Proteolytic activity in leg ulcer exudate

E.\p ncrmulnl 1993: 2: 29 }7 Prinlvd ill nciiiiwrk . .ill rixlil.\ n

ExpeHiitental Dermatology ISSN 0906-6705

Proteolytic activity in leg ulcer exudate Palolahti M, Lauharanta .1, Stephens RW, Kuuseht P, "Valieri A. Proteolytic activity iti leg ulcer exudate. Exp Dermatol 1993: 2: 29-37. © Munksgaard, 1993

Abstract: Twenty-Uve leg ulcer exudate samples from 17 patients with chronic non-healing venous leg ulcer were analyzed for proleolytic activity usitig radial caseinolysis procedures and zymographic analysis, and for fibronectin fragtnentation using imtnunoblotting technology. Caseinolytic activity was detected iti 21 of the 25 samples. A minority of thetn were inhibited (3 were totally, 6 partially inhibited) by aprotitiin, a serine proteinase inhibitor, suggesting thai proleinase(s) other than plasmin were also responsible for the caseinolysis. In zymographie analysis, 23 of the 25 samples showed positive reactions for enzyme activities cotnigratitig with plastniti and urokinase-type plasniinogen activator. Fibronectin fragtnentation, another sign of proteolytic activity, was seen in all but 2 ulcers. No correlation was seen between bacterial infection or inflammatory cells and the above parameters in the wound fluid. Acute wound lluid collected from the donor sites of patients undergoing split skin grafting was used as a control. In the control specimens no proteolytic activity was found during the days following operation. These results show that there is proteolytic activity in the chronic ulcer exudate and support the possibility that the proteolytic activity and consequent llbronectin fragmentation may be related lo the retarded epithelization and ulcer healing.

Introduction

It is generally accepted that leg ulcers are due to venous hypertension, which in turn is caused by chronic insufficiency of deep veins, cotntnunicating veins, or superficial veins of the leg (1). Various theories have been suggested for the succeeding microcirculatory changes (reviewed in 2-5), including theories in which cither the pericapillary deposition of fibrin as a diffusion baiTier (6), or the tole of white blood cells (7) is emphasized. According to the latter theory, activated white cells release proteolytic enzytnes and free radicals, which in turn result in tissue destruction. However, as the authot s mention, the real importance of the white cells in this respect is difficult to assess, since any inllammatory process caused by tissue destruction involves the action of white cells as well. Although several theories concerning the pathomechanisms involved in the formation of a leg ulcer have been proposed, there has been less discussion as to why an existing ulcer does not heal (8). We became interested in the possible role of proteolysis in delayed wound healing after the find-

Miina Palolahti', Jorma Lauharanta', Ross W. Stephens', Pentti Kuusela' and Antti Vaheri' 'Department of Virology and '^Depaittnent of Bacteriology and Immunology University of Helsinki: and ^Department ot Dermatology. University Central Hospital. Helsinki. Finland

Key words: bacteria - tibronectin - leg ulcer plasmin - plasminogen activator, proteinases urokinase - wound healing Dr. Miina Paloiahti. Department of Virology RO. Box 21 (Haartmaninkatu 3). SF-00014 University of Helsinki. Finland. Accepted for publication 22 March 1993

ing that elevated levels of proteolytic activity were found to be one of the main factors responsible for the persistence of chronic corneal ulceration (9). Analysis of tear fiuid specimens revealed hieh amounts of plasmin. a serine proteinase. and led to successful use of the proteinase inhibitor, aprotinin, in topical tteatment of corneal ulcers. These findings raised the question of whether excess proteolytic activity is also associated with therapyresistant leg ulcer, accounting for the inhibition of epithelization. In the present study, we examined proteolytic activity, particularly plastnin and plasniinogen activator activity, in leg ulcer exudate samples and have detnonstrated that proteolytic activity is also cotntnonly found in leg ulcer exudates but not in exudates frorn acute wounds. Patients and methods Patients

Seventeen patients (14 woman 67-93 years of age and 3 men 39-86 yr of age) with non-healing venous leg ulcets entered the study. Chronic venous 29

Palolahti et al. insufficiency was evident on the basis of clinical history showing thrornbosis of the deep veins of the leg or persistent edema of the legs associated with lipodermatosclerosis. Arterial disease was excluded in each case by Doppler ultrasound examination. Five patients had diabetes tnellitus. Two of them were insulin-dependent (OH, VK), 1 patient (MP) was treated with oral inedication, and 2 (MR, VH) with diet only. The duration of the present investigated ulcers varied from 2 months to more than 10 yr, the mean duration being about 2.5 yr (Table 1). Wound fluid exudate samples The exudate specimens were collected from the ulcers in the morning, before application of any topical treatment, using blunt glass microcapillaries. The first specirnen was usually taken on the 1st or 2nd day of the treatment period and in 5 patients two or three separate samples were collected on different days. In this phase the treatment consisted of soaks, rnechanical cleaning, nonenzymatic topical preparations, and various types of dressings. The amount of exudate varied from 20 to 100 1^1. A smear specimen of the exudate was rnade and stained after air-drying using May-Grunwald and Giemsa stains for analysis of the inflatnrnatory cells. The rest of the specimen was centrifuged and stored at — 20 C until analyzed for proteinases. Control samples Leg ulcer exudates were compared with exudates from 9 patients with an acute wound, a donor site Table 1. Patient data

Patient AM CB EA ER GB IS KK LT MP MR MS MV OH PL

SN VH VK

30

Sex

Ag e

Duration of present ulcer

F F F F F F F F F F F F M M F M F

82 68 75 67 83 78 78 81 70 93 75 80 62 39 77 86 69

3.5 years 3 years 8 years 1 year 2 years 8 months 2 montlis 5 montfis 2 years 3 years 4 years 2 months 1 year > 1 0 years 2 months 1.5 years 4 years

Size of ulcer (cm^) 64

of a split skin graft. The acute wound fluid was collected frotn underneath a setnipertneable polyurethane dressing (Op-Site®, Stnith & Nephew Ltd., Hull, Hertfordshire, England) using a syringe. The samples were collected daily beginning on the operation day until the exudation ceased or when the dressing detached (Patients 6 and 9). Tbe specitnens were stored at — 20 C after clarification by low-speed centrifugation. Radial easeinolysis proeedures The proteolytic activity of the exudate samples was detertnined by the radial caseinolysis procedure (10). In this assay, 5 |il of a 5 mg protein/ml dilution of each sample was applied in a well cut in an opaque casein-containing agarose gel, and allowed to diffuse. To prepare the gel, agarose was melted at a concentration of 1.2% in 0.1 M TrisHCl buffer, pH 8.0, and nonfat dry milk setved as the source of casein; a 15% (w/v) stock solution of milk powder was made in the above buffer and added to the agarose gel, so that the final concentration of milk powder in the test plate was 0.6%. The gels were incubated at +37"C for 48 h in a humified box. Lytic zones which developed around the wells indicated proteinase activity; tbe diatneters of the zones were measured and cotnpared with the lysis produced by human plasmin (20 casein units per mg; Kabi Vitrum, Stockholm, Sweden), which served as a standard. The results are expressed as micrograms of plastnin-like activity per tnl of wound fluid. Plasminogen activator (PA) activity was measured (10) using casein-agarose gels which, in addition, contained purified plastninogen (15 U/mg; Kabi, Stockholm, Sweden) at a concentration of 2 |.ig protein/ml. Urokinase (70000 IU/mg; Calbiochetn. La .lolla, CA) was used as a standard. The results are expressed as IU per milliliter of woutid fluid. It should be noted that, as plasminogen is added in the easein-agarose gel, this assay shows direct plasmin-like activity in addition to PA activity. True PA activity was thus obtained by subtracting direct plasmin-like activity from the PA activity seen on the agarose plate. The samples were also placed on a gel containing 100 KlU/ml of aprotinin (20000 KlU/ml; Trasylol®, Bayer, Leverkusen, Germany), a broad-spectrum inhibitor of serine proteinases, in order to determine if aprotinin inhibited the developtnent of lytic zones. Zytnography Zymographic analysis was performed to determine the tnolecular weights ofproteinases in the leg ulcer

Proteolytic activity in leg uleer exudate exudates (9). The samples were diluted in nonreducing satnple buffer (0.125 M Tris-HCl, 4% sodium dodeeyl sulfate (SDS), 20% glycerol, 0.002% bromophenol blue, pH 6.8) and electrophoresed in an 8'M) sodium dodeeyl sulfate polyacrylatnide gel (SDS-PAGE). The acrylatnide gel was placed on a casein-agarose gel (with and without plastninogen) similar to that used in the radial caseinolysis procedure and the results were recorded after 24-48 h of incubation at +37 C.

of 6 samples, totally in 3. In the other II. no inhibition could be detected. Zymography was used to detertnine the molecular size of the proteinase(s). The major lysis band could be seen comigrating with the plasmin standard (M, 85 000) in 23 of the 25 exudate specimens (Fig. 1). In PA zymograms, in which the casein detector gel also contained plasminogen, PA activity corresponding to the urokinase standard (M, 52 000) was found in the same 23 specimens (Fig. 2). Two of the satnples showed no lysis in either assay.

Fibroneetin fragmentation Fibroneetin fragmentation

Fragmentation of fibronectin in exudates was detected using an immunoblotting procedure (11). The samples were diluted in reducing satnple buffer (containing 10% mercaptoethanol). Proteins were first separated by 8% SDS-PAGE and transferred to nitrocellulose sheets by electrophoresis for 1.5 h at 340 mA. Fibronectin polypeptides were detected with horseradish peroxidase-labeled rabbit anti-human fibronectin IgG (Dako, Copenhagen, Denmark) after incubation for 2 h at +37"C and addition of 4-chloronaphthol, diaminobenzidine and peroxide.

The degree of fibronectin fragmentation in the satnples was studied using itnmunoblotting. In 19 of the 22 leg ulcer satnples. fibronectin was found to be fragmented (Fig. 3). The samples of 2 patients (patients SN and VH) showed no fragtnentation at all. In one of the latter (VH) no fibtonectin fragtnentation was detected in 3 separate samples collected on different days. However, proteolytic activity was seen in tnost of these fragtnentationnegative satnples in the caseinolytic assay and zymography (Table 2).

Baeterial analysis

Inflammatory eells

Bacteria were quantitatively cultivated from samples from 11 patients. Quantitation was done by plating 10 \i\ of prediluted satnples on duplicate chocolate and minimal nutrient agar plates, which were subsequently incubated overnight at +37 C. Colony forming units (cfu) per tnilliliter of the sample were calculated by counting the bacterial colonies on plates visually.

All the exudate samples contained varying atnounts of inflamtnatory cells. The majority of them proved to be polymorphonuclear leukocytes. Their proportion varied from 56 to 93% (mean 79) of total cells, whereas mononuclear cells ranged from 7 to AA% (mean 21). There was no correlation between the white cell count and the proteolytic activity of the samples.

Results Proteolytie aetivity

Baeterial growth

Altogether 25 exudate samples were collected from 17 leg ulcer patients and tested for total net plastnin-like activity and for PA activity using radial caseinolysis procedures with and without plasminogen (Table 2). Of the 25 samples, 21 showed net plasmin-like proteolytic activity in the procedure without plasminogen. In 11 of those, the lysis was only partial (i.e. the zones remained partially opalescent). Four specimens developed no lysis. Nineteen of the 25 specitnens had varying amounts of PA activity measured by the procedure with plasminogen, 6 had none (i.e. below the detection limit of 14 mlU per tnilligram of protein). When aprotinin was added to the agarose gel, the development of the lytic zones in the gels without plasminogen was partially inhibited in the case

There were no signs of a strong baeterial infection in the leg ulcer or in the surrounding tissue in any of the patients. However, bacterial growth was detected in 9 of the 11 exudate samples tested (Table 3). The negative result indicated that the actual bacterial number was lower than 2.5 x lO' efu/ml. In 4 satnples, more than one bacterial strain was found; in 3 of these, Pseudonionas aeruginosa was one of the organisms. Staphyloeoeeus aureus and S. epidermidis were isolated in 4 and 2 samples, respectively. There was no correlation between the bacterial strain or the nutnber of the bacteria and the caseinolytic activity in the sample. Control samples Acute wound fluid samples from 9 split skin graft donor sites were analyzed by using radial caseinoly31

Palolahti et al. Tabie 2. Sample data Zymography

Caseinolysis"

Patient AM CB EA ER

GB IS

Sample

Exudation'

1 II 1 II 1

+ ++ ++ + + +-I+++

II

+++

III 1 1

+

KK LT MP MR MS MV OH

Inhibifion with aprofinin

156*

0

+++

17 112 106 13 560

++

194* 240*

++ +

+++

+ 1 11

Direct piasminlike activity" |.ig/ml wound fluid

-t- + + +++ ++ ++ ++ ++

PL SN VH

1 II

+-I-

VK

III 1

++ ++

34 42* 62* 26* 82* 50* 97 400 504 12 143* 165*

67*

(%)

Piasminogen activator acfivity" iU/ml wound tluid

-

80 80 50 30 0 100 100 0

0.392 4.592 3.640 3.332 3.584 0.420 0.392 3.024 13.440 3.136 1.764

10 ND 0 0 0 50 100 0 0 0 0

2.254 3.612 6.720 4.900 2.268 8.680 17.640 0,840 2.226 2.562 6.272

0

-

FN Plasmin

ti-PA

Iragmenfation

+ + + +

+ ++ +++ ++

+ H+++

++ +

-I--I- +

+++

+++

ND

+ + +

++ ++++ +++

.f-1±

+ +

-l- + -t--t-

+ +

++ ++

+ +

±

+++

4-

+ + +

+ +++ ++

+

±

-l--H-t-

±

++ ±

++

ND

±

+++

+++ -

+++

± +

+ ++

±

ND ++ ND ..-1-

" The amount of exudation (+,+ +.+ + + ) was estimated by the same person (JL). " Caseinolysis was measured in ^g ot plasmin-like activity and in IU ot PA acfivity ' - = below the detection limit of 0.2 (ig plasmin-like activity and 14 mill PA activity per milligram ol protein. * Lysis zone opalescenf, ND=nof done (not enough sample).

sis procedut-es for plastnin-like activity and for PA activity, zymographic analysis for plasmin and PA activity, and fibroneetin itnmunoblotting to leveal fibronectin fragmentation. Six of the satnples showed minimal net plasminlike proteolytic activity in the samples collected on the operation day. The samples collected on the following days had no activity. In the caseinolytic assay to deteet PA activity, 4 samples taken on the operation day were positive and tbe remaining samples negative. None of these samples showed lysis comigrating with the plasmin standard in the zymographic analysis. However, PA activity corresponding to the u-PA standard was seen in all the satnples. A low level of fibronectin fragmentation was seen in sotne of the satnples (Table 4). Discussion

The present study shows that proteolytic activity is found in the exudate samples from patients with non-healing venous leg ulcer. Earlier, we detnonstrated plasmin in the tear fiuid of patients with 32

therapy-resistant eorneal ulcers (9) and plasmin and fibronectin fragtnentation in acute ocular allergen exposure (12) and chronic secretory otitis tnedia (13). These findings give support to the assertion that plasminogen activation, the most important form of extracellular proteolysis associated with various physiological cellular funetions such as cell migration and tissue retnodeling (for reviews, see 14,15,16), is also a general tnechanism in tissue-destructive processes (17-18). Enhanced proteolytic activity has also been associated with certain dermatological disorders, not all of which have tissue destruction as a dominant feature. These include pemphigus (19-20), bullous pemphigoid (21), psoriasis (22, 23, 24), atopic dermatitis (25), and alopecia areata (26). Indirect evidence for proteolytic activity in venous leg ulcers has been obtained in previous studies, which have shown degradation of the matrix proteins fibronectin and vitronectin in chronic venous ulcer Huids (27, 28). So far the identity of tbe proteinases responsible for this has been obscure (28).

Proteolytic activity in leg ulcer exudate

kD

Plasmin

1

2

94 — 69 —

46

skin graft donor sites which served as a model for an acute and well-healing wound. Some of these control specimens showed PA and/or plasmin-like activity on the day of operation but not on the succeeding days. We attribute these activities to the trauma of the operation. In addition, according to zymographic analysis, lysis comigrating with the u-PA standard was found but none was seen to comigrate with plasmin standard. The u-PA positivity in the zymographic analysis not accompanied by net positivity for u-PA in the caseinolytic assay may be related to the observations of the presence of the u-PA in the wound during normal uncomplicated reepithelization (29). The direct radial caseinolysis procedure measures, in addition to plasmin, all other proteinases that cleave casein (plasmin-like activity). It also should be noted that, in this very sensitive method.

u-PA

kD

30 — 94 Figure I. Zymographic analysis for dclcnninalion of molecular weights of direct proteolytic activity. The 2 samples (.^0 jil of 5 mg protein/ml dikition) from palient HA are seen in lane 1 and 2. Plasmin = human plasmin standard (250 ng). Molecular weights of marker proteins in kilodaltons are indicated on the left.

Our results showed that proteolytic activity was present in chronic venous leg ulcer while it was not found in acute, well-healing wounds during the days following operation. The conclusion is based on results obtained by the following techniques: radial caseinolysis. which determines direct plasmin-like activity, its inhibition by aprotinin, and plasminogen activator activity; zymographic analysis of plasmin and PA activity; and fibronectin immunoblotting. The results obtained with these different techniques do not fully correlate, due to the fact that they measure different aspects of proteolytic activity. The control specimens were collected from split

69

—

46

30 — Fi};iirc 2. Zymographie analysis for determination of molecular weights of plasniinogen activator. Samples (20 (.d ol" 5 me protcitt/ml dilulioii) were tlie same as analyzed in Fig. 1. tiPA = urokinase-type plasminogen acti\ator standard (16 m i l l ) . Molecular weights of marker proteins in kilodaltons are indicated on ihe lefl.

Palolahti et al. Table 3, Bacteria in the leg ulcer exudate

kD

FN

1

2

Patient

Sample

AM CB

1 II

EA

1

Bacteria Acinetobacter sp. Staphylococcus auieus Staphylococcus epidermidis Staphylococcus aureus Pseudomonas aerugiiwsa

II GB KK LT MR MS MV PL VK

Staphylococcus aureus Esherichia coli

Enterococcus faecalis Staphylococcus epidermidis Pseudomonas aeruginosa

Proteus mirabitis Enterobacter cloacae Pseudomonas aeruginosa Staphylococcus aureus

10=" cfu/ml 5 23 3.8 24 8.7