMP) (interstitial collagenase,stromelysin, and gelatinases) (Matsubara et al, 1991; Agren et al, 1992;Young and Grinnell, 1994) and serine proteinases, e.g., plasmin andplasminogen activator (Palolahti et al, 1993; Stacey et al, 1993).Gelatinases (MMP-2 and MMP-9) occur at particularly high levels inwound fluid from chronic leg ulcers (Wysocki et al, 1993; Bullen et al,1995; Yager et al, 1996; Weckroth et al, 1996), and elastase has beenidentified as the enzyme in wound fluid responsible for fibronectindegradation (Grinnell and Zhu, 1994, 1996; Rao et al, 1995).

Palolahti et al (1993) reported that chronic wound fluid containselevated levels of a caseinolytic activity although they did not identify

Manuscript received May 21, 1997; revised September 8, 1997; accepted forpublication September 16, 1997.

Reprint requests to: Dr. Frederick Grinnell, Department of Cell Biology andNeuroscience, UT Southwestern Medical Center, Dallas, TX 75235.

Abbreviations: MMP, matrix metalloproteinase; PR-3, proteinase 3; SLPI,secretory leukoproteinase inhibitor.

0022-202X/98/$10.50 · Copyright © 1998 by The Society for Investigative Dermatology, Inc.

67

degraded the elastase substrate methoxysuccinyl-ala-ala-pro-val-p-nitroanilide at a slow rate. The abovefindings suggested that the enzyme responsible forcaseinolytic activity might be proteinase 3, an elastase-related enzyme whose physiologic functions are poorlyunderstood. Consistent with the above possibility, wefound that monoclonal antibodies against proteinase 3removed caseinolytic activity from wound fluid, andthat purified proteinase 3 had a similar caseinolyticprofile and inhibitor sensitivity to burn fluid. Key words:burn injury/elastase/wound healing. J Invest Dermatol 110:67–71, 1998

the enzyme(s) responsible for the activity. Recently, we noticed thatacute wound fluid also contains substantial caseinolytic activity. Wehave now identified the enzyme responsible for this activity as proteinase3 (PR-3), an elastase-related serine proteinase whose physiologicfunctions are poorly understood (Campanelli et al, 1990; Rao et al,1991). Details are reported herein.

MATERIALS AND METHODS

Materials Human neutrophil PR-3 was purchased from Athens Research andTechnology (Athens, GA). Human neutrophil elastase (specific activity 5 20units per mg) and methoxysuccinyl-ala-ala-pro-val-p-nitroanilide (Cat#454454) were from Calbiochem (San Diego, CA). Recombinant humansecretory leukocyte proteinase inhibitor was from R&D Systems (Minneapolis,IN). Rabbit anti-human fibronectin antibody was prepared in our laboratory.Protein G agarose was purchased from Boehringer (Indianapolis, IN). Mousemonoclonal anti-PR-3 antibody (CLB-ANCA 12.8) was purchased fromResearch Diagnostics (Flanders, NJ). Aprotinin and casein were purchased fromSigma (St. Louis, MO). Specific elastase inhibitor MDL 27,367 (MDL) (Mehdiet al, 1990) was a gift from Hoechst Marion Roussel (Cincinnati, OH). Humanplasma fibronectin was obtained from the New York Blood Center.

Wound fluid This research project was approved by the University Institu-tional Review Board. Informed consent was obtained for all procedures.Samples of burn fluid were from four patients whose burn wounds includedan extremity. Sampling of burn fluid was initiated 4–8 h after injury (day 0)and continued at 48–72 h intervals. Collections were carried out over 4 hperiods during which time patient extremities were enclosed in a sterile gloveheld in place with a noncompressive cling wrap. Wound fluid samples werecentrifuged at 22,000 3 g (Beckman J2–21 M, 20 rotor, Beckman, Fullerton,CA) for 15 min at 22°C. Wound fluid supernatants were frozen in liquid N2and stored at –70°C until used.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting SDS-PAGE and immunoblotting were carriedout as described previously (Grinnell and Zhu, 1994, 1996). Briefly, samplesfor SDS-PAGE were dissolved in reducing sample buffer containing 62.5 mMTris, 1% SDS, 10% glycerol, 0.01% bromophenol blue, pH 6.8, and 5%

68 HE ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 1. Burn fluid shows increased caseinolytic activity and intactfibronectin during the first few days after injury. Samples (60 µg) of blisterfluid (B) and burn fluid collected on the days indicated were (a) subjected tocasein zymography or (b) immunoblotted with an antibody against fibronectin.Purified human fibronectin (FN) (0.2 µg) was used as a standard forimmunoblotting.

Figure 2. The elastase inhibitor MDL blocks caseinolytic activity inburn fluid. Samples (75 µg) of burn fluid (day 3) were subjected to caseinzymography. MDL (100 µM) or aprotinin (0.2 mg per ml) were added to thereaction buffer where indicated.

mercaptoethanol, and subjected to electrophoresis on 7.5% acrylamide mini-gel (Bio-Rad Mini-Protein apparatus, Bio-Rad, Richmond, CA) at 22°C for45 min at 200 V.

For immunoblotting, polypeptides separated by SDS-PAGE were transferredto nitrocellulose paper (Schleicher and Schuell, Keene, NH) by electrophoresisat 22°C for 1 h at 100 V. The transferred proteins were incubated withpolyclonal anti-fibronectin antibodies for 2 h at 22°C followed by alkaline-phosphatase-conjugated goat anti-rabbit IgG (Bio-Rad) for 1 h at 22°C.Visualization was accomplished using the Bio-Rad alkaline-phosphatase substratekit according to the manufacturer’s instructions.

Casein zymography Samples of wound fluid were dissolved in nonreducingsample buffer containing 62.5 mM Tris, pH 6.8, 1% SDS, 12% glycerol, and0.01% bromophenol blue, and SDS-PAGE was carried out as above usingcasein-containing acrylamide gels (7.5% acrylamide and 0.1% casein). Afterelectrophoresis, gels were washed twice with 2.5% Triton-X-100 for 30 minto remove SDS, rinsed briefly with H2O, and incubated overnight at 37°C inreaction buffer containing 50 mM Tris, 150 mM NaCl (pH 7.4). At the endof the incubations, the gels were stained with 0.125% Commassie Brilliant blue.Areas of proteinase activity appeared as clear zones against a dark bluebackground. Images were captured using a Dage-MTI CCD72 camera (Dage,Michigan City, IN) and analyzed with NIH Image 1.55. Data are presented inarbitrary units.

Elastase activity Elastase activity was determined as described previously(Grinnell and Zhu, 1994, 1996). Briefly, samples of wound fluid or chromato-graphic fractions were incubated at 22°C in 1 ml of reaction buffer containing0.1 M HEPES, 0.5 M NaCl, 10% dimethylsulfoxide, and 1 mM elastasesubstrate (methoxysuccinyl-ala-ala-pro-val-p-nitroanilide), pH 7.5 (Nakajimaet al, 1979). After incubation for the times indicated, substrate degradation wasdetermined by measuring OD410 (Beckman DU-40 spectrophotometer).

Figure 3. Gel filtration chromatography shows that a single peakcontains an elastase-related enzyme and most of the caseinolytic activity.Aliquots of the fractions from gel filtration HPLC were subjected to caseinzymography (a) (15 µl) or incubated with the elastase substrate (b, c) (10 µl).Levels of caseinolytic activity were quantitated by image capture and computeranalysis and are expressed as arbitrary units. Substrate degradation at the timesindicated was determined by measuring OD410.

Fractionation of caseinolytic activity in wound fluid

Ammonium sulfate precipitation Saturated ammonium sulfate (500 µl) was addeddropwise to 2 ml of burn fluid (50 mg per ml) to obtain a concentration of20% saturation. After incubation for 30 min, the sample was subjected tocentrifugation at 3000 3 g for 30 min. This procedure was repeated twice toobtain additional fractions of 40% and 60% saturation. The pellets were washedtwice with 20%, 40%, or 60% saturated ammonium sulfate, respectively, anddissolved in 50 µl of 20 mM Tris, pH 7.4. Protein concentrations weredetermined by DC Protein Assay kit (Bio-Rad).

Gel filtration high performance liquid chromatography (HPLC) Aliquots of the 60%ammonium sulfate precipitate (µ1 mg protein) were chromatographed on aProtein-Pak 300SW gel filtration column (Waters, Milford, MA) using theWaters HPLC system (Model 510) at a flow rate of 0.5 ml per min and 20 mMTris, 1 M NaCl, pH 7.4 buffer. Fractions were collected at 0.5 min intervals.Ultraviolet absorption was continuously recorded at 280 nm. The fractionswere stored at 4°C.

Ion exchange HPLC Fractions from the gel filtration column that containedcaseinolytic activity were pooled and chromatographed on a Protein-Pak DEAE8HR ion exchange column (Waters) using the Waters HPLC system (Model510) at a flow rate of 0.2 ml per min. The gradient was 15 min of 100% bufferA (20 mM Tris, pH 7.4) followed by 60 min of 100% buffer A to 100% bufferB (20 mM Tris, 1 M NaCl, pH 7.4). Ultraviolet absorption was continuouslyrecorded at 280 nm. The fractions were stored at 4°C.

Immunoprecipitation Immunoprecipitation of PR-3 was carried out asfollows. Samples of burn fluid (300 µg) were preincubated with protein Gagarose (50 µl) for 30 min at 4°C followed by centrifugation for 20 s at12,000 3 g to preclear any proteins that nonspecifically adsorb to the beads.Supernatants were then incubated with anti-PR-3 monoclonal antibody CLB-ANCA 12.8 (1:10, 0.2 mg per ml) or mouse IgG (1:10, 0.2 mg per ml) for1.5 h and then protein G agarose (50 µl) for 1 h. The complex was collectedby centrifugation for 20 s at 12,000 3 g and washed three times with Tris-buffered saline (20 mM Tris, pH 7.5, 150 mM NaCl). The beads were extractedwith non-reducing sample buffer containing 62.5 mM Tris, pH 6.8, 1% SDS,12% glycerol, and 0.01% bromophenol blue.

RESULTS

Caseinolytic activity in acute (burn) wound fluid increasesduring the first 2 d after injury Figure 1a shows the profile ofcaseinolytic activity detected by casein zymography of burn fluidsamples collected sequentially from a single patient. In blister fluid (B)and time 0 burn fluid (collected starting 4–6 h after injury), we detected

VOL. 110, NO. 1 JANUARY 1998 PROTEINASE 3 IN WOUND FLUID 69

Figure 4. The caseinolytic profile of purified PR-3 resembles thecaseinolyic profile of burn fluid. Samples of burn fluid (BF, day 3) (75 µg)or purified PR-3 as indicated were subjected to casein zymography. In (b),MDL (100 µM) was added to the reaction buffer.

a band at µ70 kDa. Beginning at day 2 after injury, a large numberof additional caseinolytic bands became evident. These bands sometimeswere seen individually, but often occurred as a smear from 30 to100 kDa. After day 2, the overall caseinolytic profile remained relativelyconstant although there was some variation in the intensity of thebands. Similar results were observed with burn fluid samples obtainedfrom three other patients.

It should be noted that the overall protein composition of all theburn fluid samples including blister fluid resembled plasma (Young andGrinnell, 1994). Also, a similar profile of caseinolytic enzyme activitywas detected in wound fluid collected from patients with venous stasisulcers (data not shown).

The elastase inhibitor MDL 27,367 blocks burn fluid caseinolyticactivity Because zymographs depend on running SDS-PAGE undernon-reducing conditions, the correspondence between molecular sizeand migration is undependable. Moreover, complexes between pro-teinases and their inhibitors often resist dissociation by SDS samplebuffer. Therefore, the caseinolytic profile shown in Fig 1a could haveresulted from one or several enzymes.

To learn more about the identity of the proteinase(s) responsible forcaseinolytic activity in burn wound fluid, experiments were carriedout using proteinase inhibitors. For these experiments, we analyzedburn fluid obtained on day 3. Preliminary results indicated thatinhibitors of trypsin-like and chymotrypsin-like serine proteinases,leupeptin, and chymostatin, had no effect on caseinolysis. As shownin Fig 2, however, most of the caseinolytic activity was blocked byMDL (100 µM), an inhibitor of human neutrophil elastase. In addition,a caseinolytic band at µ30 kDa (Fig 2, arrow) was found to be inhibitedboth by MDL and by aprotinin (0.2 mg per ml). Inhibition by MDLsuggested that the enzyme(s) responsible for caseinolytic activity inburn fluid might be elastase.

Burn fluid fibronectin appears to be intact Previously, it hasbeen shown that unregulated levels of elastase in wound fluid resultsin fibronectin degradation (Rao et al, 1995; Grinnell and Zhu, 1996).As shown in Fig 1b, however, no degradation of fibronectin wasobserved in the burn fluid samples. Fibronectin was absent from blisterfluid, presumably bound to degraded matrix (i.e., gelatin), but couldbe detected in time 0 burn fluid (µ6 h) postburn and reached maximallevels by 2–4 d. These results indicated that if elastase or an elastase-like enzyme was present in the wound fluid samples, the enzymeconcentration was probably too low to have a detectable effect onfibronectin.

Partial purification indicates that a single elastase-relatedenzyme accounts for most of the caseinolytic activity in burnfluid To determine whether one or several enzymes were responsible

Figure 5. The inhibitor SLPI selectively blocks the caseinolytic activityof human neutrophil elastase but not PR-3 or burn fluid. Samples ofburn fluid (75 µg) (BF, day 3), the peak fraction (10 µl) from DEAE HPLC(F), purified PR-3 or human neutrophil elastase (HNE) as indicated weresubjected to casein zymography. In (b), recombinant SLPI (4 µg per ml) wasadded to the reaction buffer.

for caseinolysis, samples of burn fluid were subjected to fractionationusing a combination of ammonium sulfate precipitation followed bygel filtration and ion-exchange HPLC. Caseinolytic activity was foundmostly in the 60% ammonium sulfate fraction. Gel filtration HPLC ofthis fraction typically resulted in two protein peaks of comparable size.Based on the column calibration, the first peak corresponded toµ165 kDa, whereas the second peak corresponded to µ55 kDa (datanot shown). As shown in Fig 3a, both peaks exhibited caseinolyticactivity although most of the activity (µ80%) was in the second peak.

Because caseinolytic activity was inhibited by the elastase inhibitorMDL (see Fig 2), the gel filtration HPLC fractions also were testedfor in vitro elastase activity using the elastase-specific substrate (methoxy-succinyl-ala-ala-pro-val-p-nitroanilide). Figure 3b shows that enzymeactivity could be detected after a 1 h assay period, but was associatedwith the minor peak of caseinolytic activity. In longer assays (2–3 d),however, a second peak of enzyme activity also became evident asshown in Fig 3c. The slower developing peak of in vitro enzymeactivity co-migrated with the majority of caseinolytic activity.

When the second peak from the gel filtration HPLC column, whichcontained most of the caseinolytic activity, was further fractionated byion-exchange HPLC, several protein peaks eluted from the column,but a single peak contained all of the caseinolytic activity and the slowdeveloping elastase-like activity (data not shown) (see Fig 5).

Biochemical and immunologic studies identify the elastase-related enzyme in burn fluid as PR-3 The above results wereconsistent with the idea that a single enzyme related to elastase wasresponsible for most of the caseinolytic activity in wound fluid. Wetherefore considered the possibility that this activity might correspondto PR-3. PR-3 is a 29 kDa, elastase-like serine proteinase that hasbeen reported to degrade methoxysuccinyl-ala-ala-pro-val-p-nitroanil-ide with a substantially lower catalytic efficiency than neutrophil elastase(Rao et al, 1991).

Figure 4a shows that on casein zymographs, purified PR-3 migratedas a smear of activity comparable with that observed with burn fluid,except the µ30 kDa band present in burn fluid (arrow) was absentfrom PR-3. Figure 4b shows that addition of MDL substantiallyinhibited the caseinolytic activity of PR-3.

Another distinction between elastase and PR-3 is sensitivity tohuman secretory leukoproteinase inhibitor (SLPI), which inhibitshuman neutrophil elastase but not PR-3 (Rao et al, 1991, 1993).Figure 5 compares the activity of PR-3, human neutrophil elastase(HNE), burn fluid (BF), and partially purified burn fluid caseinolyticenzyme (F) in the absence (Fig 5a) and presence (Fig 5b) of 4 µgrecombinant SLPI per ml. Unlike the multiple bands typical of PR-3,human neutrophil elastase migrated in casein zymography as twodiscrete bands, one of which was the µ30 kDa band (Fig 5a, arrowhead).Burn fluid contained both the smear and the µ30 kDa band, whereaspartially purified burn fluid caseinolytic enzyme contained only thesmear. In the presence of SLPI, human neutrophil elastase activity andthe µ30 kDa component in burn fluid were inhibited (Fig 5b), butnot the smear of caseinolytic activity associated with PR-3 or burn fluid.

Parallel studies carried out in vitro gave similar results as above. That

70 HE ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 6. SLPI inhibition of in vitro enzyme activity indicates that burn fluid contains a mixture of PR-3 and elastase. Samples of burn fluid (day 3)(100 µg) (a), PR-3 (200 ng) (b), or human neutrophil elastase (4 ng) (c) were incubated with the elastase substrate. Recombinant SLPI (0.8 µg per ml) was addedto the reaction buffer where indicated. Substrate degradation was determined at the time indicated by measuring OD410.

Figure 7. Most caseinolytic activity can be cleared from burn fluid byimmunoprecipitation with anti-PR-3 antibodies. Samples (500 µg) ofburn fluid (BF, day 3) were pretreated with protein G agarose and then thesupernatants were incubated with the monoclonal antibody against PR-3 (1:10,0.2 mg per ml) (ANCA) or control mouse IgG (1:10, 0.2 mg per ml) (IgG)followed by protein G agarose. Aliquots of the pellets from the pretreatment(P1) and antibody/IgG (P2) incubations and the supernatants (S) after antibody/IgG treatment were subjected to casein zymography.

is, SLPI completely inhibited degradation of the elastase substrate byhuman neutrophil elastase (Fig 6c), but had no effect on degradationof the elastase substrate by PR-3 (Fig 6b). As shown in Fig 6a, SLPIinhibited activity associated with burn fluid by about 50% suggestingthat burn fluid contained a mixture of PR-3 and elastase.

Finally, experiments were carried out using an antibody against PR-3. Figure 7 shows that most of the caseinolytic activity in burn fluidwas pelleted with anti-PR-3 monoclonal antibody ANCA 12.8 andprotein-G agarose (P2), whereas little activity pelleted with protein-Gagarose alone (P1) or with control IgG.

DISCUSSION

Palolahti et al. (1993) reported that chronic wound fluid containselevated levels of a caseinolytic activity although they did not identifythe enzyme(s) responsible for the activity. As shown in this study, acutewound fluid also contains caseinolytic activity, whose timing ofappearance resembles that of the metalloproteinase MMP-9 describedpreviously (Young and Grinnell, 1994).

In the current studies, we identify the enzyme responsible for themajority of the caseinolytic activity as PR-3. Evidence implicatingPR-3 as the cause of caseinolytic activity in wound fluid includedinhibition of activity by MDL but not SLPI, slow degradation of aspecific elastase substrate, and immunoprecipitation with a PR-3-specific monoclonal antibody. Moreover, the appearance of purified

PR-3 on casein zymography—a smeared band between 40 and100 kDa—resembled the SLPI-insensitive caseinolytic profile obtainedwith burn fluid. Why PR-3, a µ29 kDa monomer (Campanelli et al,1990), should run as a higher molecular weight complex in caseinzymography remains to be determined.

In addition to the caseinolytic activity attributed to PR-3, a secondcomponent on casein zymographs was evident at µ30 kDa. Thiscomponent migrated similarly as the smaller band of human neutrophilelastase activity. The µ30 kDa-associated activity was inhibited notonly by MDL, but also by SLPI. Moreover, the µ30 kDa componentwas absent from casein zymographs of the HPLC gel filtration/ionexchange peak of caseinolytic activity and was left behind when burnfluid was immunoprecipitated by antibody against PR-3. These resultsare consistent with the idea that the µ30 kDa proteinase correspondsto a different enzyme from PR-3, most likely human neutrophilelastase, although we have not tested this possibility directly. Onediscrepancy was the ability of aprotinin to inhibit the activity of the30 kDa band. Under routine assay conditions, most investigators havefound that aprotinin cannot inhibit elastase activity even thoughaprotinin can bind to elastase (Heck et al, 1985; Fioretti et al, 1993).Given the spatial and other constraints of the zymography assay,however, binding of aprotinin may be sufficient to result in inhibitionof enzyme activity.

PR-3 is one of the three neutral serine proteinases present in theazurophilic granules of neutrophils along with elastase and cathepsinG and is secreted by degranulation after activation of neutrophils(Baggiolini et al, 1979). Although the physiologic function of theenzyme is not well understood, PR-3 has been identified as a targetantigen in autoimmune diseases (Campanelli et al, 1990; Mrowka et al,1995) and may also play a role in the pathophysiology of emphysema(Kao et al, 1988). The latter activity likely involves the ability of PR-3 to degrade extracellular matrix molecules such as fibronectin, laminin,elastin, and collagen type IV, which has been demonstrated in vitro(Rao et al, 1991). In the burn wounds we were studying, however,there did not appear to be any degradation of fibronectin, indicatingthat the enzyme levels probably were too low to have a detectable effecton fibronectin. Whether PR-3 contributes to fibronectin degradation inchronic wounds remains to be determined.

The role of PR-3 in the wound environment may depend on theability of the enzyme to regulate chemokines and growth factors. PR-3 has been shown to play a role in processing of tumor necrosis factor-α (Robache-Gallea et al, 1995) and activation of transforming growthfactor-β (Csernok et al, 1996) and interleukin-8 (Padrines et al, 1994).Factors such as transforming growth factor-β and interleukin-8 provideimportant chemotactic and mitogenic signals for inflammatory andmesenchymal cells critical to wound repair (Moulin, 1995; Martin,1997), but little is known about the mechanisms by which enzymesprocess these molecules during the healing process.

This research was supported by NIH grant GM21681. We are indebted to Drs GaryPurdue and John Hunt and Ms. Kelly Spann for their assistance with the burn patients,to Anne-Marie Bashour and Meifang Zhu for their helpful advice, and to Dr. WilliamSnell for his helpful comments and suggestions.

VOL. 110, NO. 1 JANUARY 1998 PROTEINASE 3 IN WOUND FLUID 71

REFERENCES

Agren MS, Taplin CJ, Woessner Jr JF, Eaglstein WH, Mertz PM: Collagenase in woundhealing: Effect of wound age and type. J Invest Dermatol 99:709–714, 1992

Alper JC, Tibbetts LL, Sarazen AA: The in vitro response of fibroblasts to the fluid thataccumulates under a vapor- Permeable Membrane. J Invest Dermatology 84:513–515, 1985

Baggiolini M, Schnyder J, Bretz U, Dewald B, Ruch W: Cellular mechanisms of proteinaserelease from inflammatory cells and the degradation of extracellular proteins. CibaFdn Symp 75:105–121, 1979

Bullen EC, Longaker MT, Updike DL, Benton R, Ladin D, Hou Z, Howard EW: Tissueinhibitor of metalloproteinase-1 is decreased and activated gelatinases are increasedin chronic wounds. J Invest Dermatol 104:236–240, 1995

Campanelli D, Melchior M, Fu Y, Nakata M, Shuman H, Nathan C, Gabay JE: Cloningof cDNA for proteinase 3: A serine proteinase, antibiotic, and autoantigen fromhuman neutrophils. J Exp Med 172:1709–1715, 1990

Cooper DM, Yu EZ, Hennessey P, Ko F, Robson MC: Determination of endogenouscytokines in chronic wounds. Ann Surg 219:688–692, 1994

Csernok E, Szymkowiak CH, Mistry N, Daha MR, Gross WL, Kekow J: Transforminggrowth factor-beta expression and interaction with proteinase 3 in anti-neutrophilcytoplasmic antibody-associated vasculitis. Clin Exp Immunol 105:104–111, 1996

Falanga V, Grinnell F, Gilchrest B, Maddox YT, Moshell A: Workshop on the pathogenesisof chronic wounds. J Invest Dermatol 102:125–127, 1994

Fioretti E, Eleuteri AM, Angeletti M, Ascoli F: High- performanace liquid chromatrographicseparation of aprotinin-like inhibitors and their determination in very small amounts.J Chromatog 617:308–312, 1993

Grinnell F, Zhu M: Identification of neutrophil elastase as the proteinase in burn woundfluid responsible for degradation of fibronectin. J Invest Dermatol 103:155–161, 1994

Grinnell F, Zhu M: Fibronectin degradation in chronic wounds depends on the relativelevels of elastase, α1-proteinase inhibitor, and α2-macroglobulin. J Invest Dermatol106:335–341, 1996

Grinnell F, Ho CH, Wysocki A: Degradation of fibronectin and vitronectin in chronicwound fluid: Analysis by cell blotting, immunoblotting, and cell adhesion assays.J Invest Derm 98:410–416, 1992

Heck LW, Darby WL, Hunter FA, Bhown A, Miller EJ, Bennett JC: Isolation,characterization, and amino-terminal amino acid sequence analysis of humanneutrophil elastase from normal donors. Analyt Biochem 149:153–162, 1985

Kao RC, Wehner NG, Skubitz KM, Gray BH, Hoidal JR: Proteinase 3: A distinct humanpolymorphonuclear leukocyte proteinase that produces emphysema in hamsters.J Clin Invest 82:1963–1973, 1988

Kirsner RS, Katz MH, Eaglstein WH, Falanga V: The biology of wound fluid. Wounds5:122–128, 1993

Martin P: Wound healing – aiming for perfect skin regeneration. Science 276:75–81, 1997Matsubara M, Girard MT, Kublin CL, Cintron C, Fini ME: Differential roles for two

gelatinolytic enzymes of the matrix metalloproteinase family in remodelling cornea.Dev Biol 147:425–439, 1991

Matsuoka J, Grotendorst GR: Two peptides related to platelet-derived growth factor arepresent in human wound fluid. Proc Natl Acad Sci USA 86:4416–4420, 1989

Mehdi S, Angelastro MR, Burkhart JP, Koehl JR, Peet NP, Bey P: The inhibition ofhuman neutrophil elastase and cathepsin G by peptidyl, 1,2-dicarbonyl derivatives.Biochem Biophys Res Comm 166:595–600, 1990

Moulin V: Growth factors in skin wound healing. Eur J Cell Biol 68:1–7, 1995Mrowka C, Csernok E, Gross WL, Feucht HE, Bechtel U, Thoenes GH: Distribution of

the granulocyte serine proteinases proteinase 3 and elastase in humanglomerulonephritis. Am J Kidney Dis 25:253–261, 1995

Nakajima K, Powers JC, Ashe BM, Zimmerman M: Mapping the extended substratebinding site of cathepsin G and human leukocyte elastase. J Biol Chem 254:4027–4032, 1979

Padrines M, Wolf M, Walz A, Baggiolini M: Interleukin-8 processing by neutrophilelastase, cathepsin G and proteinase-3. FEBS Letters 352:231–235, 1994

Palolahti M, Lauharanta J, Stephens RW, Kuusela P, Vaheri A: Proteolytic activity in legulcer exudate. Exp Dermatol 2:29–37, 1993

Rao NV, Wehner NG, Marshall BC, Gray WR, Gray BH, Hoidal JR: Characterizationof proteinase-3, a neutrophil serine proteinase. J Biol Chem 266:9540–9548, 1991

Rao NV, Marshall BC, Gray BH, Hoidal JR: Interaction of secretory leukocyte proteaseinhibitor with proteinase-3. Am J Respir Cell Mol Biol 8:612–616, 1993

Rao CN, Ladin DA, Liu YY, Chilukuri C, Hou ZZ, Woodley DT: Alpha 1-antitrypsinis degraded and non-functional in chronic wounds but intact and functional in actuewounds: The inhibitor protects fibronectin from degradation by chronic woundfluid enzymes. J Invest Dermatol 105:572–578, 1995

Robache-Gallea S, Morand V, Bruneau JM, et al: In Vitro processing of human tumornecrosis factor-α. J Biol Chem 270:23688–23692, 1995

Stacey MC, Burnand KG, Mahmoud-Alexandroni M, Gaffney PJ, Bhogal BS: Tissue andurokinase plasminogen activators in the environs of venous and ischaemic leg ulcers.Br J Surg 80:596–599, 1993

Tarnuzzer RW, Schultz GS: Biochemical analysis of acute and chronic wound environments.Wound Rep Reg 4:321–325, 1996

Weckroth M, Vaheri A, Lauharanta J, Sorsa T, Konttinen YT: Matrix metalloproteinases,gelatinase and collagenase, in chronic leg ulcers. J Invest Dermatol 106:1119–1124, 1996

Wysocki AB, Grinnell F: Fibronectin profiles in normal and chronic wound fluid. LabInvest 63:825–831, 1990

Wysocki AB, Staiano-Coico L, Grinnell F: Wound fluid from chronic leg ulcers containselevated levels of metalloproteinases MMP-2 and MMP-9. J Invest Dermatol 101:64–68, 1993

Yager DR, Zhang l-Y, Liang H-X, Diegelmann RF, Cohen IK: Wound fluids from humanpressure ulcers contain elevated matrix metalloproteinase levels and activity comparedto surgical wound fluids. J Invest Dermatol 107:743–748, 1996

Young PK, Grinnell F: Metalloproteinase activation cascade after burn injury: a longitudinalanalysis of the human wound environment. J Invest Dermatol 103:660–664, 1994