Effects of tissue-type plasminogen activator (t-pa) on Streptococcus sanguis infected endocardial vegetations in vitro

Transcription

1 Journal of Antimicrobial Chemotherapy (1988) 21, Effects of tissue-type plasminogen activator (t-pa) on Streptococcus sanguis infected endocardial vegetations in vitro A. G. M. Baiting", J. Thompson', J. J. Erods*, H. Mittie', E. J. P. Brominer* and R. ran Fnrtfa" 'Department of Infectious Diseases, University Hospital, Rffnsburgerweg 10, 2300 RC Leiden; b Gaubius Institute TNO, Herenstraat 5d, 2313 AD Leiden, The Netherlands The effects were studied of in-vitro exposure to tissue-type plasminogen activator (t-pa) on endocardial vegetations isolated from rabbits with Streptococcus sanguis endocarditis as well as on Str. sanguis-'vsecxcd plasma clots. For both vegetations and clots t-pa induced lysis of the infected fibrin matrix, demonstrated by a release of streptococci into the incubation medium and by an increase in the concentration of fibrin degradation products. The dots decreased in weight, whereas only partial lysis of the vegetations was obtained. Addition of plasminogen to the incubation medium did not enhance the effect of t-pa on vegetations. Fibrinolysis by t-pa did not interfere with the antimicrobial action of benzylpenicillin on the bacteria in a fibrin matrix. It is concluded that t-pa can degrade the fibrin matrices of infected endocardial vegetations and plasma dots. Reduction of vegetation size by t-pa may prove to be useful in the treatment of bacterial endocarditis. Introduction Bacterial endocarditis requires prolonged administration of high doses of antibiotics to eradicate the endocardial infection. Reduction of part or all of the vegetation by fibrinolytic treatment might contribute to the treatment of bacterial endocarditis. Unlike more conventional fibrinolytic agents such as streptokinase or urokinase (Johnson, Dewar & Aheme, 1980; Dewar et al., 1986), tissue-type plasminogen activator (t-pa) offers the advantage of having a localized effect on fibrin deposits without inducing a generalized lytic state (Emeis et al., 1985; Verstraete & Collen, 1986a, b). As the first step in a study on the use of t-pa in experimental bacterial endocarditis the in-vitro effects of t-pa and of a combination of t-pa with cloxacillin were investigated on Staphylococcus epidermidis-infected plasma clots, which were used as a model of infected endocardial vegetations (Buiting et al., 1987). Exposure to t-pa had a lytic effect on the clots which could be assessed reliably by determination of the clot weight and the release of bacteria from the clot The antibacterial action of cloxacillin on the staphylococci in the non-lysed part of the dots was not affected by t-pa. The next step was to study the effects of t-pa on infected endocardial vegetations. Before doing this in the complex situation of an intact animal, the effect of incubation with t-pa was studied on vegetations that were isolated from the heart of rabbits with /88/ S03.00/ The British Society for Antimicrobial Chemotherapy

2 610 A. G. M. Baiting et al Streptococcus sanguis endocarditis. In parallel experiments, isolated endocardial vegetations and Str. sanguis-infected plasma clots were incubated in vitro with t-pa, and the effects of t-pa in both models were compared. Micro-organism Materials and methods Streptococcus sanguis (strain NCTC 7864) was used to infect rabbits and to prepare infected plasma clots (Meddens et al., 1984). For this strain the minimal inhibitory concentration (MIC) of benzylpenicillin was mg/1, as determined in Todd-Hewitt broth (Oxoid, London, England) in a twofold dilution series. The minimal bactericidal concentration (MBQ of benzylpenicillin was mg/1, as determined by serial dilution and plating of aliquots of the MIC dilutions. Before use, an overnight culture in Todd-Hewitt broth giving an average of 1-5 x 10 8 cfu/ml was washed three times with saline and diluted to the appropriate concentration. Tissue-type plasminogen activator Human two-chain melanoma-derived t-pa (2100 IU/ml in 0-01 M phosphate buffer, 0-5M NaCl, 001% (v/v) Tween 80) was a gift of Dr J. Verheijen (Kluft et al., 1983). Aliquots of this solution were stored at 70 C. Activity was determined with a spectrophotometric assay (Verheijen et al., 1982) and expressed in international units (Gaflhey & Curtis, 1985). One international unit (IU) represents approximately 2 ng of t-pa. Other materials Bovine thrombin (Merck & Dade, Dubingen, Switzerland) was dissolved in distilled water (80 NIH units/ml) and stored at -20 C. Human plasminogen (Kabi, Stockholm, Sweden) was dissolved in distilled water (12-5 casein units (CU)/ml) and stored at -20 C. Human plasmin (Kabi) was dissolved in distilled water (20 casein units (CU)/ml) and stored at - 70 C. Sodium benzylpenicillin (Gist Brocades, Delft, The Netherlands) was dissolved in 0-9% NaCl. Plasma used for incubation of endocardial vegetations was prepared from rabbit blood anticoagulated with sodium citrate (final concentration: 13 mmol/1). The blood was centrifuged first for 15 min at 1000 g (room temperature) and then for 30 min at 20,000 g (4 Q. The plasma was stored at - 70 C. Plasma used to prepare infected clots was obtained from citrated blood by centrifugation for 15 min at 1000 g. Endocardial vegetations Infected endocardial vegetations were derived from male New Zealand White rabbits (Enki Farms, Someren, The Netherlands) weighing kg in which endocarditis had been induced as described (Durack & Beeson, 1972). In short, a polyethylene catheter was placed in the left ventricle via the left carotid artery. After 48 h, 1 ml of a

3 In-ritro effects of plaaminogai activator 611 suspension containing 10 8 live streptococci in saline was injected into a marginal ear vein. This led to infection in all rabbits, as shown by a positive blood culture 24 h after inoculation. Two days after inoculation the animals were killed by intravenous injection of sodium pentobarbital. Endocardial vegetations were harvested with sterile instruments. Microscopy of sections of vegetations stained with Mayer's haematoxyllin confirmed that adherent cardiac tissue had been removed. The isolated vegetations were weighed and either used immediately or rapidly frozen at 70 C. Concentration of fibrin degradation products The concentration of fibrin degradation products (fdp) was determined by an ELISA technique. Microtiterplates (96 wells, Dynatech ) were coated with 1 /ig/well of sheep anti-rabbit fibrinogen IgG (Cooper Biomedical, Malvern, PA) in PBS. After incubation with the diluted samples of the incubation medium, the plates were washed and incubated with peroxidase-conjugated sheep anti-rabbit fibrinogen IgG (Cooper). Finally, tetramethyl-benzidine was added for colour development. The reaction was stopped with 4-5 M H 2 SO 4 and absorption was measured at 450 nm with a Titertek multiscan photometer. The fdp concentration was assessed with the use of a normal pool rabbit serum dilution curve as standard. For the standard curve spectrophotometric extinction was log-linearly proportional to the dilution. Fdp concentrations in the medium were expressed as a percentage of the concentration of the pooled serum. All reactions were performed at 20 C. Quantitative bacteriology Bacteria were counted as colony forming units (cfu). Counting was performed in incubation media as well as homogenates of vegetations and clots by placing six 0-01 ml aliquots of a series of tenfold dilutions on sheep blood agar. The plates were incubated for at least 24 h at 37 C. To compensate for differences in the initial number of bacteria per ml of the incubation medium of vegetations or infected plasma clots, effects on the numbers of bacteria were expressed as the difference between log cfu/ml at time point x and log cfu/ml at zero time. A value greater than 0 for this difference represents a release into the medium of streptococci from the vegetations or from the infected plasma clots. For blood cultures, 1 ml samples were diluted with 1 ml Liquoid (Oxoid), inoculated into 5 ml glucose agar broth, and incubated for at least 24 h at 37 C. Growth o/str. sanguis in incubation media The growth of Str. sanguis in Todd-Hewitt broth, plasma, and 001 M phosphatebuffered saline (ph 7-40) (PBS) was measured by inoculating 10 ml aliquots of these media with approximately 1 x 10 6 bacteria in the logarithmic growth phase. During incubation for 180 min at 37 C in a shaking water bath, 0-2 ml samples were taken at 45 min intervals for cfu counting. The effect of benzylpenicillin on bacterial growth was assessed by adding this antibiotic to the medium at zero time in concentrations up to 2 mg/1. For measurement of the combined effect of benzylpenicillin and t-pa, benzylpenicillin was added to the incubation medium to a concentration of 0, 01, or 04mg/1 and t-pa to 0 or 60 IU/ml at zero time.

4 612 A. G. M. Baiting et ol Incubation of endocardia! vegetations Endocardial vegetations were divided with a sterile scalpel into parts weighing approximately 10 mg. These were placed randomly in sterile plastic tubes (120 x 17 mm, Falcon Plastics, Becton & Dickinson, Oxnard, CA) containing 2-5 ml incubation medium which was either platelet-free rabbit plasma or PBS. The tubes were closed with a cotton wool stopper and placed in a shaking 37 C water bath. At zero time and at several time-points up to 180 min (in some experiments up to 18 h), cfu were counted in 0-2 ml aliquots of the medium. The concentration of fdp was determined for vegetations incubated in PBS after 180 min. t-pa was added to the incubation medium at zero time in concentrations up to 480 IU/ml. After incubation the contents of the tubes were examined visually. Thereafter the tubes were centrifuged at 100 g for 10 min (4 Q. The pellets containing the remnants of the vegetations were homogenized in 2 ml saline in a Potter Elvejhem homogenizer after which cfu in the homogenate were counted. To assess the effect of benzylpenicillin on the streptococci in the vegetations, the antibiotic was added at 90 min to the incubation medium in a concentration of 0-1 or 0-4 mg/1. Effects of t-pa on infected plasma clots Infected plasma clots containing about 10 5 streptococci were prepared as previously described (Buiting et al., 1987, Bryant; 1979). In brief, 0-9 ml titrated rabbit plasma was mixed with 0-1 ml of a suspension of Str. sanguis and clotted with 1-2 NTH units of thrombin. Each clot was placed in a sterile plastic tube containing 2-5 ml PBS, closed with a cotton wool stopper, and placed in a 37 C shaking water bath. t-pa was added at zero time in concentrations up to 240 IU/ml. At zero time and at several time-points thereafter cfu were counted in 0-2 ml aliquots. At chosen timepoints the incubation medium was aspirated and the clots were weighed and homogenized in 2 ml saline in a Potter Eveljehm homogenizer, after which cfu were counted in the homogenates. The concentration of fdp in the medium was determined for clots incubated in PBS after 180 min. Effect of t-pa and benzylpenicillin on bacterial growth in infected plasma clots The effect of benzylpenicillin on growth of the bacteria in clots was determined by incubation of clots in Todd-Hewitt broth (Oxoid) to which benzylpenicillin had been added in concentrations up to 4 mg/1. Cfu were counted in homogenates of the clots and in the incubation medium. For determination of the combined effect of benzylpenicillin and t-pa, the latter was added to the incubation medium in concentrations of 0 or 60 IU/ml at zero time and benzylpenicillin was added in concentrations of 0, 0-1, or 0-4 mg/1 at 90 min. Statistical analysis Statistical analysis was done by a two-tailed Student's f-test, multiple regression analysis, or analysis of variance where appropriate. P values higher than 005 were considered non-significant. Values given in the text represent the mean and the standard deviation of at least four experiments.

5 In-vitro cffecte of ^"""I'l^fii actirator 613 Results Effect of t-pa and penicillin on bacterial growth in incubation media When an inoculum of Sir. sanguls in log-phase growth was added to either plasma or PBS, bacterial growth stopped, but the number of bacteria per ml did not decrease during a period of 180 min. Thus, an increase in the number of bacteria in plasma or PBS in which vegetations or infected plasma clots are incubated represents a release of bacteria from the infected fibrin matrix. When log-phase proliferating streptococci were inoculated in Todd-Hewitt broth, logarithmic growth was observed for 180 min. Addition of benzylpenicillin to the Todd-Hewitt broth caused a concentration-dependent decrease of the growth of the streptococci. To determine the combined effect of t-pa and benzylpenicillin and of each compound alone multiple regression analysis was performed of the log numbers of cfu/ml (N) with time of incubation in hours (/), t-pa concentration, and the interaction between time and benzylpenicillin concentration in mg/1 (Q as independent variables. t-pa had no significant influence on the growth of the streptococci, whereas time (P < 0-001) and the concentration of benzylpenicillin (P < 0-005) had a significant effect on the growth of the streptococci (regression equation: N = ( x C x t). In other words, the growth rate of the streptococci decreases significantly with increasing benzylpenicillin concentration, and t-pa does not influence the growth rate of the streptococci or the effect of benzylpenicillin on the growth rate of streptococci. Effects of t-pa on isolated endocardia} vegetations Inspection of tubes containing vegetations exposed to t-pa for 180 min showed an opalescent medium containing small fragments of vegetations. When no t-pa was added the medium remained clear. The non-lysed part of the vegetations after exposure to t-pa appeared to be flabbier and more frayed than vegetations incubated without t-pa. Vegetations incubated in plasma adhered firmly to the plastic surface of the tubes for at least 18 h. Addition of t-pa released the vegetations from the tube wall. No adherence of the vegetations was seen when buffer was used as medium. A decrease in weight during incubation could not be used to assess the effect of t-pa on the endocardia] vegetations, because of the influence of t-pa on their consistency and because adhering medium introduced a considerable error due to the low initial weight of the vegetations. Incubation of vegetations in PBS with t-pa for 180 min led to release of streptococci from the vegetations into the medium, as shown by a significant increase in the log numbers of cfu/ml medium relative to the increase in medium without t-pa {P < 0001 by analysis of variance). Release of bacteria was observed at a concentration as low as 15IU t-pa/ml (results not shown) and increased with increasing concentrations until a maximum was reached at a concentration of 120 IU t-pa/ml. Most of the release occurred during the first 90 min of exposure to t-pa (Figure l(a)). More frequent determination of the log numbers of cfu during the first 45 min of incubation (at time points 0, 5, 15, 30, and 45 min) revealed that during this period the log numbers of cfu/ml increased almost linearly and that the rate of the release depended significantly on the t-pa concentration (P < 0-001) (Table I). The relationship between log cfu/ml

6 614 A. G. M. Baiting et td. Table L Release of streptococci from vegetations during exposure to t-pa for 45 min Medium Intercept a Regression coefficients * c time (h) t-pa x time (IU/ml xh) PBS plasma n.s (P < 0-001) (P < 0-001) (P < 0001) and t-pa is defined by the following equation: log cfu/ml at time x log cfu/ml at time 0 = a+bx time (h) + C x t-pa (IU/ml) x time (h); where a is the intercept and b and c are the regression coefficients whose values are given in Table I. After the first 45 min the increase levelled off and after 180 min no further increase was observed. Incubation of vegetations without t-pa in plasma instead of PBS led to a small but significant (P < 0-001) release of bacteria (Table I). However, with the same concentration of t-pa the release of bacteria was similar in plasma and PBS with respect to magnitude and time course (Table I). The maximum effect of t-pa was roughly the same for both media Incubation time (min) Figure 1. Release of streptococci from endocardia! vegetations (a) and infected plasma clots (b) exposed to t-pa. Data are expressed as increase in log cfu/ml in the incubation medium (PBS). t-pa was added to the medium at I - 0. Values are means of at least four experiments. D, 240 IU/ml; A, 60 IU/ml; x, 30 IU/ml; O, control.

7 In-ritro effects of ptasminogen actirator t-pa (IU/ml) Figure 2. Effect of t-pa on the concentration of fibrin degradation products (fdp) in the medium of endocardia] vegetations incubated in PBS to which t-pa was added at t = 0. The concentration of fdp was determined after 180 min and was expressed as a percentage of the concentration in pooled serum of normal rabbits. Values are the mean and SJ>. of at least three experiments. Vegetations were not lysed completely, even after incubation for 18 h with 480 IU/ml of t-pa in plasma or buffer. Incubation with 120 and 240 IU t-pa/ml for 180 min resulted in the release of about 20% of the bacteria in both media. As a result of fibrin degradation, the concentration of fdp in the incubation medium of endocardial vegetations after exposure for 180 min to 60 or 240 IU t-pa/ml was significantly higher (P< 0002 by analysis of variance) than that of vegetations not exposed to t-pa. At 240 IU t-pa/ml the concentration of fdp in the medium was five times (5-05 ±0-68) that of controls (Figure 2). Exposure for 180 min to benzylpenicillin added to the medium at 90 min had no significant effect on the bacterial numbers in the vegetations. Neither did the combination of t-pa and benzylpenicillin have a significant influence on the bacterial numbers of vegetations (Table IT). 240 Table U. Effect of t-pa (added at t = 0) and benzylpenicillin (added at t = 90 min) on the numbers of streptococci in vegetations after incubation for 270 min. in PBS. Numbers indicate mean log cfu per vegetation from at least four vegetations (s.d. in parenthesis) t-pa (IU/ml) 0 Benzylpenicillin (mg/1) (0-89) 6-66 (1-27) 6-86 (1-38) 6-27 (1-27) 6-37 (1-59) 7-14 (0-6)

8 616 A. G. M. Baiting tt al Incubation tlmt (min) Flgnre 3. Release of streptococci from endocardia! vegetations during exposure to t-pa (O. A, D) or t-pa plus plasminogen (, A> ) Plasminogen and t-pa were added to the incubation medium (PBS) at t» 0 in concentrations of 0 or 1 CU/ml and 0, 60, or 240 IU/ml, respectively. Values are the means of at least three experiments. Influence of plasminogen and plasmin on the lysis of endocardia! vegetations To determine whether the effect of t-pa on the endocardial vegetations was limited by the content of plasminogen in the medium, 1 CU human plasminogen/ml was added to the incubation medium (PBS or plasma) containing 0, 60, or 240IU t-pa/ml. The effect of plasminogen and t-pa on the release of the streptococci was compared with the effect of t-pa alone. The addition of plasminogen to PBS did hot have a significant effect on the t-painduced release of bacteria from the vegetations (Figure 3), nor did this lead to complete lysis of the vegetations. Addition of plasminogen to plasma also had no additional lytic effect (results not shown). A control experiment showed that the human plasminogen used together with t-pa could lyse rabbit plasminogen-free fibrin (results not shown). To determine whether, complete lysis could be obtained with plasmin, vegetations were incubated for 180 min without t-pa in PBS containing 0-5 CU/ml plasmin. This led to the release of about 20% of the bacteria from the vegetations into the medium after 180 min of incubation, which is similar to the release of bacteria obtained with t-pa. However, plasmin did not give complete lysis of the vegetations either. 180 Effects of t-pa and benzylpenicillin on infected plasma clots The weight of infected plasma clots incubated in PBS in the absence of t-pa did not change over 180min (weight at t 0 = 0-717±0-057g; weight at f, 80 = ±0054 g; n.s., Student's /-test). Addition of 60 IU t-pa per ml medium gave a significant decrease of clot weight to about 80% of the weight of controls in 180 min (weight at 'iso = ±0-089 g; P < 0001, Student's t-test). Most of this effect was obtained in 90 min.

9 In-Yttro effects of pbaminogen activator 617 Incubation of the infected dots in PBS with t-pa led to a release of streptococci from the clots into the medium, as shown by a significant increase in the log numbers of cfu/ml medium as compared to the increase in the medium of clots not exposed to t-pa (P<0001 by analysis of variance) (Figure l(b)). Most of the effect occurred during the first 45 min. At 240IU t-pa/ml, about 30% of the bacteria were released from the dot into the medium after 180 min. The release of streptococci from infected plasma dots and vegetations exposed to t-pa was similar with respect to the time course and the magnitude of the release (Figure l(a), (b)). Incubation of infected plasma clots in Todd-Hewitt broth with t-pa also led to release of bacteria from the clots (results not shown). The concentration of fdp in the incubation medium was higher for dots exposed to t-pa than for dots incubated without t-pa. After 180 min of exposure to 240 IU t-pa/ml in PBS, the concentration was about three times higher than that of controls not exposed to t-pa. The number of streptococci in dots incubated in Todd-Hewitt broth increased logarithmically after a lag phase of about 90 min, from 5-24 ±0-22 to 7-23 ±0-20 log cfu/g during a period of 180 min. Addition of benzylpenicillin after 90 min of incubation resulted in a concentration-dependent decrease of the bacterial numbers per gram dot weight. To determine the combined effect of t-pa and benzylpenicillin, and of each compound alone multiple regression analysis was performed of the log numbers of cfu/g (N) with time of incubation in hours (/), t-pa concentration, and the interaction between time and benzylpenicillin concentration in mg/1 (Q as independent variables. No significant influence of t-pa was found, but the results showed a significant influence of time (P < 0-001) and of the concentration of benzylpenicillin {P < 0-001) on the growth of the streptococci in the clots (regression equation: N= 5-93+( xCx/). Table in shows the bacterial numbers per g of clot weight after 270 min of incubation. In other words, the growth rate of streptococri in infected plasma dots decreases significantly with increasing benzylpenicillin concentrations, and t-pa does not influence the growth rate of the streptococci or the effect of benzylpenicillin on the growth rate, as was also found for the vegetations. Discussion The main conclusion drawn from this study is that in-vitro exposure to tissue-type plasminogen activator results in partial lysis of infected endocardial vegetations. This lysis was not enhanced by the addition of plasminogen to the incubation medium. For the infected plasma clots, lysis as a result of exposure to t-pa was shown by a decrease in the weight of the clots and a release of bacteria into the incubation medium, which Table IIL Effect of t-pa (added at / - 0) and benzylpenicillin (added at I = 90) on the numbers of streptococci in infected plasma dots after incubation for 270 min in Todd-Hewitt broth. Numbers indicate mean log cfu/g of clot weight from at least five dots (S.D. in parenthesis) t-pa (IU/ml) 0 Benzylpenicillin (mg/1) (015) 7 30 (0-22) 5-78 (014) 6-07(0-25) 4 10 (0-17) 4 20(0-17)

10 618 A. G. M. Batting et al have been found to be reliable parameters for the lytic effect of t-pa (Buiting et al., 1987), and by an increase in the concentration of fibrin degradation products (fdp) in the medium. Similarly, lysis of vegetations as a result of exposure to t-pa was shown by a release of bacteria and an increase in the concentration of fdp in the medium. For both clots and vegetations, the lytic effect was obtained mainly during the first 90 min of exposure to t-pa. The lysis of vegetations obtained by exposure to t-pa was similar for incubation in PBS and plasma. The only difference was the small release of bacteria into plasma in the absence of t-pa, which may probably be attributed to a lytic activity already present in plasma. Because the lytic effect obtained with t-pa in PBS without plasminogen was similar to that with t-pa in plasma, it is highly probable that the t-pa-induced lysis is mediated by plasminogen already incorporated during the formation of the vegetation or clot. Despite high concentrations of t-pa and prolonged incubation, the endocardia] vegetations were not completely lysed, perhaps because the vegetations are not fully degradable by the fibrinolytic system or because a sufficient amount of plasminogen is not present at the right site. The former possibility is supported by the finding that even plasmin could not completely lyse the vegetations. The latter is contradicted by the absence of an effect of plasminogen added to the incubation medium and the similarity of the maximum effect of t-pa in plasma and in buffer, indicating that the effect of t-pa is not limited by an absence of plasminogen outside the vegetation. These findings are of particular interest in relation to the in-vivo situation. Resistance of part of the vegetation to fibrinolytic degradation would limit the value of t-pa treatment and could at least partially explain the inability of the body to eliminate infected endocardia! vegetations. The adherence of the vegetations to a plastic surface during incubation in plasma could be due to fibrin formation induced by procoagulant activity present on the vegetational surface. This is concluded from the finding that t-pa could undo this adherence and that adherence did not occur when the vegetations were incubated in buffer lacking fibrinogen. Several studies have pointed to the important role of procoagulant activity as one of the mechanisms responsible for the persistence of the endocardial vegetations (van Ginkel et al., 1979; Drake, Rodgers & Sande, 1984). Fibrin formation induced by procoagulant activity might also counteract the effect of t-pa in vivo. The addition of benzylpenicillin to broth in which infected plasma clots had been incubated led to a decrease of the bacterial numbers per gram clot weight. t-pa did not influence the effect of benzylpenicillin on bacteria in the non-lysed part of infected clots, as had been found earlier for Slaph. epidermidis and cloxacillin (Buiting et al., 1987). No effect of benzylpenicillin on numbers of streptococci in the vegetations incubated in vitro was seen. The difference in the effect of benzylpenicillin can be explained by the fact that the bacteria in the vegetations are in a resting state, whereas those in the clots are in log-phase growth. The implications of the present results for the treatment of bacterial endocarditis with t-pa in vivo can be only speculative for the time being. Although t-pa can lyse an infected vegetation to a certain extent, complete lysis may not be reached. It is not certain that a reduction of the size of the endocardia] vegetation will enhance sterilization of the vegetation by antimicrobial treatment (Lutas et al., 1986; Robbins et al., 1986). Furthermore, some possible adverse effects of lysis of a vegetation can be

11 Io-vitro effects of pumlnogen activator 619 anticipated. The in-vitro findings suggest that lysis could release bacteria and parts of the vegetation into the circulation. Such release would lead to septic or embolic complications, for instance in the central nervous system. More information on these aspects and on the place of fibrinolysis by t-pa in the treatment of bacterial endocarditis could be obtained from studies on t-pa treatment performed in the rabbit model of bacterial endocarditis. Acknowledgements This study was supported by the Netherlands Heart Foundation (Grant ). The technical assistance of W. Schmal-Bauer, M. Welling, and D. van der Keur is gratefully acknowledged. References Bryant, R. E. (1979). Antibiotic effectiveness against Staphylococcus aureus in fibrin clots and pus: model for evaluating antibiotic therapy of endocarditis. In Infectious Diseases: Current Topics, Vol. I (Gilbert, D. N. & Sanford, J. P., Eds), pp Grune & Stratton, New York. Buiting, A. G. M., Thompson, J., Emeis, J. J., Mattie, H., Brommer, E. J. P. & van Furth, R. (1987). Effects of tissue-type plasminogen activator on Staphylococcus epidermidis-wfected plasma clots as a model of infected endocardia! vegetations. Journal of Antimicrobial Chemotherapy 19, Dewar, H. A., Jones, M. R., Barnes, W. S. F. & Griffin, S. G. (1986). Fibrinolytic therapy in bacterial endocarditis: experimental studies in dogs. European Heart Journal 7, Drake, T. A., Rodgers, G. M. & Sande, M. A. (1984). Tissue factor is a major stimulus for vegetation formation in enterococcal endocarditis in rabbits. Journal of Clinical Investigation 73, Durack, D. T. & Beeson, P. B. (1972). Experimental bacterial endocarditis. I. Colonization of a sterile vegetation. British Journal of Experimental Pathology 53, Emeis, J. J., Brommcr, E. J. P., Kluft, C. & Brakman, P. (1985). Progress in fibrinolysis. In Recent Advances in Blood Coagulation, Vol. 4 (Polkr, L., Ed.), pp Churchill Livingstone, Edinburgh. Gaffney, P. J. & Curtis, A. D. (1985). A collaborative study of a proposed international standard for tissue plasminogen activator (t-pa). Thrombosis and Hemostasis 53, Johnson, C. E., Dewar, H. A. & Aherne, W. A. (1980). Fibrinolytic therapy in subacute bacterial endocarditis: an experimental study. Cardiovascular Research 14, Kluft, C, van Wezel, A. L., van der Velden, C. A. M., Emeis, J. J., Verheyen, J. H. & Wijngaards, G. (1983). Large-scale production of extrinsic (tissue-type) plasminogen activator from human melanoma cells. In Advances in Biotechnological Processes, Vol. 2 (Mizrahi, A & van Wezcl, A. L., Eds), pp Alan R. Iiss, New York. Lutas, E. M., Roberts, R. B., Devereux, R. B. & Prieto, L. M. (1986). Relation between the presence of echocardiographic vegetations and the complication rate in infective endocarditis. American Heart Journal 112, Meddens, M. J. M., Thompson, J., Mattie, H. & van Furth, R. (1984). Role of granulocytes in the prevention and therapy of experimental Streptococcus sanguis endocarditis in rabbits. Antimicrobial Agents and Chemotherapy 25, Robbins, M. J., Sociro, R., Frishman, W. H. & Strom, J. A. (1986). Right-sided valvular endocarditis: etiology, diagnosis and an approach to therapy. American Heart Journal 111, van Ginkel, C. J. W., Th6rig, L., Thompson, J. Oh, J. I. H. & van Aken, W. G. (1979). Enhancement of generation of monocyte tissue thromboplastin by bacterial phagocytosis: possible pathway for fibrin formation on infected vegetations in bacterial endocarditis. Infection and Immunity 25,

12 620 A. G. M. Butting et al Verbeijen, J. H., Mullaart, E., Chang, G. T. G., Kluft, C. & Wijngaards, G. (1982). A simple sensitive spectrophotomctric assay for extrinsic (tissue-type) plasminogen activator applicable to measurements in plasma. Thrombosis and Hemostasis 48, Veretracte, M. & Collen, D. (1986a). Thrombolytic therapy in the eighties. Blood 67, Verstraete, M. & Collen, D. (1986ft). Phannacology of thrombolytic drugs. Journal of the American College of Cardiology 8, Suppl. B, 33b-40b. (Received 6 August 1987; accepted 15 December 1987)