| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Sam and Rose Stein Institute for Research on Aging and the Theodore Gildred Cancer Center, Department of Medicine, University of California at San Diego, La Jolla, CA 92093
 |
Abstract |
|---|
 Top Abstract IntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
binding activity,
which retains the clan VHIII Fab binding specificity. In
naive adults, about 5% of peripheral B cells and >13% of splenic
IgM-secreting cells display MS binding activity, in association with
high IgM and low IgG circulating anti-MS Ab titers. Neonatal
exposure to MS elicited two distinct temporal phases of immune
responsiveness. The early phase, representing the first approximately 5
wk of life, was associated with MS-specific B cell and T cell
tolerance. Microfluorometric assays revealed that exposure caused a
dramatic MS-specific B cell clonal loss in bone marrow and spleen, but
levels normalized by about 3 wk of life. The late phase (>6 wk of age)
was associated with spontaneous priming for MS-specific T cell
responses and production of MS-specific IgG1 Abs despite long term
persistently depressed in vivo and in vitro MS-specific IgM responses.
In vivo challenge during the late phase induced high frequencies of
MS-specific IgG-secreting cells, indicating recruitment of highly
focused Ab responses that were predominantly encoded by rearrangements
of the S107 family, a member of the VHIII clan. These
studies document the immunodominance of the VH-restricted
Fab binding site on staphylococcal protein A and demonstrate the
diverse effects of a B cell superantigen on the emerging peripheral B
cell compartment. |
Introduction |
|---|
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
In contrast to conventional protein Ags, which are generally recognized by <0.1% of lymphocytes in a naive mature B cell pool, recent reports have described a group of natural proteins that can be bound by the BCR of up to 50% of human peripheral B cells (2). They have been termed B cell superantigens because they display properties comparable to those of bacterial and viral proteins with special stimulatory properties for T cells. We have hypothesized that these exogenous (and perhaps endogenous (3, 4, 5)) proteins may be capable of influencing large scale shifts in B cell clonal representation (reviewed in 6 . B cell superantigens may contribute to the abnormal B cell and Ab patterns detected in HIV infection (7, 8, 9, 10) as well as to the clonal selection of autoreactive B cells in certain autoimmune diseases (reviewed in 11 . In addition, analysis of Ab gene expression in single human B cells has provided evidence of positive selection of VH3 family-expressing clones in healthy adults (12). The in vivo clonal response to experimental challenge with a B cell superantigen has not been previously investigated.
Of putative B cell superantigens, the bacterial membrane protein,
staphylococcal protein A (SpA), is the best characterized. This 42-kDa
membrane protein, composed of five homologous 58–61 amino acid Ig
binding domains and a transmembrane domain (13, 14), is produced by
virtually every clinical isolate of Staphylococcus aureus.
Experimental systems indicate that SpA is a virulence factor (15, 16),
but the mechanism of its effect has not been defined. The immunologic
properties of SpA are linked to two distinct types of Ig binding sites
(i.e., an Fc binding site and a VH-restricted Fab
binding site) that are separate and functionally noncompetitive (17, 18).
In the human system, Fab-mediated SpA binding is restricted to the products of the human VH3 gene family, and at least 16 of the approximately 22 potentially functional VH3 gene segments can encode for an SpA binding site (19, 20, 21, 22). Moreover, 85% of adult VH3 IgM and about 65% of VH3 IgG have Fab-mediated SpA binding activity (19, 20, 23) (G. J. Silverman, unpublished observations), while every unmutated VH3 Ab from fetal liver B cells evaluated was found to bind SpA (22). VH3 IgM Abs can bind SpA with a Kd of <10-8 M, which is comparable to conventional Ab responses (24). Chain recombination studies demonstrate that L chains are relatively passive in binding (20). In vitro, SpA selectively binds to VH3-expressing B cells (2), and in the correct cytokine milieu selectively induces the differentiation of VH3 Ig-producing cells (25, 26). Fab-mediated interactions between human VH3 Ig and SpA were recently shown to induce an in vivo immune complex-mediated inflammatory response in a rabbit model (27).
SpA binding capacity correlates with conserved clan VHIII-specific sequences that include the VH framework 1 and 3 subdomains, and these Abs display a hierarchy of affinities that appears to be linked to specific VH gene segment usage (20, 24). Almost every mammalian species uses clan VHIII-related genes, and their Ab responses include this germline-linked SpA binding activity (28). In the mouse, Abs from the clan VHIII, J606, S107 (29), and large 7183 family (G. J. Silverman, manuscript in preparation) commonly display Fab-mediated binding capacity, but other related murine families may also be associated with this activity.
The current studies were initiated to investigate the outcome of immune exposure to SpA during the neonatal period. For more than 35 years, neonatal injection has been known to cause Ag-specific neonatal unresponsiveness or tolerance in which re-exposure fails to elicit a detectable Ab response (30, 31, 32). It is now believed that T cell tolerance is not an intrinsic property of the newborn immune system but is the nature of the APCs prevalent in the neonate (33, 34, 35). Immune tolerance can be the result of Th cell immune deviation (36, 37, 38, 39, 40, 41) or the induction of suppressor T cell clones (42, 43). In the B cell compartment, neonatal exposure to BCR-selective ligands can also result in profound changes in the relative representation of Ag-specific precursors (44, 45, 46, 47). The current studies examine the immediate and long term immunologic consequences of neonatal exposure of polyclonal BALB/c populations to a B cell superantigen. The results indicate that exposure elicits two distinct temporal phases in which superantigen-specific tolerance dominates the early phase while the late phase is associated with T cell priming and mixed B cell responsiveness. These studies begin to elucidate the predominant immunologic features of B cell superantigen-induced responses.
|
Methods and Materials |
|---|
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
As previously described (2), endotoxin-free recombinant SpA (Repligen, Cambridge MA) was chemically modified to create a form of SpA, termed modified SpA (MS), that retains VHIII Fab binding specificity but is devoid of Fc binding activity. For certain studies, MS was biotinylated. OVA (Sigma, St. Louis, MO), BSA (Sigma), recombinant ß-galactosidase (ß-gal; Sigma), and phycoerythrin (PE) (Molecular Probes, Eugene OR) were used as control protein immunogens.
Mice and immunizations
BALB/cJ mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and were bred at University of California-San Diego under specific pathogen-free conditions. Applying a protocol previously reported for neonatal treatment with a bacterial T cell superantigen (48), neonatal mice, beginning within 24 h of birth, were injected with 100 µg of protein in PBS i.p. every other day for the first 2 wk of life (eight times, 100 µg protein). Certain mice were challenged by s.c. injection with 200 µg of protein emulsified in an equal volume of CFA (Difco, Detroit, MI), then sacrificed 9 or 10 days later (41), and the immune response was examined. Blood was collected retro-orbitally in heparinized tubes, and cells were removed by centrifugation before storage of the plasma at -80°C. Animals were maintained in accordance with the guidelines of the University of California-San Diego Office of Animal Resources. In all experiments, mice were matched for age and sex.
T cell proliferation
Assays for Ag-induced proliferation used draining lymph node suspensions and splenocytes, which can be more reflective of general immune responsiveness (40). Mononuclear cells were cultured at 5 x 105/well in triplicate in sterile 96-well plates (catalogue no. 25860, Corning Costar, Cambridge, MA) in 0.2 ml of serum-free medium (HL-1, BioWhittaker, Walkersville, MD) supplemented with 2 mM L-glutamine without Ag or with MS, control proteins, Con A at 1 µg/ml (Sigma), or purified protein derivative (PPD) of Mycobacterium tuberculosis at 700 U/ml (gift from W. O. Weigle) for 5 days. Proliferation was assessed by the incorporation of 1 µCi of [3H]thymidine during the last 18 h of culture, and the incorporation of label was measured by liquid scintillation counting. To evaluate Ag-specific T cell anergy, certain cultures included murine rIL-2 at 50 U/ml (PharMingen, La Jolla CA) (49, 50).
Enzyme-linked assays of Ab response
A standard ELISA was used to quantify the Ab responses to MS and control Ags. Briefly, microtiter plates (catalogue no. 3690, Corning Costar) were coated overnight with protein at 5 µg/ml in PBS. After blocking with 2% BSA-PBS, serum samples diluted in block were incubated for 4 h at room temperature. The amount of bound Ab was determined by incubation with horseradish peroxidase-labeled affinity-purified goat F(ab')2 anti-mouse IgM or IgG (Jackson ImmunoResearch, West Grove, PA), with values obtained after incubation of substrate for 15 min. Calibration studies with the OVA-specific mAb, clone OVA-14 (Sigma), and the MS-specific mAb, J606, demonstrated >40-fold greater relative sensitivity for the detection of OVA binding. To compare Ab titers, values from different groups of mice were compared at sample dilutions at which the lower mean signal provided an OD of approximately 0.2. A mouse serum pool was used as a standard. For competition studies, sera were titrated to a linear portion of the binding curve (i.e., 1/2,500 for IgM anti-MS activity in naive mice and 1/12,500 for IgG anti-MS activity in postchallenge mice). The relative inhibition capacity of a competitor was determined by comparison to dilutions of the same samples without inhibitor.
Enzyme-linked immunospot (ELISPOT) assays
To quantitate the frequency of Ig- and specific Ab-secreting splenocytes, wells were precoated overnight at 4°C with 5 µg/ml of affinity-purified goat anti-mouse IgM, anti-IgG (Southern Biotechnology Associates, Birmingham, AL), MS, OVA, or 10% FBS in PBS. Plates were washed once with 0.05% Tween-20/PBS and blocked for 1 h at 37°C with HL-1 medium containing 10% FBS that was previously passed over a SpA-Sepharose column (Pharmacia, Piscataway, NJ). Freshly isolated splenocytes in 100 µl of supplemented HL-1 medium were added to triplicate wells with fourfold dilutions starting at 100,000 cells/well and incubated in 5% CO2 at 37°C for 4 h. Adapting a reported protocol (51), in certain studies 6 x 106 fresh splenocytes were first cultured in 5 ml of supplemented HL-1 medium in a 25-ml flask without and with Salmonella minnesota LPS (Sigma) and incubated for 72 h before examination of Ig secretion by ELISPOT assay. To identify the isotype or subclass of specific Ig/Ab secretors, plates were vigorously washed to remove cells, and then alkaline phosphatase-conjugated anti-mouse IgM or IgG (Jackson ImmunoResearch), anti-IgG1, or anti-IgG2a (Southern Biotechnology Associates) was incubated overnight at 4°C. Plates were washed six times before adding 100 µl of 0.6% agarose containing the substrate 5-bromo-4-chloro-3-indoyl phosphate (Sigma) to the well and were developed overnight at room temperature. When possible, values were based on the mean of replicate wells containing >10 but <100 spots/well.
To evaluate the availability of Ag-inducible Th cells, IL-2-producing cells were quantitated in a capture ELISA format employing purified paired mAb, as previously described (52). Microtiter wells were precoated overnight at 4°C with rat anti-mouse IL-2 (PharMingen), then washed and blocked with 10% FBS (SpA column depleted) in HL-1 at 37°C for at least 1 h. Freshly isolated splenocytes in supplemented HL-1 medium were incubated at varying dilutions starting at 200,000 cells/well without or with proteins at 10 µg/ml. After incubation in 5% CO2 at 37°C for 48 h, cells were removed, and plates were washed vigorously, then biotin-conjugated rat anti-mouse IL-2 (PharMingen) was added, and incubated at 4°C overnight. After washing, plates were incubated with streptavidin-alkaline phosphatase (Kirkegaard & Perry, Gaithersburg MD) at 37°C for 30 min, then plates developed and read as described above.
Flow cytometric analysis
Adapting previously reported methods (2), staining of mononuclear cells was performed in PBS containing 1% FBS that had been passed over an SpA column (Pharmacia). In multiparameter studies, Fab-mediated SpA binding was detected with biotin- or PE-labeled MS and compared with that of FITC-labeled anti-mouse IgM (clone 11/41), APC-labeled anti-B220 (clone RA3–6B2), and biotin-labeled anti-mouse IgD (clone AMS 9.1), used with streptavidin-labeled Cychrome (PharMingen). Data were acquired using a FACSCaliber (Becton Dickinson, Sunnyvale, CA) and were analyzed with CellQuest software (Becton Dickinson). Dead cells were excluded based on light scatter and propidium iodide uptake.
RT-PCR assays
Total RNA from control and MS-treated mice was isolated from
40 x 106 fresh splenic cells using Trizol reagent
(Life Technologies, Grand Island, NY). After chloroform extraction and
isopropanol precipitation, all RNA samples were individually stored in
diethylpyrocarbonate-treated H2O at -80°C and
quantitated on a GeneQuant Spectrophotometer (Pharmacia). Briefly, from
each sample, 3 ng of RNA was reverse transcribed with oligo(dT) using
the 1st Strand Synthesis Kit (Life Technologies) to make a 100-µl
stock of cDNA. For the amplification of murine genes, a previously
calibrated amount of cDNA was added to a reaction mixture containing 5
µl of 10x PCR buffer (Boehringer Mannheim, Indianapolis, IN), 2.5
µl of 10 mM dNTP (Pharmacia), and 1.5 µl of each 20 mM primer
solution (Operon, Alameda, CA). Nuclease-free H2O was added
to a final reaction volume of 50 µl. Thermocycling conditions
included a hot start of 95°C for 3 min with the addition of 2 U of
Taq polymerase (Boehringer Mannheim), then 30 cycles of 95°C for 1
min, 56°C for 1 min, and 72°C for 1 min, followed by a final 72°C
for 7 min. Products were individually stored at -20°C. Aliquots from
each PCR sample were separated on a 2% agarose gel containing 1%
ethidium bromide. Under UV light, gel images were directly digitized
through a high resolution digital camera (UVP, San Gabriel, CA) and
then quantitated using ImageQuant according to the manufacturer’s
protocol (Molecular Dynamics, Sunnyvale, CA). The signal strength of
each sample was determined by interpolation from each standard curve of
a pooled splenocyte cDNA sample. The content of ß-actin cDNA was used
to equalize the amount of cDNA used in VH- PCR
amplifications (53) and to normalize the quantitated values. Each of
the seven VH family-specific separate reactions employed
the same antisense 1 CH1-derived
oligonucleotide primer (5'-ggc tta caa tca caa tcc ctg g-3') and a
different sense FR1-derived oligonucleotide (CLIII-FR1A (7183), 5'-gtg
gag tct ggg gga ggc tta-3' (54); CLIII-FR1B (X24), 5'-gga ggt gac ctg
gtg cag cct gga-3'; CLIII-FR1C (J606), 5'-gga gga tgc ttg gtg caa cct
gga-3'; CLIII-FR1D (S107), 5'-gga gga agc ttg gta cag cct ggg-3' (55);
CLII-FR1A (Q52), 5'-gga cct gac ctg gtg cag ccc tca-3'; CLI-FR1A
(J558), 5'-ggt gaa gct tgg ggc ttc agt ga-3' (54); CLI-FR1B (Vgam3),
5'-cag atc cag ttg gtg cag tc-3'). Specificity was confirmed by
sequencing and hybridization studies (54, 55) (F. Hajjar and G.
Silverman, manuscript in preparation).
Statistical analysis
Comparisons between MS-treated and control groups used the rank sum Wilcoxon and Mann-Whitney one-tailed U tests.
|
Results |
|---|
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
In naive mice, Fab-mediated MS binding activity is high in
circulating IgM, while a much lower activity is associated with the IgG
isotype (Figs. 1
and
2). With maturation, the spontaneous levels
of anti-MS Abs in naive mice gradually increase, achieving a steady
level at 7–12 wk of age. Because this age-dependent change parallels
the natural increase in total Ig levels, it does not appear to
represent a preferential induction of anti-MS Abs.
Presumably, this unconventional germline-associated binding specificity
is prevalent among IgM Abs, reflecting the representation of clan
VHIII-related families. The lower activity among IgG Abs is
a carryover in clones selected by other Ags and probably reflects the
lower avidity of IgG binding interactions and the effects of
hypermutation (56).
|
). The early phase
following neonatal MS treatment is associated with IgM anti-MS
levels that are lower than those in naive age-matched controls
(p < 0.0001), and these levels also remain
lower in the late phase (Figs. 1 and 2). MS binding activity in the
serum of naive mice is indistinguishable from the response in PE- or
OVA-treated age-matched control mice (not shown).
|
To determine whether the two distinct temporal phases of altered IgG MS
binding levels reflect different intrinsic levels of immune
responsiveness, mice neonatally treated with MS were subsequently
challenged with the immunogen. During the early phase, challenge with
MS did not affect circulating IgM anti-MS levels in either naive or
MS neonatally conditioned mice (Fig. 2A). Early phase
challenge of neonatally MS-conditioned mice induced only a small IgG
response that was blunted compared with the induced IgG response in
naive age-matched controls (Fig. 2B).
During the late phase, adult MS challenge induced only modest increases
in IgM anti-MS activity in mice that had been neonatally treated
with MS or in age-matched naive mice (Fig. 2E).
Significantly, during the late phase, postchallenge IgG anti-MS
responses of MS neonatally conditioned mice greatly exceeded the
postchallenge responses of naive controls (p <
0.001; Fig. 2F), thus confirming recent reports that
neonatal treatment generally results in priming for Ab responses to
adult challenge (40, 41).
To compare these findings to the responses induced by a conventional
soluble protein Ag, groups of neonatal mice received equivalent
treatment with OVA (Fig. 2, C, D, G,
and H). In age-matched naive mice, the titers of IgM
anti-OVA activity were 30- to 100-fold lower than those of IgM
anti-MS activity (p < 0.0001; Fig. 2, C and G). In these control groups, neonatal
conditioning with OVA did not affect IgM anti-OVA activity during
the early phase (Fig. 2C), while it resulted in lower IgM
anti-OVA activity during the late phase (Fig. 2G)
compared with that in age-matched naive mice. With OVA challenge at 3
wk of age, about half the OVA-conditioned mice displayed OVA-specific
IgM and IgG Ab responses (Fig. 2, C and D),
proliferation, and induction cytokine production (not shown). The
remainder of this group displayed OVA-specific tolerance (see below).
Hence, OVA was less effective than MS at inducing and maintaining
Ag-specific tolerance in the early phase. During the late phase, all
OVA-conditioned mice displayed induced vigorous IgM and IgG
anti-OVA responses following adult challenge. However, after
challenge during the late phase, the OVA-specific Ab activity in
neonatally OVA-primed mice was significantly lower than the
postchallenge anti-MS Ab response in equivalently MS-treated mice
(p = 0.001; Fig. 2, G andH).
Neonatal treatment transiently decreases the frequency of MS-specific B cells
To determine whether the two different phases of immune
responsiveness reflect changes in the representation of MS-specific B
cells, cytofluorometric studies were performed (Fig. 3). Consistent with specificity studies of
human mononuclear cells (2), MS binding activity was restricted to a
subset of murine immature and mature B cells (Fig. 3). In naive adult
mice, about 5% of sIgM+ splenocytes displayed MS binding
activity (57). The proportion of splenic sIgM+ B cells with
MS binding activity varied little among individuals of the same age and
was greatest among sIgD+ peripheral B cells (not shown),
consistent with findings in the human immune system (2, 11). Binding
was detected in both Ig 
- and 
-expressing populations in a normal
distribution (not shown).
|
).
Similar losses of MS-binding mature and immature B cell subsets were
detected in the bone marrow. At this time, neonatal MS treatment also
caused modest decreases in the representation of B220+
mononuclear cells in the spleen (Fig. 3A), suggesting that
this represented MS-induced B cell clonal deletion and not modulation
of sIgM affecting the detection of MS binders. To evaluate whether residual in vivo MS might be blocking the binding of labeled MS during these in vitro staining studies, certain mice were neonatally treated with biotinylated MS. However, in vitro staining of mononuclear cells with fluorochrome-labeled streptavidin failed to detect B cell-associated biotinylated MS (not shown), suggesting that residual MS from the treatments did not significantly contribute to the observed decrease in the frequency of MS-binding B cells.
As the MS-treated animals aged, the proportion of B cells with MS
binding activity in the central and peripheral compartments reverted to
the levels present in naive mice. Within 1 wk after the last MS
treatment, or at 3 wk of age, levels of splenic MS binders were
completely normalized (Fig. 3C). These ephemeral changes
resemble the central clonal deletion described following Ag treatment
of Ag-specific monoclonal Ig transgenic mice (58, 59). However, in our
studies of the polyclonal populations of BALB/cJ mice, it remains a
distinct possibility that with the observed normalization of the
overall frequency of MS-binding B cells masks a subtle but persistent
change in specific BCR VH region expression by the B cell
repertoire.
Neonatal treatment causes a persistent decrease in the frequency of MS-specific IgM-secreting cells
The functional capacity of the B cell compartment was assessed
using ELISPOT studies that detect in vivo activated Ig-secreting cells.
Consistent with reports that this compartment of activated
IgM-secreting B cells is homeostatically autonomous (60, 61), we found
comparable frequencies of splenic IgM-secreting cells
(1300–9000/106 splenocytes) in 4- to 35-wk-old naive mice
raised under specific pathogen-free conditions (Figs. 4 and 5). In naive
mice at the early and late phase ages, B cells secreting MS-binding IgM
represented 15.9 ± 1.7 and 13.1 ± 0.8% (mean ± SEM)
of all IgM-secreting B cells, respectively. The demonstration that this
higher proportion of IgM-secreting B cells in naive mice displays MS
specificity is consistent with a greater detection sensitivity for
binding than can be attained in cytofluorometric studies (Fig. 3).
|
). Notably, in all studies mice that received neonatal treatment
with MS had a lower frequency of spontaneous IgM anti-MS secretors
than naive mice (Figs. 4 and 5) even after prolonged periods (i.e., 72
wk of age; p < 0.004). Hence, this post-treatment
decrease in MS-specific IgM-secreting cells continued throughout the
late phase despite the development of MS-specific T cell responsiveness
(see below). In mice neonatally treated with conventional protein Ags,
OVA (Fig. 4) or ß-gal (not shown), there was no affect on the
frequencies of MS-specific IgM secretors. Using these same methods,
conventional Ag-specific IgM-secreting cells in naive and neonatally
conditioned mice could not be detected (Figs. 4 and 5). Significantly,
in vitro LPS stimulation of splenic B cells caused a three- to fivefold
relative increase in the frequency of IgM secretors without affecting
the proportion of cells secreting MS-specific IgM (not shown). Hence,
despite normal frequencies of MS-binding splenic B cells, mice that
received neonatal MS treatment displayed the same reduction in
MS-specific IgM-secreting cells after in vitro LPS stimulation, a
pattern of in vitro nonresponsiveness associated with anergic B cells
(51, 62, 63, 64, 65).
|
). During the
early phase, MS challenge also failed to induce levels of anti-MS
IgG-secreting cells above the low background level (170 ±
35/106 splenocytes; not shown).
During the late phase, adult challenge with MS did induce an increase
in total IgM-secreting cells that was associated with a modest
increment in IgM anti-MS-secreting cells compared with that in
nonchallenged MS-conditioned mice (Fig. 5). However, the ratios of
MS-specific IgM-secreting cells/total IgM-secreting cells remained
depressed in these late phase neonatally conditioned mice
(p = 0.0003). Challenge with CFA alone of adult
MS-conditioned mice or naive age-matched controls also did not affect
the frequency of total IgM-secreting cells or of MS-specific
IgM-secreting cells (not shown).
Significantly, during this late phase, challenge of MS-conditioned mice
resulted in a significant induction of IgG1-secreting B cells and a
mean proportion of 91% of these cells secreting MS-specific IgG1 Ab
(Fig. 5B). By comparison, the induction of IgG2a-secreting
cells was modest (200–400/106 splenocytes), and
MS-specific IgG2a-secreting B cells were not detected (not shown).
In mice neonatally conditioned with OVA, adult challenge with OVA/CFA
resulted in a significant induction of IgG1 anti-OVA-secreting
cells (Fig. 5B), while few (<200/106
splenocytes) IgG2a-secreting cells were induced (not shown).
Significantly, in these OVA-conditioned mice only a mean of 7% of
induced IgG1-secreting cells was specific for the OVA immunogen (Fig. 5B). Hence, there was a significant difference in the mean
frequency of postimmunization OVA-specific IgG-secreting cells, which
was about 30-fold lower than that induced by equivalent treatment with
MS (p < 0.0004). These findings are consistent
with past reports that the secondary immune response to the
conventional protein Ags in adjuvant are highly degenerate (66, 67). In
contrast, the same treatment regimen with the prototype B cell
superantigen, MS, recruits essentially only MS-specific Ab-forming
cells. In these studies there were no detectable age- or sex-associated
differences in the response patterns of mice between 9 and 30 wk of
age.
T cell response to neonatal treatment with MS
To assess T cell recognition of the immunogen, Ag-specific
proliferation assays were performed on neonatally treated mice after in
vivo challenge. Splenocytes (Fig. 6A) and lymph node cells (not
shown) of 30-day-old mice displayed MS-specific nonresponsiveness, the
hallmark of tolerance (68, 69), while there was a vigorous response to
PPD, a component of the adjuvant used in the challenge immunizations.
Addition of IL-2 to cultures did not result in a proliferative response
to MS (not shown), suggesting that MS nonresponsiveness was not due to
Ag-specific T cell anergy (49, 50). Notably, in control studies
splenocytes (Fig. 6) or purified B cells (not shown) from naive animals
did not proliferate in the presence of MS, suggesting that under these
conditions MS is not a direct B cell mitogen. To evaluate the T cell
responsiveness during the late phase, neonatally treated mice were
challenged at 6 wk of age or older (Fig. 6B). Splenocytes
from these mice displayed an Ag-specific proliferative response (e.g.,
p = 0.014 at 10 µg/ml MS), indicating spontaneous in
vivo priming during the late phase.
|
Neonatally conditioned mice were also evaluated for the frequency of
splenic cytokine-secreting cells during the late phase of the response.
Even without subsequent in vivo exposure, the splenocytes of about half
the group of mice neonatally treated with MS had significantly
increased levels of MS-responsive IL-2 secretors, while all
OVA-conditioned mice contained increased levels of OVA-responsive IL-2
secretors (Fig. 7). Importantly, after
equivalent neonatal priming and adult challenge with their respective
immunogens in CFA, based on cytokine secretion we found comparable
frequencies of MS-specific Th cells in MS-treated mice and of
OVA-specific Th cells in OVA-treated mice (Fig. 7). This suggests that
a greater availability of T cell help (i.e., clonal size of Ag-specific
T effector cells) was not responsible for the higher frequency of
inducible immunogen-specific IgG-secreting cells detected during
late phase-associated secondary anti-MS responses.
|
To further characterize the B cell response to MS challenge, based
on the documented IgG1 bias of the induced response we evaluated the
expressed 1 gene rearrangements for evidence of
VH-mediated clonal restriction. For these studies, we
developed an RT-PCR based system to compare splenocyte samples for
their relative content of mRNA from each of seven different
VH families that are representative of the three clans of
murine VH genes. These gene families are cumulatively the
source of >85% of the expressed Ig repertoire of BALB/c mice
(71, 72, 73, 74).
In general, for any sample the cumulative level of VH-1
gene expression paralleled the relative frequencies of IgG1-secreting
cells in the splenocyte samples (Figs. 8 and
9). As expected, splenic samples from naive
mice contained very low levels of VH-1 mRNA. Mice that
did not receive both a priming and an adult challenge of Ag also
displayed very low levels of detectable 1 rearrangement mRNA. By
comparison to naive mice, in each sample from a mouse neonatally
conditioned with OVA and then challenged as an adult (n
= 4), there was at least a 10-fold mean increase in the expression of
all seven VH families, indicating a
non-VH-selective stimulation. Significantly, the
VH gene expression of mice that were MS conditioned and
challenged (n = 6) uniformly demonstrated much greater
levels of expression of the S107 family (VHIII clan),
representing 40-fold mean higher levels than those in the OVA-treated
and challenged mice (p < 0.005) and >400-fold
mean higher levels than those in naive mice. In comparisons of these
MS-treated and the OVA-treated mice, the MS responders also displayed
lower expression levels of the six other VH families, which
were statistically significant for the Q52 (p
< 0.005), 7183 (p < 0.005), and X24 families
(p < 0.005), indicating selective B cell
recruitment in the anti-MS response. These data document the
dominance of a clan VHIII-restricted component in the
cellular recruitment into the secondary immune response to a
bacterial B cell superantigen.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
Within the current investigations of SpA, an experimental Ag studied for decades, we provide the first characterization of the clonal response following in vivo challenge with this B cell superantigen. Both cytofluorometric studies and serologic assays document that the frequency of B cells in the naive repertoire capable of Fab-mediated SpA interactions is orders of magnitude greater than that for conventional protein Ags. The specific IgG Ab response after challenge of primed adults is also of higher magnitude. To independently survey the functional capacity of the B cell compartment, we quantitated spontaneous Ig/Ab-secreting cells. We found that a mean of 13% of the IgM-secreting cells in naive animals displays Fab-mediated SpA binding activity, presumably reflecting the composition of the primary B cell repertoire (60, 61). Notably, during secondary immune responses to MS >90% of IgG1-secreting cells are MS specific, while in naive mice IgG-secreting cells are infrequent and MS-specific IgG-secreting cells are undetectable under the same conditions.
The IgG response to MS contrasts with results for OVA, which appears to induce a highly degenerate response, containing only a small proportion of Ag-specific IgG-responding B cells, confirming earlier reports for other soluble homogeneous conventional protein Ags (66, 67). Calibration experiments of the immunospot assay demonstrated that it is unlikely that sensitivity thresholds (e.g., the epitopic heterogeneity or precoating density) artifactually limited the detection of OVA-specific Ig secretors. Similarly, we found that after equivalent neonatal priming and adult challenge with another conventional protein immunogen, ß-gal, <20% of the IgG secretors are Ag specific (not shown). Therefore, in studies designed to juxtapose MS-induced findings with responses to conventional homogeneous soluble protein Ags, we believe that the results fairly highlight the potent ability of SpA to induce highly focused Ab responses.
The current studies address the fundamental question of the role
of the clan VHIII-restricted Fab binding site in active
immune responses to SpA. To sample the specific repertoire we could
have evaluated panels of MS-specific hybridomas, but there was concern
in light of recent reports that this method can rescue B cells that are
anergic in vivo, which are therefore not representative of the active
in vivo response (65). In support of our more direct approach, we found
that >90% of IgG1-forming cells induced by secondary immunization
recognize SpA, and >80% of these MS-specific IgG Abs can be blocked
by the binding of a VHIII Ig from a naive host. Hence, most
of this MS-induced response appears to be directed at the site
responsible for binding of VHIII-encoded Abs. Therefore, we
surveyed for VH-1 expression and discovered that the
secondary responses to MS recruited an overwhelming predominance of
S107 rearrangements, which have been shown to generally encode for SpA
binding activity (29). These data document that the clan
VHIII-restricted Fab binding site of SpA is dominant during
active immune responses.
Epitope dominance resulting in the recruitment of a highly restricted Ab response has been described in several systems, but most often this pattern has been associated with structurally simpler Ags (e.g., carbohydrates, haptens, or peptides). In the few reports of highly restricted responses to protein Ags, the dominant B cell epitopes are generally redundant determinants in functionally multivalent proteins (75, 76). Hence, while the focused B cell response elicited by SpA can be considered as another type of epitope dominance, the clan VHIII restriction of binders and the high frequency of potential responders are very different from those of other characterized Ags. Taken together, these immunologic properties of a very high binder frequency in the naive repertoire, immunodominance of the Fab binding site, and recruitment of a VH-restricted B cell responders that dominates the immune response embody the hallmarks of a B cell superantigen.
MS is a superantigen, but a B cell is still a B cell
Aside from the unconventional structural correlates
(VH dependence) of binding that are responsible for the
high frequency of potential responders, in most important respects the
cellular correlates for immune responsiveness to this B cell
superantigen do not appear to be distinct from those of conventional
Ags (i.e., SpA binders display a range of affinities, and T lymphocytes
play a central role in determining the outcome of exposure).
Preliminary findings in S107-1 libraries suggest that secondary
anti-MS responses are associated with the expression of clonally
restricted gene sets representing independent S107 gene rearrangements,
which presumably were selected based on their relative affinity within
the accessible B cell pool (G. J. Silverman, manuscript in
preparation).
The current studies also suggest that this immunogen can induce a range of Ag-specific B cell clonal fates among the specific B cells in polyclonal lymphocyte populations, findings that previously could only be considered in transgenic monoclonal Ig systems. Hence, our studies begin to illustrate how a B cell superantigen can be exploited for the dissection of the events that contribute to active immune responses and stage-dependent influences on the formation of the mature B cell repertoire. Offering another attraction for the future development of model systems, SpA can be manipulated to alter intrinsic Fab binding affinity and effective valency.
Neonatal treatment with a B cell superantigen causes shifts in tolerance and immunologic memory
During the early phase following neonatal treatment we found an
absence of proliferative, cytokine, or Ab responses to Ag re-exposure.
In other reports it has been shown that Ag-specific tolerance can
affect all types of effector Th cells (42, 43, 70). The inducible
tolerance associated with the neonatal period has been ascribed to an
age-associated immune defect in which Ag is presented in the context of
inadequate costimulatory signals (33, 34, 35, 39, 77). Ag-specific
tolerance can be mediated by dendritic cells (78) or by resting B cells
(79, 80, 81), which are especially susceptible to tolerization or negative
selection during early development (82) and which clearly display MS
binding capacity at exceptionally high frequency (Fig. 3).
Consistent with earlier (30, 31) and more recent reports (40, 41), our studies demonstrate that the eventual outcome of neonatal Ag conditioning is an enhanced adult capacity for Ag-specific Ab and proliferative responses, indicating in vivo priming and immunologic memory. We found that adult challenge induces an MS-specific IgG1 Ab response, suggesting that neonatal exposure to a B cell superantigen, such as conventional self and non-self Ags (36, 37, 38, 40, 41), leads to a Th2-biased response. Notably, the duration of complete tolerance to MS could also be prolonged by weekly i.p. instillations of 100 µg of MS in alum (not shown) or by the substitution of native SpA that has higher intrinsic Fab binding affinity (G. J. Silverman, unpublished observations), which is predictable based on classical models of neonatal tolerance (30, 83).
Our studies of the late phase of the response demonstrate that despite the same high frequency of peripheral MS-binding B cells as that detected in naive mice, neonatal MS treatment results in lower IgM anti-MS circulating Ab levels and the persistent loss of in vitro spontaneous anti-MS IgM secretion. In transgenic Ig systems, this loss of spontaneous Ig-secreting cells of a defined Ag specificity has been used as a marker of clonal functional inactivation (i.e., anergy) (63, 64, 65), although in the MS neonatal treatment system other regulatory or trafficking-related processes have not been ruled out.
The late phase of the MS response is also associated with spontaneous priming of MS-specific T cell responses and production of IgG anti-MS Abs. Presumably, these divergent effects on the IgM and IgG responses, demonstrated to persist for more than a year after the last infusion, reflect the continued influence of an Ag that may be retained in lymphoid tissue for years (84). While partial tolerance has been previously reported following neonatal exposure to mixed immunogens, this was interpreted as representing nonresponsiveness to certain antigenic determinants at the same time as other determinants are the targets of active immune responses (31). In contrast, our studies characterize the response to a homogeneous Ag that is recognized by B cells at very high frequency via a VHIII-restricted Fab binding site.
Based on the transgenic anti-HEL/HEL tolerance model, MS-specific partial anergy would be predicted to involve B cell clones expressing BCR with higher binding affinity (85). Functional inactivation of MS-specific clones could occur during the continuous generation of immature/transitional B cells in the bone marrow, which are highly sensitive to BCR-mediated tolerance induction and negative selection (86, 87), or their newly emergent counterparts in the periphery that have similar sensitivity (58, 82, 88). A possibly related process is responsible for the sensitivity to tolerization reportedly associated with memory B cell differentiation (89, 90, 91, 92, 93, 94). Hence, we question whether the active anti-MS response truly reflects the products of B lymphocytes expressing BCR with the highest binding affinity. Significantly, recent rapid amplification of cDNA ends-PCR studies have demonstrated that neonatal MS treatment results in a persistent loss of VH 7183 family expression in the primary B cell pool (J. V. Nayak and G. J. Silverman, manuscript in preparation), possibly due to MS-induced clonal loss. The further dissection of these phenomena will therefore include a more rigorous definition of the relative affinities of the superantigen binding activities conveyed by different murine Ab genes.
B cell superantigens and repertoire development
During the neonatal stage of development, BCR-mediated signals apparently affect the rate at which the maturing B lymphocytes fill the periphery. In a recent report maternal IgG was shown to act as a nonselective BCR-interacting agent on the emerging B cell compartment, resulting in enhanced cellular expansion in the periphery. Concurrently, this influence delayed the spontaneous differentiation of evolving B cell clones to the Ig-secreting stage (95). In a more clonally restricted manner, specific encounters with cross-reacting Ag or their anti-idiotypic surrogates during early immune development have been reported to shift the Ab V gene/clonal representation in subsequent Ag-specific responses (44, 45, 46, 47). Our studies suggest that neonatal exposure to a bacterial product conveys a VH-selective stimulation/differentiation of certain superantigen-specific B cell clones, concurrent with the functional inhibition or down-regulation of others. As discussed, Ag persistence is almost certainly responsible for the observed prolonged effects on the B cell compartment, while the range of BCR affinities for the B cell superantigen probably contributes to the different clonal outcomes, including the continual or recurrent "tickling" of certain clones (96, 97). Hence, we wonder whether the current results predict that neonatal B cell superantigen exposure can bias the accessible pool of B cell responders for subsequent exposures to conventional ligands, and whether an equivalent natural process may be responsible for evidence in humans and mice of unconventional VH family-directed selection (12, 98). Recent reports of endogenous proteins with B cell superantigen-like properties (3, 4, 5) warrant further investigation of these postulated processes.
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Gregg J. Silverman, Department of Medicine-0663, University of California at San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0663. E-mail address:
3 Current address: University of Pittsburgh School of Medicine, Pittsburgh Cancer Institute, BST 1013E, 211 Lothrop St., Pittsburgh, PA 15213.
4 Current address: Klinische Forschergruppe fur Rheumatologie, Universitat Freiburg, Breisacher Str. 64, 79106 Freiburg i. Breisgau, Germany.
5 Abbreviations used in this paper: BCR, B cell receptor for Ag; SpA, staphylococcal protein A; MS, chemically modified staphylococcal protein A that retains Fab binding activity; ß-gal, ß-galactosidase; PE, phycoerythrin; PPD, purified protein derivative; ELISPOT, enzyme-linked immunospot; sIgM, surface IgM.
Received for publication May 8, 1998. Accepted for publication July 8, 1998.
|
References |
|---|
|
TopAbstractIntroductionMethods and MaterialsResultsDiscussionReferences |
|---|
-toxin-deficient mutants of Staphylococcus aureus isolated by allele replacement. Infect. Immun. 55:3103.-globulin. J. Immunol. 87:147.
