A heat stable factor (SF) in normal human serum which is cytotoxic to Landschutz ascites tumor cells has been described. The cytotoxic action requires the presence of human or rabbit complement but guinea pig complement is ineffective. The substance is found in the beta-globulin fraction and has characteristics similar to C′4 of human complement. It is selectively absorbed on human and mouse tumor cells and human placenta but not by normal human liver, kidney or heart. SF inhibits the action of heterologous antibody on ascites and HeLa cells. We are indebted to Dr. A. Rimon, Marcus Memorial Blood Bank Institute, Jaffa, for the Cohn fractions, and to Dr. Esther Tenenbaum of this Department for a generous supply of HeLa cells and Chang cells.

THE mechanism of cell destruction in allergic inflammation is not well understood. Two explanations for production of cell damage have, however, been proposed: (a) that the antigen-antibody reaction causes, in some non-specific manner, permeability changes in cells which result in the release of active mediators of inflammation (histamine, etc.) and that the latter are directly involved in the allergic process1; (b) that the antigen-antibody reaction activates proteolytic enzymes (plasmin, proteases) that injure the cells2–4.

SEVERAL studies were undertaken in order to reproduce in laboratory animals cardiac lesions which would resemble those appearing in human beings suffering from the sequelae of streptococcal infections (Schultz, 1936 ; Smith, Morgan and Mudd, 1940; Gross, Cooper and Philips, 1941; Robinson, 1947; Schultz, 1947; Murphy and Swift, 1949; Murphy, 1949; Clawson, 1950; Robinson, 1951; Murphy, 1952; Glaser, Thomas, Morse and Darnell, 1956). These studies describe the histopathological changes obtained in laboratory animals following inoculation by various routes of haemolytic streptococci, of their products and of a number of other micro-organisms. As to the mechanism involved in the production of cardiac lesions, two main theories have been proposed: (a)allergic phenomena,or production ofa uto- antibodies to the heart muscle (Rich and Gregory,1944; Rich,1946; Cavelty, 1947a and 1947b), (b) direct action of haemolytic streptococci and their toxins on the heart muscle (McLeod, 1953; Kellner and Robertson, 1954a and 1954b). Although both theories have been supported by some experimental evidence, the exact mechanism has not yet been clarified. This study describes the histo-pathological changes obtained in rabbits as a result of single intramyocardial injections of haemolytic streptococci and their cell-free extract, as well as of other bacterial species, both related and not related to haemolytic streptococci. The possible mechanism involved, and especially the role of trauma to the myocardium, in inducing cardiac lesions, is described.

STUDIES on the cytotoxic action of antibodies, plasmin and streptococcal haemolysins on Ehrlich ascites tumour cells (E-cells) have been recently described (GCinsburg, 1959; Ginsburg and Grossowicz, 1960; Ginsburg and Ram, 1960). It was found that the cytopathogenic changes induced in the tumour cells by streptococcal haemolysins were morphologically very similar to those caused by anti-Ehrlich ascites tumour rabbit serum (AE), the action of the latter is known to be complement dependent (Flax, 1956; Goldberg and Green, 1959). It is known that both the haemolytic as well as the cytopathogenic effect caused by streptococcal haemolysins can be inhibited by antihaemolytic agents such as cholesterol, lecithin, congo red and trypan blue (Van Heyningen, 1950; Bernheimer, 1954; Ginsburg, 1959; Ginsburg and Grossowicz, 1960). On the other hand, the cytotoxic and haemolytic effect of antibodies can be inhibited by agents known to destroy or inhibit complement action, chelating agents, plasmin, heparin etc. (Gorrill and Hobson, 1952; Pillemer, Ratnoff, Blum and Lepow, 1953; Mayer, 1 958; Ginsburg and Ram, 1960). The present study shows that the cytotoxic action of antibodies produced in the rabbit against Ehrlich ascites tumour cells can be inhibited by various phospholipids and some of their constituents as well as by agents known to inhibit streptococcal haemolysins. In addition the inhibition of immune haemolysis by phospholipids will be described.

IT has been recently shown that various strains of Streptococcus pyoqenes group A possess a cell-bound haemolysin (CBH) which can be released from the streptococcal cells into the surrounding medium by some surface active materials (Tween 40, Tween 80, Triton) by crystalline albumin, but not by sonlic energy (Ginsburg and Grossowicz, 1957 ; Ginsburg, 1958; Ginsburg and Grossowicz, 1959). This haemolytic factor was designated as streptolysin " D " (SLD) D signifying detergent. SLD is distinguished from streptolysin 0 (SLO) streptolysin S (SLS) (Todd, 1938) and from the intracellular haemolysin (IH) described by Schwab (Schwab, 1956) on the bases of different substrate requirements for formation, sensitivity to U.N. irradiation and to sonic energy (Ginsburg, 1958; Ginsburg and Gros- sowicz, unpublished). Besides haemolysing RBC of various animal species both cell-free and cell- bounid SLD, are also capable of injuring and killing various mamimalian cells in vitro (Ehrlich ascites tumour cells, fibroblasts, amnion cells, leucocytes) (Gins- buirg, 1958 ; Ginsburg and Grossowicz, 1959). The purpose of the present study is to show that Ehrlich ascites tuilour cells damaged by different streptococcal haemolysins may be disintegrated bv various proteolytic enzymes which by themselves are not lethal

1) Defined media which allow heavy growth of 3 strains of group A streptococci have been developed. The medium consists of either a) 21 amino acids, glutamine, 6 vitamins, salts, purines, pyrimidines and glucose or b) 13 amino acids and 4 vitamins. 2) Cysteine is important both as an essential amino acid and as a reducing agent. As an amino acid only a small amount (10 μg ml) is needed and this can be substituted by an equivalent amount of cystine. As a reducing agent it can be replaced by ascorbic acid and less effectively by thiomalic or thioglycollic acids. Concentration of cysteine was critical for initiation of growth from small inocula. With less than 5 × 106 ml cells at least 350 μg ml cysteine HCl are needed for obtaining visible growth. 3) Pyridoxal was necessary in a medium of 13 amino acids and 3 vitamins (nicotinic acid, pantothenate, riboflavin) whereas in a complete medium (22 amino acids) no need for pyridoxal was found. Pyridoxal could be replaced by L- or DL-alanine.

During our study of the action of water-soluble poly-a-amino acids on blood clotting (1), it was observed that the basic poly-amino acids: poly-lysine (2), poly-ornithine (3) and poly-arginine (3), retard fibrinolysis of human clotted blood. A more detailed analysis of this phenomenon was therefore undertaken. Fibrinolytic activity of oxalated human plasma was induced by mixing the plasma with a suspension of ,- hemolytic streptococci (4, 5), by treatment with a cellfree broth containing streptokinase (6), or by shaking the plasma with chloroform (6, 7). The activated plasma was then treated with the poly-amino acids (prepared in this laboratory), and a fibrin clot obtained by the addition of thrombin. The final mixtures were incubated at 370 C for 15-24 hr to determine -if lysis occurred. When the preparations used did not interfere with fibrinolysis, dissolution of the clot occurred. Inhibition of fibrinolysis was indicated by the maintenance of the fibrin clot, obtained as described *above, for 24 hr. A typical experiment with poly-L-lysine and a streptococei-activated plasma is described below. Oxalated human plasma (0.5 ml) was mixed with a suspension of 1-hemolytic streptococci (0.4 ml), and the mixture added to 1 ml saline solution containing 40y poly-L-lysine hydrochloride. Clotting was induced by adding 4 units of bovine thrombin (Upjohn Company) in 0.1 ml saline with vigorous shaking. The clot was incubated at 370 C for 24 hr. No visual change in the clot was observed. In the control experiment where the 1 ml poly-lysine solution was substituted by saline, a complete lysis of the clot was evident within 30 min. The fibrinolytic activity of plasma activated by Ihemolytic streptococci was not inhibited either by the neutral poly-DL-alanine (8) or by the acidic poly-Laspartic (9) and poly-D-glutamic (10) acids up to concentrations of 500y/ml final test mixture. The basic poly-a-amino acids, poly-DL-lysine hydrochloiide (average chain length n = 35) (2), poly-DL-ornithine hydrochloride (n = 30) (3), and poly-DL-arginine sulfate (n = 30) (3), on the other hand, prevented fibrinolysis at concentrations greater than 30y40y/ml test mixture. In the presence of the basic poly-amino acids, fibrinolysis Was inhibited equally well when the streptococcal culture suspension was replaced (in the test mixture) by a cell-free supernatant containing streptokinase. Furthermore, it has been demonstrated that the fibrinolytic activity of chloroform-treated plasma and of menstrual blood was also inhibited by relatively low concentrations of poly-DL-lysine and poly- DL-arginine. It thus seems justified to assume that the basic poly-amino acids are capable of inhibiting hydrolysis of fibrin by plasma fibrinolysin (plasmin) under the experimental conditions used. Preliminary experiments indicated that the average molecular weight of the basic poly-amino acids plays a profound role in ther determination of their antifibrinolytic properties. L-lysine monomer, as well as L-lysyl-L-lysine (11), did not show any antifibrinolytic activity up to a concentration of 750y/ml. A tetra-Llysine showed slight antifibrinolytic activity at 750y/ ml, whereas poly-L-lysine of average chain length n= 7, 35, and 100 showed distinct antifibrinolytic activity at concentrations of 500y, 40y, and 35y/ml test mixture, respectively. No great difference was observed in the antifibrinolytic activity of poly-L-, poly-D-, and poly-DL-lysine of similar average molecular weights. In our previous study on the action of water-soluble poly-amino acids on blood clotting (1), it was demonstrated that the acidic poly-amino acids, poly-D-glutamic acid and poly-L-aspartic acid, as well as heparin, are capable of neutralizing the anticoagulant activity of the basic poly-amino acids. A similar relationship was found to hold for the antifibrinolytic effect of the basic poly-amino acids. Heparin, as well as poly-Laspartic acid (n = 50), was found to obviate the antifibrinolytic activity of poly-lysine. The neutralization of the antifibrinolytic activity of the basic poly-amino acids occurred when approximately equivalent concentrations of the basie -and acidic poly-amino acids were applied. The ability of the basic synthetic peptides to inhibit rev,ersibly the. proteolytic activity of fibrinolysin resembles, the antiproteolytic properties of some natural peptides, such as the pepsin,inhibitor and the' trypsin inhibitors (12). The interaction of the natural as well as the synthetic peptides with the different proteolytic enzymes is probably determined in both cases by some specific groups present in the enzyme and the inhibitor, as well as by the electrostatic forces prevailing between the enzyme and the relatively high molecular weight inhibitor. Further studies with the 'synthetic amino acid polymers may contribute to our basic knowledge of the mode of actlon of naturally naturally occurring polypeptides on enzyme behavior.