Complement Activation in Relation to Protective Function of Antibodies to Escherichia coli O Antigens S. AHLSTEDT & L. A. HANSON Department of Immunology, Institute of Medical Microbiology; and Department of Pediatrics; University of Gothenburg, Gothenburg, Sweden
Ahlstedt, S. & Hanson, L. A. Complement Activation in Relation to Protective Function of Antibodies to Escherichia coli O Antigens. Scand. J. Immunol. 4, 511-515, 1975. . The relation between complement activation in an indirect hemolysis assay of £. coli O antibodies and their protective function in experimentally infected mice was analyzed. The antibody amounts and avidities were estimated in antisera and in IgG and IgM antibody fractions by the ammonium sulphate precipitation (ASP) technique of Farr. The protection obtained was due to both IgG and IgM antibodies and showed a close relation to the quantity of antibodies measured by the ASP technique. No link between complement activation capacity of the antibodies and their protective function could be seen. In the system used, the IgM antibodies showed a correlation between antibSdy avidity and complement activating efficiency, whereas the IgG antibodies activated complement poorly. The results do not indicate a major function of the complement in antibody-mediated protection against E. coli in this experimental system in mice. Dr. S. Ahlstedt, Institute of Medical Microbiology, University of Gothenburg, Guldhedsgatan 10, S 413 46 Gothenburg, Sweden
Infections caused by Escherichia coli are prevented by antibodies of the IgG and the IgM classes against the somatic O antigen of the infecting strain (2, 13, 14). We have previously shown that the protective function of the antibodies was influenced by their avidity: antibodies of high avidity were more efficient than those of low avidity (1, 2, 4). However, the functional basis for this difference in efficiency is unknown. Recently we noticed that, at least for IgM antibodies, efficiency in complement activation causing indirect hemolysis was related to the antibody avidity (9). Since complement activation also can result in enhanced chemotaxis, phagocytosis, and bacteriolysis, increasing antibody avidity might add to protective capacity.
In the present study we investigated whether the complement activating capacity of IgM and IgG antibodies against £. coli O antigen relates to their protective function. MATERIALS AND METHODS Antigens. The antigen used for immunization of rabbits was formalin-killed £. colt bacteria of the O6:K13:H1 serotype. A crude lipopolysaccharide (LPS) antigen consisting of a heat extract of E. coli O6:K2:Hl (10) was used in the indirect hemolysis (IHL) method for determining complement activation of antibodies as well as in the indirect hemagglutination (IHA) method. The antigen used for antibody determinations by the ammonium
512
S. Ahlstedt & L. A. Hanson
Table I. Spearman rank correlation coefficients for antibody variables determined in whole sera* Variables
«
IHL-Protection Prot./ASP-Avidity IHA-Protection ASP-Protection IHI^ASP IHL-IHA ASP-IHA IHL-Avidity ASP-Avidity IHL/ASP-Avidity IHA-Avidity
35 40 40 40 66 66 66 66 66 66 66
r
0.243 0.257 0.426 0.619 0.396 0.436 0.324 0.070 -0.009 0.143 -0.079
P
>0.05 >0.05 0.05 >0.05
* Primary data can be obtained from the authors. IHL = indirect hemolysis; IHA = indirect hemagglutination; ASP = ammonium sulphate precipitation.
sulphate precipitation (ASP) technique, a modified Farr assay, was a purified 0 6 LPS (18). Labeling with i^sj (AB Atomenergi, Nykoping, Sweden) was performed with«the chloramine-T method (12), as described by Gupta & Reed (7). Antisera. Fifteen outbred rabbits weighing 2-3 kg were immunized intravenously with formalin-killed E. coli O6 bacteria and bled as previously described (4). In total, samples from 66 bleedings were analyzed. Separation of IgM and IgG antibodies. Gel filtration of 17 antisera with a wide range of antibody titers and avidities through Agarose Bio-gel® A-5m (Bio-Rad Laboratories, Richmond, Calif.) was done as previously described (11) to separate IgM and IgG antibodies. Presence of IgM in the absence of IgG and vice versa in tbe fractions was ascertained by double diffusion-in-gel analysis (20), using antisera specific for these immunoglobulin classes (Nordic Laboratories, Tilburg, the Netherlands). Indirect hemolysis (IHL) assay. The efficiency of tbe antibodies in activating complement was tested in an indirect hemolysis system. Erythrocytes were coated with a crude heatextract of LPS as previously described (10). Tbe antiserum or antibody fraction was serially twofold diluted and incubated with
coated sheep erythrocytes and excess of guinea pig serum as the source of complement. The reciprocal antibody dilution giving 50% hemolysis was registered as the IHL titer (15). Ammonium sttlphat-e precipitation (ASP) technique, a modified Farr assay, was performed using the purified labeled LPS O6 antigen as previously described (3, 6). The antibody titers were expressed as the reciprocal of the serum dilution causing precipitation of 20% of the antigen. Calculatiofi of avidity Ko. The relative avidity of the antisera, expressed as the intrinsic average association constant, was calculated as previously described (3, 6). Indirect hemagglutination (IHA) technique. IHA titers were determined on antisera, as in previous work (5). Protection assays were performed to evaluate the capacity of the antisera to protect intraperitoneally infected mice (13). A detailed description has been given elsewhere (4). Activity per antibody unit was calculated as the ratio between the titer of IHL or protection and the ASP titer (4). The statistical tests used were the Spearman rank correlation test and the Wilcoxon rank sum test (8). RESULTS delation between protective efficiency and serological characteristics of the atitisera Comparison of the IHL titers, reflecting capacity to activate complement, and the protection titers for antibodies in antisera showed that there was no relation between these variables (Table I). In addition, analysis of the ratio values 'IHL titers/protection titers' for antibody groups with Ko ^ 0.25, 0.25 ^ Ko ^ 0.45, and Ko ^ 0.45 showed no significant differences between the groups. The protective capacity per antibody unit (protection titer/ASP titer) did not show any relation to antibody avidity (Table I), indicating that there was no influence of tbis variable in the avidity range studied. In contrast, the protection titers were highly correlated to the ASP and IHA titers (Table I).
Complement and Antibodies to E. coli O Antigens
513
Table II. Spearman rank correlation coefficients for antibody variables determined in IgM antibody fractions*
Table III. Spearman rank correlation coefficient for antibody parameters determined in IgG antibody fractions*
Variables
n
Variables
n
r
P
IHL-Protection Prot./ASP-Avidity IHA-Protection ASP-Protection IHL-ASP IHL-IHA ASP-IHA IHL-Avidity ASP-Avidity IHL/ASP-Avidity IHA-Avidity
17 16 17 17 17 17 17 17 17 17 17
Prot./ASP-Avidity ASP-Avidity Protection-ASP
14 17 14
0.09 O.OU 0.713
> 0.05 > 0.05 < 0.01
r
P
0.076 0.141 0.256 0.528 0.238 0.249 0.542 0.608 -0.347 0.754 0.192
>0.05 >0.05 >0.05 0.05 >0.05 0.05
* Primary data can be obtained from the authors. ASP = ammonium sulphate precipitation.
body unit ('protective titer/ASP titer') of the analyzed IgM and IgG antibody fractions as tested by the Wilcoxon rank sum test (Table IV). DISCUSSION The protective immune mechanisms against Gram-negative bacterial infections are important to elucidate, since such infections are very common, like urinary tract infections, or have a serious prognosis, like sepsis and meningitis. Cell-mediated immunity does not seem to provide any significant protection against E. coli bacteria causing experimental pyelonephritis in rats and rabbits (16). In contrast, antibodies against E. coli O antigens are efficiently protective against such experimental pyelonephritis and against intraperitoneal infections in mice (1, 2, 4, 13). The mode of function of these protective antibodies is not obvious, however. Thus it is unknown whether activation of complement, resulting in enhanced phagocytosis, inflammation, and bacteriolysis, is involved. This study does not provide any evidence of a relation between the protective and complement-activating capacities of the antibodies. This statement must be limited to the IgM antibodies, however, since the IgG antibodies studied were so poor in complement activation that their capacity could not be evaluated. On the other hand, the protection obtained with the antisera obviously was due to both the IgG and IgM antibodies. Thus the protective titers were closely related to the antibody quantities determined by the ASP technique for the antisera as well
as the IgG and IgM antibody fractions. Further, the protective activity per antibody unit (4) was similar for the IgG and IgM antibodies. It is not surprising that a bacteriolytic activity of complement-activating E. coli antibodies is of no significance for protection, because E. coli causing infections in humans is very frequently resistant to the bactericidal activity of serum (17). On the other hand, the antibody-induced complement-mediated promotion of phagocytosis and inflammatory reactions was expected to be more important. Again, there was no evidence of this, which agrees with the fact that IgG and IgM antibodies are equally protective, although only IgM was efficient in complement activation. It could also be argued, however, that complement may on its own, through the alternative pathway and effects on C5 (19), induce chemotaxis, immune adherence, and anaphylatoxin release, thereby enhancing phagocytosis and inflammation. It should be added that the experimental model used is complex, and the possibility cannot be excluded that an additional effect of complement activation may have been lost among the many defense factors at work. ACKNOWLEDGEMENTS The skilful technical assistance of Mrs. Birgitta Andersson and Miss Lena Beijer is gratefully acknowledged. The work was supported by the Medical Research Council (215) and the Valter, Ellen and Lennart Hesselman Foundation for Scientific Research.
REFERENCES 1. Ahlstedt, S., Holmgren, J. & Hanson, L. A. Significance of amount and avidity of £. coli O antibodies for manifestation of their serological and protective properties. A preliminary study. Int. Arch. Allergy 42, 826, 1972. 2. Ahlstedt, S., Holmgren, J. & Hanson, L. A. The primary and secondary antibody response to Escheriehia coli 06 lipopolysaccharide analysed at the humoral and cellular level. Amount and avidity of the antibodies in relation to protective capacity. Immunology 24, 191, 1973. 3. Ahlstedt, S., Holmgren, J. & Hanson, L. A. The validity of the ammonium sulphate precipitation
Complement and Antibodies to E. coli 0 Antigens technique of estimation of antibody amount and avidity. Immunology 25, 917, 1973. 4. Ahlstedt, S., Holmgren, J. & Hanson, L. A. Protective capacity of antibodies against E. coli O antigen with special reference to the avidity. Int. Arch. Allergy 46, 470, 1974. 5. Ahlstedt, S., Jodal, U., Hanson, L. A. & Holmgren, J. Quantitation of Escherichia coli O antibodies by direct and indirect agglutination in comparison with a radio immuno assay. Int. Arch. Allergy 48, 455, 1975. 6. Ahlstedt, S. & Kaijser, B. Experimental Escherichia coli O6 pyelonephritis in rabbits. Effect on O6 antibody quantity and avidity of prior immunization with E. coli O2 bacteria. Immunology 28, 841, 1975. 7. Gupta, J. D. & Reed, C. E. The direct reaction between Salmonella enteritidis endotoxin and antibody. Measurement of 7S and 19S antibody in normal, tolerant and immune sera. /. Immunol. 98, 1093, 1967. 8. Hajek, J. & Sidak, Z. Theory of the Rank Tests. Academic Press, New York & London, 1967. 9. Hanson, L. A., Ahlstedt, S., Kaijser, B., Goldblum, R. & Holmgren, J. Protection or tissue damage by the antibody response induced by crossreacting antigens to Escherichia coli O and K antigens. In The Immune System and Infectious Diseases, IV Convocation of Immunology, Buffalo, 1974. In press. 10. Holmgren, J. Studies of methods for quantitation of agglutinins and precipitins to Escherichia coli O and K antigen. Int. Arch. Allergy 37, 380, 1970. 11. Holmgren, J., Svennerholm, A. M. & Ouchterlony, 0. Experimental studies on cholera immunization. 1. The response of neutralization and vibriocidal antibodies after immunization with culture filtrate material from V. cholerae. Acta path, microbiol. scand. Sect. B. 80, 429, 1972. Received 27 February 1975 Received in revised form 7 May 1975
515
12. Hunter, V. M. & Greenwood, F. C. Preparation of iodine-131 labeled human growth hormone with specific activity. Nature (Lond.) 194, 495, 1962. 13. Kaijser, B., Holmgren, J. & Hanson, L. A. The protective effect against E. coli of O and K antibodies of different immunoglobulin classes. Scand. J. Immunol. 1, 27, 1972. 14. Kaijser, B. & Oiling, S. Analysis of the antibody response and its protective capacity in experimental, hematogenous pyelonephritis caused by Escherichia coli in rabbits. /. infect. Dis. 128, 41, 1973. 15. Mayer, M. M. Complement and complement fixation, p. 133 in Kabat, E. A. & Mayer, M. M. (eds.). Experimental lmmunochemistry, 2nd ed. Charles C Thomas, Publisher, Springfield, 111., 1961. 16. Miller, T. B. & North, J. D. K. Host response in urinary tract infections. Kidney Int. 5, 179, 1974. 17. Oiling, S., Hanson, L. A., Holmgren, J., Jodal, U., Lincoln, K. & Lindberg, U. The bactericidal effect of normal human serum on E. coli strains from normals and from patients with urinary tract infections. Infection 1, 24, 1973. 18. 0rskov, F., 0rskov, I., Jann, B., Jann, K., MulIerSeitz, E. & Westphal, O. lmmunochemistry of Escherichia coli O antigens. Acta path, microbiol. scand. 71, 339, 1967. 19. Snyderman, R., Shin, H. S., Phillips, J. K., Gewurz, H. & Mergenhagen, S. E. A neutrophil chemotactic factor derived from C'5 upon interaction of guinea pig serum with endotoxin. /. Immunol. 103, 413, 1969. 20. Wadsworth, C. A slide microtechnique for the analysis of immune precipitates in gel. Int. Arch. Allergy 10, 355, 1957.
Ahlstedt, S. & Hanson, L. A. Complement Activation in Relation to Protective Function of Antibodies to Escherichia coli O Antigens. Scand. J. Immunol. 4, 511-515, 1975. . The relation between complement activation in an indirect hemolysis assay of £. coli O antibodies and their protective function in experimentally infected mice was analyzed. The antibody amounts and avidities were estimated in antisera and in IgG and IgM antibody fractions by the ammonium sulphate precipitation (ASP) technique of Farr. The protection obtained was due to both IgG and IgM antibodies and showed a close relation to the quantity of antibodies measured by the ASP technique. No link between complement activation capacity of the antibodies and their protective function could be seen. In the system used, the IgM antibodies showed a correlation between antibSdy avidity and complement activating efficiency, whereas the IgG antibodies activated complement poorly. The results do not indicate a major function of the complement in antibody-mediated protection against E. coli in this experimental system in mice. Dr. S. Ahlstedt, Institute of Medical Microbiology, University of Gothenburg, Guldhedsgatan 10, S 413 46 Gothenburg, Sweden
Infections caused by Escherichia coli are prevented by antibodies of the IgG and the IgM classes against the somatic O antigen of the infecting strain (2, 13, 14). We have previously shown that the protective function of the antibodies was influenced by their avidity: antibodies of high avidity were more efficient than those of low avidity (1, 2, 4). However, the functional basis for this difference in efficiency is unknown. Recently we noticed that, at least for IgM antibodies, efficiency in complement activation causing indirect hemolysis was related to the antibody avidity (9). Since complement activation also can result in enhanced chemotaxis, phagocytosis, and bacteriolysis, increasing antibody avidity might add to protective capacity.
In the present study we investigated whether the complement activating capacity of IgM and IgG antibodies against £. coli O antigen relates to their protective function. MATERIALS AND METHODS Antigens. The antigen used for immunization of rabbits was formalin-killed £. colt bacteria of the O6:K13:H1 serotype. A crude lipopolysaccharide (LPS) antigen consisting of a heat extract of E. coli O6:K2:Hl (10) was used in the indirect hemolysis (IHL) method for determining complement activation of antibodies as well as in the indirect hemagglutination (IHA) method. The antigen used for antibody determinations by the ammonium
512
S. Ahlstedt & L. A. Hanson
Table I. Spearman rank correlation coefficients for antibody variables determined in whole sera* Variables
«
IHL-Protection Prot./ASP-Avidity IHA-Protection ASP-Protection IHI^ASP IHL-IHA ASP-IHA IHL-Avidity ASP-Avidity IHL/ASP-Avidity IHA-Avidity
35 40 40 40 66 66 66 66 66 66 66
r
0.243 0.257 0.426 0.619 0.396 0.436 0.324 0.070 -0.009 0.143 -0.079
P
>0.05 >0.05 0.05 >0.05
* Primary data can be obtained from the authors. IHL = indirect hemolysis; IHA = indirect hemagglutination; ASP = ammonium sulphate precipitation.
sulphate precipitation (ASP) technique, a modified Farr assay, was a purified 0 6 LPS (18). Labeling with i^sj (AB Atomenergi, Nykoping, Sweden) was performed with«the chloramine-T method (12), as described by Gupta & Reed (7). Antisera. Fifteen outbred rabbits weighing 2-3 kg were immunized intravenously with formalin-killed E. coli O6 bacteria and bled as previously described (4). In total, samples from 66 bleedings were analyzed. Separation of IgM and IgG antibodies. Gel filtration of 17 antisera with a wide range of antibody titers and avidities through Agarose Bio-gel® A-5m (Bio-Rad Laboratories, Richmond, Calif.) was done as previously described (11) to separate IgM and IgG antibodies. Presence of IgM in the absence of IgG and vice versa in tbe fractions was ascertained by double diffusion-in-gel analysis (20), using antisera specific for these immunoglobulin classes (Nordic Laboratories, Tilburg, the Netherlands). Indirect hemolysis (IHL) assay. The efficiency of tbe antibodies in activating complement was tested in an indirect hemolysis system. Erythrocytes were coated with a crude heatextract of LPS as previously described (10). Tbe antiserum or antibody fraction was serially twofold diluted and incubated with
coated sheep erythrocytes and excess of guinea pig serum as the source of complement. The reciprocal antibody dilution giving 50% hemolysis was registered as the IHL titer (15). Ammonium sttlphat-e precipitation (ASP) technique, a modified Farr assay, was performed using the purified labeled LPS O6 antigen as previously described (3, 6). The antibody titers were expressed as the reciprocal of the serum dilution causing precipitation of 20% of the antigen. Calculatiofi of avidity Ko. The relative avidity of the antisera, expressed as the intrinsic average association constant, was calculated as previously described (3, 6). Indirect hemagglutination (IHA) technique. IHA titers were determined on antisera, as in previous work (5). Protection assays were performed to evaluate the capacity of the antisera to protect intraperitoneally infected mice (13). A detailed description has been given elsewhere (4). Activity per antibody unit was calculated as the ratio between the titer of IHL or protection and the ASP titer (4). The statistical tests used were the Spearman rank correlation test and the Wilcoxon rank sum test (8). RESULTS delation between protective efficiency and serological characteristics of the atitisera Comparison of the IHL titers, reflecting capacity to activate complement, and the protection titers for antibodies in antisera showed that there was no relation between these variables (Table I). In addition, analysis of the ratio values 'IHL titers/protection titers' for antibody groups with Ko ^ 0.25, 0.25 ^ Ko ^ 0.45, and Ko ^ 0.45 showed no significant differences between the groups. The protective capacity per antibody unit (protection titer/ASP titer) did not show any relation to antibody avidity (Table I), indicating that there was no influence of tbis variable in the avidity range studied. In contrast, the protection titers were highly correlated to the ASP and IHA titers (Table I).
Complement and Antibodies to E. coli O Antigens
513
Table II. Spearman rank correlation coefficients for antibody variables determined in IgM antibody fractions*
Table III. Spearman rank correlation coefficient for antibody parameters determined in IgG antibody fractions*
Variables
n
Variables
n
r
P
IHL-Protection Prot./ASP-Avidity IHA-Protection ASP-Protection IHL-ASP IHL-IHA ASP-IHA IHL-Avidity ASP-Avidity IHL/ASP-Avidity IHA-Avidity
17 16 17 17 17 17 17 17 17 17 17
Prot./ASP-Avidity ASP-Avidity Protection-ASP
14 17 14
0.09 O.OU 0.713
> 0.05 > 0.05 < 0.01
r
P
0.076 0.141 0.256 0.528 0.238 0.249 0.542 0.608 -0.347 0.754 0.192
>0.05 >0.05 >0.05 0.05 >0.05 0.05
* Primary data can be obtained from the authors. ASP = ammonium sulphate precipitation.
body unit ('protective titer/ASP titer') of the analyzed IgM and IgG antibody fractions as tested by the Wilcoxon rank sum test (Table IV). DISCUSSION The protective immune mechanisms against Gram-negative bacterial infections are important to elucidate, since such infections are very common, like urinary tract infections, or have a serious prognosis, like sepsis and meningitis. Cell-mediated immunity does not seem to provide any significant protection against E. coli bacteria causing experimental pyelonephritis in rats and rabbits (16). In contrast, antibodies against E. coli O antigens are efficiently protective against such experimental pyelonephritis and against intraperitoneal infections in mice (1, 2, 4, 13). The mode of function of these protective antibodies is not obvious, however. Thus it is unknown whether activation of complement, resulting in enhanced phagocytosis, inflammation, and bacteriolysis, is involved. This study does not provide any evidence of a relation between the protective and complement-activating capacities of the antibodies. This statement must be limited to the IgM antibodies, however, since the IgG antibodies studied were so poor in complement activation that their capacity could not be evaluated. On the other hand, the protection obtained with the antisera obviously was due to both the IgG and IgM antibodies. Thus the protective titers were closely related to the antibody quantities determined by the ASP technique for the antisera as well
as the IgG and IgM antibody fractions. Further, the protective activity per antibody unit (4) was similar for the IgG and IgM antibodies. It is not surprising that a bacteriolytic activity of complement-activating E. coli antibodies is of no significance for protection, because E. coli causing infections in humans is very frequently resistant to the bactericidal activity of serum (17). On the other hand, the antibody-induced complement-mediated promotion of phagocytosis and inflammatory reactions was expected to be more important. Again, there was no evidence of this, which agrees with the fact that IgG and IgM antibodies are equally protective, although only IgM was efficient in complement activation. It could also be argued, however, that complement may on its own, through the alternative pathway and effects on C5 (19), induce chemotaxis, immune adherence, and anaphylatoxin release, thereby enhancing phagocytosis and inflammation. It should be added that the experimental model used is complex, and the possibility cannot be excluded that an additional effect of complement activation may have been lost among the many defense factors at work. ACKNOWLEDGEMENTS The skilful technical assistance of Mrs. Birgitta Andersson and Miss Lena Beijer is gratefully acknowledged. The work was supported by the Medical Research Council (215) and the Valter, Ellen and Lennart Hesselman Foundation for Scientific Research.
REFERENCES 1. Ahlstedt, S., Holmgren, J. & Hanson, L. A. Significance of amount and avidity of £. coli O antibodies for manifestation of their serological and protective properties. A preliminary study. Int. Arch. Allergy 42, 826, 1972. 2. Ahlstedt, S., Holmgren, J. & Hanson, L. A. The primary and secondary antibody response to Escheriehia coli 06 lipopolysaccharide analysed at the humoral and cellular level. Amount and avidity of the antibodies in relation to protective capacity. Immunology 24, 191, 1973. 3. Ahlstedt, S., Holmgren, J. & Hanson, L. A. The validity of the ammonium sulphate precipitation
Complement and Antibodies to E. coli 0 Antigens technique of estimation of antibody amount and avidity. Immunology 25, 917, 1973. 4. Ahlstedt, S., Holmgren, J. & Hanson, L. A. Protective capacity of antibodies against E. coli O antigen with special reference to the avidity. Int. Arch. Allergy 46, 470, 1974. 5. Ahlstedt, S., Jodal, U., Hanson, L. A. & Holmgren, J. Quantitation of Escherichia coli O antibodies by direct and indirect agglutination in comparison with a radio immuno assay. Int. Arch. Allergy 48, 455, 1975. 6. Ahlstedt, S. & Kaijser, B. Experimental Escherichia coli O6 pyelonephritis in rabbits. Effect on O6 antibody quantity and avidity of prior immunization with E. coli O2 bacteria. Immunology 28, 841, 1975. 7. Gupta, J. D. & Reed, C. E. The direct reaction between Salmonella enteritidis endotoxin and antibody. Measurement of 7S and 19S antibody in normal, tolerant and immune sera. /. Immunol. 98, 1093, 1967. 8. Hajek, J. & Sidak, Z. Theory of the Rank Tests. Academic Press, New York & London, 1967. 9. Hanson, L. A., Ahlstedt, S., Kaijser, B., Goldblum, R. & Holmgren, J. Protection or tissue damage by the antibody response induced by crossreacting antigens to Escherichia coli O and K antigens. In The Immune System and Infectious Diseases, IV Convocation of Immunology, Buffalo, 1974. In press. 10. Holmgren, J. Studies of methods for quantitation of agglutinins and precipitins to Escherichia coli O and K antigen. Int. Arch. Allergy 37, 380, 1970. 11. Holmgren, J., Svennerholm, A. M. & Ouchterlony, 0. Experimental studies on cholera immunization. 1. The response of neutralization and vibriocidal antibodies after immunization with culture filtrate material from V. cholerae. Acta path, microbiol. scand. Sect. B. 80, 429, 1972. Received 27 February 1975 Received in revised form 7 May 1975
515
12. Hunter, V. M. & Greenwood, F. C. Preparation of iodine-131 labeled human growth hormone with specific activity. Nature (Lond.) 194, 495, 1962. 13. Kaijser, B., Holmgren, J. & Hanson, L. A. The protective effect against E. coli of O and K antibodies of different immunoglobulin classes. Scand. J. Immunol. 1, 27, 1972. 14. Kaijser, B. & Oiling, S. Analysis of the antibody response and its protective capacity in experimental, hematogenous pyelonephritis caused by Escherichia coli in rabbits. /. infect. Dis. 128, 41, 1973. 15. Mayer, M. M. Complement and complement fixation, p. 133 in Kabat, E. A. & Mayer, M. M. (eds.). Experimental lmmunochemistry, 2nd ed. Charles C Thomas, Publisher, Springfield, 111., 1961. 16. Miller, T. B. & North, J. D. K. Host response in urinary tract infections. Kidney Int. 5, 179, 1974. 17. Oiling, S., Hanson, L. A., Holmgren, J., Jodal, U., Lincoln, K. & Lindberg, U. The bactericidal effect of normal human serum on E. coli strains from normals and from patients with urinary tract infections. Infection 1, 24, 1973. 18. 0rskov, F., 0rskov, I., Jann, B., Jann, K., MulIerSeitz, E. & Westphal, O. lmmunochemistry of Escherichia coli O antigens. Acta path, microbiol. scand. 71, 339, 1967. 19. Snyderman, R., Shin, H. S., Phillips, J. K., Gewurz, H. & Mergenhagen, S. E. A neutrophil chemotactic factor derived from C'5 upon interaction of guinea pig serum with endotoxin. /. Immunol. 103, 413, 1969. 20. Wadsworth, C. A slide microtechnique for the analysis of immune precipitates in gel. Int. Arch. Allergy 10, 355, 1957.
