Hideyuki Ikematsu MD, Department of Clinical Research, Hara-Doi Hospital, 6-40-8, Aoba Higashiku Fukuoka, 813–8588, Japan. E-mail: firstname.lastname@example.org
To investigate age-related alterations in human humoral immunity, we analysed Ig heavy chain variable region genes expressed by peripheral B cells from young and aged individuals. Three hundred and twenty-seven cDNA sequences, 163 µ and 164 γ transcripts with VH5 family genes, were analysed for somatic hypermutation and VHDJH recombinational features. Unmutated and mutated µ transcripts were interpreted as being from naive and memory IgM B cells, respectively. In young and aged individuals, the percentages of naive IgM among total µ transcripts were 39% and 42%, respectively. D and JH segment usage in naive IgM from aged individuals was similar to that from young individuals. The mutational frequencies of memory IgM were similar in young and aged individuals. γ transcripts, which are regarded as being from memory IgG B cells, showed a significantly higher mutational frequency (7·6%) in aged than in young individuals (5·8%) (P < 0·01). These findings suggest that VHDJH recombinational diversity was preserved, but that the accumulation of somatic mutations in the IgG VH region was increased in aged humans. The accumulation of somatic mutations in IgG B cells during ageing may imply that an age-related alteration exists in the selection and/or maintenance of peripheral memory B cells.
Human antibodies are composed of heavy and light chains, and the heavy chain variable regions are crucial for determining affinity to antigens . The heavy chain variable region genes are generated by the recombination of three genomically separated genes: VH, D, and JH. D to JH recombination is the initial stage performed during B cell development, followed by VH to DJH recombination. The heavy chain variable region genes produce much of the diversity by: VHDJH recombination, N- and P-nucleotide addition on recombination and somatic hypermutation after recombinational events . In addition, class-switching of Ig constant region genes increases the diversity of human antibodies. In the primary Ag-driven immune response IgM antibodies are produced initially, then high-affinity IgG antibodies are produced at the late stage [3,4].
Susceptibility to infectious diseases influences human longevity. Infectious disease mortality, due particularly to pneumonia, has been increasing in aged humans . One of the causative factors is ‘immunosenescence’, the alteration of the host defence mechanism caused by ageing. The study of human immunosenescence is less advanced than that of mice and humoral immunity, a defence system against infectious agents, in aged humans has not been investigated fully to date.
Whether or not Ig mutation levels are influenced by ageing has not been precisely determined. This issue is important for clarifying one aspect of humoral immunity in aged humans. Mutation levels of Ig variable region genes isolated from peripheral B cells in young and aged humans have been reported by some investigators. The analysis of heavy chain VH6 family genes has shown a significantly lower mutational frequency in aged individuals than in young ones . Similar results have also been found for light chain Vκ genes . On the other hand, comparable or higher mutational levels in aged compared to young individuals has been shown in the analysis of heavy and light chain genes, although age-related alterations of Ig mutation levels in these studies are difficult to discuss due to the low numbers of individuals [8,9]. Recently, mutational levels of µ transcripts with VH6 family genes have been reported to be higher in aged individuals (2·8%) than in young ones (2·4%), although the differences were not statistically significant .
To clarify qualitative changes in Ig genes elicited by ageing, we compared Ig heavy chain variable region genes expressed by peripheral B cells isolated from young and aged individuals. Both µ and γ transcripts were examined in this study. Age-related alterations of humoral immunity, in which µ and γ isotypes have a differential role, can be evaluated more precisely by the independent analysis of these isotypes. We chose VH5 family genes for Ig variable segment gene analysis because the limited numbers and lower levels of polymorphism in these genes, compared to other VH gene families, allowed us to carry out precise sequence analysis [11,12]. Somatic mutation levels in each Ig isotype were analysed and compared between young and aged individuals. VHDJH recombinational regions were also analysed to compare recombinational characteristics between young and aged individuals.
MATERIALS AND METHODS
Registered in this study were 10 healthy young (five men and five women) and 10 healthy aged (four men and six women) individuals (Table 1). Individuals with hypergammaglobulinaemia or monoclonal gammopathy were not included. Viral infections including HIV-1, HTLV-1, HBV and HCV were not detected in any individuals. Age in the young ranged from 19 to 30 years, with a mean age of 25·2 years. Age in the aged ranged from 65 to 92 years, with a mean age of 78·3 years. Blood sampling was performed after informed consent was obtained from the study participants. This study was approved by the ethics committee of Hara-Doi Hospital. Peripheral blood mononuclear cells (PBMC) were isolated from heparinized venous blood by density centrifugation over an LSM solution (Organon Teknika-Capple, Durham, NC, USA). PBMC were then prepared to a cell concentration of 1 × 106 cells/ml and preserved at –80°C until mRNA extraction.
Table 1. . Clinical data and results of Ig heavy chain variable region gene analysis
No. of analysed sequences
Mutational frequency, %
IgM (unmutated/ mutated)
IgG (unmutated/ mutated)
n.a.: Not available.
mRNA extraction and cDNA synthesis
mRNA was extracted from PBMC with the QuickPrep micro mRNA purification kit (Amersham Pharmacia Biotech Ltd, Little Chalfont, UK) according to the manufacturer's instructions. Briefly, PBMC were disrupted in a solution buffer containing guanidinium thiocyanate and N-lauroyl sarcosine. After microcentrifugation, the homogenate containing mRNA was incubated with oligo (dT)-cellulose. Bound mRNA was then recovered with an elution buffer containing 10 mm Tris-HCl and 1 mm EDTA. Using the extracted mRNA, first-strand cDNA was synthesized using random hexamer primers and reverse transcriptase as described previously . The synthesized cDNA was incubated with 2 units of ribonuclease H (Life Technologies, Rockville, MD, USA) at 37°C for 20 min.
Polymerase chain reaction (PCR) amplification of the expressed Ig VHDJH gene segments
One-tenth of the first-strand cDNA reaction product was used as a template for PCR. The PCR reaction was performed in 50-µl using a GeneAmp XL PCR kit containing 2 units of rTth DNA polymerase, XL (Perkin-Elmer Biosystems, Foster City, CA, USA). To amplify the expressed VHDJH-Cµ and Cγ gene segments utilizing VH5 genes, a VH5 gene family-specific sense primer was used together with an antisense Cµ or Cγ primer. The oligonucleotide primers were as follows: a 5′-VH5 primer encompassing the VH5 family leader sequence (5′-ATGGGGTCAAC CGCCATCCTCGCCCT-3′); a 3′-Cµ primer consisting of a sequence (5′-AGACGAGGGGGAAAAGGGTT-3′); and a 3′-Cγ primer consisting of a sequence (5′-TAGTCCTTGACCAG GCAGCC-3′). Thirty cycles of amplification were performed. Each cycle consisted of a denaturing step at 93°C for 1 min, an annealing step at 62°C for the Cµ primer or 64°C for the Cγ primer for 1 min, and an extension step at 72°C for 1 min, followed by a 10-min final extension. The PCR-amplified products were fractionated on a 2% agarose gel containing 1 µg/ml ethidium bromide. The amplified DNA was visualized by exposure to UV light.
Cloning and sequencing of the expressed Ig VHDJH gene segments
The PCR product of the expressed Ig VHDJH gene segments was ligated into a vector (Novagen Inc, Madison, WI, USA) using a DNA ligation kit, version 1 (TaKaRa Shuzo Co. Ltd, Kyoto, Japan) according to the manufacturer's protocol. Competent cells (Novagen Inc, Madison, WI, USA) were then transformed by the recombinant vectors, and the transformed bacterial clones were selected in culture medium containing ampicillin. The plasmid DNA was purified from a culture of a selected bacterial colony using QIAprep Spin Miniprep Kits (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. Dideoxy sequencing was performed using the double-stranded plasmid DNA and an ABI PRISM™ 310 Genetic Analyser (PE Applied Biosystems, Foster City, CA, USA). The VHDJH gene segments were sequenced from independent bacterial clones derived from each PCR product of a single blood sample.
Analysis of DNA sequencing data
VH5 family genes, VH251 and VH32, were designated by comparison with the germline sequences shown in the current human Ig gene v-base database (http:www.mrc-cpe.cam.ac.ukimt-docpublicINTRO.htlm). Utilized JH family genes were also identified by comparison with the germline sequences found in the v-base. Some sequences were excluded from the analysis of CDR length and/or JH segment assessment because they could not be identified. The D segment genes utilized were identified in each sequence by maximal alignment with 27 germline D segments, as reported by Corbett et al. . When one sequence showed homology in more than seven continuous nucleotides with germline D segment configurations, the sequence was assigned as a utilized D segment. When this condition was not satisfied, the sequence was regarded as an unidentifed D segment. This definition was based on the finding that, in some cases, more than two germline D segments were identified for one sequence unless this definition was used. Obtained sequences were compared with VH251 or VH32 germline configurations to identify somatic point mutations.
The χ2 test was used to compare the differences of the analysed data between the young and aged. The Mann–Whitney U-test was used to compare the differences in the distribution of the analysed data between the young and aged.
Analysis of the expressed VHDJH gene segments
The VHDJH gene segments of VH5 family genes were sequenced from 10 independent clones derived from a single blood sample. A total of 380 sequences were obtained. In two aged individuals, the sequencing data of µ transcripts were not obtained due to the lack of sufficient mRNA extraction product. No identical VHDJH recombinations were shared by different individuals. When identical sequences were detected in a single PCR product, they were treated as one sequence for analysis. Three hundred and twenty-seven sequences with unique VHDJH recombinations, 163 µ and 164 γ transcripts, were subjected to analysis. B cells with unmutated Ig variable region genes have been designated as naive B cells [15–17]. B cells with mutated Ig genes have been regarded as memory B cells. In this study, µ transcripts with no mutations were defined as IgM naive, and those with mutations were defined as IgM memory. The prevalence of IgM naive among total µ transcripts from aged individuals (42%) was equivalent to that from young individuals (39%) (Table 1). Because IgG B cells can be considered to be memory B cells, whole γ transcripts were analysed as memory IgG.
Frequency of somatic hypermutation
Individual mutational frequencies (number of mutated nucleotides per number of total analysed bases from each individual) are listed in Table 1. The individual mutational frequencies of IgM memory and IgG in the young ranged from 1·4 to 6·2% and from 4·0 to 7·7%, respectively, similar to a previous report . These mutational frequencies did not correlate with age, gender, serum IgM levels or serum IgG levels (Table 1). The mean levels of individual mutational frequencies of IgM memory in the young and aged were 3·3% and 3·9%, respectively, resulting in no significant difference (Fig. 1). The mean level of individual mutational frequencies of IgG in the aged (7·6%) was significantly higher than that in the young (5·9%) (P < 0·05, Fig. 1).
Mutational frequencies for all individuals (number of mutated nucleotides per number of total analysed bases from all tested individuals) were compared between the young and aged (Table 2). In IgM memory, the mutational frequencies of the overall region in the young and aged were 3·4% and 3·9%, respectively, and resulted in a statistically significant difference (P < 0·05).
Table 2. . Mutational frequency in IgM and IgG from young and aged humans
No. of analysed sequences
Mutational frequency, % (no. of mutations/No. of total analysed bases)
The mutational frequencies of IgM memory were higher in the aged than in the young throughout the VH region, but the differences were not statistically significant except for FR1. In IgG, the mutational frequency of the overall region in the aged [7·6%, 1753 mutations/22 932 base pairs (bp)] was significantly higher than that in the young (5·8%, 1463 mutations/25 284 bp) (P < 0·01). In the CDRs (CDR1 and CDR2), the mutational frequency in the aged (11·3%, 584 mutations/5148 bp) was significantly higher than that in the young (8·9%, 508 mutations/5676 bp) (P < 0·01). Both regions of CDR1 and CDR2 in the aged were mutated more frequently than those in the young (P < 0·05 and P < 0·01, respectively). In the FRs (FR1, FR2 and FR3), the mutational frequency in the aged (6·6%, 1169 mutations/17 784 bp) was significantly higher than that in the young (4·9%, 955 mutations/19 608 bp) (P < 0·01). Each FR region, except for FR2, in the aged was mutated more frequently than that in the young (P < 0·01).
Features of replacement to silent mutation ratio (R/S ratio)
The R/S ratios of analysed sequences were compared between the young and the aged (Table 3). In IgM memory, the R/S ratio of CDRs (CDR1 and CDR2) in the aged (3·9) was higher than that in the young (2·6), but the difference was not statistically significant. The R/S ratios of FRs (FR1, FR2 and FR3) in IgM memory showed no significant differences between the young and aged. In IgG, the R/S ratio of CDRs in the aged (2·2) was lower than that in the young (2·5), although no statistically significant differences were obtained. The R/S ratios of CDR1 and CDR2 in the aged were also lower than those in the young. There were no significant differences in the R/S ratios of FRs between the young and aged.
Table 3. . R/S ratio in IgM and IgG from young and aged humans
The CDR3 length distributions in IgM naive, IgM memory and IgG are summarized in Table 4. In IgM naive in the young, CDR3 lengths of 14–18 amino acids were most frequent (51·4%), followed by 9–13 amino acids (24·3%). In the aged, CDR3 lengths of 9–13 amino acids were most frequent (60·7%), followed by 14–18 amino acids (21·4%). The CDR3 length of the aged was significantly shorter than that of the young (P < 0·05). In both IgM memory and IgG, the CDR3 length distributions were not statistically different between the young and aged.
Table 4. . Distribution of CDR3 length in IgM and IgG from young and aged humans
No. of analysed sequences
CDR3 length in amino acids, no. of sequences (%)
a These data were cited from healthy adult peripheral B cell analysis reported by Yamada et al. .
The D segment usages in IgM naive, IgM memory and IgG are summarized in Table 5. One hundred and thirty-six (41·6%) of 327 analysed sequences had no assigned D segments, and the frequency of non-assigned D segments was similar to the finding reported by Corbett et al. . Two sequences with D–D fusion were detected. One, IgM naive, included D5–12 and D6–13 and the other, IgM memory, included D3–16 and D1–26. The presence of inverted D segments was not examined. In IgM naive, the most frequently used D segment in the young was D3–10 (16·2%), followed by D2–15 (10·8%) and D3–22 (10·8%). Preferential usage of D3–10 has been reported by Corbett et al. . The most frequently used D segment in the aged was D3–10 (14·3%), followed by D6–19 (10·7%), D3–3 (7·1%) and D2–15 (7·1%). There was no statistically significant bias in the distribution of D segment usage between the young and aged. Similarly, in both IgM memory and IgG, the distribution of D segment usage did not differ statistically between the young and aged.
Table 5. . D and JH segment usage in IgM and IgG from young and aged humans
No. of sequences (%)
aNames of the 27 human functional D segments are as in Corbett et al. . The 27 D segments are aligned on the basis of their relative positions in the D locus. 1–1 is the most VH locus-proximal and 7–27 the most JH locus-proximal. bNot detected indicates sequences in which no D segment was able to be assigned in this study. cThese data include sequences in which D–D fusion was detected. D–D fusion was treated as two identified D segments. dThese data were cited from rearranged heavy chain sequence analysis reported by Corbett et al. . eThese data were cited from healthy adult peripheral B cell analysis reported by Yamada et al. . n.a.: Not available.
The JH segment usage in IgM naive, IgM memory and IgG is summarized in Table 5. In IgM naive, the most frequently used JH segment in the young was JH4 (40·5%), followed by JH6 (27·0%) and JH5 (13·5%). Preferential usage of the JH4 segment has been reported in normal adults . The most frequently used JH segment in the aged was JH4 (46·4%), followed by JH3 (25·0%) and JH6 (17·9%). There was no statistically significant bias in the distribution of JH segment usage between the young and aged. Similarly, in both IgM memory and IgG, JH segment usage was not significantly different between the young and the aged.
The overall mutational frequency of the VH5 family genes observed in this study was 4·4%, similar to previously reported mutational frequencies of various VH genes derived from peripheral blood . The features of the CDR3 region in the VH5 family genes analysed were similar to those reported previously for various Ig genes (Tables 4 and 5). The mutational frequencies of Ig VH genes may differ in various types of peripheral tissue; however, the mutational frequency obtained by this study appeared to be similar to that of total spleen B cells  and VH5 B cells derived from gut . It is possible that VH5 B cells in the periphery undergo a similar developmental process to other B cells.
B cells with different Ig isotypes have distinct biological characteristics and different roles in vivo. Thus, analysis of VH genes by isotype may be informative. The mutation levels of µ and γ Ig transcripts were analysed separately in this study. In addition, we analysed unmutated and mutated µ transcripts separately, which may have improved the sensitivity of the analysis.
Unmutated µ transcripts are regarded as being derived from naive B cells producing IgM antibody (naive IgM B cells). The analysis of naive IgM B cells gives rise to an accurate evaluation of the differences in junctional diversity between young and aged humans. The CDR3 length of naive IgM B cells was statistically shorter in aged individuals than in young individuals (Table 4). D and JH segment usage in naive IgM B cells from aged individuals was similar to that in young individuals (Table 5). N- and P-nucleotide lengths in naive IgM B cells showed little difference between the two groups (data not shown). Although it is not clear in this study whether or not TdT activity diminishes during ageing, it is possible that the VHDJH recombinational diversity generated in the early stage of B cell development is equivalent in young and aged individuals.
Biased CDR3 length distribution or JH segment usage has been used as a marker for particular B cell populations in human fetal liver B cells, peripheral B cells in autoimmune disease patients and neoplastic B cells [21–23]. In mutated γ transcripts (memory IgG B cells), CDR3 length distribution and JH segment usage in aged individuals was similar to that in young individuals (Tables 4 and 5). This finding suggests that IgG B cells in aged individuals are unlikely to be derived from a specific population. The lack of difference in the recombinational features between IgG B cells from young and aged individuals may imply that the alterations resulting in elevated mutational frequencies in aged individuals were produced after the recombination process in B cell development.
In aged mice, both the extent of somatic hypermutation and the level of affinity to antigens in specific antibodies after some vaccinations have been shown to be diminished significantly compared that seen in young mice [24–26]. In aged humans, both lower amounts of and lower affinities to antigens for specific antibodies after vaccination have been reported [24,25]. On the other hand, compatible humoral responses to vaccines in young and aged humans have been described recently , suggesting that the mechanism of Ag-driven affinity maturation in B cells might not always be impaired by ageing. In this study, the mutated µ transcripts of aged individuals were not lower in mutational frequency than those of young individuals (Table 2, Fig. 1). This finding suggests that the somatic hypermutation machinery in B cells might not be impaired by ageing.
B cells have been reported to survive longer than was expected previously [28,29]. We have demonstrated that the progeny of a human B cell clone producing a specific autoantibody underwent ongoing somatic mutations and persisted in vivo for over 2 years , suggesting that human memory B cells with specific antibodies may survive over a period of years and can accumulate somatic point mutations. Ig VH genes of plasma cells derived from intestinal lamina propria have been reported to be more highly mutated in older individuals than in younger ones . In addition, clonally related plasma cells were detected at different bowel sites  suggesting the persistent dissemination followed by the induction of somatic mutations into Ig genes. In this context, mutated Ig VH genes in aged humans are derived possibly from long-lived B cells. γ transcripts (IgG B cells) from aged individuals showed significantly higher mutational frequencies compared to those from young individuals (Table 2, Fig. 1). Although the possibility cannot be denied that almost all mutations found in the Ig genes of aged humans were introduced into naive B cells by newly elicited antigenic immune responses, these results might be attributed to the accumulation of additional somatic mutations in Ig genes during the long-survival process of memory B cells. IgG antibodies from aged individuals were 1·8% higher in mutational frequency than those from young individuals (7·6%versus 5·8%, Table 2), and we calculated that five point mutations per one antibody were added to the VH region of aged individuals. The increased mutation levels in IgG antibodies from aged humans may reflect an ongoing accumulation of somatic mutations in memory B cells that survive over a period of years.
In general, monospecific high-affinity antibodies show nucleotide substitution features of a high R/S ratio in the CDRs and a low R/S ratio in the FRs . Mutated µ transcripts displayed these characteristics of R/S ratios in aged individuals (Table 3), suggesting that the Ag-driven affinity maturation process in IgM B cells is maintained in aged humans. It is, however, well known that affinity maturation in IgG B cells is more important to the humoral immunity.
Accumulation of additional point mutations in IgG antibodies during ageing may affect their affinity. It should be noted that the R/S ratio of CDRs in IgG from aged individuals appears to be lower than that found in young individuals (2·2 versus 2·5, Table 3), although the difference was not statistically significant. On the other hand, the mutation level of CDRs in IgG from aged individuals was significantly higher than that found in young individuals (11·3%versus 8·9%, Table 2). Thus, it is possible that the elevated mutation level found in aged humans is caused by a mechanism that is not involved in the Ag-driven affinity maturation process. The affinity of IgG antibodies might be decreased by ongoing induction of Ig gene mutations during ageing. Further investigation will be necessary to address this issue. Recently, Banerjee et al. have reported preserved somatic hypermutation machinery and an impaired selection process during ageing using germinal centre B cells from young and aged humans . Their data support our results obtained from peripheral IgG B cells (postgerminal centre B cells).
In conclusion, Ig VH genes in peripheral B cells from aged humans were mutated more frequently than those of young humans. Interestingly, the R/S ratio of CDRs in the IgG VH region was lower in aged humans than in young ones. It is possible that the elevated mutation level found in aged humans may not be caused by a change in the actual mechanism of somatic hypermutation itself, but rather by ongoing induction of Ig gene mutations during ageing. Increased accumulation of somatic mutations into Ig VH genes, which may be not related to Ag-driven affinity maturation, could possibly affect the process of B cell selection and maintenance in the life of peripheral memory B cells.