Immune Response to Pneumococcal Polysaccharides 4 and 14 in Elderly and Young Adults: Analysis of the Variable Heavy Chain Repertoire

Healthy young adults (<30 years old) and elderly volunteers (>65 years old) were asked to participate in the present study. Each individual was questioned concerning prior pneumococcal vaccination, medications, previous illness, and present health. In addition, we obtained complete blood counts, comprehensive chemistry profiles (including renal and liver functions and serum albumin), total B-cell and T-cell subset counts, and total IgG, IgM, and IgA levels in all study candidates. Individuals previously immunized with the pneumococcal vaccine and any individual considered to be immunocompromised on the basis of medication (chemotherapy, steroid preparations, and immunosuppressive agents, including anti-tumor necrosis factor alpha agents), with previous and/or present illness (previous pneumococcal disease, splenectomy, autoimmune disease, end-stage renal or liver disease, human immunodeficiency virus positivity, organ transplant, or cancer), or with abnormal complete blood count, chemistries, B or T cells, or immunoglobulin levels did not qualify. All informed consent forms and questionnaires were approved by the Institutional Review Board of the Medical College of Ohio.

Participating volunteers were then immunized with the 23-valent PPS vaccine (Pneumovax-Merk & Co., Inc., West Point, PA). Blood samples were obtained preimmunization and 6 weeks postimmunization, and the responses to PPS4 and -14 were determined by enzyme-linked immunosorbent assay (ELISA) (15). In addition, B cells specific for PPS4 and PPS14 were isolated from postimmune sera.

Biotinylated polysaccharides were used to select PPS4- and PPS14-specific B cells. The biotinylation procedure was performed as previously described (20). Briefly PPS4, -14, and -23F (American Type Culture Collection) were dissolved in 0.2 M sodium bicarbonate and reacted with cyanogen bromide. Biotin hydrazide (Pierce Chemical Co., St. Louis, MO) was added to the activated polysaccharide solution and reacted for 2 h at room temperature. The solution was dialyzed overnight at 4°C against phosphate-buffered saline. Biotinylated polysaccharides were filter sterilized and stored at −80°C. Streptavidin-coated immunomagnetic beads (Dynal, Lake Success, NY) were labeled with selected biotinylated PPS. Beads were washed with 0.1% bovine serum albumin-phosphate-buffered saline and then mixed with 50 μg of biotinylated PPS overnight at 4°C. The labeled immunomagnetic beads were washed five times and used to select B cells specific for PPS4 and PPS14.

Peripheral blood lymphocytes obtained 6 weeks postvaccination were isolated by using Ficoll Histopaque gradient (Sigma, St. Louis, MO). Adherent cells were removed, and Bcells were washed in RPMI 1640-1% newborn calf serum (Sigma). Immunomagnetic beads coated with irrelevant PPS23F were added, followed by incubation at 4°C for 1 h. Cross-reactive B cells were separated from the population and discarded. Further depletion of cross-reactive B cells was achieved by cross incubating the B cells. Cells for PPS4-specific isolation were first depleted with 25μg of PPS14-coated beads. Likewise, cells for PPS14 isolation were depleted with PPS4 beads. A total of 25 μg of PPS4-coated immunomagnetic beads or PPS14-coated beads was added independently to each of two aliquots, followed by incubation at 37°C for 30 min with 10 μg of cell wall polysaccharide (CWPS)/ml to block nonspecific binding. Polysaccharide-specific B cells were isolated by magnetic separation and washed five times with RPMI 1640-5% newborn calf serum. RNA was extracted from the isolated B cells by using lysis binding buffer (Dynal, Inc.) with 5 μl of RNasin. cDNA was prepared as per manufacturer's instructions with oligo(dT₂₅) beads (Dynabeads mRNA Direct Kit).

Single B-cell clones with specificity for PPS14 were isolated as described above and expanded by using a cell culture system described by Lagerkvist et al. and Weitkamp et al. (16, 44). Selected B cells were diluted to 1 cell/10 μl and plated in 96-well culture plates containing 5.0 × 10⁴ irradiated EL-4-B5 mouse thymoma cells (kindly provided by R. H. Zubler) in complete RPMI medium. After sorting, the feeder cell media was supplemented with 100 U of recombinant human interleukin-2, 5 ng of phorbol myristate acetate/ml, and 10% T-cell replacing factor (supernatant from pokeweed mitogen activated human T cells). The B cells were incubated for 7 days at 37°C in an atmosphere of 5% CO₂. After 7 days, 100 μl of supernatant was removed, and 1.0 × 10⁴ irradiated CD40L transfected fibroblastic L cells (kindly provided by Y.-J. Liu) were added with the above media with the addition of 5 ng of recombinant human interleukin-4/ml. The B cells were cultured for a total of 3 weeks with the addition of the CD40L L cells again at day 14. B-cell colonies were tested by ELISA for the ability to secrete immunoglobulin on days 14 and 21. Nunc-Maxisorp (Nalge Nunc International, Rochester, NY) plates were coated with goat anti-human immunoglobulin (H+L; Southern Biotechnology Associates, Birmingham, AL) and blocked prior to the addition of B-cell culture supernatants. Captured immunoglobulin was detected by using goat anti-human immunoglobulin (H+L) AP (Southern Biotechnology Associates). A serial dilution of purified human immunoglobulin (Biodesign International, Kennebunk, ME) ranging from 1 ng/ml to 100 μg/ml was included as a standard control on each plate, along with supernatants from wells containing no B cells (negative control) or pooled selected B cells (positive control). All B-cell colonies found to be secreting immunoglobulin on day 14 were tested for specific anti-PPS14 antibody production on day 21 by using the standard pneumococcal ELISA method (10). Nunc-Maxisorp plates were coated with PPS14, PPS4, or PPS23F and then washed and blocked. Specific antibody detection was achieved by using the AmpliQ (DakoCytomation, Ltd., Cambridgeshire, United Kingdom) amplifying detection system according to the manufacturer's specifications. Supernatants from wells containing no B cells were included as negative controls, and supernatants from wells containing pooled selected B cells were included as positive controls. We performed the expansion experiments twice, with similar results on each occasion.

The cDNA obtained from PPS-selected Bcells was used as a template in the PCR to generate heavy-chain libraries. The primer sets used to generate heavy-chain products were as described by Welschof et al. (46). First-round PCR primers were complementary to the 5′ untranslated region of immunoglobulins and 3′ IgG constant region. Second-round PCR primers were complementary to the 5′ end of the V_H gene and 3′ JH chain. PCR amplification conditions consisted of 32 cycles of 94°C for 45 s, 65°C for 30 s, and 72°C for 45 s. PCR controls consisted of cDNA obtained from unselected B cells, to confirm that all gene families were being amplified consistently, determined bygel electrophoresis. Amplification products were purified by using the GeneClean gel extraction kit (Bio 101, La Jolla, CA) and ligated into the Zero Blunt Cloning Vector system (Invitrogen, San Diego, CA). The ligated plasmids were transformed into Top10 Escherichia coli by chemical transformation. Heavy-chain libraries were plated on Luria broth (LB)-kanamycin at low density and grown overnight at 37°C.

Individual E. coli clones were selected and streaked onto a LB-kanamycin master plate and grown overnight at 37°C. These clones were lifted onto nylon filters and fixed by UV exposure for 5 min. Nylon filters were blocked in prehybridization buffer (7% sodium dodecyl sulfate, 0.272 Na₂HPO₄ [pH 7.2], 1% bovine serum albumin) for at least 1 h at 65°C prior to hybridization overnight at 65°C to a [³²P]ATP-labeled probe with specificity for the framework III region (46). Filters were washed twice with low-stringency wash buffer (0.1% sodium dodecyl sulfate and 1× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate]) at 55°C for 3 min each time. Sequence analysis was performed on selected clones (MWG Biotech, High Point, NC) with primers complementary to the vector. The resultant sequences were compared to germ line sequences by using VBASE DNAPLOT (http://vbase.mrc-cpe.cam.ac.uk ).

The percent gene usages of heavy chain against PPS4 and PPS14 were calculated for each age group and each volunteer. The Fisher exact test and Pearson chi-square value were used to determine significant differences in gene usage and oligoclonality; the Student t test was used to detect differences in mutational frequencies between age groups. A P value of ≤0.05 was considered significant. Statistical calculations were performed with SPSS 11.5.1.

We have performed studies to validate the B-cell isolation technique with PPS14-coated paramagnetic beads, by culturing selected cells in vitro. It should be emphasized that extensive negative selection was performed on isolated B cells. B cells that bound to uncoated immunomagnetic beads, streptavidin, CWPS, or cross-reactive contaminants present in the PPS preparations were depleted by incubation with PPS23F-labeled immunomagnetic beads, followed by incubation with PPS4-labeled immunomagnetic beads. Finally, any remaining binding to CWPS was blocked by the inclusion of free (unlabeled) CWPS in the positive selection step with the PPS14-coated beads. To confirm the PPS14 specificity of the selected B cells, we cultured single B cells as described in Materials and Methods. B-cell culture supernatants were tested for total immunoglobulin production at days 14 and 21 and for anti-PPS14 production at day 21. At day 14, 44 of 576 single B-cell colonies were producing immunoglobulin. By day 21, 14 single B-cell colonies were producing immunoglobulin. All single B-cell clones that were still producing immunoglobulin at day 21 were secreting immunoglobulin specific for PPS14 and failed to bind to PPS4 or PPS23F. B cells selected and cultured on two separate occasions demonstrated similar results. The efficiency of this B-cell expansion system, indicating successful in vitro expansion of ca. 3% of single Bcells, was consistent with previous work by Weitkamp et al. (43). Since all successfully cultured B cells produced anti-PPS14 antibody, these results suggest that our selection method accurately selects B cells expressing anti-PPS specific antibody.

PPS4- and PPS14-specific lymphocytes were obtained from 20 young (11 female, 9 male) and 20 elderly volunteers (10 female, 10 male). A total of 364 heavy chains with specificity for PPS4 were obtained from 15 immunized young adults and 325 heavy chains from 15 immunized elderly adults. In response to PPS14, 305 heavy chains were obtained from 19 young adults and 291 from 14 elderly adults. The V_H3 gene family comprised >90% of the total sequences isolated in response to PPS4 and PPS14 in both age groups. The V_H1, V_H4, and V_H5 gene families were isolated from both young and elderly in response to both polysaccharides; however, these gene families constituted <10% of the total heavy-chain repertoire. The V_H4 and V_H5 gene families were represented in numbers too small to allow for meaningful analysis. Variable heavy gene, complementarity-determining region 3 (CDR3) length/composition, D gene, J chain, percent identity to germ line sequence, and somatic mutation frequencies were all determined on all isolated sequences.

As shown in Fig. 1, the most commonly used heavy-chain gene loci in response to PPS4 were in the following order of usage: V_H3-07, V_H3-74, and V_H3-30, representing 50.9% of total variable heavy-chain gene usage in young adults. In all, 80 sequences were identified that utilized V_H3-07, 73 sequences utilized V_H3-74, and 32 sequences that utilized V_H3-30. Other V_H3 family genes were isolated with variations in percent usage between volunteers. Members of the V_H1 gene family were also used and represented 7.8% of the total V_H sequences isolated. The most common gene families expressed in response to PPS4, V_H3-07 (22%) and V_H3-74 (20%), were significantly overexpressed compared to gene locus expression found in unselected CD19⁺ peripheral Bcells: 4.2% and 1.4%, respectively (5).

In response to PPS14, V_H3-48 and V_H3-33 gene loci in order of use were the dominant heavy chains used and represented 48.4% of isolated sequences (Fig. 1). Other V_H3 gene loci, namely, V_H3-11, V_H3-21, and V_H3-30, accounted for 38.5% of the heavy-chain sequences obtained from young adults. The total numbers of sequences identified were 99, 49, 25, 15, 6, and 2 for gene families V_H3-48, V_H3-33, V_H3-30, V_H3-21, V_H-11 and V_H3-07, respectively. The V_H1 gene family was expressed in 6.3% of total sequences obtained from young adults. The most common gene families expressed in response to PPS14, V_H3-48 (32.4%) and V_H3-33 (16%), were significantly overexpressed compared to gene locus expression found in unselected CD19⁺ peripheral B cells: 1.4 and 2.8%, respectively (5). A similar pattern of gene locus usage was maintained when the gene usage of individual donors was investigated (Table 1).

Oligoclonality is defined as a response limited to one or few gene loci, with one gene locus representing at least 50% of the isolated sequences. Overall, most of the young donors demonstrated an oligoclonal response in response to PPS4 (87%) and PPS14 (84%).

In elderly adults, the predominant heavy-chain gene usage in response to PPS4 and PPS14 was V_H3-30, representing 26.7 and 28%, respectively (Fig. 1) corresponding to 87 and 81 identified sequences. In response to PPS4, V_H3-21 and V_H3-07 were also commonly expressed at 16 and 10.8%, corresponding to 52 and 35 sequences, respectively, with the V_H1 gene family representing 5.9% of heavy-chain gene usage. The variable heavy-chain loci V_H3-07 and V_H3-11 were commonly used in response to PPS14 by the elderly and were identified in 36 and 27 of the isolated sequences, respectively. The V_H1 gene family represented a small percentage (8.2%) of the total heavy-chain sequences isolated in the elderly. In the elderly population the V_H3-30 gene locus was overexpressed in response to both PPS4 and PPS14 (26.7 and 28%, respectively) compared to unselected CD19⁺ B cells (8.5%) (5). A similar pattern of gene locus usage was maintained when the gene usage of individual donors was investigated (Table 2). Overall, 63% of the elderly donors demonstrated oligoclonality in response to PPS4, whereas 60% of elderly donors showed oligoclonality in response to PPS14.

Comparison of young and elderly revealed significant differences in heavy chain gene usage. Both V_H3-07 (P < 0.01) and V_H3-74 (P < 0.01) were more frequently expressed in the young compared to the elderly (Fig. 1). In contrast, the V_H3-21 and V_H3-30 genes were significantly more common in elderly than young individuals (P < 0.01). No significant difference was found in V_H1 gene family expression between age groups.

When the immune response was investigated at the donor level (Tables 1 and 2), it was found that significantly more young donors (P < 0.05) utilized one predominant gene locus (87%) compared to the elderly (63%).

As shown in Fig. 1, with aging, a significant shift in gene usage occurred with decreased expression of V_H3-48 and V_H3-33. Elderly adults demonstrated a significant increase in V_H3-30, V_H3-11, and V_H3-07 heavy chain gene usage compared to young adults (P < 0.01 for all). Despite the predominance of the V_H3-30 locus in response to both PPS4 and PPS14 in the elderly, no identical CDR3 sequences were isolated between these polysaccharides (Tables 1 and 2). A predominant (>50%) gene locus was expressed in 84% of young adults compared to 60% of elderly adults (P < 0.05).

The CDR3 length of the isolated sequences from the diverse array of heavy chain genes ranged from 7 to 24 amino acids in response to both polysaccharides and age groups. To facilitate data comprehension a representative sample of sequences of common heavy chain gene families in response to PPS4 and PPS14 is shown in Tables 3 and 4. These tables present examples of the most commonly expressed genes from both young and elderly and do not contain all isolated sequences for logistical reasons. As shown in Tables 3 and 4, a wide variety of CDR3 sequences were identified in response to both polysaccharides. The mean CDR3 lengths are shown in Table 5. Significant differences in CDR3 length between young and elderly were noted in V_H3-30 in response to PPS4 but not PPS14. Similarly, the CDR3 lengths of the V_H1 gene loci, V_H1-69 and V_H1-02, were significantly shorter in the young versus the elderly (P < 0.05).

Although the CDR3 regions were overall diverse, a recurrent motif was noted in the CDR3s associated with the V_H3-07 gene locus consisting of the amino acid motif PNR in response to PPS14. The use of the 3-16 and 3-22 D genes introduced a substantial number of hydrophobic residues in the CDR3 regions associated with the V_H3-30 and V_H3-21 sequences isolated in response to PPS4 and the V_H3-30 and V_H3-07 sequences isolated in response to PPS14. In response to PPS4 (Table 5), there were significant differences (P < 0.001 and P < 0.05) in the expression of 3-16 and 3-22 D genes between young (33 and 13%, respectively) and elderly (6.8 and 29%, respectively) individuals. In the response to PPS14, these two D genes were also utilized; however, in contrast to PPS4, there was no significant difference in 3-22 D gene or 3-16 D gene usage in elderly (36 and 7.2%, respectively) compared to young (35 and 5.9%, respectively) subjects. Overall, despite identical D gene usage between individuals and polysaccharides, unique CDR3 sequences were observed as a result of the recombination of V, D, and J chains; N insertions; and deletions.

Analysis of J (joining)-chain sequences in all age groups demonstrated the presence of all known J chains; however, the J_H3 and J_H4 chains were predominant (data not shown). There was a significant difference (P < 0.05) in JH4 usage between age groups (68% young and 46% elderly) in response to PPS4 but not in response to PPS14 (63 and 59%, respectively). In contrast, there was no significant difference in JH3 usage between the age groups in response to either PPS4 or PPS14.

All heavy chain sequences were analyzed for percent identity to germ line and mutational frequencies in the CDRs and framework regions (FR) as summarized in Table 5. Sample sequences are shown in Tables 3 and 4. Overall, the heavy chain sequences were 87.3 to 100% identical to germ line sequences. In response to PPS4, the predominant heavy chain gene in young adults (locus V_H3-74) was highly mutated (90.8%) compared to germ line (Tables 3 and 4). Although only three V_H3-74 sequences were isolated from the elderly these sequences were 99.3% identical to germ line. Mutational frequencies in locus V_H3-74 were significantly higher (P < 0.05) in young compared to elderly subjects in both the FR and CDRs. The predominant heavy-chain genes in the elderly, V_H3-30 and V_H3-21, demonstrated >98% identity to germ line. The mutational frequencies in heavy chain gene families V_H3-30, V_H1-69, and V_H1-02, however, were similar between age groups.

In response to PPS14, both age groups showed >95% identity to the germ line in the most commonly used heavy chain genes, V_H3-48 and V_H3-33, for the young, and V_H3-30, for the elderly. As shown in Table 4, both age groups demonstrated similar mutational frequencies in V_H1-69, V_H1-18, and V_H1-02. There was also no significant difference in mutational frequency between young and elderly in the V_H3-30 gene products, the predominant sequence expressed in the elderly. The G-to-A substitution present in the CDR1 of V_H3-30 in both young and elderly adults is a result of allelic variation. Of the 19 identified V_H3-30 alleles in the IMGT gene database (11), 11 express an adenosine at this position and 8 express a guanidine; thus, this has not been considered a mutation in our calculations. Overall, the V_H3 gene locus is polymorphic (36), with heterogeneity resulting from the diversity of haplotypes rather than from allelic variation. In young adults, the predominant heavy chain gene (locus V_H3-48) demonstrated significantly higher (P < 0.01) mutational frequency in the CDRs compared to elderly expressed V_H3-48. It should be noted that this was not the case for the V_H3-33 gene products (Tables 3 and 4). However, the elderly showed a significantly higher (P < 0.05) mutational frequency in the CDRs in V_H3-11 and V_H3-21 genes compared to young subjects. Overall, the young group showed a greater mutational frequency in CDR1 and CDR2 compared to the elderly in the response to both PPS4 and PPS14; however, there was no significant difference between the young and elderly overall. Within all gene families, mutations were more pronounced in the CDRs than in the FRs in all age groups and in response to both polysaccharides. Analysis of mutational hotspots (RGYW) demonstrated that mutation rates within the hotspots were V_H locus dependent; however, there was no significant difference in mutational frequency within these regions between young and elderly.

We have previously analyzed the PPS-specific antibody concentrations, opsonophagocytic activity, and antibody avidity in the response to PPS4 and PPS14 in all study participants (15). Given the wide variability in gene locus usage between and within individuals, a significant association between the usage of a specific gene locus and measures of serum activity, avidity, and opsonophagocytosis could not be defined. This was especially true for opsonophagocytosis, where activity is multifactorial and related to antibody concentration, isotype, and avidity. However, when we compared the dominant utilization of specific gene loci with avidity, as defined by the molarity of sodium thiocyanate required to elute 50% of bound antibody (29) on an individual level (shown in Tables 1 and 2), specific associations were apparent.

In response to PPS4, usage of the locus V_H3-74 was associated with high avidity, and usage of the locus 3-07 was associated with medium to high avidity. In the young response to PPS14 exclusive expression of V_H3-33 was associated with moderately high avidity. However, the combination of V_H3-33 (63%) with V_H3-53 (25%) resulted in low-avidity measurements in serum. Similarly, exclusive expression of V_H3-74 (100%) resulted in high avidity, whereas in combination with V_H3-21 (29%) it resulted in moderate avidity measurements. Exclusive expression of V_H3-64 was associated with high avidity in one individual.

In the elderly response to PPS4, two gene loci, V_H3-09 and V_H3-21, were clearly associated with high avidity. One gene locus, V_H3-23, was associated with low-avidity serum measurements. In the elderly response to PPS14 two gene families, V_H3-11 and V_H3-07 were consistently associated with high-avidity serum measurements. In contrast, the most common gene locus isolated from elderly, V_H3-30, was associated with multiple other gene loci and was therefore associated with low to high avidity serum measurements to both PPS4 and PPS14.