Rhinoviruses A and C elicit long-lasting antibody responses with limited cross-neutralization

Abstract

Rhinoviruses (RVs) can cause severe wheezing illnesses in young children and patients with asthma. Vaccine development has been hampered by the multitude of RV types with little information about cross-neutralization. We previously showed that neutralizing antibody (nAb) responses to RV-C are detected twofold to threefold more often than those to RV-A throughout childhood. Based on those findings, we hypothesized that RV-C infections are more likely to induce either cross-neutralizing or longer-lasting antibody responses compared with RV-A infections. We pooled RV diagnostic data from multiple studies of children with respiratory illnesses and compared the expected versus observed frequencies of sequential infections with RV-A or RV-C types using log-linear regression models. We tested longitudinally collected plasma samples from children to compare the duration of RV-A versus RV-C nAb responses. Our models identified limited reciprocal cross-neutralizing relationships for RV-A (A12-A75, A12-A78, A20-A78, and A75-A78) and only one for RV-C (C2-C40). Serologic analysis using reference mouse sera and banked human plasma samples confirmed that C40 infections induced nAb responses with modest heterotypic activity against RV-C2. Mixed-effects regression modeling of longitudinal human plasma samples collected from ages 2 to 18 years demonstrated that RV-A and RV-C illnesses induced nAb responses of similar duration. These results indicate that both RV-A and RV-C nAb responses have only modest cross-reactivity that is limited to genetically similar types. Contrary to our initial hypothesis, RV-C species may include even fewer cross-neutralizing types than RV-A, whereas the duration of nAb responses during childhood is similar between the two species. The modest heterotypic responses suggest that RV vaccines must have a broad representation of prevalent types.

Keywords: cross-neutralization; duration; neutralizing antibodies; rhinovirus; vaccine.

Figure 1.

Analysis of sequential RV-A illnesses by linear regression modeling. (A) The expected frequency of heterotypic sequential infections is on the x-axis while the observed frequency is on the y-axis. Statistically significant coefficients in those models are shown by different point shapes. Significant negative (−) coefficients, shown by squares (one-way) or triangles (reciprocal), indicate that we observed less sequential infections than expected and potentially cross-neutralization between paired RV types, whereas significant positive (+) coefficients, shown by diamonds, may indicate a decreased likelihood of finding an antigenic relationship. Color represents the follow-up RV type in illness sequences. (B) The expected and observed frequencies of re-infections with the same virus type were plotted as in panel A to validate the statistical model. RV, rhinovirus.

Figure 2.

Analysis of sequential RV-C illnesses by linear regression modeling. (A) The expected frequency of heterotypic sequential infections is on the x-axis while the observed frequency is on the y-axis. Statistically significant coefficients in those models are shown by different point shapes. Significant negative (−) coefficients, shown by squares (one-way) or triangles (reciprocal), indicate that we observed less sequential infections than expected and potentially cross-neutralization between paired RV types, whereas significant positive (+) coefficients, shown by diamonds, may indicate a decreased likelihood of finding an antigenic relationship. Color represents the follow-up RV type in illness sequences. (B) The expected and observed frequencies of re-infections with the same virus type were plotted as in panel A to validate the statistical model. RV, rhinovirus.

Figure 3.

VP1 phylogenetic tree showing relationships between antigenically cross-reactive RV-A and RV-B serotypes. A neighbor-joining phylogenetic tree was constructed using MEGA version 7. The evolutionary distances were computed using the p-distance method. The distances are presented in the units of the number of amino acid differences per site. All major nodes are labelled with bootstrap values (500 replicates, with its value more than 75%). Branch lengths are proportional to amino acid similarity (p-distance). The analysis included 88 out of 90 RV-A and RV-B types previously tested for cross-reactivity using rabbit reference sera because RV-A1a and A1b were reassigned as a single type RV-A1, and RV-87 was reassigned as enterovirus D68. The tree branch representing RV-C types was condensed. Enteroviruses (EV-A71 and EV-D68) were included as an outgroup. Filled and empty shapes of the same color show reciprocal and one-way antigenic cross-reactivity, respectively. Circles and squares indicate cross-reactivity with one and ≥two heterologous RV type(s), respectively. RV, rhinovirus; EV, enterovirus.

Figure 4.

VP1 pairwise distances (p-distances) between antigenically cross-reactive RV-A types. The p-distances between VP1 amino acid sequences of 68 RV-A types previously tested for cross-reactivity using rabbit reference sera were computed using MEGA version 7. RV-A types with cross-reactivity (one-way or reciprocal) were compared with the remaining types (not cross-reactive) using one-way ANOVA on ranks and multiple comparisons were done using Dunn’s method (SigmaPlot 13, Systat Software). ***, p<0.001 vs. not cross-reactive types. RV, rhinovirus; VP1, viral protein 1.

Figure 5.

VP1 phylogenetic tree showing relationships between RV-A and RV-C types identified to reciprocally cross-react by statistical modeling. A neighbor-joining phylogenetic tree was constructed in MEGA version 7 as in Figure 3 using reference sequences (Table S6) of all currently assigned RV types (n = 169). The tree branch representing RV-B types was condensed. Circles and squares of the same color indicate predicted cross-reactivity with one and ≥two heterologous RV type(s), respectively. Green and black triangles show the presence or absence of cross-reactivity with RV-C2 mouse serum, respectively. RV, rhinovirus; VP1, viral protein 1.

Figure 6.

Both RV-A and RV-C neutralizing antibody (nAb) responses are very likely to persist from one age to the next. Heatmaps show the presence of nAbs to indicated RV-A (blue) and RV-C (red) types in plasma (ages 2, 10, 12, 14, 16 and 18 years; each row represents serial sampling from the same subjects) from COAST study participants (n=10). Odds ratio (OR) refers to the odds of finding a continued nAb response to RV type from one age to the next (OR 30.7, 95% CI 14.3–65.8). RV, rhinovirus; VP1, viral protein 1; COAST, Childhood Origins of ASThma.