Vaccination-induced changes in human B-cell repertoire and pneumococcal IgM and IgA antibody at different ages

Abstract

It is well known that older people are more susceptible to morbidity and mortality from infectious diseases, particularly from pulmonary diseases such as pneumococcal pneumonia where vaccines do not provide efficient protection as in younger populations. We have previously shown that the B-cell repertoire in the old is reduced and hypothesise that this may contribute to the impaired humoral responses of the elderly. Here, we investigated the repertoire and antibody responses to winter vaccination in two age groups, aged 18-49 and 65-89. We found that the serum IgM and IgA pneumococcal responses were significantly impaired in the older group, with no difference in IgG levels. IGHM spectratype analysis seems to be the most promising in terms of its predictive ability for vaccine responses. Spectratypes showed a clear change in the repertoire at day 7 after vaccination, with a return to the baseline levels at day 28. The changes at day 7 reflected expansion of IGH sequences that have smaller, more hydrophilic, CDR3 regions, and these changes were attenuated in the older group. The older group was more likely to have spectratypes indicative of a reduced diversity at day 0 and day 28. On average, the baseline repertoire in the older group was comprised of larger CDR3 regions than in the younger group. In conclusion, IgA and IgM responses are significantly impaired in the elderly pneumococcal response and are likely key mediators of protection. Hydrophilicity and/or small size of the IGH CDR3 appear to be important in these responses.

Fig. 1

B-cell CDR3 spectratypes show individual variability and show challenge-related changes. (a) Spectratypes of samples from one individual before, and at 7 and 28 days after, receiving winter vaccination for influenza and pneumonia, showing CDR3 size distribution histograms for the IGHM, IGHA and IGHG immunoglobulin genes in peripheral blood B cells. The best fit Gaussian distribution curve for each sample, as determined based on the mean and SD values, is overlaid. The X-axis represents the CDR3 size in nucleotides, and the Y-axis is the proportion of the CDR3 sequences at each size. (b) Spectratype data show that diversity of repertoire decreases with age and with change from IGHM to switched isotypes on day 0 before challenge. High correlation to the underlying Gaussian distribution (CGD) is an indicator of a high level of diversity in the population. The Y-axis shows the per cent of spectratypes that have a correlation value of over 0.9 in each group. Variables were compared using chi-squared test with Bonferroni post-test, and significant values for comparisons between isotypes are indicated using * and between ages are indicated by †. (c) CDR3 spectratypes can follow a Gaussian distribution but may still differ between individuals. Examples of IGHM spectratypes from three different young individuals before vaccination are shown, with differing mean, SD, skew, kurtosis and CGD values. (d) Challenge and age-related changes in repertoire diversity as indicated by the spectratype correlation to the baseline CGD in the young and old age groups of different isotypes before, and at 7 and 28 days after vaccination. Groups were compared using Mann–Whitney u-test. (e) Challenge and age-related changes of the mean CDR3 length. Groups were compared using Mann–Whitney u-test. Error bars are SEM, *P < 0.01, **P < 0.001 and ***P < 0.0001.

Fig. 2

Challenge-induced B-cell clonality changes. Clonality was determined by dividing the overall number of sequences by the number of different clonotypes (defined by CDR3 region identity). Data represent day 0, 7 and 28 for (a) six young samples (b) and six old samples. Data are plotted separately to indicate interindividual variation, as indicated by different bar shading.

Fig. 3

Challenge results in selection of smaller, more hydrophilic, CDR3 regions. (a) Virtual spectratypes were created from sequencing data by plotting the distribution of different sized CDR3 regions for all the IGH sequences obtained. Curves represent day 0 (circles), day 7 (triangles) and day 28 (squares). The underlying distribution, without the clonal expansion effect, can be seen by plotting the distribution of CDR3 size when only one sequence from each clonotype is counted (Fig. S1). The mean CDR3 sizes for all sequences are shown in (b), while the mean CDR3 sizes for unique clonotypes are shown in (c). (d) For smaller sequences, where the CDR3 size was 42bp or less, the Grand Average Hydrophobicity values (GRAVY index) were plotted for sequences that were found only once (single) and that were found in related clones where there were more than three sequences per clone (clonal). Only one example sequence from each clone was included to avoid bias because of clonal expansion. (e) Illustrates the difference in average GRAVY values for single vs. clonal sequences at each different CDR3 size. At all sizes, the values for clonal sequences are lower and more hydrophilic, than those for single sequences. Comparisons were performed by Mann–Whitney u-test. Error bars are SEM, *P < 0.01, **P < 0.001 and ***P < 0.0001.

Fig. 4

Antibody responses to vaccination. ELISA was performed using a pool of seven pneumococcal polysaccharides (4, 6B, 9, 14, 18C, 19F and 23F) and specific for (a) IgG, (b) IgA or (c) IgM antibodies. The levels of IgG antinuclear antibody were also determined at each time point (d), and we found a significant correlation between the mean CDR3 size of the IGHM spectratype at day 0 and the levels of IGANA at day 0 (e). Comparisons were performed by Mann–Whitney u-test. Correlation by Pearson's test. Error bars are SEM, *P < 0.01, **P < 0.001 and ***P < 0.0001.

Fig. 5

IGHM spectratype before vaccination has some predictive ability. The increase in IgM pneumococcal polysaccharide-specific antibody at day 28 shows a significant correlation with the CGD index of the IGHM spectratype. Samples in the young group are plotted as open circles, those in the old group as solid triangles. Correlation by Pearson's test.