INFECTION AND IMMUNITY, Apr. 1976, p. 1023-1029 Copyright © 1976 American Society for Microbiology
Vol. 13, No. 4 Printed in U.SA.
Production of Highly Cross-Reactive HemagglutinationInhibiting Influenza Antibodies in Ferrets NIC MASUREL AND JOACHIM DRESCHER* Department of Virology, Medical Faculty Rotterdam, Rotterdam, The Netherlands, and Institute of Virology, Medizinische Hochschule Hannover, Hannover-Kleefeld, Germany* Received for publication 15 October 1975
Ferrets were sequentially infected at time intervals of 3 weeks with different influenza virus A strains. It was found that secondary infection can result in the appearance of antibodies highly cross-reacting with a virus strain closely related to the strain of first infection. Such highly cross-reacting antibodies were designated as HCR antibodies. Evidence is presented that HCR antibodies were not antineuraminidase antibodies and, in addition, were not specifically oriented to the strain with which the crossing was observed. When using time intervals of 3 weeks between infections, no increase after secondary infection of antibodies oriented to the strain of first infection was recorded. However, when time intervals of 4 to 5 months between infections were used, secondary infections resulted in an increase of antibodies oriented to the strain of first infection ("original antigenic sin") but not in the appearance of HCR antibodies. In addition, antibodies combining specifically with both infecting strains, designated as doubly specific antibodies, were found. Thus, the conclusion was reached that the original antigenic sin phenomenon and the appearance of HCR antibodies are mutually exclusive events.
Primary infection with influenza virus can condition the antibody response to reinfection with a heterologous but antigenically related virus strain. This so-called "doctrine of the original antigenic sin" (DOS) has been described both for humans andfor other animals (1, 4, 711). Conditioning of the antibody response is recognized by an increase, after reinfection, of antibodies oriented to the strain of first infection. This is at least partially due to the formation of so-called doubly specific antibodies, found to be capable of combining specifically with the strains of both first and secondary infection (5, 7). Furthermore, the antibody response to vaccination with influenza virus can be increased by previous heterotypic infection or vaccination (12). In this paper, evidence is presented that sequential infection of ferrets with different influenza virus strains can result in the appearance after reinfection of antibodies that have a high potential of cross-reaction with strains closely related to, but not identical with, the strain of first infection. Such antibodies were designated as highly cross-reacting (HCR) antibodies. More specifically, it was found that primary infection of ferrets with virus strain "1," followed by reinfection with virus strain "2," can result in the appearance of antibodies that have
a much higher cross-reaction with virus strain "3" than have antisera of comparable antibody titers against strains 1 and 2 obtained after repeated infections with these strains. To examine whether the time interval between infections influences the appearance of HCR antibodies, ferrets were infected at different time intervals. The antibody response was measured by means of hemagglutination inhibition (HI) tests. However, since HI can be caused by both antihemagglutination and antineuraminidase (AN) antibodies (3), sera where HCR antibodies were found were tested for AN antibodies by use of a modification of the photometric antibody concentration unit (ACU) method (J. Drescher and U. Desselberger, Arch. Virol., in press). MATERIALS AND METHODS Virus. The strains of influenza virus used were: A/swine/USA/36 (Hswl Ni), A/WS/England/33 (HO N1), A/Nederland/i/49 (Hi N1), A/PR 8/USA/34 (HO Ni), A/equine/Milford/2/63 (Heq2 Neq2), A/equine/ Miami/l/63 (Heq2 Neq2), A/Sing/i/57 (H2 N2), and A/AA/1/65 (H2 N2) and the recombinant A/Aichi/68A/Bel/34 (H3 Ni). Egg-adapted virus purified by adsorption on and elution from bovine erythrocytes was used for serological tests. Infection of ferrets. Ferrets were sequentially
1023
1024
INFECT. IMMUN.
MASUREL AND DRESCHER
(H3)-A/Bel (Ni) carrying A/Aichi/68 (H3) hemagglutinin and A/Bel/34 (Ni) neuraminidase. The following R values were obtained: 0.251 (A/swine), 0.283 (A/WS), 0.310 (A/PR8), and 0.450 (A/Nederland). RESULTS
infected with each combination of strains A/swine, A/WS, A/PR 8, and A/Nederland. In control experiments, ferrets were repeatedly infected with the same strain of virus. Each animal received 1 ml of virus (100 hemagglutinating doses) intranasally in light ether narcosis. The time interval between infections was 3 weeks in one series of experiments and 4 to 5 months in the second series. Blood samples were drawn prior to each infection and 3 weeks after the last infection. Prior to first infection, none of the animals showed detectable antibody titers against the virus strains used. Antibody titration. Sera were tested repeatedly by means of the photometric HI test (6). Antibody titers were expressed in terms of the reciprocal highest serum dilution (d5,) yielding binding of 50% of 90 to 110 hemagglutinin concentration units (HCUs) (4) of virus. The specificity of antibody titers was examined by determining the ACU titers of sera by means of the photometric ACU method (2) and testing whether or not the reaction of virus with antisera could be described by the isotherm previously established for the reaction of virus with homologous antihemagglutinin antibodies. In addition, sera were tested by use of the hemagglutination inhibition (HI) pattern test. Titers were expressed in terms of the reciprocal highest serum dilution yielding 50% inhibition of three agglutinating doses of virus. Testing of sera for the presence of doubly specific and monospecific antibodies was carried out as described previously (5). Testing for AN antibodies. Sera showing a high degree of cross-reaction with heterologous strains (i.e., virus strains not used for infection of the serum donor) were reacted under the conditions of the photometric ACU method with 100 to 150 HCUs of the corresponding heterologous virus strains. The results obtained were evaluated to determine the constant R, describing the reaction of AN antibodies with neuraminidase as described previously (Drescher and Desselberger, in press). As controls, sera oriented to the virus strains A/ swine, A/WS, A/PR 8, and A/Nederland were tested in like manner, using the recombinant A/Aichi
Determination of the cross-reactivity of antisera. The cross-reactivity of antisera obtained after infection with different influenza virus strains ("polyvalent antisera") was compared with that of antisera containing comparable levels of homologous antibodies raised by infections with the same strain of virus ("monovalent antisera"), as follows. For all monovalent antisera, the d5,, values found with a strain of heterologous virus were divided by the d5, values found with homologous virus, and the geometric means of these ratios were called "cross-reaction indexes" (CRI). Using the titers obtained by means of HI pattern tests, the analogous ratios were computed and called CR1'. Table 1 lists the values found. The d50 values and HI titers against heterologous virus of the polyvalent antisera that could be expected if they had the same cross-reactivity as found with monovalent antisera (d5..x\). and HI,,,,., respectively) were calculated as follows: d50e., = d.5o-CRIj + d50,2,CRI2,4 + d50,3CRI3,4 (equation 1). d={},,, d501,2, and d=,,5(: are the d50 values obtained when testing the polyvalent antiserum with virus strains 1, 2, and 3. CRI,,4, CRI2,4, and CRI:.4 are the CRI values for the cross-reaction with virus strain 4 of antibody oriented to virus strains 1, 2, and 3. By analogy, the HI titers expected were cal4
culated as follows:
HI2CRI'2,4
HIVXP.
=
HI 1CRI'14 +
HI3CRI'3.4
(equation 2). HI1, HI2, and HI3 the HI titers of the polyvalent antiserum obtained for its reaction with virus strains 1, 2, and 3, and CRI'l4, CRI'24, and + are
TABLE 1. Values of the index of cross-reaction Cross-reaction index with virus strain Antibody oriented to virus strain
A/swine/30
A/WS/33
A/PR 8/34
A/Ned./49
CR!a
CRI'b
CRI
CRI'
CRI
CRI'
CRI
CRI'
1.0
1.0
0.02
0.0021
Vol. 13, No. 4 Printed in U.SA.
Production of Highly Cross-Reactive HemagglutinationInhibiting Influenza Antibodies in Ferrets NIC MASUREL AND JOACHIM DRESCHER* Department of Virology, Medical Faculty Rotterdam, Rotterdam, The Netherlands, and Institute of Virology, Medizinische Hochschule Hannover, Hannover-Kleefeld, Germany* Received for publication 15 October 1975
Ferrets were sequentially infected at time intervals of 3 weeks with different influenza virus A strains. It was found that secondary infection can result in the appearance of antibodies highly cross-reacting with a virus strain closely related to the strain of first infection. Such highly cross-reacting antibodies were designated as HCR antibodies. Evidence is presented that HCR antibodies were not antineuraminidase antibodies and, in addition, were not specifically oriented to the strain with which the crossing was observed. When using time intervals of 3 weeks between infections, no increase after secondary infection of antibodies oriented to the strain of first infection was recorded. However, when time intervals of 4 to 5 months between infections were used, secondary infections resulted in an increase of antibodies oriented to the strain of first infection ("original antigenic sin") but not in the appearance of HCR antibodies. In addition, antibodies combining specifically with both infecting strains, designated as doubly specific antibodies, were found. Thus, the conclusion was reached that the original antigenic sin phenomenon and the appearance of HCR antibodies are mutually exclusive events.
Primary infection with influenza virus can condition the antibody response to reinfection with a heterologous but antigenically related virus strain. This so-called "doctrine of the original antigenic sin" (DOS) has been described both for humans andfor other animals (1, 4, 711). Conditioning of the antibody response is recognized by an increase, after reinfection, of antibodies oriented to the strain of first infection. This is at least partially due to the formation of so-called doubly specific antibodies, found to be capable of combining specifically with the strains of both first and secondary infection (5, 7). Furthermore, the antibody response to vaccination with influenza virus can be increased by previous heterotypic infection or vaccination (12). In this paper, evidence is presented that sequential infection of ferrets with different influenza virus strains can result in the appearance after reinfection of antibodies that have a high potential of cross-reaction with strains closely related to, but not identical with, the strain of first infection. Such antibodies were designated as highly cross-reacting (HCR) antibodies. More specifically, it was found that primary infection of ferrets with virus strain "1," followed by reinfection with virus strain "2," can result in the appearance of antibodies that have
a much higher cross-reaction with virus strain "3" than have antisera of comparable antibody titers against strains 1 and 2 obtained after repeated infections with these strains. To examine whether the time interval between infections influences the appearance of HCR antibodies, ferrets were infected at different time intervals. The antibody response was measured by means of hemagglutination inhibition (HI) tests. However, since HI can be caused by both antihemagglutination and antineuraminidase (AN) antibodies (3), sera where HCR antibodies were found were tested for AN antibodies by use of a modification of the photometric antibody concentration unit (ACU) method (J. Drescher and U. Desselberger, Arch. Virol., in press). MATERIALS AND METHODS Virus. The strains of influenza virus used were: A/swine/USA/36 (Hswl Ni), A/WS/England/33 (HO N1), A/Nederland/i/49 (Hi N1), A/PR 8/USA/34 (HO Ni), A/equine/Milford/2/63 (Heq2 Neq2), A/equine/ Miami/l/63 (Heq2 Neq2), A/Sing/i/57 (H2 N2), and A/AA/1/65 (H2 N2) and the recombinant A/Aichi/68A/Bel/34 (H3 Ni). Egg-adapted virus purified by adsorption on and elution from bovine erythrocytes was used for serological tests. Infection of ferrets. Ferrets were sequentially
1023
1024
INFECT. IMMUN.
MASUREL AND DRESCHER
(H3)-A/Bel (Ni) carrying A/Aichi/68 (H3) hemagglutinin and A/Bel/34 (Ni) neuraminidase. The following R values were obtained: 0.251 (A/swine), 0.283 (A/WS), 0.310 (A/PR8), and 0.450 (A/Nederland). RESULTS
infected with each combination of strains A/swine, A/WS, A/PR 8, and A/Nederland. In control experiments, ferrets were repeatedly infected with the same strain of virus. Each animal received 1 ml of virus (100 hemagglutinating doses) intranasally in light ether narcosis. The time interval between infections was 3 weeks in one series of experiments and 4 to 5 months in the second series. Blood samples were drawn prior to each infection and 3 weeks after the last infection. Prior to first infection, none of the animals showed detectable antibody titers against the virus strains used. Antibody titration. Sera were tested repeatedly by means of the photometric HI test (6). Antibody titers were expressed in terms of the reciprocal highest serum dilution (d5,) yielding binding of 50% of 90 to 110 hemagglutinin concentration units (HCUs) (4) of virus. The specificity of antibody titers was examined by determining the ACU titers of sera by means of the photometric ACU method (2) and testing whether or not the reaction of virus with antisera could be described by the isotherm previously established for the reaction of virus with homologous antihemagglutinin antibodies. In addition, sera were tested by use of the hemagglutination inhibition (HI) pattern test. Titers were expressed in terms of the reciprocal highest serum dilution yielding 50% inhibition of three agglutinating doses of virus. Testing of sera for the presence of doubly specific and monospecific antibodies was carried out as described previously (5). Testing for AN antibodies. Sera showing a high degree of cross-reaction with heterologous strains (i.e., virus strains not used for infection of the serum donor) were reacted under the conditions of the photometric ACU method with 100 to 150 HCUs of the corresponding heterologous virus strains. The results obtained were evaluated to determine the constant R, describing the reaction of AN antibodies with neuraminidase as described previously (Drescher and Desselberger, in press). As controls, sera oriented to the virus strains A/ swine, A/WS, A/PR 8, and A/Nederland were tested in like manner, using the recombinant A/Aichi
Determination of the cross-reactivity of antisera. The cross-reactivity of antisera obtained after infection with different influenza virus strains ("polyvalent antisera") was compared with that of antisera containing comparable levels of homologous antibodies raised by infections with the same strain of virus ("monovalent antisera"), as follows. For all monovalent antisera, the d5,, values found with a strain of heterologous virus were divided by the d5, values found with homologous virus, and the geometric means of these ratios were called "cross-reaction indexes" (CRI). Using the titers obtained by means of HI pattern tests, the analogous ratios were computed and called CR1'. Table 1 lists the values found. The d50 values and HI titers against heterologous virus of the polyvalent antisera that could be expected if they had the same cross-reactivity as found with monovalent antisera (d5..x\). and HI,,,,., respectively) were calculated as follows: d50e., = d.5o-CRIj + d50,2,CRI2,4 + d50,3CRI3,4 (equation 1). d={},,, d501,2, and d=,,5(: are the d50 values obtained when testing the polyvalent antiserum with virus strains 1, 2, and 3. CRI,,4, CRI2,4, and CRI:.4 are the CRI values for the cross-reaction with virus strain 4 of antibody oriented to virus strains 1, 2, and 3. By analogy, the HI titers expected were cal4
culated as follows:
HI2CRI'2,4
HIVXP.
=
HI 1CRI'14 +
HI3CRI'3.4
(equation 2). HI1, HI2, and HI3 the HI titers of the polyvalent antiserum obtained for its reaction with virus strains 1, 2, and 3, and CRI'l4, CRI'24, and + are
TABLE 1. Values of the index of cross-reaction Cross-reaction index with virus strain Antibody oriented to virus strain
A/swine/30
A/WS/33
A/PR 8/34
A/Ned./49
CR!a
CRI'b
CRI
CRI'
CRI
CRI'
CRI
CRI'
1.0
1.0
0.02
0.0021
