Long-term kinetics of adult human antibody repertoires

Abstract

In healthy humans, antibody repertoires change during ontogeny and senescence. The dynamics of antibody repertoires among adults over a longer period of time in one and the same individual has, however, not been extensively studied. In this study we analysed peripheral blood samples from five healthy adults, taken over a period of 10 weeks and once 9 years later. A competitive, quantitative polymerase chain reaction (PCR) was developed to investigate short and long-term variations in V_H gene family repertoires. Serum antibody levels to common self and non-self antigens were determined in samples taken at the same time-points as the cell samples to analyse possible correlations between molecular and serological expression profiles. We found a high degree of stability in the V_H gene family repertoire over time as well as between individuals with a Caucasian background. A specific change in the usage of primarily the V_H3 and V_H5 gene families was observed in one individual at one time-point. The deviating pattern resembled the V_H gene family utilization pattern observed in naturally activated B lymphocytes. The fluctuations in V_H3 and V_H5 gene family expression correlated with the presence of rheumatoid factor in serum. We discuss the possible influence of polyclonal, transient stimulation of B cells on V_H gene repertoires, as measured in circulating B cells.

Introduction

While the potential B-cell repertoire of a human is very large, only a fraction is actually expressed in the adult individual at one time-point. Three main mechanisms influence the process shaping the peripheral repertoire of B cells. First, the potential diversity in lymphocyte repertoires is created largely by a well-defined, developmentally regulated process in the bone marrow through recombination of the genes coding for the variable region of the heavy (V_H) and the light (V_L) chains¹ with insertion of templated and non-templated nucleotides at the junctions.² Second, several selection processes take place based on the antigen-binding part of the antibody molecule expressed on the cell surface.^3,4 This way autoreactive clones are deleted in the bone marrow,⁵ excluded from access to survival signals in the peripheral organs, or induced into an anergic state.^5,6 Third, antigen-driven responses during life induce the proliferation of specific clones, introduce somatic diversification by mutation and generate a memory pool of lymphocytes.^7,8 Together, these processes result in alterations of the antibody repertoire during B-cell development as well as during ontogeny and senescence.⁹

At a serum level, the immunoglobulin M (IgM), as well as the immunoglobulin G (IgG), reactivity pattern differentiates until 1–2 years of age, after which it is relatively stable. Reactivity patterns change from expressing a high degree of homogeneity among neonates to being more diverse in children, young adults and elderly individuals. Self-reactive repertoires are more conserved between individuals and in time than antibacterial repertoires.^10–16 At a molecular level, evidence has accumulated that expression of the V_H gene family is not random, nor a simple reflection of genomic complexity.^17,18 The genes coding for the V_H region are divided into seven families, based on nucleotide sequence homology,^14,15 and are located on chromosome 14q32.33.¹⁶ Restricted V_H gene family usage has been shown early in fetal development,^19,20 in malignant B cells,²¹ in CD5⁺ B1 B cells²² and in autoantibody repertoires.²³ Adult peripheral repertoires show a V_H gene family utilization pattern that correlates approximately with their genomic complexity, with the V_H3 gene family most frequently used.^17,18,24 However, some of the smaller gene families are over-represented, while only a few members of the V_H3 family are preferentially used.^25,26

Most studies concerning V_H gene family repertoires in healthy adults are based on samples obtained from one or more persons at one time-point. The reported deviation of one individual at one time-point from the normally observed repertoire in the work of Huang et al.,²⁷ which could not be reproduced in samples from the same individual taken 1 year later,²⁸ raises the question of how stable the repertoire is during the lifetime of a healthy individual, and to what extent it is influenced by fluctuations in the environment or health state of the individual at the time of sampling. The lack of information on this subject makes it hard to assess the significance of restricted or deviating V_H gene repertoires reported in ontogeny, senescence or autoimmune disease. Therefore, we developed a DNA-based competitive polymerase chain reaction (PCR) to quantify V_H gene family usage in peripheral blood lymphocytes (PBL) at consecutive time-points. We also performed enzyme-linked immunosorbent assay (ELISA)-based determinations of common serum antibody reactivities to evaluate possible correlations between V_H gene family utilization and serological status.

Materials and methods

Donor samples

Peripheral blood samples from five healthy individuals (two male, three female, all of Caucasian origin) were obtained during a 10-week period at 1-week intervals, and once ≈ 9 years after the first sample was taken. The ages of the individuals at the time the first sample was taken varied between 32 and 48 years. From four of the individuals three time-points were analysed: week 1, week 9 or 10, and 9 years. From one individual, called ‘B’, all 11 time-points were analysed.

Serum, isolation of mononuclear cells and DNA preparation

Whole blood was allowed to clot and, after centrifugation (1000 g, 10 min), the serum thus obtained was aliquoted and frozen at −20° until required for analysis. Mononuclear cells were isolated from heparinized blood on a Lymphoprep gradient (Nycomed™; Pharma AS, Oslo, Norway), washed twice in RPMI (Gibco-BRL, Life Technologies, Ltd, Paisley, UK) without supplements, frozen in fetal calf serum (FCS) with 10% dimethylsulphoxide (DMSO) and stored in liquid nitrogen. For preparation of genomic DNA, the cells were thawed, washed in RPMI and lysed in digestion buffer [100 mm NaCl, 10 mm Tris–HCl (pH 8·0), 0·5% sodium dodecyl sulphate (SDS) and 0·1 mg/ml proteinase K] for 12 hr at 37°. After extraction with saturated (6 m) NaCl, the DNA was precipitated with ethanol and dissolved in 1× TE (10 mm Tris–HCl, pH 7·6, 1 mm EDTA, pH 8·0).

Competitor constructs and PCR primers

A quantitative PCR technique was developed based upon competition for reagents between a target sequence and an engineered competitor sequence. Although originally designed for colorimetric quantification,^29,30 in this study the two competitive fragments were distinguished by a size difference of 20–30 bp and quantified by densitometry. The constructs containing a competitive sequence were made by PCR amplification of the genomic target gene followed by cloning of the fragment into a vector plasmid. The cloning procedure is illustrated in Fig. 1. A small intron (the LacO sequence, 23 bp) was incorporated into the fragment during amplification, by adding this sequence to one of the primers. Thus, the genomic target gene and the competitive fragment contained the same primer sequences as well as an intervening sequence with >90% similarity, and were therefore expected to amplify with equal efficiencies.³¹ Six vector plasmids containing a competing sequence, one for each of the main V_H families (V_H7 was detected by the V_H1-directed primers), were constructed by running two separate amplifications, using the primers shown in Table 1. The two products from the first and the second PCR were cloned together in a two-step process into a pSp72 vector (Promega, Madison, WI) and control-sequenced.

Figure 1.

Construction of the competitor templates. V_H FR3=framework region 3 of the V_H genes (residues 66–92; see ref. 51). LacO=Lac operon sequence, 23 additional nucleotides introducing a length difference.⁵² Black bars above the sequence represent the forward primers, bars below the sequence represent the reverse primers (sequences are given in Tables 1 and 2). (a) Chromosomal DNA and the primers used for cloning the competitor template. (b) The competitor template and the position of the primers used for the quantitative polymerase chain reaction (PCR).

Table 1. Sequences and restriction sites for the oligonucleotides used in constructing the competitor templates.

Target region	Restriction site	Nucleotide sequence
Forward primers, product ‘A’
″VH1 leader region	EcoRI	ACTGGACCTGGAGGGTCTT
″VH2 leader region	EcoRI	TTCTCCACAGGGGTCTTATC
″VH3 leader region	EcoRI	ATGGAGTTTGGGCTGAGCTG
″VH4 leader region	EcoRI	ATGAAACACCTGTGGTTCTTC
″VH5 leader region	EcoRI	TGGGGTCAACCGCCATCCT
″VH6 leader region	EcoRI	TGTCTGTCTCCTTCCTCATC
Reverse primers, product ‘A’
″VH FR3 (cons. seq.)	BamHI	TSTCTCGCACAGTAATACAYG
″VH FR3 (VH2)	Sal1	TSTCTCGCACAGTAATACAYG
″VH FR3 (VH6)	BamHI	TCTCTCTCACAGTAATACACAG
Forward primers, product ‘B’
″VH FR3+LacO seq.	BamHI	AATTGTTATCCGCTCACACAATTCCRTGTATTACTGTGCGAGA
″VH FR3 (VH2)+LacO seq.	Sal1	AATTGTTATCCGCTCACACAATTCCRTGTATTACTGTGCGAGA
″VH FR3 (VH6)+LacO seq.	BamHI	AATTGTTATCCGCTCACACAATTGCTGTGTATTACTGTGAGAGA
Reverse primer, product ‘B’
″JH (cons. seq.)	HindIII	TYACCTGAGGAGACRGTGA

All sequences are given from 5′ to 3′. Product ‘A’ and ‘B’ refer to those shown in Fig. 1.

Cons. seq., consensus sequence for the seven different V_H families; FR3, framework region 3 of the V_H heavy genes (residues 66–92; ref. 51); LacO seq., 23 additional nucleotides introducing a length difference,⁵² shown in italics; leader region, leader region of the V_H gene.

Competitive PCR and quantification

A 160-ng sample of genomic DNA was co-amplified together with different concentrations of the competitor template in a 10-µl reaction volume with 0·3 µm of a V_H family [framework region 1 (FRI)]-specific oligonucleotide primer and a 0·3-µm mix of the JH1-6 oligonucleotides (see Table 2). Taq polymerase (1 U) was used in the supplier buffer together with 175 µm of each dNTP (Perkin-Elmer Biosystems, Stockholm, Sweden). For each of the six different competitor templates at least five different concentrations of competitor were run per person per time-point. The two products were quantified by means of densitometry (CCD Camera, software DIANA II and Tina 2·0 g; Fujifilm Sverige AB, Stockholm, Sweden), and the data were plotted as log₁₀(competitor concentration) against log₁₀[ratio(target ÷ competitor)]. The number of target DNA gene copies was then calculated by extrapolating from the intersection of the curves, where the amounts of target and competitor are equal [log₁₀(ratio[target ÷ competitor])=0]. Data were corrected for the difference in length between the genomic and the competitor templates. Each individual data point is the result of one experiment, unless stated otherwise.

Table 2. Sequences for the oligonucleotides used as primers for competitive polymerase chain reaction (PCR).

Target region	Nucleotide sequence
VH1 FR1	TGG TGC AGT CTG GGG CTG A
VH2 FR1	GAA ACC CAC ACA GAC CCT CA
VH3 FR1	AGG TGC AGC TGG TGG AGT C
VH4 FR1	ACC CTG TCC CTC ACC TGC
VH5 FR1	GTC TCT GAG GAT CTC CTG TA
VH6 FR1	CTC TCA CTC ACC TGT GCC A
JH1,4,5	TGA CCT GAG GAG ACG GTG A
JH2	TGA CCT GAG GAG ACA GTG A
JH3	CTT ACC TGA AGA GAC GGT GA
JH6	TTA CCT GAG GAG ACG GTG A

All sequences are given from 5′ to 3′.

FR1, framework region 1 one of the V_H gene (residues 1–30; see ref. 51)

Antigens and ELISA reactions

Microtitre plates (96-well) were coated with the relevant antigen, 50 µg/ml in phosphate-buffered saline (PBS), by incubation for 2 hr at room temperature or at +4° overnight. The antigens used were purified Fc fragments, double-stranded (ds)DNA, thyroid hormone receptor (THR), thyroglobulin (TG), cytomegalovirus (CMV), herpes simplex virus (HSV), pneumococcal polysaccharides (PPS) 3 and 6, tetanus toxin (TT) and teichon acid. All antigens were obtained from the Swedish Institute for Infectious Disease Control (SMI, Solna, Sweden), apart from the Fc fragments which were purified from commercial Endobulin (IgG content at least 99·5%) (Immuno Sweden AB, Solna, Sweden) by repeated affinity chromatography with anti-κ and -λ antibodies, following papain digestion and separation on a protein G column. After saturation with 1% gelatin in PBS or 0·5% bovine serum albumin (BSA) in PBS (200 µl/well) for 30 min at room temperature, the plates were washed three times in PBS. Serum was serially diluted in PBS containing 1% gelatin. After incubation (2 hr at room temperature or 4° overnight) and washing, the plates were incubated with the relevant conjugates: αΗIg, αΗIgG or αΗIgM, labelled with peroxidase (Dakopatts AB, Älvsjö, Sweden). After washing, the substrate ortho-phenyldiamine (OPD) in buffer solution (0·049 m citric acid, 0·11 m Na₂HPO₄, pH 5·0) was added. The reactions were stopped with 10% SDS and the plates read at 450 nm in a microplate reader (Bio-Tek Instruments Inc., Winooski, Vermont). The ELISA reactions were performed directly after the initial 10-week period. A serum sample for the 9-year time-point was not taken.

Data analysis and statistics

The relative values for V_H gene family usage were calculated by dividing the value obtained for the usage of one V_H gene family by the sum of the values for all V_H gene families, multiplied by 100. Competitor starting concentrations were determined in pmol and transformed to number of template copies before quantification and analysis. For multiparameter correlations of serum antibody levels and V_H gene family usage, the linear Pearson correlation was used.

Results

Competitive quantitative PCR for V_H gene family analysis

V_H gene family usage at a DNA level was analysed using a competitive, quantitative PCR. A protocol involving competitive amplification of a target sequence together with various known concentrations of a cloned version, was adapted from Lundeberg et al.²⁹ As shown in Fig. 2, the distribution of V_H gene family utilization is the same using absolute values (number of template molecules) and relative values (% of total V_H gene family utilization): V_H3>V_H4>V_H1>V_H5>V_H2/V_H6. Single experiments were run several times, using a new template aliquot as well as new dilutions of primers and competitor, obtaining very similar results (data not shown). Interassay variation, measured as coefficient of variation (CV) was defined as follows and varied between 10 and 60%:

Figure 2.

V_H gene family utilization in five healthy individuals. Data shown represent the distribution of the mean values for V_H gene family expression measured at three to 11 different time-points for each individual. The whiskers include the minimal and maximal mean values, whereas the boxes include 25–75% percentile values. The squares indicate medium values. (a) Mean absolute values of V_H gene family utilization in number of template molecules. (b) Mean relative values of V_H gene family utilization, as percentage of total V_H gene family utilization in each sample.

$$\text{CV}=[\text{(standard deviation}÷\text{mean value)}\times 100]\text{.}$$

The magnitude of the CV correlated negatively to the magnitude of family utilization, such that the CV varied between 10 and 20% between different individuals and different time-points for the V_H gene family most frequently used (V_H3).

V_H gene family usage in healthy individuals

Peripheral blood samples were taken from five healthy individuals at three time-points: week 1, week 9 or 10, and 9 years after the first sample. For four of the five individuals the V_H gene family repertoire at the different time-points is shown in Fig. 3. Ranges for V_H gene family utilization are given in Table 3, for each individual as well as for the compiled data.

Figure 3.

Relative V_H gene family utilization in peripheral blood lymphocytes (PBL) from four healthy individuals at three different time-points. V_H gene family utilization is shown as percentage of total V_H gene family utilization. To facilitate comparison between the different time-points, the different V_H gene families are connected with lines. Circles indicate the first time-point (week 1), squares indicate the second time-point (week 9 or 10), whereas the triangles indicate samples taken 9 years after the first sample.

Table 3. Ranges and mean values of V_H gene family utilization.

	VH1	VH2	VH3	VH4	VH5	VH6
Ranges
″Within individuals
″″B (n = 11)	1·9–16·1	0·35–1·4	25·4–88·9	1·2–26·0	1·0–34·7	ND–2·6
″″A (n = 3)	0·9–9·1	0·3–2·2	67·1–79·6	10·5–14·6	2·7–10·7	0·2–4·2
″″C (n = 3)	2·3–9·7	0·2–0·6	71·0–91·7	1·0–15·6	0·2–6·1	0·1–1·3
″″H (n = 3)	3·3–6·2	0·2–1·7	67·0–76·9	13·5–19·3	2·8–3·8	ND–8·7
″″L (n = 3)	8·2–24·0	0·7–2·1	59·3–82·4	2·7–18·2	2·0–3·1	1·7–1·9
″All individuals (mean values)	4·6–16·3	0·3–1·6	60·7–81·9	7·8–16·0	2·6–10·7	0·4–3·7
″All values	0·9–24·0	0·2–2·2	25·4–91·7	1·0–26·0	0·2–34·7	ND–8·7
Mean values (±SD)
″Within individuals
″″B (n = 11)	11·1 (±4·1)	0·7 (±0·3)	60·7 (±21·5)	16·0 (±8·7)	10·7 (±11·8)	0·8 (±0·9)
″″A (n = 3)	4·6 (±4·2)	1·3 (±1·0)	74·1 (±6·4)	12·4 (±2·1)	5·6 (±4·4)	2·0 (±2·1)
″″C (n = 3)	7·2 (±5·0)	0·3 (±0·3)	81·9 (±12·3)	7·8 (±8·5)	3·2 (±1·6)	0·4 (±0·7)
″″H (n = 3)	4·9 (±1·4)	0·7 (±0·9)	71·9 (±4·9)	15·6 (±3·2)	3·2 (±0·5)	3·7 (±4·5)
″″L (n = 3)	16·3 (±7·9)	1·6 (±0·7)	67·2 (±13·1)	10·4 (±7·8)	2·6 (±0·5)	1·8 (±0·1)
″All individuals (mean values)	8·8 (±4·4)	0·9 (±0·5)	71·1 (±7·9)	12·4 (±3·5)	4·9 (±3·4)	1·8 (±0·3)
″All values	9·7 (±5·5)	0·8 (±0·6)	67·0 (±17·9)	13·9 (±7·6)	7·2 (±9·1)	1·4 (±1·9)

ND, non-determinable.

Kinetics of V_H gene usage in one healthy subject during 10 consecutive weeks

Compared to the mean values of week 1 and 9 years, one of the five individuals (‘B’) – not presented in Fig. 3– showed a marked decrease in V_H3 gene family usage at week 10 (from 84·7 to 27·5%), in combination with a rise in V_H4 (from 5·7 to 25·5%) and V_H5 (from 4·1 to 30·9%). From this individual, samples taken during 10 consecutive weeks were analysed and the results are shown in Fig. 4. Figure 4(a) shows patterns of V_H gene family utilization, comparable to that presented in Fig. 3 for the other individuals. Figure 4(b) shows the values for all V_H families over time. Apart from the data obtained at week 10, all other time-points analysed showed a V_H gene family usage similar to week 1 and 9 years, as well as to the data obtained from the other individuals. To exclude the possibility of an artefact, the PCR for the sample from week 10 was run twice more, once preparing new DNA from another aliquot of frozen cells, and once using a different aliquot of primers; almost identical results were obtained: 24·5%, 27·5% and 30·1%, respectively, for V_H3 gene family usage (Fig. 4b).

Figure 4.

Relative V_H gene family utilization in peripheral blood lymphocytes (PBL) from one healthy individual at 10 consecutive weeks and once 9 years later. V_H gene family utilization is shown as percentage of total V_H gene family utilization. (a) Patterns of V_H gene family utilization. Circles indicate the time-points week 1 to week 9. The diamond indicates one time-point (week 10) in which the repertoire deviates from the mean. Dotted lines indicate values obtained with repeated polymerase chain reaction (PCR) amplifications for the same time-point sample. (b) V_H gene family utilization over time, given in weeks (w) or years (yr).

Correlations of V_H gene usage with serum immunoglobulin levels towards foreign antigens and autoantigens

Serum samples obtained during a 10-week period, at the same time-points as the PBL, were analysed for concentrations of antibodies against common foreign antigens and autoantibodies. The anti-PPS6 antibody levels correlated positively with antibodies against DNA (IgG) in individual H and negatively in individuals B (IgG and IgM) and L (IgG only). There was also a positive correlation between anti-PPS6 and TG antibodies in individuals A (P < 0·005) and H (P < 0·05). Individuals A and H showed positive correlations between HSV and TG antibodies (P < 0·05), while there was a negative correlation between HSV and TG (IgG and IgM) antibodies in individual C (P < 0·05). In Fig. 5 the kinetics of serum autoantibody levels are shown for the 10-week period in individual B. A negative correlation was demonstrated between serum rheumatoid factor (RF) antibodies and V_H3 gene family usage (r = − 0·84, P = 0·002), as well as a positive correlation between serum levels of RF antibodies and V_H5 gene family usage (r = 0·64, P = 0·047). Moreover, a negative correlation was found between antibodies against THR of IgG class and V_H6 gene usage (r = − 0·76, P = 0·011) and between THR-IgG and V_H1 gene family usage (r = − 0·68, P = 0·028). With three time-points available on V_H gene family usage for the other four individuals, no general correlations between serum antibodies and V_H gene usage were found.

Figure 5.

Serum antibody levels against common autoantigens as measured by enzyme-linked immunosorbent assay (ELISA) in one healthy individual (B) at 10 consecutive weeks. ♦, RF; ▪, anti-DNA antibodies; ▴, anti-thyroglobulin (TG); •, anti-thyroid hormone receptor (THR) (IgG), where filled symbols indicate IgM antibodies, and open symbols IgG antibodies. (The level of anti-TG IgG antibodies at week 1 is >40 OD units.)

Discussion

Several methods have been established to determine V_H gene usage, from copy DNA libraries and in situ hybridization, to PCR-based methods. The method used in the present study, a quantitative PCR based on competition for reagents between two DNA target sequences, allows for a rapid and highly reproducible analysis of V_H gene family usage in samples from several individuals at more than one time-point. Differences in amplification efficiency between a plasmid and a genomic target (as indicated in ref. 29) should be similar for the different V_H gene families, and the method was therefore designed to be semiquantitative, comparing the relative frequencies of V_H gene family utilization in a given sample. The magnitude of variation between the different assays correlated negatively to the level of utilization of the V_H gene family, such that the CV was very large (up to 70%) for families that are utilized on a level close to the detection limit of the method (20–50 copies), such as V_H2 and V_H6. Therefore, the values should always be considered in relation to the utilization of the other V_H families. This implies that a nested PCR approach is preferable for absolute quantification of single genomic targets present at very low levels. By using the nested approach, sensitivity can be increased to detection of 10 template copies,³² while reducing the variations to <20%.³³ A deviating pattern of V_H gene family utilization, which was observed in one single sample, could be demonstrated repeatedly in aliquots of the same sample using different sets of primers and correlated well with a published case report.²⁷ This shows that the stable pattern we generally found was not introduced by the primers used for amplification, nor by any other part of the method, but rather indicates the genetic control of V_H gene family utilization, as discussed below.

Recently, Khosaka and co-workers suggested a strict genetic control of the V_H gene family repertoire in peripheral blood, based on their studies in monozygotic twins, concordant and discordant for rheumatoid arthritis (RA).³⁴ The observed V_H gene repertoire in our study is very similar to that reported by Kohsaka and co-workers for IgG expressing PBL (99% correlation), as well as to V_H gene family usage in the available repertoire – i.e. the small resting lymphocytes – in PBL (93–98% correlation), as determined by in situ hybridization.^18,35 Data from a more recent, larger study on healthy individuals, in which reverse transcription (RT)–PCR on µ-transcripts was used, showed 95% correlation with our results.³⁶ Together, present and previously published data indicate a very strictly controlled V_H gene repertoire in circulating small lymphocytes.

As one of the individuals (‘B’) showed a marked decrease in V_H3 gene family usage in the sample from week 10 as compared to week 1 and 9 years, eight additional weekly samples (week 2 until week 9) were analysed for this individual. All these samples showed a ‘normal’ V_H gene family utilization pattern, as described above. The V_H gene utilization pattern observed in the week-10 sample appeared to be similar (67% correlation) to the pattern observed by Huang and co-workers in one of two samples (Am-1), taken at a 6-month interval from one individual.^27,28 As mutational rates in the IgM cDNA of this sample were higher than observed in the sample showing a ‘normal’ V_H gene expression pattern, the authors hypothesized that the skewing of the repertoire reflected a recent exposure to antigen.²⁸ Interestingly, a fluctuation in the utilization of mainly the V_H3 gene family – in connection with changes in the utilization of the V_H1 and V_H4 gene families – has been described during the early phase of human immunodeficiency virus (HIV) infection.^36–38 In our sample the observed decrease in V_H3 was associated with an increase in V_H5 gene usage. This deviation in the pattern of V_H gene family usage resembles not only the case study mentioned above, but also the V_H gene family repertoire observed in actual repertoires in PBL (activated B cells and plasma cells),^18,39,40 indicating a (transient) activation of peripheral B cells in the week-10 sample for individual B in this study and in sample Am-1 in the study of Huang and co-workers. The fact that deviations can occur in the repertoire of recirculating B cells is neither surprising nor new. However, the fact that the pattern of deviation is similar in two completely independent cases, using different methods, and moreover resembles the repertoire observed in activated peripheral B cells, may indicate a possible underlying genetic or cellular mechanism.

To further investigate antibody specificities in actual repertoires, serum samples taken at the same time-points as the PBL samples were analysed for concentrations of antibodies against common foreign antigens and autoantibodies. As the PCR method used distinguishes the antibodies based on a nucleotide sequence conserved within the V_H families, it is expected to measure mainly framework contributions to antibody specificities. Fine-tuned antibody specificities, determined by the complementarity-determining region 3 (CDR3) regions of the heavy and the light chain will not be distinguished, with the possible exception of a massively expanded clone with a CDR3 region specific for the challenging antigen. Some common foreign and autoantigens, however, are known to be recognized by the framework region of antibodies, for example superantigens and RF. A significant, negative correlation was shown for individual B between RF in serum and V_H3 gene family usage in the periphery, as well as a positive correlation between RF and V_H5 gene usage. This is in agreement with the notion that most peripheral cells are newly generated and that activated cells disappear from the periphery to become antibody-producing cells. The V_H gene families normally used in anti-immunoglobulin antibodies (RF) are mainly V_H3, sporadically V_H4 and V_H1, but not V_H5.^41–45 After infection or vaccination, serum RF levels are shown to be transiently increased.^46,47 Activation of RF-expressing B lymphocytes in the absence of T-cell help has been shown to lead to peripheral deletion of the RF-expressing clones within 2–3 days.⁴⁸ General T-cell-independent activation can be caused by polyclonal activators. In mice, as well as in humans, staphylococcal protein A selectively stimulates B cells that express V_H families belonging to the V_HIII clan, to which the human V_H3 gene is assigned.^49,50 Together, existing data indicate that infection with a pathogen expressing a V_H3 family-restricted polyclonal stimulator, might be able to cause the temporal skewing of the repertoire and the correlation with RF in serum.

In summary, we have found that the V_H gene family repertoire is remarkably stable over time, as well as similar among individuals with a Caucasian background. We describe a ‘normal’ stable V_H family repertoire, which is genetically controlled, only partly reflects genomic complexity of the V_H gene families^18,34 and is probably expressed by small resting B lymphocytes. In addition, we propose the existence of an ‘activated’ repertoire, characterized by a low V_H3 and a high V_H5 expression as compared to the normal V_H gene repertoire. This form of repertoire skewing is present in naturally activated,¹⁸ as well as in transiently activated, peripheral B-lymphocyte repertoires (this study and refs ²⁷ and ²⁸).

Abbreviations

CV

coefficient of variation

FR

framework region

PPS

pneumococcal polysaccharide

RA

rheumatoid arthritis

RF

rheumatoid factor

THR

thyroid hormone receptor

TG

thyroglobulin

TT

tetanus toxoid