.4PMIS 105 25-30. 1997 C o n v r w h t 0 A P M I S I997 ., Y

Priiired I ~ I Denmork AN rights reserved LWW8

ISSN 0903-4641

Plasminogen activators and inhibitors in peritoneal tissue

LENA HOLMDAHL,’ MARTEN FALKENBERG,’ MARIE-LOUISE IVARSSON’ and BO RISBERG’

‘Department of Surgery, Ostra Hospital, University of Goteborg, Goteborg, and 2Department of Surgery, Malmo University Hospital, University of Lund, Lund, Sweden

Holmdahl L.. Falkenberg, M., Ivarsson, M.-L. & Risberg, B. Plasminogen activators and inhibitors in peritoneal tissue. APMIS 105: 25-30, 1997.

Serosal trauma elicits an inflammatory response which leads to the deposition of fibrin at injured sites, the residuals of which appear to be essential in excessive tissue repair and formation of intraabdominal adhesions. Local plasminogen activity may modulate this early phase of tissue repair. The present study was undertaken to investigate the distribution and cellular expression of plasminogen activators and their inhibitors in human peritoneal normal and inflamed tissue. Tissue-type plasminogen acti- vator (t-PA) was expressed in subserosal capillary walls, and in normal mesothelium, but not in inflammation. Immunoreactivity for the plasminogen activator inhibitor type 1 (PAI-1) was present in normal mesothelium, and substantially increased in inflammation, where, in addition, immunoreac- tivity was found throughout the submesothelial tissue. This PAI-1 was partly co-localized with macro- phages, as was the urokinase plasminogen activator (u-PA), suggesting an involvement of these cells in peritoneal tissue fibrinolysis. Inflammation or abrasion of the mesothelium during surgery is likely to cause a depletion of the local t-PA source and expose the potentially PAI-1 -containing submesothel- ial tissue, thus promoting persistence of fibrin and formation of adhesions.

Key words: Plasminogen activators; plasminogen activator inhibitors; peritoneum; wound healing.

Lena Holmdahl, Department of Surgery, Ostra Hospital, University of Goteborg, S-416 85 Goteborg. Sweden.

A major clinical problem related to serosal tissue repair is the formation of adhesions. The dominant cause of abdominal adhesions is prior surgical trauma, followed by infection (14, 23). It has been proposed that peritoneal insults in- duce the formation of adhesions through a cen- tral common pathway, in which tissue fi- brinolytic capacity is an important denomi- nator (1). Support for the hypothesis that fibrinolytic activity may have a pivotal role in the early formation of adhesions has recently been presented (6).

Several studies using the fibrin slide technique have demonstrated that the peritoneum pos-

Received February 26, 1996. Accepted October 5, 1996.

sesses fibrinolytic activity measured as plas- minogen activation (PA) (12, 13), and that PA is reduced by inflammation (20).

Likewise, evidence of an inhibitor of fibrino- lysis has been presented (15, 16), later con- firmed to be plasminogen activator inhibitor type 1 (PAI-1) (22).

The production of PAI-1 has been suggested to occur in the mesothelial cells and submeso- thelial capillary vascular endothelium (1 7), which was recently confirmed (24). Rabbit post- surgical peritoneal macrophages ( 10) and hu- man mesothelial cells (21) in culture release acti- vators and inhibitors of plasmin generation. However, little is known about the cellular dis- tribution and localization of the fibrinolytic components in peritoneal tissue in vivo, and hence the major cellular sources are as yet un-

25

HOLMDAHL et ( I / .

identified. The aim of the present study was to localize the expression of the individual com- ponents both in inflamed and normal human peritoneum.

MATERIALS AND METHODS

Tissue collection and preparation During laparotomy, human peritoneal tissue lined

by mesothelium was obtained from patients with per- itonitis (n=2, appendicitis), and from patients with- out any evidence of inflammatory reaction in the ab- dominal cavity (n= 3, one abdominal wall hernia, two colonic resections). The excised tissue was kept on ice and immediately embedded in OCT medium (Miles Inc., Diagnostics Division, Elkhart, IN, USA), snap- frozen in liquid nitrogen, and stored at -70°C until sectioned. Serial frozen sections were cut and mounted on slides coated with chrom alum gelatin. After overnight drying, the sections were fixed in ace- tone at room temperature for 10 min and stored at -70°C until further processed.

hzinunohistochemical procedure The immunohistochemical procedure followed a

protocol consisting of incubation with primary anti- bodies, secondary antibodies, a tertiary antibody complexed with alkaline phosphatase, addition of a chromogenic substrate, and subsequent counterstain- ing, as previously described (3). Briefly, the cryostat sections were brought to room temperature, and again fixed in acetone. Consecutive sections were in- cubated in a humid chamber for 1 h at room tem- perature with one of the following: mouse mono- clonal antibodies (IgG,) directed against human t-PA (PAM 3, Biopool, Ume5, Sweden, 1 : l O ) u-PA (MUK 1, Biopool, 1:20) PAI-1 (MA1 11 Biopool, 1:40), macrophages (CD 68, DAKO A/S, Glostrup, Denmark, 1:60), and as a negative control isotype - matched monoclonal antibodies against Aspergillus niger glucose oxidase (Mouse IgG, negative control, Code No. X 93 1, DAKO, 1 : 10). All dilutions were in phosphate-buffered saline (PBS) containing 5% bo- vine serum albumin (BSA). As an additional control sections were processed with the primary antibodies omitted. Following washing three times, the sections were incubated at room temperature with an anti- body directed against mouse IgG (Rabbit Anti- Mouse Immunoglobins, Code No. Z 412, DAKO, 1:40) for 30 min in a humid chamber, and washed with 0.05 M Tris-buffered saline, pH 7.6 (TBS). The secondary antibody was diluted in PBS containing 5% BSA and 10% human serum to prevent unspecific reactivity. The next step included incubation with al- kaline phosphatase and mouse monoclonal anti-alka- line phosphatase complexes (APAAP Mouse Mono-

clonal, Code No. D 651, DAKO, 1:40) in TBS for 30 min.

Specimens were thereafter washed with TBS before adding the chromogenic alkaline phosphatase sub- strate. The chromogenic substrate was prepared as follows: to 12 mg naphthol AS-MX phosphate (Sigma N 4875, Sigma, St Louis, MO, USA) 1.2 ml dimethylformamide was added, followed by 59 ml Tris buffer O.lM, pH 8.2, and 90 p1 1 M levamisole to inhibit staining of endogenous alkaline phosphatases. The substrate was prepared in advance and stored in aliquots at -20°C. Immediately before use, Fast Red TR (Sigma F-1500) was added. After 30 min of in- cubation, sections were again washed with TBS. Fi- nally, the sections were counterstained with hema- toxylin, and mounted.

RESULTS

Analysis of specimens revealed that t-PA was present in the mesothelium under normal con- ditions (Fig. 1A) and substantially reduced in inflammation. Expression of t-PA could not be detected in the submesothelial tissue, except for in the capillary vessel walls.

In non-inflamed peritoneum u-PA was ex- pressed in the mesothelial cell layer (Fig. lB), whereas in peritonitis it was found throughout the submesothelium and co-localized to a large extent with the immunoreactivity of macro- phages.

PAI-1 was present in both mesothelium and submesothelial capillary vascular walls in all samples (Fig. lC), but in peritonitis, partially co-localized with a positive staining for macro- phages (Fig. lD), the expression of PAI-1 was intensified and widely distributed in the subme- sothelial tissue (Fig. 1E).

Sections processed with antibodies serving as a negative control, or with omission of primary antibodies, did not show any immunoreactivity for the probed proteins (Fig. 1F).

The presence of fibrinolytic factors in normal and inflamed peritoneal tissue is summarized in Table 1.

DISCUSSION

This is, to our knowledge, the first study to loc- alize the individual components of the fi- brinolytic system in human peritoneal tissue under normal conditions and in peritonitis.

26

FIBRINOLYSIS IN PERITONEAL TISSUE

TABLE 1. E,xpression of plusniinogen activators and their inhibitor in normal and influmed peritoneal tissue. (+) denotes present, ( -) denotes not detectable

Factor Non -inflamed tissue Inflamed tissue Meso- Capillaries Macro- Submeso- Meso- Capillaries Macro- Submeso- thelium phages thelium thelium phages thelium

- - + - - t-PA + + U-PA + PAI-1 + +

+ + + + + + +

- - -

- -

Normal mesothelium expressed both t-PA, u- PA and PAI-1, but immunoreactivity for t-PA was reduced in inflammation. In peritonitis, ex- pression of u-PA and PAI-1 was also found in the submesothelial tissue, partially co-localized with immunoreactivity for macrophages.

Human peritoneum has been shown to pos- sess fibrinolytic activity, as indicated by plas- minogen activator activity (PAA) in peritoneal biopsies (12, 13) and gelatin peritoneal surface imprints (17) when using a fibrin slide tech- nique. Later, t-PA antigen was detected in per- itoneal homogenates (20, 22), and was demon- strated to account for 95% of the PAA in perito- neal tissue extracts (7).

An inhibitor of plasmin formation present in inflamed peritoneum has been demonstrated (15, 16). More recently, PAI-1 was identified in homogenates from inflamed peritoneal tissue, but not in normal samples (22), and proposed to cause the reduction in fibrinolytic capacity found during peritonitis (22). The gene ex- pression of PAI-1 was confined to mesothelial cells and submesothelial vessel walls in inflamed tissue (24).

The results of immunohistochemistry rely on the specificity of the antibodies. The antibodies were chosen because they have been used pre- viously (2, 4). Using paraffin-embedded perito- neal tissue and different antibodies (polyclonal and monoclonal), preliminary data from a sub- sequent study support the present findings on localization and cellular distribution of plas- minogen activators and inhibitors.

Findings from the present study partly op- pose the gene expression previously reported, by showing a widespread immunoreactivity for PAI-1 in the submesothelial tissue. One expla- nation may be that the immunostaining for PAI-1 was non-specific. However, since control experiments involving both omission of primary

antibodies and an irrelevant mouse antibody of the same subclass (IgG,) were negative, this is not likely. Another explanation is that the PAI- 1 gene expression was transitional in the in- flammatory process, and not present at the time of sampling. PAI-1 has upon synthesis been as- sociated with the extracellular matrix (9), where it is bound to vitronectin (25) and stored until used, which may explain the discrepancy in the expression of PAI-I mRNA and protein.

Macrophages have a multitude of functions in inflammation and tissue repair, and have been demonstrated to produce PAI-1 and u-PA in vitro (11, 18). In vivo, macrophages have shown immunoreactivity for u-PA (2, 19). Therefore, it is possible that the immunoreactiv- ity for PAI-1 and u-PA found co-localized with macrophages is at least partially expressed by macrophages. It is also possible that macro- phages release inflammatory cytokines, includ- ing tumor necrosis factor alpha and trans- forming growth factor beta, that induce a re- lease of PAI-1 from tissue fibroblasts in the vicinity (8). Furthermore, human fibroblasts may express potentially active u-PA (19), which in conjunction with polymorphonuclear neutro- phils ( 5 ) contribute to the immunoreactivity of

Although immunohistochemically present, PAI-1 has not been detectable in small samples of normal tissue (7, 22). In our own preliminary results with extracts from a larger amount of tissue, PAI-1 antigen and activity, and u-PA ac- tivity were detectable (unpublished obser- vation). In this preliminary study, the level of tPA in inflamed tissue was 0.11, uPA 1.9, and PAI-1 8.2 of that of non-inflamed tissue, which harmonizes with the present findings.

The lack of expression of t-PA in the subme- sothelial tissue suggests that the main function of t-PA in the peritoneum is to aid clearance of

27

U-PA.

HOLMDAHL et al.

Fig. 1. Frozen sections of human peritoneal tissue immunostained using the APAAP technique. Expression of t-PA (A, X20), u-PA (B. X20), and PAI-1 (C, X20) in the mesothelial lining and submesothelial tissue of non-inflamed peritoneum. In inflammation, partly co-localized with macrophages (4 X lo), there was an intense immunoreactivity of PAI-1 in the submesothelial tissue (E, X 10). Control incubation shown in panel F (X10).

fibrin deposits in the cavity proper. The acti- vation of plasminogen to plasmin in extracellu- lar matrix proteolysis is more likely attributed to u-PA, possibly regulated by tissue macro- phages. The role of u-PA in clearing intracavita- ry fibrin deposits is to be determined.

28

Based on the present study, we propose that damage and abrason of the mesothelium during surgery reduces the fibrinolytic capacity by de- leting the t-PA source, and by exposing the po- tentially PAI- l -containing submesothelial tis- sue, especially in peritonitis. This finding em-

FIBRINOLYSIS IN PERITONEAL TISSUE

phasizes the role of the mesothelium in the regulation of fibrin residues in the peritoneal cavity, and thus in the early formation of in- traabdominal adhesions.

This study was supported by grants from the Su,edish Medicul Research Council (Project 00660), University of Goteborg, the Swedish Society of Medicine. and the Gothenburg Medicul Society.

REFERENCES

1. Buckman, R. F., Woods, M., Surgent, L. & Ger- vin, A. S.: A unifying pathogenetic mechanism in the etiology of intraperitoneal adhesions. J. Surg. Res. 20: 1-5, 1976.

2. Buo, L., Lyberg, T., Jorgensen, L., Johansen, H. T. & Aasen, A. 0.: Location of plasminogen ac- tivator (PA) and PA inhibitor in human colorec- tal adenocarcinomas. APMIS 101: 235-241, 1993.

3. Cordell, J. L. , Fulini, B. , Erher, W. N . , Ghosh, A. K., Abduluzis, Z., MacDonald, S., Pulford, K. A . F., Stein, H. & Muson, D. Y.: Immunoen- zymatic labelling of monoclonal antibodies using immune complexes of alkaline phospha- tase and monoclonal anti-alkaline phosphatase (APAAP complexes). J. Histochem. Cytochem. 32: 219-229, 1984.

4. Fulkenberg, M. , Tjurnstrom, J. , Ortenwall, P.. Olausson, M. & Risberg, B.: Localization of fi- brinolytic activators and inhibitors in normal and atherosclerotic vessels. Thromb. Haemost.

5. Heipel, J. M . & Ossowski, L.: Human neutro- phi1 plasminogen activator is localized in speci- fic granules and is translocated to the cell sur- face by exocytosis. J. Exp. Med. 164: 826-840, 1986.

6. Holmdahl, L.. Al-Juhreen, M. & Risberg, B.: The role of fibrinolysis in the formation of

7.5: 933-938, 1996.

postoperative adhesions. Wound Rep. Reg. 7: 171-176, 1994.

7. Holmdahl, L., Eriksson, E. & Risberg, B.: Fi- brinolysis in human peritoneum during surgery. Surgery 119: 701-705, 1996.

8. Idell, S., Zwieb, C., Boggnrunr. J.. Holiduy, D.. Johnson, A. R. & Raghu, G.: Mechanisms of fibrin formation and lysis by human lung fibroblasts: influence of TGF-beta and TNF-al- pha. Am. J. Physiol. 263: L487-L4g4, 1992.

9. Knudsen, B. S., Hurpel. P. C. & Nuchman, R. L.: Plasminogen activator inhibitor is associated with the extracellular matrix of cultured bovine smooth muscle cells. J. Clin. Invest. 80: 1082- 1089, 1987.

10. Kuraoku, S.. Cumpeau, J. D., Rodgers, K. E., Nukumurn, R. M. & diZerega, G. S.: Effects of interleukin- 1 (IL- 1) on postsurgical macrophage secretion of protease and protease inhibitor ac- tivities. J. Surg. Res. 52: 71-78, 1992.

11. Lundgren, C. H. , Sawu, H., Sobel, B. E. & Fujii, S.: Modulation of expression of monocytei macrophage plasminogen activator activity and its implications for attenuation of vasculopathy. Circulation 90: 1927-1934, 1994.

12. Merlo, G., Fausone, G., Barbero, C. & Custug- nu, B.: Fibrinolytic activity of the human per- itoneum. Eur. Surg. Res. 12: 433-438, 1980.

13. Myhre-Jensen, O., Bergman Lursen, S. & Ast- rup, T.: Fibrinolytic activity in serosal and syn- ovial membranes. Arch. Pathol. Lab. Med. 88; 623-630, 1969.

14. Perry, J. F., Smith, A. & Yonehiro, E. G.: Intes- tinal obstruction caused by adhesions: A review of 388 cases. Ann. Surg. 135: 810-816, 1995.

15. Pugutch, E. M. J., Foster, E. A , , MucFarhne, D. E. & Poole, J. C. F.: The extraction and sep- aration of activators and inhibitors of fibrino- lysis from bovine endothelium and meso- thelium. Br. J. Haematol. 18: 669-681, 1970.

16. Pugutch, E. M. J. & Poole, J. C. F.: Inhibitor of fibrinolysis from mesothelium. Nature 221:

17. Rafterj'. A , : Method for measuring fibrinolytic activity in a single layer of cells. J. Clin. Pathol. 34: 625-629, 1981.

18. Sukselu, O., Hovi, T. & Vuheri, Z.: Urokinase- type plasminogen activator and its inhibitor se- creted by cultured human monocyte-macro- phages. J. Cell. Physiol. 122: 125-132, 1985.

19. Schufer, B. M . , Muier, K., EickholL U. , Todd, R. F. & Krumer, M. D.: Plasminogen activation in healing human wounds. Am. J. Pathol. 144:

20. Thompson, J. N., Puterson Brown, S., Hur- bourne, T., Whawell, S. A., Kulodiki, E. & Dud- ley, H. A. F.: Reduced human peritoneal plas- minogen activating activity: possible mechanism of adhesion formation. Br. J. Surg. 76: 382-384, 1989.

21. van Hinsbergh, V. W. M., Kooistru, T., Scheffer, M . A , , win Bockrl, J. H. & van Muijen, G. N . P. : Characterization and fibrinolytic properties of human omental tissue mesothelial cells. Comparison with endothelial cells. Blood 7.5:

22. Vipond, M. N., Whuwell. S. A., Thompson, J . N. & Dudley, H. A , : Peritoneal fibrinolytic ac- tivity and intra-abdominal adhesions. Lancet 335: 1120-1 122, 1990.

23. Weibel. M . & Mujno, A . G.: Peritoneal ad- hesions and their relation to abdominal surgery. A postmortem study. Am. J. Surg. 126: 345- 353, 1973.

269-270, 1969.

1269-1280, 1994.

1490-1497, 1990.

29

HOLMDAHL r / d

24. Wlwwell, S. A,, Wung, Y., Fleming, K. A . , Thompson, E. M. & Thompson, J. N.: Localiz- ation of plasminogen activator inhibitor-I pro- duction in inflamed appendix by in situ mRNA hybridization. J. Pathol. 169: 67-71, 1993.

25. Wimun, B., Almquist, A"., Sigurdurdottir, 0. & Linduhl, T.: Plasminogen activator inhibitor 1 (PAI) is bound to vitronectin in plasma. FEBS Lett. 242: 125-128, 1988.

30