Vol. 18, No. 3 Printed in U.S.A.
INFECrION AND IMMUNrry, Dec. 1977, p. 612-616 Copyright i 1977 American Society for Microbiology
Effect of Serum from Various Animal Species on Erythrocyte Attachment of Endotoxins and Other Bacterial Antigens MARTIN PRAINO AND ERWIN NETER* Department of Microbiology, State University of New York at Buffalo and Children's Hospital of Buffalo, Buffalo, New York 14222 Received for publication 31 May 1977
Lipopolysaccharide 0 antigens (endotoxins) and other bacterial antigens readily attach to erythrocytes in vitro. This attachment is prevented by certain mammalian and avian sera. In this study, the inhibitory capacity of sera from lower animals was compared with that of higher animals for a total of 30 species. Antigens and the corresponding antisera included both crude 0 antigens and purified lipopolysaccharide preparations, the common enterobacterial antigen from Escherichia coli 014, the Vi antigen from Citrobacter ballerup, the polyribose-phosphate antigen from Haemophilus influenzae type b, and the crude teichoic acid antigen from Staphylococcus aureus. Antigen and serum mixtures were incubated at 37°C for 30 min and used for erythrocyte modification; failure of hemagglutination by homologous bacterial antiserum provided evidence of inhibitory capacity. Sera from the classes Mammalia and Aves were very strong inhibitors; those of Reptilia and Osteichthyes were moderate in activity, displaying variation within the classes; those of Amphibia and Chondrichthyes were minimal inhibitors; and those, of Merostomata, Crustacea, and Lamellibranchiata displayed questionable or no inhibitory capacity. Inhibitory sera were active with all antigens tested. The findings suggest evolution of inhibitory factors consistent with the theory of two diverging lines of animal phylogeny based on embryological criteria and closely parallel the observations of an endotoxinaltering capacity in vertebrate sera that is not found in invertebrate sera or hemolymph. It is known that various bacterial antigens sera (7, 9, 14), human plasma fractions III-0, IVattach to erythrocytes of many animal species 1, IV-6, IV-7, and albumin (7, 14), bovine serum in vitro. When the modified erythrocytes are albumin (14), antibiotics such as polymyxin B incubated with homologous antibodies to the (10), proteins such as hemoglobin, histone, and adsorbed antigen, hemagglutination occurs (3, protamine (11), and various lipids including lipid 6, 7, 13). Among the various bacterial antigens A, as well as cell membrane components of leuknown to attach to erythrocytes are crude 0 kocytes and platelets (18). Substances that have antigens and lipopolysaccharides (LPS) of gram- no inhibitory effect include human plasma fracnegative bacteria, the common enterobacterial tions II and II (14), certain antibiotics such as antigen (CA), the Vi antigen, and the heteroge- bacitracin, penicillin, and erythromycin (10), netic teichoic acid antigen from certain gram- proteins such as casein and egg albumin, amino positive microorganisms. Erythrocytes may acids, deoxyribonucleic acid, and ribonucleic serve as a useful model to demonstrate interac- acid (19). tion of cell membranes and certain toxins, beIn 1969, Springer et al. (20) demonstrated that cause attachment to host cells is probably a an erythrocyte membrane component inhibits necessary step leading to cellular toxicity. For attachment of endotoxins and the CA to erythexample, Hill and Weiss (4) showed that in A- rocytes and probably functions as the erythrostrain mice a relationship exists between the cyte receptor. A degree of specificity was establethality of Sablonella endotoxin and its ability lished by the observation that 0 and CA antito attach to their erythrocytes. gens are inhibited by the receptor in small Addition to the above antigens of various sub- amounts. Far larger quantities are required to stances before incubation with erythrocytes re- inhibit attachment of other antigens such as the sults in inhibition of attachment and thus of Vi antigen, two stearoyl derivatives of group A hemagglutination. Among the many known in- streptococcus, and crude teichoic acid antigen hibitors ofattachment are mammalian and avian from certain gram-positive organisms. The sub612
VOL. 18, 1977
ERYTHROCYTE ATTACHMENT OF ENDOTOXINS
stance was appropriately termed "LPS receptor" and later characterized as lipoglycoprotein, rich
in neuraminic acid (19). More recently, von Eschen and Rudbach (22) demonstrated that sera from certain, but not all, animal species possess factors that detoxify endotoxin, are heat stable, and are inactivated by CA2" ions. These factors are present in sera of vertebrates, but not in invertebrate sera or hemolymph, and may have evolved at a time corresponding to the rise of the sharks (class Chondrichthyes). The objectives of the present study were threefold: (i) to determine whether sera from lower animal classes have the capacity to inhibit attachment to erythrocytes of various bacterial antigens, (ii) to compare inhibitory capacity of sera from lower and higher animals, and (iii) to learn whether a correlation exists between the inactivation of endotoxin and the inhibition of cell attachment by sera from various animal species. MATERIALS AND METHODS Inhibitory capacity of sera from various animal species was determined and quantitated using the bacterial hemagglutination procedure (7). Normal sera. Samples of sera from the classes Mammalia, Aves, Reptilia, Amphibia, Osteichthyes, Chondrichthyes, Merostomata, Crustacea, and Lamellibranchiata (2) were obtained through the courtesy of the following: Jon Rudbach, University of Montana, Missoula; Michael Sigel, University of Miami, Miami, Fla.; Elias Cohen, Roswell Park Memorial Institute, Buffalo, N.Y.; Dorothy Pollock, Transfusion Service, the Children's Hospital of Buffalo, Buffalo, N.Y.; Jack Cory, Rocky Mountain Laboratory, Hamilton, Mont.; the Buffalo Zoological Gardens, Buffalo, N.Y.; and the animal facilities of the State University of New York at Buffalo and the Children's Hospital of Buffalo, Buffalo, N.Y. Other sera were purchased from Alan Boyden, the Serological Museum, Rutgers University, New Brunswick, N.J.; Colorado Serum Co., Inc., Denver, and the Grand Island Biological Co., Grand Island, N.Y. Antigens. Crude preparations of 0 antigens wereprepared from Sabnonella cholerae-suis and Escherichia coli according to the method described previously by Neter et al. (12). Briefly, organisms were grown on brain veal agar (Difco) in Kolle flasks for 18 h at 37°C, The growth was harvested with 25 ml of phosphate hemagglutination buffer (Difco, pH 7.3), heated for 1 h in a water bath at 100°C, and centrifuged at 22,500 x g for 15 min at 2°C. The resulting supernatant, termed heat-killed supernatant, was decanted and kept frozen until used. Purified LPS from Salmonella montevideo and E. coli 0113, obtained by the phenol-water extraction procedure of Westphal et al. (23), were kindly supplied by Otto WestphaL Max-Planck Institut fur Immunbiologie, Freiburg, Germany, and Jon Rudbach, re-
spectively.
613
The CA from E. coli 014 heat-killed supernatant was prepared in a manner similar to that of crude 0
antigen described above. Vi antigen of Citrobacter ballerup was prepared in the following manner: a colony of the organism from blood agar was inoculated into 6 ml of brain heart infusion broth (Difco) and grown for 4 h in a water bath at 37°C; the broth was transferred to a Kolle flask containing brain veal agar and grown for 18 h at 37°C. The organisms were harvested as described above, and the supernatant was kept frozen for later use. The polyribose-phosphate antigen of Haemophilus influenzae type b was prepared by growing the organisms on chocolate Mueller-Hinton agar plates (150 by 15 mm) for 18 h at 37°C, harvesting the growth with 5 ml of phosphate hemagglutination buffer with 0.5% Formalin, allowing the suspension to stand at 22°C for 4 h, and centrifuging at 22,500 x g for 15 min at 2°C. The supern Lant, containing the antigen, was stored for testing later. The teichoic acid antigen of Staphylococcus aureus was prepared as previously described by Neter et al.
(8).
Unless stated otherwise, crude antigens for modification of erythrocytes were used in a final dilution of 1:10, and purified LPS was used in a final concentration of 12.5 ,ug/ml. Antisera. Antisera to the 0 antigens of S. choleraesuis and E. coli were obtained from patients with the respective infections. Additional 0 and Vi antisera were purchased from Lederle Diagnostics, Pearl River, N.Y. Antisera to CA and teichoic acid antigens were obtained by repeated injections of ethanol-soluble CA from E. coli 07 and S. aureus, respectively, into albino rabbits. Antiserum to H. influenzae type b was obtained from the Commonwealth of Massachusetts, Department of Public Health, Division of Biological Laboratories, Boston. Procedures. Sera to be studied as inhibitors of antigen attachment to erythrocytes were first tested for the presence of antibodies to both the respective antigen and the indicator erythrocytes (human blood group 0) and, when indicated, absorbed as follows. Undiluted serum, 1 to 2 ml, was mixed with the sediment obtained from an equal amount of antigen-modified erythrocytes (see below). The mixtures were incubated at 37°C for 30 min and centrifuged at 1,500 x g for 5 min. If agglutination was observed, the procedure was repeated. Sera whose antibodies could not be removed by three absorptions or that lysed the erythrocytes were not used. Briefly, antigen and test serum, unheated or heated at 56°C for 30 min, in appropriate dilutions, were incubated at 37°C for 30 min. Buffer instead of serum was used for control purposes. The mixtures were then added to the 'sediment of thrice-washed, 2.5% erythrocytes and incubated for 30 min at 37°C. The erythrocytes were then washed three times, resuspended, and, in 0.2-ml amounts, added to equal amounts of homologous antiserum in twofold serial dilutions. The degree of hemagglutination was read and recorded after centrifugation at 1,500 x g for 2 min. A decrease in the titer of the antiserum against the homologous bacterial antigen when compared to
614
INFECT. IMMUN.
PRAINO AND NETER
classes Mammalia and Aves exhibited strong inhibitory capacity. Among the lower vertebrate classes, not previously studied, sera from Reptilia and Osteichthyes displayed moderate inhibitory capacity, whereas those from Amphibia and Chondrichthyes were minrimal in their ability to block antigen attachment to erythrocytes. Sera or hemolymph from all invertebrate classes investigated, namely, Merostomata, Crustacea, and Lamellibranchiata, exhibited little or no inhibitory capacity. Differences in the inhibitory capacity of sera from vertebrates were further substantiated by
the control indicates inhibitory capacity. A 2-fold difference was considered questionable and is listed as ± in Table 1; a 4-fold difference as minimal and listed as 1; an 8-fold difference as moderate and listed as 2; a 16- to 32-fold difference as strong and listed as 3; a >32-fold difference as very strong and listed as 4.
RESULTS Table 1 summarizes the results of the study of the comparative capacity of sera from various animal species to inhibit the attachment of several bacterial antigens to erythrocytes. In accord with previous reports (7, 9, 14), sera from the
TABLE 1. Inhibitory capacity of serum from various animal species on erythrocyte attachment of various bacterial antigens Bacterial antigen tested
Animal serum tested (class and common name)
P
erae-
suis
Vertebrates Mammalia Human Fetal calf Rabbit Guinea pig Aves Turkey Chicken Duck Goose Grouse Pigeon
Reptilia Log turtle Snapper Caiman Painted turtle
Amphibia Bullfrog Lamper eel Siren Cryptobranchus Osteichthyes Carp Conger eel Blue marlin Chondrichthyes Shark Barn door skate Invertebrates Merostomata Horseshoe crab Crustacea English lobster Shore crab Lamellibranchiata
E. coli
monte-
video
4 4
4 4
CA, E. coli
H. influenzae
S. au-
014
Vi C ballerup
type b
4 4
4 4
4 4
4 4
4
4
4
4
4
3-4
LPS S.
to
capac-
ity/class
reus
4
4
3 3 3
3-4
3-4 2-3
2 1 2 3
2-3 3
0 1-2 1-2
0 2 2
2-3 3
3
3
3
3
1
0
±
1
0
±
i,
±
2
2-3
2
±
2-3
4
2-3 2-3
1 ±
2
1-2 0
±
±
1
± ±
± ±
2
2
1
3
1-2
±
0
i
i
1
1
0
0
±
±
±
1 1
0 0
±
±
0
± ±
0-i:
0
1-2
± 1 1 1 1 1-2 ±-l Denotes degree of inhibitory capacity, none (0) to very strong (4) (see Materials and Methods).
Oyster
a
4a 4 4 3-4
-ed-
±-1
VOL. 18, 1977
ERYTHROCYTE ATTACHMENT OF ENDOTOXINS
titration. Highly inhibitory sera, such as mammalian and avian sera, are active in dilutions of approximately 1:1,000, reptilian and osteichthyian sera are active in dilutions of 1:100, and amphibian and chondrichthyian sera are active in a dilution of 1:10 or are even inactive at this dilution. As far as the invertebrates are concerned, most sera were inactive at a dilution of 1:10. Further studies revealed that serum or hemolymph from Limulus (horseshoe crab) and oyster, even when used undiluted, had no significant inhibitory capacity. To determine whether, in a more sensitive system, inhibitory capacity can be detected with serum specimens from invertebrates and lower vertebrates, experiments were carried out utilizing small amounts of antigen for erythrocyte modification. Crude S. cholerae-suis 0 antigen was used in a dilution of 1:150, instead of 1:10. Serum from four out of five invertebrate species showed no increase and the remainder showed only minimally enhanced inhibition of antigen attachment. As far as the lower vertebrates are concerned, a slight increase in the prevention of erythrocyte modification was effected by serum from four out of five species. The serum from the bullfrog exhibited no activity, even under these conditions. It can be seen from the results shown in Table 1 that inhibitory capacity of animal sera, regardless of phylogenetic origin, is nonspecific; i.e., it is independent of antigen used. To evaluate the possible role of complement and other heat-labile factors in inhibitory activity, all sera were also tested after being heated at 560C for 30 min. The results indicate that inactivation of sera did not affect the inhibitory
capacity. DISCUSSION It has been known for many years that certain bacterial antigens become attached to the cell surface of erythrocytes and that pretreatment with tannic acid and similar substances is not required. Antigen attachment is detected by agglutination by homologous antiserum. Springer et al. in 1970 (20) demonstrated that erythrocyte membranes contain a substance that binds LPS of gram-negative bacteria and the CA, but not, in similar amounts, certain other bacterial antigens (20, 24). The isolated erythrocyte membrane component was termed LPS receptor. Numerous other substances, including sera of higher animals, also inhibit antigen attachment to erythrocytes (7, 9-11, 14, 18). The present experiments were suggested by the observation of von Eschen and Rudbach (22), who reported
615
Mammalia
Aves Reptilia
it
Amphibia
Merostomata
I Osteichthyes
Chondrichthyes
\I/ *II! Bilateral
Phyla
FIG. 1. Probable main lines of animal phylogeny (adapted from Dodson's Evolution [1]).
striking differences in the capacity of sera from various animal species to inhibit toxicity of endotoxin. This study was undertaken to determine the capacity of such sera to inhibit antigenic modification of erythrocytes by LPS and other bacterial antigens. This investigation has revealed marked differences in the capacity of sera from various animal species to inhibit attachment of various bacterial antigens to erythrocytes. This activity was greater in sera from higher animal species than in those from lower animal species. The inhibitory activity was not destroyed by heating at 56°C for 30 min. Inhibition is nonspecific since, before testing, antibodies against the various bacterial antigens were removed by absorption. Information is not yet available as to the factors present in the sera of all species tested responsible for inhibition of antigen attachment. It appears that the degree of inhibition closely parallels the evolution of two diverging lines of animal phylogeny based on embryological criteria (Fig. 1). The chordate line culminating in the class Mammalia may have evolved these serum factors at an early stage, possibly at the time corresponding to the rise of the higher fish or sharks. Sera from the class Osteichthyes as a group were stronger inhibitors than those from the class Chondrichthyes, although both may have evolved at a similar time phylogenetically from a placoderm ancestor. Presumably, Osteichthyes gave rise to Amphibia and higher animals, whereas the Chondrichthyes are thought to be a "dead end" in evolution (1). The shark
616
PRAINO AND NETER
was the lowest animal species whose serum displayed minimal inhibitory activity of all sera tested. Comparison of the inhibitory capacity of sera on attachment of bacterial antigens to erythrocytes and their ability to detoxify endotoxin, as reported by von Eschen and Rudbach (22), revealed some striking similarities. Sera and hemolymph from invertebrates have virtually no effect, and sera from higher animal species were highly active in both systems. In both studies variation in the effectiveness of members of several classes of animals was observed. Heating of serum at 56°C did not abolish the activity of serum from various animal species in either system, indicating that the entire complement system is not involved. However, it should be kept in mind that, as shown by Johnson and Ward (5), C5 and C6, even in the absence of C3 activation, may play a role in endotoxin detoxification. In addition, other serum proteins may contribute to detoxification (15-17, 21, 25). The relationship of any of these factors to inhibition of antigen attachment to erythrocytes remains to be elucidated. ACKNOWLEDGMENT This investgtion was supported by Public Health Service grant AI 00658 and training grant 5 TOl Al 00442-05 from the National Institute of Allergy and Infectious Diseases. We also gratefully acknowledge the excellent technical assistnce of Helga von Langendorff.
LrFERATURE CffE 1. Dodson, E. 0. 1960. Evolution: process and product. Reinhold Publishing Corp., New York. 2. Grzimek, L C. B. 1972-1974. Grzimek's animal life encyclopedia. Van Nostrand Reinhold Co., New York. 3. Hayes, L 1951. Specific serum agglutination of sheep erythrocytes sensitized with bacterial polysaccharides. Aust. J. Exp. Biol. Med. Sci. 29:51-62. 4. Hill, G. J., and D. W. Weiss. 1964. Relationships between susceptibility of mice to heat-killed salmonellae and endotoxin and the affinity of their red blood cells for killed organi, p. 422-427. In M. Landy and W. Braun (ed.), Bacterial endotoxina Institute of Microbiology, Rutgers, The State University, New Bnswick, N.J. 5. Johnson, K. J., and P. A. Ward. 1972. The requirement for serum complement in the detoxification of bacterial endotoxin. J. Immunol. 108:611-616. 6. Keogh, E. V., E. A. North, and ML F. Warburton. 1948. Adsorption of bacterial polysaccharides to erythrocytes. Nature (London) 161:687-688. 7. Neter, K. 1956. Bacterial hemagglutination and hemolysis. Bacteriol. Rev. 20:166-188. 8. Neter, E., H. Anai, and E. A. Gorzynid 1960. The
INFECTr. IMMUN. effects of staphylococcal extract and of viable staphylococci on dermal reactivity of the rabbit to epinephrine. J. Infect. Dis. 107:332-340. 9. Neter, E., E. A. Gorzynski, H. Anzai, and V. Bokkenheuser. 1962. Hemagglutination by mixtures of human gamma globulin and heterogenetic bacterial antigen. J. inmunol. 88:411-417. 10. Neter, E., E. A. Gorzynski, 0. Westphal, and 0. Luderitz. 1958. The effects of antibiotics on enterobacterial lipopolysaccharides (endotoxins), hemagglutination and hemolysis. J. Immunol. 80:66-72. 11. Neter, E., E. A. Gorzynsk, 0. Westphal, 0. Luderitz, and D. J. Klumpp. 1958. The effects of protamine and histone on enterobacterial lipopolysaccharides and hemolysis. Can. J. Microbiol. 4:371-383. 12. Neter, E., 0. Westphal, 0. Liideritz, and E. Gorzynski. 1956. The bacterial hemagglutination test for the demonstration of antibodies to Enterobacteriaceae. Ann. N.Y. Acad. Sci. 66:141-156. 13. Neter, E., 0. Westphal, 0. Luderitz, E. A. Gorzynski, and E. Eichenberger. 1956. Studies of enterobacterial lipopolysaccharides. Effects of heat and chemicals on erythrocyte-modifying, antigenic, toxic and pyrogenic properties. J. Immunol. 76:377-385. 14. Neter, E., D. A. Zak, N. J. Zalew8ki, and L F. Bertram. 1952. Inhibition of bacterial (Escherichia coli) modification of erythrocytes. Proc. Soc. Exp. Biol. Med. 80:607-610. 15. Rosen F. S., R. C. Skarnes, IL Landy, and M. J. Shear. 1958. Inactivation of endotoxin by a humoral component. HI. Role of divalent cation and a dialyzable component. J. Exp. Med. 108:701-711. 16. Skarnes, R. C. 1966. The inactivation of endotoxin after interaction with certain proteins of normal serum Ann. N.Y. Acad. Sci. 133:644-662. 17. Skarnes, R. C. 1970. Host defense against bacterial endotoxemia mechanism in normal animals. J. Exp. Med. 132:300-316. 18. Springer, G. F., and J. C. Adye. 1975. Endotoxin-binding substances from human leukocytes and platelets. Infect. Immun. 12:978-986. 19. Springer, G. F., J. C. Adye, A. Bezkorovainy, and J. R. Murthy. 1973. Functional aspects and nature of the lipopolysaccharide-receptor of human erythrocytes. J. Infect. Dis. 128:5202-5212. 20. Springer, G. F., S. V. Huprikar, and E. Neter. 1970. Specific inhibition of endotoxin coating of red cells by a human erythrocyte membrane component. Infect. Immun. 1:98-108. 21. Stauch, J. E., and A. G. Johnson. 1959. The alteration of bacterial endotoxins by human and rabbit serum. J. Immunol. 82:252-263. 22. von Eschen, K. B., and J. A. Rudbach. 1974. Inactivation of endotoxin by serum: a phylogenetic study. J. Infect. Dis. 129:21-27. 23. Westphal 0., 0. Liideritz, and F. Bister. 1952. Ober die Exraktion von Bakterien mit Phenol/Wasser. Z. Naturforwch. Teil B 7:148-155. 24. Whang, H. Y., E. Neter, and G. F. Springer. 1970. Specificity of an erythrocyte membrane receptor for bacterial antigens. Z. Immunitaetsforsch. 140:298-303. 25. Yoshioka, IL, and S. Konno. 1970. Characteristics of endotoxin-altering fractions derived from normal serum.Ill. Isolation and properties of horse serum a2macroglobulin. Infect. Immun. 1:431-439.
INFECrION AND IMMUNrry, Dec. 1977, p. 612-616 Copyright i 1977 American Society for Microbiology
Effect of Serum from Various Animal Species on Erythrocyte Attachment of Endotoxins and Other Bacterial Antigens MARTIN PRAINO AND ERWIN NETER* Department of Microbiology, State University of New York at Buffalo and Children's Hospital of Buffalo, Buffalo, New York 14222 Received for publication 31 May 1977
Lipopolysaccharide 0 antigens (endotoxins) and other bacterial antigens readily attach to erythrocytes in vitro. This attachment is prevented by certain mammalian and avian sera. In this study, the inhibitory capacity of sera from lower animals was compared with that of higher animals for a total of 30 species. Antigens and the corresponding antisera included both crude 0 antigens and purified lipopolysaccharide preparations, the common enterobacterial antigen from Escherichia coli 014, the Vi antigen from Citrobacter ballerup, the polyribose-phosphate antigen from Haemophilus influenzae type b, and the crude teichoic acid antigen from Staphylococcus aureus. Antigen and serum mixtures were incubated at 37°C for 30 min and used for erythrocyte modification; failure of hemagglutination by homologous bacterial antiserum provided evidence of inhibitory capacity. Sera from the classes Mammalia and Aves were very strong inhibitors; those of Reptilia and Osteichthyes were moderate in activity, displaying variation within the classes; those of Amphibia and Chondrichthyes were minimal inhibitors; and those, of Merostomata, Crustacea, and Lamellibranchiata displayed questionable or no inhibitory capacity. Inhibitory sera were active with all antigens tested. The findings suggest evolution of inhibitory factors consistent with the theory of two diverging lines of animal phylogeny based on embryological criteria and closely parallel the observations of an endotoxinaltering capacity in vertebrate sera that is not found in invertebrate sera or hemolymph. It is known that various bacterial antigens sera (7, 9, 14), human plasma fractions III-0, IVattach to erythrocytes of many animal species 1, IV-6, IV-7, and albumin (7, 14), bovine serum in vitro. When the modified erythrocytes are albumin (14), antibiotics such as polymyxin B incubated with homologous antibodies to the (10), proteins such as hemoglobin, histone, and adsorbed antigen, hemagglutination occurs (3, protamine (11), and various lipids including lipid 6, 7, 13). Among the various bacterial antigens A, as well as cell membrane components of leuknown to attach to erythrocytes are crude 0 kocytes and platelets (18). Substances that have antigens and lipopolysaccharides (LPS) of gram- no inhibitory effect include human plasma fracnegative bacteria, the common enterobacterial tions II and II (14), certain antibiotics such as antigen (CA), the Vi antigen, and the heteroge- bacitracin, penicillin, and erythromycin (10), netic teichoic acid antigen from certain gram- proteins such as casein and egg albumin, amino positive microorganisms. Erythrocytes may acids, deoxyribonucleic acid, and ribonucleic serve as a useful model to demonstrate interac- acid (19). tion of cell membranes and certain toxins, beIn 1969, Springer et al. (20) demonstrated that cause attachment to host cells is probably a an erythrocyte membrane component inhibits necessary step leading to cellular toxicity. For attachment of endotoxins and the CA to erythexample, Hill and Weiss (4) showed that in A- rocytes and probably functions as the erythrostrain mice a relationship exists between the cyte receptor. A degree of specificity was establethality of Sablonella endotoxin and its ability lished by the observation that 0 and CA antito attach to their erythrocytes. gens are inhibited by the receptor in small Addition to the above antigens of various sub- amounts. Far larger quantities are required to stances before incubation with erythrocytes re- inhibit attachment of other antigens such as the sults in inhibition of attachment and thus of Vi antigen, two stearoyl derivatives of group A hemagglutination. Among the many known in- streptococcus, and crude teichoic acid antigen hibitors ofattachment are mammalian and avian from certain gram-positive organisms. The sub612
VOL. 18, 1977
ERYTHROCYTE ATTACHMENT OF ENDOTOXINS
stance was appropriately termed "LPS receptor" and later characterized as lipoglycoprotein, rich
in neuraminic acid (19). More recently, von Eschen and Rudbach (22) demonstrated that sera from certain, but not all, animal species possess factors that detoxify endotoxin, are heat stable, and are inactivated by CA2" ions. These factors are present in sera of vertebrates, but not in invertebrate sera or hemolymph, and may have evolved at a time corresponding to the rise of the sharks (class Chondrichthyes). The objectives of the present study were threefold: (i) to determine whether sera from lower animal classes have the capacity to inhibit attachment to erythrocytes of various bacterial antigens, (ii) to compare inhibitory capacity of sera from lower and higher animals, and (iii) to learn whether a correlation exists between the inactivation of endotoxin and the inhibition of cell attachment by sera from various animal species. MATERIALS AND METHODS Inhibitory capacity of sera from various animal species was determined and quantitated using the bacterial hemagglutination procedure (7). Normal sera. Samples of sera from the classes Mammalia, Aves, Reptilia, Amphibia, Osteichthyes, Chondrichthyes, Merostomata, Crustacea, and Lamellibranchiata (2) were obtained through the courtesy of the following: Jon Rudbach, University of Montana, Missoula; Michael Sigel, University of Miami, Miami, Fla.; Elias Cohen, Roswell Park Memorial Institute, Buffalo, N.Y.; Dorothy Pollock, Transfusion Service, the Children's Hospital of Buffalo, Buffalo, N.Y.; Jack Cory, Rocky Mountain Laboratory, Hamilton, Mont.; the Buffalo Zoological Gardens, Buffalo, N.Y.; and the animal facilities of the State University of New York at Buffalo and the Children's Hospital of Buffalo, Buffalo, N.Y. Other sera were purchased from Alan Boyden, the Serological Museum, Rutgers University, New Brunswick, N.J.; Colorado Serum Co., Inc., Denver, and the Grand Island Biological Co., Grand Island, N.Y. Antigens. Crude preparations of 0 antigens wereprepared from Sabnonella cholerae-suis and Escherichia coli according to the method described previously by Neter et al. (12). Briefly, organisms were grown on brain veal agar (Difco) in Kolle flasks for 18 h at 37°C, The growth was harvested with 25 ml of phosphate hemagglutination buffer (Difco, pH 7.3), heated for 1 h in a water bath at 100°C, and centrifuged at 22,500 x g for 15 min at 2°C. The resulting supernatant, termed heat-killed supernatant, was decanted and kept frozen until used. Purified LPS from Salmonella montevideo and E. coli 0113, obtained by the phenol-water extraction procedure of Westphal et al. (23), were kindly supplied by Otto WestphaL Max-Planck Institut fur Immunbiologie, Freiburg, Germany, and Jon Rudbach, re-
spectively.
613
The CA from E. coli 014 heat-killed supernatant was prepared in a manner similar to that of crude 0
antigen described above. Vi antigen of Citrobacter ballerup was prepared in the following manner: a colony of the organism from blood agar was inoculated into 6 ml of brain heart infusion broth (Difco) and grown for 4 h in a water bath at 37°C; the broth was transferred to a Kolle flask containing brain veal agar and grown for 18 h at 37°C. The organisms were harvested as described above, and the supernatant was kept frozen for later use. The polyribose-phosphate antigen of Haemophilus influenzae type b was prepared by growing the organisms on chocolate Mueller-Hinton agar plates (150 by 15 mm) for 18 h at 37°C, harvesting the growth with 5 ml of phosphate hemagglutination buffer with 0.5% Formalin, allowing the suspension to stand at 22°C for 4 h, and centrifuging at 22,500 x g for 15 min at 2°C. The supern Lant, containing the antigen, was stored for testing later. The teichoic acid antigen of Staphylococcus aureus was prepared as previously described by Neter et al.
(8).
Unless stated otherwise, crude antigens for modification of erythrocytes were used in a final dilution of 1:10, and purified LPS was used in a final concentration of 12.5 ,ug/ml. Antisera. Antisera to the 0 antigens of S. choleraesuis and E. coli were obtained from patients with the respective infections. Additional 0 and Vi antisera were purchased from Lederle Diagnostics, Pearl River, N.Y. Antisera to CA and teichoic acid antigens were obtained by repeated injections of ethanol-soluble CA from E. coli 07 and S. aureus, respectively, into albino rabbits. Antiserum to H. influenzae type b was obtained from the Commonwealth of Massachusetts, Department of Public Health, Division of Biological Laboratories, Boston. Procedures. Sera to be studied as inhibitors of antigen attachment to erythrocytes were first tested for the presence of antibodies to both the respective antigen and the indicator erythrocytes (human blood group 0) and, when indicated, absorbed as follows. Undiluted serum, 1 to 2 ml, was mixed with the sediment obtained from an equal amount of antigen-modified erythrocytes (see below). The mixtures were incubated at 37°C for 30 min and centrifuged at 1,500 x g for 5 min. If agglutination was observed, the procedure was repeated. Sera whose antibodies could not be removed by three absorptions or that lysed the erythrocytes were not used. Briefly, antigen and test serum, unheated or heated at 56°C for 30 min, in appropriate dilutions, were incubated at 37°C for 30 min. Buffer instead of serum was used for control purposes. The mixtures were then added to the 'sediment of thrice-washed, 2.5% erythrocytes and incubated for 30 min at 37°C. The erythrocytes were then washed three times, resuspended, and, in 0.2-ml amounts, added to equal amounts of homologous antiserum in twofold serial dilutions. The degree of hemagglutination was read and recorded after centrifugation at 1,500 x g for 2 min. A decrease in the titer of the antiserum against the homologous bacterial antigen when compared to
614
INFECT. IMMUN.
PRAINO AND NETER
classes Mammalia and Aves exhibited strong inhibitory capacity. Among the lower vertebrate classes, not previously studied, sera from Reptilia and Osteichthyes displayed moderate inhibitory capacity, whereas those from Amphibia and Chondrichthyes were minrimal in their ability to block antigen attachment to erythrocytes. Sera or hemolymph from all invertebrate classes investigated, namely, Merostomata, Crustacea, and Lamellibranchiata, exhibited little or no inhibitory capacity. Differences in the inhibitory capacity of sera from vertebrates were further substantiated by
the control indicates inhibitory capacity. A 2-fold difference was considered questionable and is listed as ± in Table 1; a 4-fold difference as minimal and listed as 1; an 8-fold difference as moderate and listed as 2; a 16- to 32-fold difference as strong and listed as 3; a >32-fold difference as very strong and listed as 4.
RESULTS Table 1 summarizes the results of the study of the comparative capacity of sera from various animal species to inhibit the attachment of several bacterial antigens to erythrocytes. In accord with previous reports (7, 9, 14), sera from the
TABLE 1. Inhibitory capacity of serum from various animal species on erythrocyte attachment of various bacterial antigens Bacterial antigen tested
Animal serum tested (class and common name)
P
erae-
suis
Vertebrates Mammalia Human Fetal calf Rabbit Guinea pig Aves Turkey Chicken Duck Goose Grouse Pigeon
Reptilia Log turtle Snapper Caiman Painted turtle
Amphibia Bullfrog Lamper eel Siren Cryptobranchus Osteichthyes Carp Conger eel Blue marlin Chondrichthyes Shark Barn door skate Invertebrates Merostomata Horseshoe crab Crustacea English lobster Shore crab Lamellibranchiata
E. coli
monte-
video
4 4
4 4
CA, E. coli
H. influenzae
S. au-
014
Vi C ballerup
type b
4 4
4 4
4 4
4 4
4
4
4
4
4
3-4
LPS S.
to
capac-
ity/class
reus
4
4
3 3 3
3-4
3-4 2-3
2 1 2 3
2-3 3
0 1-2 1-2
0 2 2
2-3 3
3
3
3
3
1
0
±
1
0
±
i,
±
2
2-3
2
±
2-3
4
2-3 2-3
1 ±
2
1-2 0
±
±
1
± ±
± ±
2
2
1
3
1-2
±
0
i
i
1
1
0
0
±
±
±
1 1
0 0
±
±
0
± ±
0-i:
0
1-2
± 1 1 1 1 1-2 ±-l Denotes degree of inhibitory capacity, none (0) to very strong (4) (see Materials and Methods).
Oyster
a
4a 4 4 3-4
-ed-
±-1
VOL. 18, 1977
ERYTHROCYTE ATTACHMENT OF ENDOTOXINS
titration. Highly inhibitory sera, such as mammalian and avian sera, are active in dilutions of approximately 1:1,000, reptilian and osteichthyian sera are active in dilutions of 1:100, and amphibian and chondrichthyian sera are active in a dilution of 1:10 or are even inactive at this dilution. As far as the invertebrates are concerned, most sera were inactive at a dilution of 1:10. Further studies revealed that serum or hemolymph from Limulus (horseshoe crab) and oyster, even when used undiluted, had no significant inhibitory capacity. To determine whether, in a more sensitive system, inhibitory capacity can be detected with serum specimens from invertebrates and lower vertebrates, experiments were carried out utilizing small amounts of antigen for erythrocyte modification. Crude S. cholerae-suis 0 antigen was used in a dilution of 1:150, instead of 1:10. Serum from four out of five invertebrate species showed no increase and the remainder showed only minimally enhanced inhibition of antigen attachment. As far as the lower vertebrates are concerned, a slight increase in the prevention of erythrocyte modification was effected by serum from four out of five species. The serum from the bullfrog exhibited no activity, even under these conditions. It can be seen from the results shown in Table 1 that inhibitory capacity of animal sera, regardless of phylogenetic origin, is nonspecific; i.e., it is independent of antigen used. To evaluate the possible role of complement and other heat-labile factors in inhibitory activity, all sera were also tested after being heated at 560C for 30 min. The results indicate that inactivation of sera did not affect the inhibitory
capacity. DISCUSSION It has been known for many years that certain bacterial antigens become attached to the cell surface of erythrocytes and that pretreatment with tannic acid and similar substances is not required. Antigen attachment is detected by agglutination by homologous antiserum. Springer et al. in 1970 (20) demonstrated that erythrocyte membranes contain a substance that binds LPS of gram-negative bacteria and the CA, but not, in similar amounts, certain other bacterial antigens (20, 24). The isolated erythrocyte membrane component was termed LPS receptor. Numerous other substances, including sera of higher animals, also inhibit antigen attachment to erythrocytes (7, 9-11, 14, 18). The present experiments were suggested by the observation of von Eschen and Rudbach (22), who reported
615
Mammalia
Aves Reptilia
it
Amphibia
Merostomata
I Osteichthyes
Chondrichthyes
\I/ *II! Bilateral
Phyla
FIG. 1. Probable main lines of animal phylogeny (adapted from Dodson's Evolution [1]).
striking differences in the capacity of sera from various animal species to inhibit toxicity of endotoxin. This study was undertaken to determine the capacity of such sera to inhibit antigenic modification of erythrocytes by LPS and other bacterial antigens. This investigation has revealed marked differences in the capacity of sera from various animal species to inhibit attachment of various bacterial antigens to erythrocytes. This activity was greater in sera from higher animal species than in those from lower animal species. The inhibitory activity was not destroyed by heating at 56°C for 30 min. Inhibition is nonspecific since, before testing, antibodies against the various bacterial antigens were removed by absorption. Information is not yet available as to the factors present in the sera of all species tested responsible for inhibition of antigen attachment. It appears that the degree of inhibition closely parallels the evolution of two diverging lines of animal phylogeny based on embryological criteria (Fig. 1). The chordate line culminating in the class Mammalia may have evolved these serum factors at an early stage, possibly at the time corresponding to the rise of the higher fish or sharks. Sera from the class Osteichthyes as a group were stronger inhibitors than those from the class Chondrichthyes, although both may have evolved at a similar time phylogenetically from a placoderm ancestor. Presumably, Osteichthyes gave rise to Amphibia and higher animals, whereas the Chondrichthyes are thought to be a "dead end" in evolution (1). The shark
616
PRAINO AND NETER
was the lowest animal species whose serum displayed minimal inhibitory activity of all sera tested. Comparison of the inhibitory capacity of sera on attachment of bacterial antigens to erythrocytes and their ability to detoxify endotoxin, as reported by von Eschen and Rudbach (22), revealed some striking similarities. Sera and hemolymph from invertebrates have virtually no effect, and sera from higher animal species were highly active in both systems. In both studies variation in the effectiveness of members of several classes of animals was observed. Heating of serum at 56°C did not abolish the activity of serum from various animal species in either system, indicating that the entire complement system is not involved. However, it should be kept in mind that, as shown by Johnson and Ward (5), C5 and C6, even in the absence of C3 activation, may play a role in endotoxin detoxification. In addition, other serum proteins may contribute to detoxification (15-17, 21, 25). The relationship of any of these factors to inhibition of antigen attachment to erythrocytes remains to be elucidated. ACKNOWLEDGMENT This investgtion was supported by Public Health Service grant AI 00658 and training grant 5 TOl Al 00442-05 from the National Institute of Allergy and Infectious Diseases. We also gratefully acknowledge the excellent technical assistnce of Helga von Langendorff.
LrFERATURE CffE 1. Dodson, E. 0. 1960. Evolution: process and product. Reinhold Publishing Corp., New York. 2. Grzimek, L C. B. 1972-1974. Grzimek's animal life encyclopedia. Van Nostrand Reinhold Co., New York. 3. Hayes, L 1951. Specific serum agglutination of sheep erythrocytes sensitized with bacterial polysaccharides. Aust. J. Exp. Biol. Med. Sci. 29:51-62. 4. Hill, G. J., and D. W. Weiss. 1964. Relationships between susceptibility of mice to heat-killed salmonellae and endotoxin and the affinity of their red blood cells for killed organi, p. 422-427. In M. Landy and W. Braun (ed.), Bacterial endotoxina Institute of Microbiology, Rutgers, The State University, New Bnswick, N.J. 5. Johnson, K. J., and P. A. Ward. 1972. The requirement for serum complement in the detoxification of bacterial endotoxin. J. Immunol. 108:611-616. 6. Keogh, E. V., E. A. North, and ML F. Warburton. 1948. Adsorption of bacterial polysaccharides to erythrocytes. Nature (London) 161:687-688. 7. Neter, K. 1956. Bacterial hemagglutination and hemolysis. Bacteriol. Rev. 20:166-188. 8. Neter, E., H. Anai, and E. A. Gorzynid 1960. The
INFECTr. IMMUN. effects of staphylococcal extract and of viable staphylococci on dermal reactivity of the rabbit to epinephrine. J. Infect. Dis. 107:332-340. 9. Neter, E., E. A. Gorzynski, H. Anzai, and V. Bokkenheuser. 1962. Hemagglutination by mixtures of human gamma globulin and heterogenetic bacterial antigen. J. inmunol. 88:411-417. 10. Neter, E., E. A. Gorzynski, 0. Westphal, and 0. Luderitz. 1958. The effects of antibiotics on enterobacterial lipopolysaccharides (endotoxins), hemagglutination and hemolysis. J. Immunol. 80:66-72. 11. Neter, E., E. A. Gorzynsk, 0. Westphal, 0. Luderitz, and D. J. Klumpp. 1958. The effects of protamine and histone on enterobacterial lipopolysaccharides and hemolysis. Can. J. Microbiol. 4:371-383. 12. Neter, E., 0. Westphal, 0. Liideritz, and E. Gorzynski. 1956. The bacterial hemagglutination test for the demonstration of antibodies to Enterobacteriaceae. Ann. N.Y. Acad. Sci. 66:141-156. 13. Neter, E., 0. Westphal, 0. Luderitz, E. A. Gorzynski, and E. Eichenberger. 1956. Studies of enterobacterial lipopolysaccharides. Effects of heat and chemicals on erythrocyte-modifying, antigenic, toxic and pyrogenic properties. J. Immunol. 76:377-385. 14. Neter, E., D. A. Zak, N. J. Zalew8ki, and L F. Bertram. 1952. Inhibition of bacterial (Escherichia coli) modification of erythrocytes. Proc. Soc. Exp. Biol. Med. 80:607-610. 15. Rosen F. S., R. C. Skarnes, IL Landy, and M. J. Shear. 1958. Inactivation of endotoxin by a humoral component. HI. Role of divalent cation and a dialyzable component. J. Exp. Med. 108:701-711. 16. Skarnes, R. C. 1966. The inactivation of endotoxin after interaction with certain proteins of normal serum Ann. N.Y. Acad. Sci. 133:644-662. 17. Skarnes, R. C. 1970. Host defense against bacterial endotoxemia mechanism in normal animals. J. Exp. Med. 132:300-316. 18. Springer, G. F., and J. C. Adye. 1975. Endotoxin-binding substances from human leukocytes and platelets. Infect. Immun. 12:978-986. 19. Springer, G. F., J. C. Adye, A. Bezkorovainy, and J. R. Murthy. 1973. Functional aspects and nature of the lipopolysaccharide-receptor of human erythrocytes. J. Infect. Dis. 128:5202-5212. 20. Springer, G. F., S. V. Huprikar, and E. Neter. 1970. Specific inhibition of endotoxin coating of red cells by a human erythrocyte membrane component. Infect. Immun. 1:98-108. 21. Stauch, J. E., and A. G. Johnson. 1959. The alteration of bacterial endotoxins by human and rabbit serum. J. Immunol. 82:252-263. 22. von Eschen, K. B., and J. A. Rudbach. 1974. Inactivation of endotoxin by serum: a phylogenetic study. J. Infect. Dis. 129:21-27. 23. Westphal 0., 0. Liideritz, and F. Bister. 1952. Ober die Exraktion von Bakterien mit Phenol/Wasser. Z. Naturforwch. Teil B 7:148-155. 24. Whang, H. Y., E. Neter, and G. F. Springer. 1970. Specificity of an erythrocyte membrane receptor for bacterial antigens. Z. Immunitaetsforsch. 140:298-303. 25. Yoshioka, IL, and S. Konno. 1970. Characteristics of endotoxin-altering fractions derived from normal serum.Ill. Isolation and properties of horse serum a2macroglobulin. Infect. Immun. 1:431-439.
