| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Immunology and Bacteriology, University of Glasgow, Western Infirmary, Glasgow, Scotland; and
Neurosciences Group, Institute of Molecular Medicine, John Radcliffe Hospital, Headington, Oxford, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The muscle-type AChR is a pentameric glycoprotein, the fetal isoform
comprising two and one 
, 
, and 
subunit; from 
32-wk
gestation, the subunit is replaced by an 
(3). The
anti-AChR response is polyclonal with autoantibodies directed
against various AChR epitopes (4, 5, 6, 7). However, one
extracellular region on the 
subunit, known as the main immunogenic
region, is a major target for anti-AChR Abs (3). These
Abs are clearly pathogenic as their removal by plasmapheresis markedly
improves symptoms (8). Moreover, transfer of Ab across the
placenta from MG mothers causes transient neonatal MG in 10–15% of
babies (9), and MG can be induced experimentally by
transfer of MG serum Abs or monoclonal anti-AChR Abs into
laboratory animals (3, 10).
The thymus clearly plays an important role in the autoimmune response
in MG (11, 12). The thymic medulla of most patients with
early onset MG (EOMG, age <40 years) is colonized by lymph node-like T
cell areas containing AChR-specific helper T cells and germinal centers
(GC). Plasma cells, which spontaneously secrete anti-AChR
autoantibodies in vitro, are also present and are selectively activated
in vivo (1, 5, 13, 14, 15). Thymectomy results in a fall in
serum anti-AChR titer that correlates with clinical improvement
(2, 5, 16), and anti-AChR autoantibodies have been
produced from hybridomas (17) and Fab libraries derived
from thymus tissue (6, 7). We have observed that 20%
of these GC contain many "plasmablasts" producing anti-AChR and
50% have AChR trapped on their follicular dendritic cells (FDC) (H.
Shiono and N. Willcox, manuscript in preparation). It has been
proposed that the formation of thymic GC may initially be provoked by
rare medullary myoid cells that express native AChR, the only
extramuscular cell type to do so (18, 19). The relative
contribution of the thymus to the production of anti-AChR-secreting
plasma cells compared with the secondary lymphoid organs (lymph nodes
and spleen) is unknown, but the above evidence suggests a major
role.
Therefore, we tested the hypothesis that GC in the thymuses of EOMG patients are undergoing Ag-driven clonal expansion, somatic hypermutation, and selection, resulting in export of anti-AChR-producing plasma cells in a process comparable with the normal GC response in a secondary lymphoid organ. We have analyzed the expressed VH gene repertoire and clonal origins of B cells in thymic GC containing AChR-specific B cells. The results show that these B cells are undergoing Ag-driven clonal proliferation and somatic hypermutation of their Ig V genes and that the rearranged VH gene sequences expressed by these cells show a remarkable degree of heterogeneity that may reflect a protracted response of the immune system to chronic autoantigenic stimulation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A 32-year-old female developed MG and was thymectomized 31 mo later when her serum anti-AChR titer was 40.7 nM (AChR bound per liter of serum), typical for the EOMG patient subgroup (5, 13, 14, 15). The symptoms of the patient had significantly improved at 4 mo and again at 10 mo postoperatively. Thymic tissue was embedded in Tissue-Tek OCT and frozen at -70°C. Serial frozen sections (6- to 8-µm thick) were cut with a cryostat and mounted on slides coated with 2% 3-amino-propyltriethoxy silane (Sigma, Poole, U.K.). Sections were air-dried, fixed in acetone for 10 min, and stored at -70°C with desiccant.
Immunohistochemistry
Sections were stained with mouse mAbs specific for B cells (anti-CD20; DAKO, Carpinteria, CA), T cells (anti-CD3; Dako), proliferating cells (Mib-1 anti-proliferating cell nuclear antigen (anti-PCNA); Santa Cruz Biotechnology, Santa Cruz, CA), FDC (Wue-2), and plasma cells (Wue-1; A. Greiner, University of Würzburg), followed by rabbit anti-mouse Ig (DAKO) and the alkaline phosphatase anti-alkaline phosphatase (APAAP) complex (DAKO). Immune complexes containing APAAP were detected by incubation with new fuchsin substrate, and the sections were counterstained with Mayer’s hematoxylin (Sigma).
Detection of AChR-specific B cells
AChR was solubilized from the AChR -transfected human
rhabdomyosarcoma cell line TE671 (20). This cell line
expresses >80% adult AChR with <20% of the fetal isoform. Extracts
were diluted in 0.05% Triton X-100 in PBS, and 1 ml was incubated with
3 x 106 cpm of
125I--bungarotoxin (BuTx, 2000 Ci/mmol;
Amersham, Little Chalfont, U.K.) for 90 min. Immunoprecipitation
showed that 30–40% of AChR were saturated with
125I-BuTx. Controls for the specificity of
125I-labeled BuTx-AChR binding included the
use of tonsil tissue from non-MG patients, thymus tissue from
clinically typical MG patients with undetectable serum anti-AChR
Abs, and the preincubation of 10x excess unlabeled BuTx (Sigma)
with AChR before incubation with 125I-labeled
BuTx. Frozen sections were fixed with acetone and incubated with
125I--BuTx-labeled AChR or
125I--BuTx alone for 90 min at room
temperature. After washing with PBS, the sections were fixed with 4%
paraformaldehyde and dehydrated in a graded ethanol series. Dried
slides were dipped in melted LM-1 emulsion (Amersham), exposed for 6
days at 4°C, developed, and counterstained with hematoxylin.
Microdissection of GC B cells and DNA extraction
GC that contained B cells specific for AChR were stained with anti-CD20, and sections were overlaid with Scott’s tap water substitute (0.35% NaHCO3, 0.2% MgSO4). GC were accurately excised using sterile blood lancets controlled by Nikon (Melville, NY) Narishige micromanipulators under a Nikon Diaphot inverted microscope at x100 magnification. The excised tissue was digested in 30 µl of proteinase K (0.7 mg/ml; Boehringer Mannheim, Indianapolis, IN) at 50°C for 1 h. Heating to 95°C for 10 min inactivated the enzyme, and the DNA was stored at -20°C.
Amplification and cloning of rearranged heavy chain V genes
A nested PCR system was used to amplify the rearranged heavy
chain V genes. In the first round, genes were amplified with primers
complementary to the VH leader sequences and the
3' end of the JH gene segment-intron. In the
second round of PCR, VH family-specific primers
complementary to the start of the VH framework
region 1 and each JH segment were used. The
primers were designed to amplify all functional, rearranged heavy chain
genes and were based on previously published VH
primer sequences and the human VBASE directory of Ig genes (21, 22). Table I
shows a list of
amplification and sequencing primers. We have shown that these primers
amplify members from each VH family from PBL DNA
under the conditions described below (G. Sims and G. Rowley,
unpublished observations).
|
F' cells (Invitrogen, San Diego, CA). Sequencing and analysis of rearranged VH genes
Plasmid DNA from clones containing gene inserts was prepared using QIAprep spin mini-prep kits (Qiagen, Chatsworth, CA), precipitated, washed thoroughly, and resuspended in 10 mM Tris-HCl, pH 8.5. The VH genes were sequenced using ABI automated cycle sequencing (Applied Biosystems, Foster City, CA) in both directions using primers complementary to sequences flanking the cloning site. Sequences were compared with the human VBASE directory of immunoglobulin genes (22) using DNAPLOT (W. Müller, Institut für Genetik, Köln, Germany) to identify the best matching germline gene segments. The nomenclature for the V, D, and J gene segments adopted here and the definitions of the complementarity determining regions (CDRs) have been previously described (23, 24, 25, 26). Clonally related V genes were identified on the basis of identical V-D-J gene rearrangements and CDR3 regions. The pattern of mutations for each sequence was compared with other sequences from the same GC and germline genes to identify hybrid sequences derived from recombinant VH gene segments. Genealogical trees showing the relationships between B cell clones were constructed by analysis of shared and unshared mutations using phylogenetic analysis using parsimony (27).
Determination of PCR error rate
Multiple replicate sequencing of germline rearrangements and
hybridomas revealed 7 base changes in 34 VH genes
and 1 base change in 7 VH genes, respectively.
Therefore, taking the most conservative estimate, the PCR error rate is
less than one mutation per four VH gene segments
(or 1/1200 bp), which corresponds to 1.1 x
10-5 mutations/bp/cycle. This is similar to
previous estimates of the polymerase error rate for the amplification
of VH genes using nested PCR
(28, 29, 30).
Statistics
The distribution of VH and
JH family usage was assessed using
2 analysis. Data with low expected frequencies
were excluded. The probabilities for individual
VH families and JH genes
were assessed using two-tailed analyses with compensation for multiple
tests using the methods of Bonferroni and Holm.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunohistochemistry was used to identify GC-like structures in
the thymuses of five EOMG patients. Staining of serial sections
revealed that the medulla contained a large number of B cell clusters
covering 10–20% of the area of each section; most thymic GC were
indistinguishable from the GC in tonsil tissue from healthy
individuals. These GC were encapsulated by a follicular mantle (FM) of
densely packed, small CD20+ B cells (Fig. 1A) with crescents of
CD3+ T cells just within the FM, mainly around
the apex of the light zone (Fig. 1B). The FDC formed a dense
reticulated network throughout the light and dark zone, even extending
into the FM (Fig. 1C). Proliferating cells were more
frequent within the GC than the surrounding FM, and light and dark
zones could usually be distinguished (Fig. 1D). Some plasma
cells were associated with the GC; most were clustered at the border of
the GC and the FM. No GC were detected in thymuses from two non-MG
patients who had undergone heart surgery.
|
BuTx/AChR labeling
As observed in other EOMG thymuses (H. Shiono and N.
Willcox, manuscript in preparation),
125I-BuTx alone gave diffuse labeling in
50% of the GC, with a distribution similar to that of FDC. It was
not seen in tonsils or in the thymuses of two otherwise typical but
"seronegative," i.e., without detectable anti-AChR Abs, MG
patients (31). Because this labeling was also blocked by
the cholinergic drug carbamyl choline, this strongly implies that many
thymic GC contain intact AChR trapped on FDC. All four GC examined in
this study specifically bound 125I-BuTx.
With 125I-BuTx-AChR (which includes 30% of
free 125I-BuTx), 20% of GC, including GC
C, showed multiple discrete, moderately positive "centrocytes" and
intensely stained plasmablasts (Fig. 1
E). The latter were
also clearly present in GC A, B, and D often toward the periphery of
the GC, although AChR+ centrocytes were not
readily detectable (Fig. 1F). Approximately 50% of thymic
GC were entirely negative. The proportion of GC exhibiting these
patterns of binding varied between patients (H. Shiono and N.
Willcox, manuscript in preparation). We also consistently saw
occasional individual or clustered heavily labeled cells outside
follicles in the nearby T cell areas where staining for plasma cells
with Wue-1 had been noted. Because of limited sensitivity, this method
probably detects cells with a high concentration of internal Ab more
efficiently, especially the plasma cells known to be present in EOMG
(5, 6, 7). It may be biased toward certain epitopes, and the
stage of B cell differentiation and affinity of the Ab are also likely
to be important. Finally, we found no
125I-BuTx-AChR binding cells in thymic GC from
two seronegative MG patients or in normal tonsil GC (data not
shown).
Amplification of heavy chain V genes from thymic GC
To determine whether thymic GC that contain
anti-AChR-producing B cells are actively engaged in a typical
GC-type response, four GC (A, B, C, and D) and the FM area from GC A
were analyzed. The GC were isolated from sections of an EOMG thymus by
microdissection, and the rearranged V genes were amplified, cloned, and
sequenced. Details of the relative AChR and BuTx staining and the
numbers of rearranged heavy chain V genes examined are summarized in
Table II.
|
Among the 315 sequences examined from the four AChR+ GC, 29 were nonfunctional sequences that included two sets of clonally related sequences, which were presumably the nonfunctional partners of successful V-D-J rearrangements. An additional 24 sequences appeared to be recombinant hybrid sequences composed of the 5' region of one VH gene linked to the 3' region of a separate V-D-J rearrangement. Every effort was made to detect recombinant sequences, as some of these are likely to be PCR artifacts. Recently it has been shown that some apparently hybrid V gene sequences from expanding B cell clones in tonsil tissue are derived from receptor revision events (32). Here we find 12 of the hybrid genes appeared to be functional, and up to half of these may be a result of receptor revision, as the CDR3 region of one V-D-J has recombined with a different upstream VH gene. However, because the nonfunctional genes are not subject to Ag selection and many of the hybrid genes may be PCR artifacts that would otherwise bias the V gene repertoire, these genes were omitted from further analyses.
Functional V gene rearrangements
Two hundred and sixteen sequences derived from 61 independent
functional V gene rearrangements were isolated from the four GC, and 46
sequences derived from 24 functional V genes were obtained from the FM
of GC A (Table II). In each GC we found a diverse assemblage of V genes
using a variety of VH, DH,
and JH gene segments with variable degrees of
somatic mutation. There was no evidence to suggest that B cell clones
from one GC directly seed a response in a neighboring GC, because no
related V genes were found in more than one GC despite their close
proximity. Moreover, with GC A, no rearrangement was found in both the
GC and the surrounding FM. Although we found a total of 18 sets of
clonally related sequences identified by their unique
VH and CDR3s with evolving mutations (see below),
the majority of sequences were either unique or differed by single
substitutions. Considering that the PCR error rate was estimated to be
less than one misincorporation per four VH genes,
related sequences that averaged less than one difference per
VH gene were assumed to be identical, whereas
sequences with more differences were considered to be significantly
mutated members of clonally related sets. This conservative approach
probably ignores some early B cell clones and so may underestimate the
true diversity.
V gene usage
Fig. 2A shows the
VH gene family usage of the functional
independent V gene rearrangements for the FM of GC A, and the GC A, B,
C, and D. Members of a clone were only counted once. The distribution
was similar in the GC and the FM. In each case,
VH3 clearly predominated, and at
least one VH5 rearrangement was also
present, but no VH6 or
VH7 genes were found. The
frequencies of the VH gene families for the
combined GC are shown in Fig. 2B. These frequencies differ
significantly from the expected distribution deduced from their
relative occurrence in the germline (p <
0.01). The VH3 family was
over-represented at the expense of
VH4 genes. The 61 independent
functional V-D-J rearrangements isolated from the GC used a total of 25
different germline VH genes. Among these, nine
were used on three or more occasions, accounting for 65% of the total
(Fig. 2C). Seven of these genes belong to the
VH3 family and are largely
responsible for its over-representation; this is even more marked in
the clonally related sequences (Table III). There was also a clear bias toward
the JH4 gene segment at the expense
of JH2 and
JH1 (Fig. 2D). Five of
the seven different DH gene families were found
in V genes from each GC with DH3 and
DH6 being the most commonly used;
however, in many cases the probable D segment could not be deduced and
in others no D segment was apparent.
|
|
There was no obvious conservation in the size or composition of the CDR3, which ranged from 3 to 21 aa, with 12–18 aa most common. At the 3' end of the CDR3, FDI, FDY, NWFDP, and YYGMDV motifs were common, representing the 5' regions of the JH3, JH4, JH5, and JH6 genes, respectively. Glycine, serine, and tyrosine residues were also common in the CDR3 regions. Indeed, 13 of the 61 functional rearrangements had significant runs of these residues; in the germline they are encoded by most DH3 genes in reading frame (RF) 2, most DH5 genes in RF3 and DH6 segments in RF1, and are probably not unusual.
Somatic mutation in FM and GC B cells
Fig. 3 shows that most of the
functionally rearranged V genes were mutated whether they came from GC
or the FM. Each GC exhibited a similar mutation profile with
considerable variation in the number of mutations per V gene,
reflecting remarkable clonal heterogeneity. For example, the
VH genes isolated from GC B had an average of 8.7
mutations, but contained both unmutated sequences and heavily mutated
genes with up to 52 mutations. The ratios of replacement to silent
mutations (R:S ratios) in the CDRs also varied greatly (Table III),
tending to be higher in the more mutated sequences. Rather
unexpectedly, the majority of VH genes in the FM
of GC A had >11 mutations (mean = 14.1); only 25% were
essentially unmutated (0–2 mutations/VH).
|
Related sequences sharing the same V-D-J rearrangement, but
differing by several nucleotides over the VH
region, were identified in each GC and GC A (FM). In total, 27
functional gene rearrangements were found to have two or more related
sequences. Although the members of nine sets of related V genes
differed by only one or fewer base changes per gene and could therefore
be ascribed to PCR error, eighteen sets of related sequences exhibited
significantly higher mutation rates and were therefore identified as
members of B cell clones (Table III). In three cases, 8C, 9C, and 15D,
related sequences that share common mutations were identified from two
or more separate sections (one cell thick). Because the sections were
amplified independently and each cell only has a single copy of each
DNA rearrangement, these sequences must be derived from different
members of a proliferating B cell clone. The possibility that the
members of these 18 B cell clones could have proliferated elsewhere and
then migrated to the same GC is extremely remote. Therefore, these
results demonstrate that B cell clonal proliferation and somatic
hypermutation are taking place in the thymic GC of MG patients.
Thymic GC are composed of many small B cell clones
Table III details the properties of the functional, clonally
related sequences isolated from the GC (1–6 clones each) and the FM (2
clones). None of the GC contained a single dominant clone. All of the
clones were small with no more than five different members isolated,
and none spanned more than five single-cell sections. Because we only
sampled a fraction of each GC, which spanned from 50 to 120 sections,
there are almost certainly many more clones present than we have
detected, underestimating the true heterogeneity of GC B cell
clones.
Memory and naive B cells proliferate in situ
To illustrate the relationships between clonally related B cells,
six clones isolated from GC C and D, which contained AChR-specific B
cells, as described previously, are displayed as genealogical trees in
Fig. 4. Both GC contained B cell clones
(11C and 17D) expressing a rearranged, unmutated V gene and mutated
variants, which suggests that these clones are recently derived from
naive B cells, whereas the earliest deduced VH
precursors for the other clones contained from 5 to 26 mutations.
Examination of all the B cell clones shows that 6/18 are derived from
naive B cells expressing essentially germline V genes (0–2 mutations),
whereas the remainder are formed from progenitors with a variable
number of VH mutations (Fig. 4C).
Although we cannot be certain that we have isolated the original
progenitor cell in each case, at least some of the 12 clonally related
sets of V genes that did not include a rearranged, unmutated V gene are
most likely to be derived from a progenitor memory B cell. This
demonstrates that both memory and naive B cells are stimulated by Ag in
thymic GC of MG patients to undergo clonal proliferation and mutate
their Ag receptors. Because we find no evidence to suggest that B cells
from one GC directly seed a neighboring GC, we predict that the
responding B cells are derived from the pool of circulating PBLs.
Notably, the majority of sequences isolated from the GC were not
members of any of the 18 clones identified (Table II), which suggests
that a significant proportion of the GC occupants are quiescent
bystander B cells.
|
The 18 B cell clones used a combination of ten different
VH gene segments, eight of which were members of
the VH3 family. Three very similar
VH3 genes,
VH3-30.1,
VH3-30.3, and
VH3-33 isolated from GC A, B, C, and
D, contributed to seven of the B cell clones, suggesting a preference
for AChR binding by this set (Table III). There is also evidence for a
common selection process among three
VH5-51 B cell clones from different
GC (Table IV). Three replacements
recurred in VH5-51 clones isolated
from GC A (FM), C, and D. At aa32 and aa52 there are common
tyrosine-to-phenylalanine (Y
F) substitutions and in CDR2 there is a
recurring aa64QE (glutamine to glutamic acid) substitution.
Moreover, the CDR3 lengths are similar and show similar amino acid
compositions in two cases. These observations suggest that a common
selection process is acting upon independent B cell clones from
separate GC.
|
subunit (6) (Table IV). In
CDR1 there is a recurring aa35ND (asparagine to aspartic acid)
substitution, and in CDR2 at aa54 and aa55 there are consecutive
serine-to-glycine (SG) substitutions and an aa57IT
(isoleucine-to-threonine) replacement. To assess the frequency of the
‘convergent’ replacements among 100 similar sequences, we performed a
basic local alignment search tool (BLAST) search on the anti-AChR
AB5 DNA sequence. The occurrence of the codons encoding aa35D, aa54G,
aa55G, and aa57T were 0, 49, 3, and 12%, respectively. However, it was
inevitable that similar sequences derived from other
VH3 genes would also be retrieved.
The high frequencies of aa54G and aa57T were found because these
residues are encoded by 16/21 and 13/21 of the other
VH3 germline genes, respectively. No
retrieved sequences had aa35D, and none of the sequences with aa55G
were derived from the VH3-48
germline gene. Therefore, the convergent replacements (and aa35ND
and aa55SG in particular) are unlikely to have occurred by chance.
This suggests that a common selection process is acting independently
in different individuals, and B cells in myasthenic GC are
undergoing clonal proliferation, somatic hypermutation, and affinity
selection for AChR binding. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, A–D). AChR was detected on the dendritic processes of
FDC in 50% of these GC, and many of them contained AChR-specific
centrocytes and plasmablasts (Fig. 1, E–F; H. Shiono and N.
Willcox, manuscript in preparation). To study their role in the
autoimmune response, we analyzed the VH gene
repertoire of thymic GC containing AChR-specific B cells. The results
of this study led to the following novel conclusions: 1) the B cell
repertoire in individual GC is strikingly heterogeneous, less than half
of the rearranged VH genes being expressed by B
cell clones as defined by shared V-D-J rearrangements and CDR3; 2) GC B
cells exhibited a wide range of mutation frequencies in their
VH genes; 3) both naive and memory B cells in GC
are stimulated by Ag to proliferate and mutate their Ig V genes; 4)
each GC contains a large number of small B cell clones, no single clone
dominating the response; and 5) the same replacement mutations are
selected for in some B cell clones from independent GC and
anti-AChR Abs from different patients, demonstrating a common
selection process. These results provide strong evidence for ongoing
Ag-driven B cell proliferation and selection in thymic GC of MG
patients, providing a rich source of AChR-specific plasma cell
precursors. The VH gene repertoire in thymic GC
In an earlier study, Guigou et al. (34) demonstrated
expression of multiple VH and
VK gene families in GC of MG patients by in situ
hybridization. To determine whether Ag-specific B cells in these
GC-like structures are undergoing a typical GC response, we analyzed
the Ig V gene sequences from four GC that stained positively for
AChR-specific B cells and for AChR trapped on dendritic processes of
FDC. Our results show that these thymic GC are composed of a remarkably
heterogeneous population of B cells. Their V gene usage was
characteristic of normal adult IgM+ and
IgG+ PBL repertoires derived from single-cell PCR
(29, 35) with over-representation of
JH4 and the
VH3 family (mainly due to the high
frequency of a small number of VH3
genes) and under-representation of
JH1,
JH2, and the
VH4 family (Fig. 2). There was also
considerable variability in the extent of somatic mutation (Fig. 3).
This is in contrast to the V gene repertoire of thymic B cells isolated
from children undergoing heart surgery, which are mostly unmutated with
higher frequencies of VH4,
VH6, and
JH2 genes (36). Unlike
the fetal-like repertoire of these thymic B cells, the myasthenic GC B
cells are characteristic of a normal memory B cell population.
Furthermore, most FM B cells from the tonsil and spleen are also
unmutated (28, 37, 38), whereas we find that mutated B
cells are prevalent in both the FM and the GC in MG. Together our
results suggest that there is extensive migration of the memory B cell
population into thymic GC during a chronic immune response.
We isolated numerous small B cell clones from the GC. Clonally related
V genes were isolated from different sections of the same GC on several
occasions (Table IV). Because each section was amplified independently,
these related V genes could not be derived from the same B cell and
therefore must be derived from a proliferating clone. This is
consistent with the PCNA staining, which showed that B cell
proliferation is taking place within the GC (Fig. 1D).
Sequence analysis revealed that the proliferating B cell clones are
undergoing somatic hypermutation in thymic GC of MG patients.
An unusual feature of these GC is the absence of a dominant B cell
clone. In mice, the GC response to haptens is oligoclonal, often
derived from approximately three progenitor B cells, one or two clones
becoming dominant as the response progresses (39, 40, 41, 42). In
normal human lymphoid tissue, studies have been limited to single cell
analysis of GC of unknown specificity from a single section. Kuppers et
al. (28) found five clones from two tonsil GC, which
accounted for 13/20 of the functional VH genes
examined, and Roers et al. (43) found three clones in a
cervical lymph node, which involved only 8/18 functional
VH genes. In contrast, our extensive analysis of
over 300 VH sequences from multiple sections of
four GC indicates that the autoimmune response in EOMG is heterogeneous
with numerous B cell clones derived from a mixture of naive and mutated
B cell progenitors (Table III, Fig. 4). This heterogeneity correlates
with the well known polyclonal nature of the anti-AChR serum Abs in
MG (3, 4, 5, 6, 7). No dominant B cell clones were identified in
the myasthenic GC. Because many clones were derived from mutated
progenitor cells, and each clone contained only small numbers of B
cells, the potential for further somatic hypermutation and selection of
high affinity memory B cells may be restricted. Similar analyses of
ectopic GC from the synovium in rheumatoid and reactive arthritis have
also revealed multiple proliferating B cell clones (44, 45). Both polyclonal GC and GC with dominant clones were
isolated from the salivary glands of patients with Sjogren’s syndrome
(46). In each case the GC B cells used a different set of
V genes from those expressed in the MG GC; however, the nature
of the stimulatory Ag is unknown.
Ag-driven selection of B cell clones
The variable gene usage, and size and composition of the CDR3, of
VH genes expressed by B cell clones in GC with
specificity for AChR indicates a polyclonal response. The presence of
AChR-specific centrocytes and plasmablasts in GC C indicate that this
GC is mounting a strong response to the autoantigen. GC A, B, and D
also contained AChR-specific plasmablasts, although Ag-specific
centrocytes were not detected. The former are presumably readily
detectable due to high expression of intracellular Ab, whereas
detection of Ag-specific centrocytes is likely to depend on the
affinity of Ag for the low density cell surface receptors. However, we
cannot rule out the possibility that some B cell clones may be
responding to other Ags. Nevertheless, the extensive BuTx staining
does demonstrate that AChR is abundant in the FDC network of each GC,
and there is evidence to suggest that two B cell clones derived from GC
D share a common selection process with an anti-AChR Ab isolated
from a different patient.
Despite the polyclonal nature of the response, common CDR amino acid
replacements were found among the
VH5-51 clones isolated from GC A
(FM), C, and D, which suggests that some clones from different GC are
responding to the same Ag (Table IV). Even more strikingly, we
identified two independent VH3-48 B
cell clones that exhibited a series of CDR amino acid replacements also
found in a Fab (AB5) with specificity for the subunit of AChR,
independently cloned from a different patient (6). None of
these replacements were apparent among nonfunctional frameshift V
genes, which suggests that the mutations were not introduced by a
natural bias in the mutation machinery (30, 47, 48). This
seems even less likely when the pattern and frequency of these
mutations is considered. There are five consecutive AGT codons in the
CDR2 of the VH3-48 germline gene
that encode serine residues. Comparisons of the
VH3-48 clones and the anti-AChR
Fab AB5 reveal that there are no replacements of the first serine
residue and five different amino acid replacements for the second and
third (none of which are serine-to-glycine, SG), whereas five of the
six replacements at aa54 and aa55 convert serine to glycine (SG,
Table IV). Moreover, the low frequency of convergent replacements
(particularly aa35ND and aa55SG) in similar V genes also suggests
that some of the B cell clones have undertaken a similar selection and
affinity maturation process as the AB5 AChR -specific Ab. These
results suggest that B cells with specificity for AChR undergo
Ag-driven clonal proliferation, somatic mutation, and affinity
maturation in the thymic GC of EOMG patients. A series of Fabs and
single-chain variable fragments are now being analyzed to map the
epitope specificities of these B cells and to examine the relationship
between mutation and affinity maturation.
These results highlight the role of thymic GC in trapping autoantigen and generating AChR-specific plasma cells from a heterogeneous population of naive and Ag-specific memory B cells. This is consistent with observations of spontaneous anti-AChR secretion by thymic plasma cells in vitro, and the reduction in anti-AChR titer and clinical improvement following thymectomy in EOMG (13, 16). The localization of GC may be linked to expression of autoantigen in the thymus. Rare muscle-like myoid cells in the thymic medulla express the fetal form of the AChR (11, 18), which is often recognized by the Abs produced by these patients (3, 6). We hypothesize that these myoid cells may provoke GC formation as a result of attack on their AChRs by autoantibodies produced during the early stages of development of MG (19). The description of thymic GC presented here is the first detailed account of a GC response to a defined Ag in humans, and provides the first direct evidence that ectopic GC are responsible for maintaining an autoimmune response through selection of specific self-reactive B cells.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Gary P. Sims, Department of Immunology and Bacteriology, University of Glasgow, Western Infirmary, Glasgow G11 6NT, Scotland, U.K. E-mail address: gs29x{at}udcf.gla.ac.uk
3 Abbreviations used in this paper: MG, myasthenia gravis; AChR, acetylcholine receptor; BuTx, -bungarotoxin; CDR, complementarity determining region; FDC, follicular dendritic cell; FM, follicular mantle; GC, germinal center; RF, reading frame; EOMG, early onset MG; PCNA, proliferating cell nuclear antigen; APAAP, alkaline phosphatase anti-alkaline phosphatase.
Received for publication January 16, 2001. Accepted for publication June 12, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. C. Nacionales, J. S. Weinstein, X.-J. Yan, E. Albesiano, P. Y. Lee, K. M. Kelly-Scumpia, R. Lyons, M. Satoh, N. Chiorazzi, and W. H. Reeves B Cell Proliferation, Somatic Hypermutation, Class Switch Recombination, and Autoantibody Production in Ectopic Lymphoid Tissue in Murine Lupus J. Immunol., April 1, 2009; 182(7): 4226 - 4236. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ragheb, R. Lisak, R. Lewis, G. Van Stavern, F. Gonzales, and K. Simon A Potential Role for B-Cell Activating Factor in the Pathogenesis of Autoimmune Myasthenia Gravis Arch Neurol, October 1, 2008; 65(10): 1358 - 1362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Weinstein, D. C. Nacionales, P. Y. Lee, K. M. Kelly-Scumpia, X.-J. Yan, P. O. Scumpia, D. S. Vale-Cruz, E. Sobel, M. Satoh, N. Chiorazzi, et al. Colocalization of Antigen-Specific B and T Cells within Ectopic Lymphoid Tissue following Immunization with Exogenous Antigen J. Immunol., September 1, 2008; 181(5): 3259 - 3267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. I. Leite, M. Jones, P. Strobel, A. Marx, R. Gold, E. Niks, J. J.G.M. Verschuuren, S. Berrih-Aknin, F. Scaravilli, A. Canelhas, et al. Myasthenia Gravis Thymus: Complement Vulnerability of Epithelial and Myoid Cells, Complement Attack on Them, and Correlations with Autoantibody Status Am. J. Pathol., September 1, 2007; 171(3): 893 - 905. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Jiang, J. Foley, N. Clayton, G. Kissling, M. Jokinen, R. Herbert, and M. Diaz Abrogation of Lupus Nephritis in Activation-Induced Deaminase-Deficient MRL/lpr Mice J. Immunol., June 1, 2007; 178(11): 7422 - 7431. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Bradshaw, A. Orihuela, S. L. McArdel, M. Salajegheh, A. A. Amato, D. A. Hafler, S. A. Greenberg, and K. C. O'Connor A Local Antigen-Driven Humoral Response Is Present in the Inflammatory Myopathies J. Immunol., January 1, 2007; 178(1): 547 - 556. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Le Panse, G. Cizeron-Clairac, J. Bismuth, and S. Berrih-Aknin Microarrays Reveal Distinct Gene Signatures in the Thymus of Seropositive and Seronegative Myasthenia Gravis Patients and the Role of CC Chemokine Ligand 21 in Thymic Hyperplasia J. Immunol., December 1, 2006; 177(11): 7868 - 7879. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Bellan, S. Lazzi, M. Hummel, N. Palummo, M. de Santi, T. Amato, J. Nyagol, E. Sabattini, T. Lazure, S. A. Pileri, et al. Immunoglobulin gene analysis reveals 2 distinct cells of origin for EBV-positive and EBV-negative Burkitt lymphomas Blood, August 1, 2005; 106(3): 1031 - 1036. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Svendsen, C. B. Andersen, N. Willcox, A. J. Coyle, R. Holmdahl, T. Kamradt, and L. Fugger Tracking of Proinflammatory Collagen-Specific T Cells in Early and Late Collagen-Induced Arthritis in Humanized Mice J. Immunol., December 1, 2004; 173(11): 7037 - 7045. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Mehr, H. Edelman, D. Sehgal, and R. Mage Analysis of Mutational Lineage Trees from Sites of Primary and Secondary Ig Gene Diversification in Rabbits and Chickens J. Immunol., April 15, 2004; 172(8): 4790 - 4796. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Shiono, Y. L. Wong, I. Matthews, J.-L. Liu, W. Zhang, G. Sims, A. Meager, D. Beeson, A. Vincent, and N. Willcox Spontaneous production of anti-IFN-{alpha} and anti-IL-12 autoantibodies by thymoma cells from myasthenia gravis patients suggests autoimmunization in the tumor Int. Immunol., August 1, 2003; 15(8): 903 - 913. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Nzula, J. J. Going, and D. I. Stott Antigen-driven Clonal Proliferation, Somatic Hypermutation, and Selection of B Lymphocytes Infiltrating Human Ductal Breast Carcinomas Cancer Res., June 15, 2003; 63(12): 3275 - 3280. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-P. Armengol, C. B. Cardoso-Schmidt, M. Fernandez, X. Ferrer, R. Pujol-Borrell, and M. Juan Chemokines Determine Local Lymphoneogenesis and a Reduction of Circulating CXCR4+ T and CCR7 B and T Lymphocytes in Thyroid Autoimmune Diseases J. Immunol., June 15, 2003; 170(12): 6320 - 6328. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Bauer, M. Zemlin, M. Hummel, S. Pfeiffer, J. Karstaedt, G. Steinhauser, X. Xiao, H. Versmold, and C. Berek Diversification of Ig Heavy Chain Genes in Human Preterm Neonates Prematurely Exposed to Environmental Antigens J. Immunol., August 1, 2002; 169(3): 1349 - 1356. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
