Search

The antigenic and genetic characteristics of the haemagglutinins of influenza type B viruses isolated since 1988 during periods of both widespread activity (1990/1991) and sporadic activity (1989/1990) were examined using microneutralization tests and direct RNA sequencing. During 1989/1990, influenza B viruses representative of two distinct lineages antigenically and genetically related to either B/Victoria/2/87 or B/Yamagata/16/88 were isolated, and a minor drift variant of B/Yamagata/16/88, B/Hong Kong/22/89, was identified. In 1990/1991, B/Hong Kong/22/89- or B/Yamagata/16/88-like viruses accounted for the majority of the influenza virus isolates in most countries. Sequence analysis of the HA1 domains of representative viruses confirmed the continued existence of two main lineages among recent strains of influenza B virus and identified unique amino acid changes that could account for the altered antigenic reactivity of some variants. Sequence analysis of the HA2 domains of some of the recent influenza B viruses allowed for a comparison of the evolutionary rates and patterns between the HA1 and HA2 domains.

Monoclonal antibodies raised against the separated hemagglutinin subunits (HA1 and HA2) of influenza A/Vic/3/75 (H3N2) virus were tested against a large panel of human and avian strains. The epitopes recognized by most antibodies were conserved among subtype H3 viruses, but reactivity of some antibodies with members of other subtypes was also observed. Particularly, the H4 virus reacted with most antibodies directed against the HA2 subunit. These results are discussed in terms of sequence similarities between subtypes and application of these antibodies as subtyping reagents.

The protective efficacy of influenza hemagglutinin expressed from recombinant vaccinia virus was compared with that induced by inactivated or infectious influenza vaccines. Intraperitoneal and intranasal routes of vaccination were compared. All the vaccines except the intranasally administered, inactivated vaccine induced detectable levels of neutralizing and hemagglutination-inhibiting antibodies in the serum of mice at 28 days postvaccination. Immunization with any of the intranasally administered vaccines reduced the amount of influenza virus nucleoprotein antigen in lungs after challenge with a homologous, mouse-adapted strain of influenza virus. Intraperitoneally administered vaccines failed to provide such protection. These results indicated that the route of vaccine administration may be the most critical factor for inducing protective immunity. The results also showed that in this mouse model a recombinant DNA-based vaccine could provide protection equivalent to that provided by conventional attenuated and inactivated influenza vaccines.

Mice and guinea pigs were immunized with the haemagglutinin (HA) of influenza B-USSR/100 virus, either in Syntex Adjuvant Formulation-1 (SAF-1) or in saline. Antibody titres were determined by ELISA, haemagglutination inhibition and virus neutralization. Animals immunized with HA in SAF-1 had significantly higher antibody titres than did animals immunized with HA in saline. Both 3-week-old and 13 1/2-month-old mice had greater and more uniform antibody responses to HA in SAF-1 than to HA in saline.

Antibodies to the H3 hemagglutinin of influenza A virus could be specifically measured by single radial hemolysis (SRH) when test antigens were recombinant viruses containing the relevant H3 hemagglutinin antigen and irrelevant Neq1 neuraminidase of A/equine/Prague/1/56 virus. Antibodies to influenza B virus could also be measured by the SRH technique. Antibody rises to influenza A or B virus measured by SRH agreed with results of hemagglutination inhibition (HI) tests for about 80% of the sera tested, including sera from volunteers receiving killed influenza vaccine and sera from patients naturally infected with influenza. Correlation between antibody titers measured by SRH and HI was also good. Antibodies to the N2 neuraminidase of influenza A virus could be specifically measured by SRH when test antigens were recombinant viruses containing the relevant N2 neuraminidase antigen and irrelevant Heq1 hemagglutinin of A/equine/Prague/1/56 virus. The SRH test for neuraminidase antibodies was more strain specific than was the SRH test for hemagglutinin antibodies. Probably for this reason, agreement between neuraminidase antibody determinations in human sera by the SRH test and by the neuraminidase inhibition test was poorer than agreement between the SRH test for hemagglutinin antibodies and the HI test.

The sequence of the HA1 subunit region of the haemagglutinin gene of influenza A/USSR/90/77, and A/Brazil/11/78, A/Lackland/3/78, A/England/333/80 and A/India/6263/80 was determined by dideoxy-sequencing methods using total virion RNA and specific oligonucleotide primers for reverse transcriptase. These 1977-1980 strains share a minimum of 85% amino acid sequence homology with influenza A/PR/8/34. Most of the surface amino acid substitutions which occurred during the evolution of A/PR/8/34 to A/USSR/90/77 and subsequently in the 1978-1980 strains are located in the 4 antigenic sites previously defined by an analysis of laboratory-selected mutants of A/PR/8/34. We deduce an evolutionary pathway for the 1977-80 strains and suggest their different epidemic properties may be a consequence of only a few amino acid changes.

Surveys of the antigenic properties of a wide range of variants of the H3N2 (Hong Kong) influenza virus subtype have revealed complex patterns of variants cocirculating during each of the main epidemic eras of the subtype. We determined hemagglutinin (HA) gene sequences for 14 isolates chosen to give the wildest possible spread of variant types. The addition of these data to existing HA gene sequence information for other variants provides a comprehensive picture of HA gene evolution during antigenic drift among H3N2 subtype viruses. The data reveal the existence of multiple evolutionary pathways during at least one period of development of the subtype and strikingly demonstrate that amino acid changes are limited to a small number of locations on the HA molecule during antigenic drift. The occurrence of sequential amino acid changes at key positions within these variable regions suggests that the HA structure has remained constant during subtype evolution so that only limited possibilities remain for further antigenic drift among H3N2 viruses.

Mouse monoclonal antibodies specific for human immunoglobulin isotypes were investigated for use in an isotype-specific enzyme immunoassay for detection of antibody to influenza type A hemagglutinin (H1 and H3). The monoclonal antibody reagents were compared with isotype-specific, hyperimmune rabbit antisera from the National Institutes of Health. Endpoint titers for immunoglobulin G (IgG) obtained with the two reagents were within fourfold of each other 84% of the time (79 of 84) and within eightfold of each other 95% of the time (89 of 94). Regression analysis of the data gave a multiple correlation coefficient (r2) of 0.77 and a Spearman rank value of 0.83 (P less than 0.001). For IgA reagents, endpoint titers agreed within fourfold 77% of the time (88 of 114) and within eightfold 92% of the time (105 of 114). The r2 was 0.73, and Spearman rank was 0.83 (P less than 0.001). IgM antibody was detected in only 17 of 114 sera by either monoclonal or polyclonal reagents. Of these sera, 14 (82%) gave titers with the two reagents that were within fourfold of each other. A similar number of fourfold titer rises were detected with each reagent in paired sera showing hemagglutination inhibition titer rises. Monoclonal antibody reagents detected 27 IgA, 29 IgG, and 6 IgM rises, while polyclonal antisera detected 26 IgA, 31 IgG, and 7 IgM rises. These results show that monoclonal antibodies specific for human immunoglobulin isotypes are suitable as reagents for diagnostic assays. The advantages of monoclonal antibodies are their high degree of specificity and the ability to be standardized and produced in unlimited quantities. Moreover, the availability of immunoglobulin subclass- and allotype-specific monoclonal antibodies will enable a more detailed analysis of the antibody response to influenza as well as other infectious agents.

We have determined nucleotide sequences of the HA1 portion of the hemagglutinin (HA) gene of the parental A/USSR/90/70 (H1N1) virus and its eight variants selected in vitro with six monoclonal antibodies to study antigenic determinants. The HA1 gene of one of the variants (B-1-23) was cloned in bacteria and its nucleotide sequence was determined by the Maxam-Gilbert method. The nucleotide sequence of the variant was confirmed by the dideoxy chain termination method. The gene sequences of the other viruses were determined by the latter method. Three variants with reduced reactivity in HI test only with the selecting antibodies possessed one amino acid substitution. On the other hand, most other variants which had the changed reactivity to multiple antibodies in HI test possessed more than one substitution. Comparison of the amino acid sequences of the HA1 molecule, deduced from the nucleotide sequences, suggested that the monoclonal antibodies W18 and 264 reacted with epitopes located on the area involving amino acid residues 125C and 189-190, respectively, whereas, the antibodies 22 and 70 reacted with epitopes involving amino acid residues 129, 132, and 157. The epitope recognized by antibody 110 overlapped with that of W18, and the epitope recognized by antibody 385 was located on the area involving at least amino acid residues 129, 159, and 189, which overlapped with some of the above epitopes. The sequence analysis with B-1-23 variant selected with antibody 264 clearly showed that in A/USSR/77 viruses, a single substitution at amino acid residue 190 effectively changes the epitope and caused a significant antigenic variation detectable by postinfection ferret sera.

B/Singapore/222/79-like influenza viruses isolated from three patients during the winter of 1981 to 1982 and cultured in either embryonated hens' eggs or MDCK cells were studied. Sequence analysis indicated that the haemagglutinin (HA) genes of the six virus preparations contained at least four distinct HA1 sequences which differed by up to six amino acids. Only one pair of viruses had amino acid differences between the egg- and MDCK cell-derived viral subpopulations and this change did not affect a glycosylation site. Mice infected with previously described recombinant vaccinia viruses expressing either the egg- or MDCK cell-derived HA of B/England/222/82 developed neutralizing antibodies against all of the 1982 type B viruses and were protected against intranasal challenge with these viruses. Therefore, in this model system, the minor sequence variation between the HAs of egg- and MDCK cell-derived influenza B/England/222/82 virus had no detectable effect on the induction of cross-protection.

Evolution of the antigenic structure of influenza virus hemagglutinin with the antigenic formula HA1 was studied by the determination of the capacity for interaction with monoclonal antibody and aminoacid substitutions in the protein. Consecutive changes in protein epitopes were found in isolates accumulating from the time of isolation of H1 influenza viruses in 1977.

To compare the immune response induced by influenza hemagglutinin (HA) variants differing by a single amino acid, the genes for each HA variant were cloned and expressed in vaccinia virus. These variant HA genes have been previously described as being present exclusively in either egg-derived or MDCK cell-derived subpopulations of influenza B/England/222/82 virus. By using this approach we were able to vaccinate animals with homogeneous preparations of these viral antigens and thus circumvent the problem of heterogeneity within RNA virus stocks. Immunization and challenge experiments in mice indicated that even though vaccination with the recombinant vaccinia viruses induced different levels of cross-reactive neutralizing antibodies, mice vaccinated with either recombinant vaccinia virus were protected from infection with either subpopulation of influenza virus. Results with this model system support the view that influenza vaccines prepared with egg-derived virus should be protective against microvariants of virus that grow preferentially in MDCK and possibly other mammalian cells.

Results of analysing antigenic variation in the haemagglutinin (HA) molecule of naturally occurring influenza A (H1N1) viruses from 1977 to 1979 with monoclonal antibodies were found to be dependent in some instances on the test system used. In several instances A/USSR/90/77 HA-specific monoclonal antibodies had sharply reduced haemagglutination-inhibition (HI) titres with variant virus although they bound to the variant and A/USSR/90/77 HAs with similar efficiencies as judged by titration in a sensitive and accurate solid-phase immunofluorimetric assay. In another instance, the converse situation was observed: monoclonal antibodies having a reduced efficiency of binding to the HA of a variant virus nevertheless had comparable HI titres with the variant and with A/USSR/90/77. The chemical basis and epidemiological significance of these observations remain to be elucidated. Nevertheless, the finding that the reaction of monoclonal antibodies can, in some cases, be markedly dependent on the test system employed is of significance for the efficient design and correct interpretation of immunochemical studies which employ monoclonal antibodies to investigate the basis for variation in influenza strains.

Nucleotide sequence analysis of the region of the haemagglutinin gene coding for the HA1 domain of the protein was performed on 19 human influenza A strains of H1 subtype representative of the two epidemic periods from 1977-1983 and from 1950-1957. The amino acid changes relative to A/USSR/90/77 are summarised and are consistent with the view that variation in these field strains involves changes largely at the Sb and Ca antigenic sites previously characterised in laboratory mutants of the haemagglutinin of influenza A/PR/8/34. The Sa and Cb sites are less variant and are probably masked by carbohydrate side chains. We discuss the significance of other amino acid changes which do not correspond to previously defined antigenic sites. We also define the "mainstream" amino acid changes characteristic of the divergent evolutionary pathways of the 1950-1957 and 1977-1983 periods and note that the rate of evolution is faster in the earlier period.