J Clin Pathol: first published as 10.1136/jcp.2.1.73 on 1 February 1949. Downloaded from J. clin. Path. (1949), 2, 73.

BACTERIAL FIBRINOLYSIN, ITS POSSIBLE THERAPEUTIC APPLICATION IN TUBERCULOUS MENINGITIS BY I. A. B. CATHIE From the Hospital for Sick Children, Great Ormond Street, London

(RECEIVED FOR PUBLICATION, NOVEMBER 12, 1948)

A constant feature at autopsy in cases dying an action more of a lytic than of a preventive from tuberculous meningitis, whether treated with nature, and eventually it was decided to investigate streptomycin or not, is a gelatinous exudate at the the possibilities of the fibrinolysin produced by base of the brain. Frequently this is found in some strains of bacteria. association with, or causing, hydrocephalus, and The fibrinolysin produced by pathogenic strains once the exudate has been formed it is difficult to of haemolytic streptococci has been examined by visualize how streptomycin treatment alone could Tillett and Garner (1933), Lancefield and Hare achieve its resolution. Even were streptomycin (1935), Colebrook and others (1935), and by others, able to penetrate and sterilize the exudate the most and a few preliminary experiments with Group A copyright. likely end result would be its organization to haemolytic streptococci showed that when grown fibrous tissue with an increase in the conditions in broth they produced varying amounts of a sub- leading to hydrocephalus. stance which was capable of lysing human fibrin The known propensity of cerebrospinal fluid ctots. The variability of yield of lysin from dif- from cases of tuberculous meningitis to form ferent strains of streptococcus, and even of day-to- spider's web clots on standing which give the day cultures from the same strain, was overcome reactions of fibrin, combined with the fibrinous when Dr. C. Lack drew my attention to the strain nature of the exudate, has led to the consideration of Strep. pyogenes known as H.64, a strain known http://jcp.bmj.com/ of other forms of therapy in combination with for its abundant pr-oduction of fibrinolysin. Since streptomycin which might inhibit fibrin formation then only this organism has been used, and it has in the thecal space. Of the possible anticoagulants, given consistently good yields of lysin. heparin has received the most attention. The draw- back that it forms a deposit with streptomycin may be avoided by spaced dosage or by the method of Preparation of Fibrinolysin St. Hil and others (1948), in which heparin is The-method of preparation of the lysin has been on October 2, 2021 by guest. Protected added to streptomycin until no further deposition essentially that described by Garner and Tillett occurs, after which the supematant fluid is (1934), with slight modifications, and consists of removed by decantation. This supernatant fluid adsorbing fibrinolysin from a broth culture on to' contains all the streptomycin used, and heparin aluminium hydroxide, from which it is eluted with may be added to it without the formation of a a phosphate buffer solution. deposit. Preparatio- of alunina-50 g. of aluminium sul- Our experience of this combination has been phate (Al,SO4),.18 H,O (Willstiitter and Kraut, 1924) disappointing as judged by the results seen in in 150 ml. of distilled water are heated at 55o C. until several patients at autopsy, and, indeed, although dissolved. 500 ml. of a 15 per cent solution of it is that heparin may prevent the deposition ammonia are warmed to 55°, mixed with the alumina, likely and stirred vigorously for 30 minutes at 55 to 60°. of further fibrin it is improbable that it would have The resulting precipitate is washed repeatedly with any effect upon the fibrinous exudate, which pre- distilled water followed by decantation until no further sumably is already present when the diagnosis of smell of ammonia remains. Usually three or four tuberculous meningitis is made. For this reason washes are necessary. It is then washed with 100 ml. some other means was sought which would have of the ammonia solution to decompose any traces of J Clin Pathol: first published as 10.1136/jcp.2.1.73 on 1 February 1949. Downloaded from 7474 A. B. CATIMIE basic sulphate. The precipitate is finally *ashed two combined with the excess calcium, producing an or three times with distilled water and either centri- opalescence through which the behaviour of a fuged or filtered through a Buchner funnel using No. 1 clot was difficult to watch. Further, it was found Whatman filter paper and negative pressure. Filtra- obtained from different patients gave tion is preferred to centrifuging as the resultant paste that plasma use. results not invariably repeatable. As an outcome is drier. The-alumina is now ready for of these observations the following method of Broth cuktre.-Strep. pyogenes H.64 is heavily seeded into 500 ml. of heart digest broth containing assay has been temporarily adopted. 2 per cent NaCl and 0.05 per cent glucose and incu- From the same donor at the same time each day bated for eighteen hours. Much longer incubation 0.4 ml. of blood is taken directly into 0.1 ml. of does not give appreciably greater yields of fibrinolysin. 3.8 per cent sodium citrate in a B.S.R. pipette, and After passage through a clarifying filter the broth is the plasma so obtained is used in a dilution,pf 1:20 ready for adsorption. in saline. To a row of tubes each containing 0.5 ml. of diluted plasma are added one drop of thrombin Adsorption and elution of fibrinolysiL - The solution (containing approximately 4 units of alumina paste is added to approximately 500 ml. of thrombin) and 0.5 ml. of serial dilutions of fibrino- the broth culture and the two are agitated together lysin. In the higher strengths of lysin no clot may while the temperature is raised to 37° C. in a water- appear, while in the greater dilutions a clot is formed bath. The mixture is maintained at this temperature within about two minutes, which is lysed in a period for one hour with frequent stirring. It is then filtered of time which lengthens as the lysin is more dilute. through a Buchner funnel, and the deposit, on which Either the end-point may be taken as the dilution of the fibrinolysin is adsorbed, is washed by resuspen- fibrinolysin which inhibits clot formation, or the tubes sion twice each in 250 ml. volumes of saline and may be incubated at 370 C. for one hour, the end point distilled water, removing the fluid after each washing being in that tube which has completely lysed the clot by filtration. After the last washing the paste should in this period. It is essential to include a control clot be as dry as possible. with saline, as occasionally the naturally occurring The deposit is now resuspended in 100 ml. of M/10 plasma fibrinolysin (Macfarlane and Pilling, 1946) will

phosphate buffer solution (35.8 g. Na2HPO4 in 950 ml. copyright. of distilled water, adjusted to pH 7.3 with concen- lyse the clot formed in it. trated HCI, and made up to 1,000 ml.), thoroughly Not all preparations of lysin prevent visible stirred, and then incubated at 37° C. for thirty minutes. clot formation, although the first elution from The alumina is then removed by filtration, and the aluminium hydroxide is more likely to have this fibrinolysin-containing filtrate is finally sterilized by property than the later ones. For this reason the means of a bacterial filter. dilution capable of lysing a standard clot in one Although the first elution contains more lysin than hour gives a more precise figure for comparison do subsequent ones, second and third elutions are only of slightly less potent and are well worth preparing. of potency with other batches fibrinolysin. http://jcp.bmj.com/ Later elutions are much weaker than the first few. Typical figures for such a titration of lysin are: All manipulations of the alumina are facilitated with an ordinary household egg-whisk. -Dilution of lysin 4 8 12 16 20 24 28 Properties of Fibrinolysin Batch The solution prepared by the above technique Time for A 3 7 19 45 1 clot to lyse consists of a somewhat variable amount of fibrino- in minutes on October 2, 2021 by guest. Protected lysin in phosphate buffer. Attempts to remove the B NC NC 3 8 21 49 115 phosphate by dialysis have been unsuccessful. When dialysed with a cellophane membrane NC = No clot seen. against frequent changes of distilled water there was still phosphate present after two weeks, and there In these examples the end point was taken as the was some loss of potency of the lysin. Against dilution capable of dissolving its clot in one hour, running tap water a deposit formed in the solution and with batch A it was 16, while with B it was 24. which, as the process was accompanied by total Batch B was therefore arbitrarily assumed to be loss of lytic activity, presumably contained the 24/16 as potent as batch A. Adjustments of fibrinolysin. potency may be made either by dilution with dis- Assay of fibrinolysin.-Originally clots were tilled water or by concentration by low tempera- prepared by the addition of an excess of calcium ture distillation. to plasma obtained from roufine blood sedimen- Preparations of fibrinolysin vary in potency tation rate samples of blood. In the stronger within considerable limits, although obtained by concentrations of lysin, however, the phosphate exactly the same method, so that some form of J Clin Pathol: first published as 10.1136/jcp.2.1.73 on 1 February 1949. Downloaded from BACTERIAL FIBRINOLYSIN 75 comparable assay is essential. Of successive elu- normal meninges were used, and these remained tions from the same alumina the fifth elution is unaltered after incubation for a month. usually about one-tenth as strong as the first. Clots formed on standing in the cerebrospinal Method of action.-A preparation of fibrino- fluid from many cases of tuberculous meningitis lysin was taken which permitted the initial forma- were incubated with fibrinolysin with disappoint- tion of a clot in all dilutions with the standard ing results. In a few cases lysis took place, but plasma and thrombin. The time for clots to form in many no effect was seen, contrasting markedly and disappear was carefully noted. Then plasma with the lysis obtained with control clots prepared and lysin were incubated together for 30 minutes, from plasma. In view of the good results with after which thrombin was added and the appear- the post-mortem exudates, however, it was ance and disappearance of clots were timed. Simi- decided to investigate the clinical application of larly thrombin and lysin were incubated, plasma fibrinolysin. was added, and clot behaviour was timed. In both 1 ml. of a trial batch of fibrinolysin was injected these experiments the times for appearance and into each of six normal thecas, and gave rise to disappearance of the clots were the same as when cell counts of 500-800 polymorphs and protein plasma, thrombin, and lysin were added simul- rises of 0-10 mg. after 24 hours. This effect is taneously. They were taken to indicate that lysin similar to the pleocytosis and protein rise caused has no effect upon either thrombin or fibrinogen, by streptomycin in a normal theca. 1 ml. of phos- and that when fibrinolysin is present in sufficient phate buffer without fibrinolysin was given into a strength to prevent visible clot formation it is able normal theca and no alteration was seen in the to dissolve the clot as it is formed. cerebrospinal fluid after 24 hours, indicating that By varying the amounts of plasma or lysin in the irritant effect was due to the fibrinolysin and methods of assay, the action of the lysin can be not to the phosphate. Many injections have been shown to be quantitative. made intrathecally in cases of tuberculous mening- itis, and usually a small increase in the existing Specificity of fibrinolysin.-It has been said copyright. (Tillett and Garner, 1933) that many strains of cell count has been seen after the first three or Strep. pyogenes produce a fibrinolysin specific for four injections, again similar to the effect of human fibrin. The fibrinolysin under discussion, streptomycin. In a typical case the cells rose from however, will dissolve fibrin clots prepared from 200 to 280 per c.mm. 24 hours after fibrinolysin rabbit, mouse, and guinea-pig plasma, although had been injected. After fibrinolysin had been the period is slightly longer than with human given for four days no further increase in the cell fibrin. The only failure has been with plasma from count was seen, the greatest count being 390. a guinea-pig with a heavy tuberculous infection. Thereafter the cell count fell steadily, and the http://jcp.bmj.com/ fibrinolysin appeared to have no adverse effect on Stability of fibrinolysin. -The remarkable it. The cell response was polymorph in type, and stability of fibrinolysin has been emphasized by reversion to the original 95 per cent lymphocytosis Garner and Tillett (1934). As judged by the time was seen after two weeks. taken to lyse a standard clot there is no loss of Both the hydrochloride and calcium chloride potency after incubation at 370 C. for 14 days. complex of streptomycin produce a deposit with Similarly, after heating for 45 minutes at 950 C., fibrinolysin in phosphate buffer, much as they do followed by storage in the cold, at room tempera- with heparin. But a mixture of the fibrinolysin on October 2, 2021 by guest. Protected ture, or at 370 C., there is no loss of activity in solution and streptomycin sulphate remains clear, 14 days. Activity is completely destroyed by auto- and only this salt has been used. for one hour at 15 lb. and acid claving pressure by The variability of different preparations of hydrolysis. fibrinolysin has been noted, and one particular yield, while not being outstandingly lytic, caused Therapeutic Application gross increases in the cerebrospinal fluid cell count Several pieces of fibrinous exudate were taken of the two patients to whom it was given, one count from the base of the brain of cases dying of tuber- rising from 300 to 2,000 per c.mm. in 24 hours. culous and coliform meningitis and subjected to No sequelae to the reactions were seen, but the the action of undiluted fibrinolysin. In each case batch of fibrinolysin was immediately withdrawn lysis of the exudate occurred in from two to seven and another substituted for it. No other batch has days, depending upon the amount of exudate used. had this toxic effect. In the tuberculous cases a very fine mesh of fibrous As the result of trial and error the amount of tissue was left. As controls pieces of brain and fibrinolysin to be used therapeutically has been J Clin Pathol: first published as 10.1136/jcp.2.1.73 on 1 February 1949. Downloaded from 76- 1 -AiC,?WE chosen as 100 times the amount- capable of lysing culous mngitis, while control plama clots are the standard clot in one hour. Thus, whtn a prep- lysed. Such cerebrospinal fluid has been tested aration lyses clots in a dilution of 1: 200, 0.5 ml. without success for evidence of antifibrinolytic are given diluted wi-th the dose of streptomycin activity. The observation, taken with the failure to and 5-10 ml. of normal saline. When the strength lyse plasma clots from a tuberculous guinea-pig, of the fibrinolysin is less, proportionately greater suggests that there may be some qualitative differ- amounts are given. With regard to frequency ence between normal fibrin and that in tuber- of administration, some cases have been given culosis, although the meningeal exudate is fairly -4ibrinolysin intrathecally every time the strepto- readily lysed. mycin was given, while other cases have received Fibrinolysin has been used almost entirely for it daily for the first fortnight of treatment only. tuberculous meningitis, but its possibilities in other Fibrinolysin has been given by the lumbar route, conditions with undesirable fibrin deposition have intracysternally, and into the- lateral ventricles not been overlooked. In plastic peritonitis and without apparent ill-effect. pneumococcal empyema, for example, it might have a therapeutic place, and such cases are being Discussion investigated as they are encountered. It is possible, With regard to the various ways of demonstra- too, that more potent, or more specific, fibrinolysin ting potency, both absolute and relative, of different may be obtained from organisms other than the yields of fibrinolysin, the following quotation from streptococcus, but as yet only this organism has Macfarlane (1948) seems apposite: "The real or been used. imaginary components of a theoretical mechanism As with the treatment of tuberculous meningitis can be separated to the investigator's satisfaction with streptomycin alone, a long period of observa- with little difficulty and allowed to interact in tion will be necessary to evaluate the efficacy of endless permutations and combinations. Each fibrinolysin as an adjuvant in this condition. experiment suggests another; always the intangible- Several cases have been under treatment withcopyright. it solution seems just within reach and the experi- for varying periods of time as yet too short to menter is led deeper and deeper into his own, often warrant any other conclusion than that it may be unjustified, interpretations of his findings." With given intrathecally without danger, and that pre- this observation in mind, only the crudest method liminary results have been sufficiently encouraging of assay is mentioned here, and the many experi- to justify an extended trial of the preparation. ments carried out with varying amounts of com- The term fibrinolysin is used here to denote the ponents have been omitted. bacterial product in the presence of which lysis of

Twelve batches of 100 ml. of fibrinolysin have a fibrin clot occurs. It is probable that this lysishttp://jcp.bmj.com/ now been prepared, and while no two batches have depends upon a compound enzyme system and been strictly comparable in strength there has been that the bacterial factor is an activator of pro- a gradual increase of potency over the series. fibrinolysin, or plasminogen (Lack, 1948). Thus, in the first batch prepared activity was demonstrable only at a dilution of 1: 20, batch 12 Summary was active with the same clot at 1: 320, while the The preparation and assay of fibrinolysin from activity of the remaining batches fell between these

haemolytic streptococci are described. on October 2, 2021 by guest. Protected figures. The one toxic batch prepared was dis- Investigation of some of its properties and mode carded with its toxicity unexplained; it was bac- of action are outlined. teriologically sterile, and with the same operators, The therapeutic possibilities of fibrinolysin, par- material, and apparatus for every batch prepared ticularly with regard to tuberculous meningitis, are no useful line of investigation suggested itself. briefly indicated. It is unlikely that fibrinolysin prepared by this method does not contain varying amounts of other I wish to thank Mr. G. W. Cecil for much pains- streptococcal toxins. Some of these may be taking technical assistance. destroyed by heat without detriment to the REFERENCES fibrinolysin, but as the preparation seems to have Colebrook, L., Maxted, W. R., and Johns, A. M. (1935). J. Path. Bact., 41, 521. so little deleterious effect on tuberculous meninges Garner, R. L., and Tillett, W. S. (1934). 1. exp. Med., 60, 239. this step has not been thought necessary. Lack, C. H. (1948). Nature, 161, 559. Lancelield R. C., and Hare, R. (1935). J. exp. Med., 61, 335. An interesting and disappointing observation Macfarlane, R. 0. (1I48). J. clin. Path., 1. 113. Macfarlane, R. G., and Pillg J. (1946). Lacet, 2, 562. has been the infrequency with which fibrinolysin St. HWil, C. A., Riley, C., and Gifford, J. H. (1948). J. clin. Pqth. has been able to lyse the fibrin clots formed on 1, 157. Tillett, W. S., anid Garner, R. L. (1933); J. exp. Mod., 66, 485. standing in the cerebrospinal fluid from tuber- wilattter, R., and Kraut, H. (1924). Ber. dtsch. chem. Ges., 57,1082.