The TCR Vb signature of bacterial superantigens spreads with stimulus strength
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
International Immunology, Vol. 18, No. 10, pp. 1433–1441 ª The Japanese Society for Immunology. 2006. All rights reserved.
doi:10.1093/intimm/dxl076 For permissions, please e-mail: journals.permissions@oxfordjournals.org
The TCR Vb signature of bacterial superantigens
spreads with stimulus strength
Martin Llewelyn1,2, Shiranee Sriskandan1, Nadia Terrazzini2, Jonathan Cohen2 and
Daniel M. Altmann1
1
Department of Infectious Diseases, Faculty of Medicine, Imperial College London W12 0NN, UK
2
Division of Medicine, Brighton and Sussex Medical School, Medical Research Building, University of Sussex,
Falmer BN1 9PS, UK
Keywords: bacterial, human, superantigens, T cells, TCRs
Abstract
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Superantigens (Sags) induce large-scale stimulation of T lymphocytes by a mechanism distinct from
conventional antigen presentation, involving direct MHC binding and stimulation of TCR families
based on Vb gene usage. The specific Vb targets of a given Sag have, since the earliest studies in
murine models, been considered a hallmark of that toxin. Bacterial Sags are implicated in the
aetiology of a wide range of human diseases, although their role has been most clearly defined in
toxic shock syndrome. While Sags have been defined by the Vb-specific changes in T cell repertoire
they induce, human studies of in vitro stimulation or analysis of cells from infected patients have
produced inconsistent findings. Here we have evaluated the contribution of HLA allelic
polymorphisms and strength of stimulus to this response. We show that there are differences in
binding and presentation of the staphylococcal Sag, staphylococcal enterotoxin A (SEA), by different
HLA-DR alleles. We also show that the TCR Vb response, previously thought to be a fixed property
defining a given Sag, varies with stimulus strength such that a broader repertoire of response is seen
at higher concentrations or following presentation by high-binding class II types. Responses of
human Vb8 and Vb1 to SEA, Vb5 to SEB and of Vb12 and Vb13 to streptococcal pyrogenic exotoxin A
are absolutely dependent on stimulus strength. These findings have important implications for
heterogeneity in the response to Sags and the consequent differences in susceptibility to severe toxic
shock.
Introduction
Superantigens (Sags) are powerful T cell stimulatory proteins specific T cell response to Sag may also be important in the
that cross-link the variable regions of particular TCR b chains aetiology in diseases with a putative Sag aetiology, such as
(TCR Vb) with MHC class II molecules on the surface of Kawasaki’s disease, atopic eczema and autoimmunity (8–10).
antigen-presenting cells (APCs) (1, 2). T cell stimulation by It is unclear why such heterogeneity in the response to Sag
Sag is characterized by Vb-specific changes in T cell reper- exposure should exist. Both genetic factors and factors
toire (3). For example, in mice SEB stimulates T cells bear- related to the nature of Sag exposure are likely to be important.
ing Vb3, 7, 8.1, 8.2, 8.3 and 17, while in man Vb3, 12, 14 It was thought that MHC class II polymorphisms were
and 17 are targeted and in contrast toxic shock syndrome unimportant with respect to responses to Sag compared with
toxin (TSST)-1 stimulates Vb3, 10, 15 and 17 in mice and Vb2 in the presentation of conventional peptides bound in the groove
man (4–6). of the heterodimer (11). However, Kotb et al. (12) recently
The best-characterized role for Sags in human disease showed an influence of HLA genotype on disease suscep-
remains that of the Sag exotoxins of Staphylococcus aureus tibility to septic shock following infection with group A
and Streptococcus pyogenes which are believed to trigger the streptococci. A mechanism for these observations was pro-
staphylococcal and streptococcal toxic shock syndromes (7). vided by our demonstration that HLA class II polymorphisms
While the magnitude of T cell response may be important in can determine the magnitude of the T cell response to
toxic shock syndrome, qualitative differences in the TCR Vb- Sags (13). In the course of these experiments, we noted that
Correspondence to: M. Llewelyn; E-mail: m.j.llewelyn@bsms.ac.uk Received 17 August 2005, accepted 11 July 2006
Transmitting editor: M. Feldman Advance Access publication 7 August 20061434 TCR Vb signature of bacterial superantigens
HLA class II polymorphisms associated with greater T cell Table 1. B-LCL genotypes and serotypes
proliferation to a recombinant streptococcal Sag, streptococ-
cal pyrogenic exotoxin A (SPEA), were associated with greater Cell line DRB1* DQA1* DQB1* Serotype
involvement of certain Vb types in the T cell response (13). We
TOK 1502 0103 0601 DR15 DQ6
set out to study this effect in more detail specifically looking at PGF 15011 01021 0602 DR15 DQ6
how Sag stimulus strength impacts on the Vb-specific re- SCHU 1501 01021 0602 DR15 DQ6
sponse. We looked at the influence of HLA class II in view of WT46 1302 01021 0604 DR13 DQ6
the HLA associations which exist for autoimmunity, Kawasaki’s HOR 1302 0102 0604 DR13 DQ6
IDF 11 0501 0301 DR11 DQ7
disease and toxic shock (8, 12, 14). We looked at the influence TISI 1103 0505 0301 DR11 DQ7
of concentration since during the course of streptococcal or SWEIG 1101 0505 0301 DR11 DQ7
staphylococcal infection, different compartments of the im- PRIESS 0401 0301 0302 DR4 DQ8
mune system may encounter widely ranging concentrations of WT51 0401 0301 0302 DR4 DQ8
Sag depending on the site of infection (15). BOLETH 0401 0301 0302 DR4 DQ8
To widen the base of our earlier observations, we began by
determining the influence of HLA-DR b chain polymorphisms
on the binding and presentation of the Sag staphylococcal transformed bare lymphocyte syndrome cell line was used as
enterotoxin A (SEA). Whereas SPEA is prototypic of a family of a class II negative control (17).
Sags which interact with the class II molecule through the
a chain, SEA is prototypic of a family of Sags which interact with Preparation of human PBMCs and T cells
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
the class II molecule through both a high-affinity interaction PBMCs were obtained by Ficoll gradient separation. Class II
with the class II b chain and a lower affinity interaction with the negative T cells were purified by negative selection using
a chain which allows cross-linking (16). Whereas SPEA is a method modified from Lavoie et al. (18). A total of 5 3 106
predominantly presented by HLA-DQ, SEA is predominantly cells were incubated with L243 (anti-HLA-DR), WR18 (anti-
presented by HLA-DR. First, we show that SEA binding is pan-class II), Leu14 (anti-CD14) and Leu19 (anti-CD19),
greater to HLA-DR4 and -DR15 than to DR11 and that T cell followed by two rounds of depletion using anti-mIg DynalTM
proliferation is greater in response to SEA presented by DR4 beads (Dynal Biotech, Oslo, Norway) according to manufac-
than DR11. For SEA, SEB and SPEA, a relationship exists turer’s instructions. Depletion of all HLA class II-expressing
between Sag concentration and Vb repertoire of T cell re- cells was confirmed by failure of purified T cells to proliferate in
sponse such that the response is narrow at the lowest con- response to Sag stimulation unless co-incubated with APCs.
centrations used and progressively broadens as concentration
rises. Importantly though, the Vb specificity of response is Flow cytometric binding assays
preserved with only a small minority of Vb types responding
even at the highest concentrations studied. For SEA and A total of 5 3 105 cultured B cells were incubated with bio-
SPEA, which bind polymorphic regions of the class II molecule, tinylated SEA and washed and after binding were visualized
we show that class II types modulate the effect of Sag using extravidin–PE (Sigma, Poole, UK), by FACS (Facscalibur,
concentration such that class II types associated with greater Cellquest software, Becton-Dickinson, Oxford, UK). A total of
binding also support a broader range of Vb T cell response 20 000 cells falling within a healthy lymphocyte gate were
at any given Sag concentration. Thus, in the setting of human analysed. SEA binding was measured as mean fluorescence
exposure to bacterial Sags, there is likely to be diversity in intensity (MFI) of cells incubated with biotinylated SEA and
the TCR expansions, depending on HLA type and the exact extravidin–PE divided by MFI of cells incubated with unbiotinyl-
nature of the exposure. This may explain heterogeneity in the ated SEA and extravidin–PE. Level of HLA-DR expression was
immunological sequelae of Sag exposure and may explain measured using the mouse mAb L243 which recognizes an
the difficulty in finding consistent Vb-specific changes in T cell epitope on the monomorphic DR a chain, and FITC-labelled
repertoire following clinical exposure to even individual Sags. anti-mouse Ig second layer. DR expression for each cell line
Our findings also give additional insights into the purpose Sags was then expressed as a percentage of the highest DR-
serve in S. aureus and S. pyogenes. expressing cell line. To correct the amount of SEA binding
by a cell line for its level of DR expression, the measured
SPEA binding was divided by the percentage of highest DR
Methods expression and multiplied by 1000.
Toxins
Purified T cell stimulation assays
Recombinant SPEA was expressed in transformed Escher-
ichia coli BL21 as described previously (15). Unconjugated To circumvent the possibility of Sag interactions with donor
and biotinylated purified SEA and SEB and biotinylated re- HLA class II expressed on activated T cells, the role of
combinant SPEA were purchased from Toxin Technology HLA class II in Sag presentation was studied as follows.
(Sarasota, FL, USA). B-lymphoblastoid cells were incubated in the presence of
different concentrations of Sag, washed once and then
B cell lines fixed using 1% PFA. Purified T cells were then incubated
A panel of HLA class II homozygous B-lymphoblastoid cell with B-lymphoblastoid cells. Negative controls were fixed
lines (B-LCLs) was used for Sag-binding assays and for Sag B-lymphoblastoid cells without Sag pre-incubation. At 48 h,
presentation to T cells in proliferation assays (Table 1). An EBV- wells were pulsed with 1 lCi of 3H-thymidine. At 72 h, cellsTCR Vb signature of bacterial superantigens 1435
were harvested using a betaplate harvester and radioactive with SEA presented by HLA-DR4 or HLA-DR11 homozygous
incorporation measured by scintillation counting (Wallac, B-LCLs. Maximum levels of proliferation were approximately
Milton Keynes, UK). twice as high for SEA presented by HLA-DR4 as for pre-
sentation by HLA-DR11. SEA was around three log more
TCR repertoire usage assays potent a stimulator of proliferation in the presence of HLA-DR4
than in the presence of HLA-DR11. Levels of proliferation to
PBMCs from HLA typed donors were stimulated using Sag
SEA presented by DR15 were similar to DR4 and again greater
or PHA, with recombinant human IL-2 (Sigma, Poole, UK) 20
than DR11 (data not shown). No differences in T cell pro-
iu ml1 added at 72 h. After 7 days, cells were harvested,
liferation were observed in either the absence of stimulus or
stained with anti-CD4–FITC and anti-Vb–PE and analysed by
the presence of PHA.
FACS (Facscalibur, Cellquest software, Becton-Dickinson).
For each Vb, percentages of CD4-positive lymphocytes falling
within a resting lymphocyte gate for unstimulated cells or Influence of Sag concentration on the Vb repertoire of the
a blasting-lymphocyte gate for stimulated cells were recorded. T cell response
Our observation that the magnitude of the T cell response to
Sag is determined by its affinity for the HLA class II involved in
Results
presentation caused us to question whether Sag stimulus
Influence of HLA-DR type on SEA binding strength might determine the Vb repertoire of response. We
approached this question first by determining whether Sag
Binding of SEA to APCs is at the class II b chain predominantly
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
concentration influences the Vb response. We studied a range
of HLA-DR but also HLA-DQ. Although other cell surface
of Sags exemplifying different modes of interaction with HLA
factors, particularly co-stimulatory pathways, are involved in
class II, SPEA (which binds to the polymorphic DQ a chain),
T cell activation by Sags, binding at the cell surface is primarily
SEB (which binds the non-polymorphic DR a chain) and SEA
mediated by HLA class II (1). Since DR expression is ~10-fold
(which binds to the DR a and b chains).
higher than DQ (19), observed binding to B-LCLs reflects
Looking first at SEB and SPEA responses, in view of our
predominantly binding to HLA-DR. Binding of SEA to HLA-DR
observation that HLA class II influences responses to SPEA
was assessed using a panel of HLA class II homozygous
(13), HLA-DQ3 homozygous donors were used. For SEB, the
B-LCLs (Fig. 1). Despite showing comparable and, in some
same donors were used but since SEB interacts with HLA
instances, lower levels of DR expression, cell lines homo-
class II through the non-polymorphic HLA-DR a chain, an
zygous for HLA-DR4 and HLA-DR15 showed significantly
influence of HLA class II haplotype would not in any case be
greater SEA binding than cell lines expressing HLA-DR11
expected.
(Fig. 1A). When expressed as SEA bound corrected for level
For each Sag, three or more separate analyses were
of DR expression (Fig. 1B), this difference was statistically
performed using three or more donors. In every case, it was
significant.
observed that the TCR Vb repertoire of response depended on
Sag concentration. Data from representative studies of T cell
Influence of HLA-DR type on SEA presentation Vb repertoire response to SEB and SPEA are shown in Fig. 3.
The ability of high- and low-binding HLA-DR types to present PHA stimulation produced no changes in T cell repertoire
SEA to purified HLA class II negative donor T cells was compared with unstimulated cells (data not shown). At the
assessed. As was previously noted for SPEA, class II-Sag- lowest Sag concentrations, a narrow Vb repertoire of response
binding differences were associated with differences in the was seen. At higher concentrations, the repertoire was
magnitude of the T cell response such that high-binding HLA- expanded to include additional TCR Vb types. This was not
DR types were associated with greater T cell proliferation than a generalization of the Sag response to include all Vbs since
the low-binding HLA-DR11. Figure 2 shows a representative the Vb repertoire of responding T cells remained markedly
experiment in which T cells from a single donor are stimulated skewed towards a few Vb types. Since proportions of the total
400
SEA binding corrected for
200
level of DR expression
A B
(fluorescence shift)
300
SEA binding
150
100 200
50 100
0 0
0 25 50 75 100 DR11 DR4 DR15
HLA - DR expression (% of highest) HLA-DR serotype
Fig. 1. Binding of biotinylated SEA to B cell lines expressing HLA-DR. Bindings of SEA 5 lg ml1 to cell lines expressing HLA-DR4 (WT51,
BOLETH, PRIESS) (n), HLA-DR11 (IDF, TISI, SWEIG) (s), HLA-DR15 (PGF, TOK, SCHU) (¤) and a class II negative cell line (bare lymphocyte
syndrome) (d) are shown (A). SEA binding corrected for level of DR expression is also shown (B) to allow statistical analysis. P values comparing
binding to HLA-DR4 or HLA-DR15 with HLA-DR11 by t-test, HLA-DR0401, P ¼ 0.019; HLA-DR15, P ¼ 0.002.1436 TCR Vb signature of bacterial superantigens
5 HLA-DQA1*03/05 homozygous donors. Here we have shown
that involvement of Vb13.1 in response to SPEA is, in part,
4 related to concentration or stimulus strength (Fig. 3). To ex-
Stimulation Index
plore this relationship in more detail, we observed the Vb-
3 specific T cell response of purified T cells to SPEA or SEA
presented by HLA homozygous B-LCLs. In each case,
2
presentation by HLA types associated with high binding
(DQA1*01 in the case of SPEA and DR4 in the case of SEA)
was associated with Vb responses which were not seen when
1
Sag presentation was by low-binding HLA types (DQA1*05
in the case of SPEA and DR11 in the case of SEA). Specifically,
0
0.01 0.1 1 10
in the case of SPEA, Vb13.1 and Vb3 responses are seen at
[SEA] ng/ml
500 ng ml1 SPEA in the presence of HLA-DQA1*01 but
not DQA1*05. In the case of SEA, while Vb5 is absent from
Fig. 2. Influence of HLA-DR on proliferation of T cells in response to the T cell response at both 1 and 10 pg ml1 SEA, the propor-
SEA. Mean responses of purified single donor T cells to SEA presented tion of Vb9 T cells is increased above base line when present-
by HLA-DR4 homozygous APCs (WT51, BOLETH, PRIESS) (d) and ation is by DR4 but not DR11 (Fig. 5).
homozygous APCs HLA-DR11 (TISI, IDF, SWEIG) (s) are shown
(61SD).
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Discussion
T cell repertoire are measured, the proportions of those Vbs, Since the immunological behaviour of the bacterial Sags was
which respond at the lowest concentrations, fall as concen- first characterized by the laboratories of Kappler, Marrack and
tration rises and other Vb types are drawn into the response. others several years ago, there has been a growing appreci-
The Vb3 and Vb17 responses to SEB in Fig. 3(A) show this. ation of the role played by Sag-induced T cell activation in
Because the response to Sags is Vb specific, non-responsive a number of human disease states including toxic shock,
Vb types are essentially absent from blasting lymphocytes Kawasaki’s disease and autoimmunity (2, 9, 21, 22). Our
analysed by FACS. Vb types which are absent from the current findings help considerably in terms of explaining why
response at low concentrations but appear in the response at disease processes associated with Sag exposure are so un-
high concentrations are therefore Sag-responsive Vb types predictable and heterogeneous in comparison to the results of
even though they may at a particular concentration be only experimental studies.
present at frequencies comparable with PHA-stimulated cells. Our observations of differential SEA binding and present-
The Vb13 response to SPEA in Fig. 3(B) would be an example. ation by HLA-DR types extend our earlier observations re-
Such Vbs are nevertheless clearly Sag responsive since their garding SPEA presentation by HLA-DQ. Furthermore, they
contribution to the total population of T cell blasts has in- have allowed us to undertake a detailed analysis of variations
creased with Sag concentration. Some of the Vb responses in Vb repertoire following exposure to three different Sags
noted have not been described before, for example, the (SEA, SEB and SPEA) which exemplify different modes of
response of Vb5 to SEB and Vb3 to SPEA. Furthermore, interaction with HLA class II. Sags interact with the class II
responses of some established TCR Vb targets were some- molecule in a variety of configurations but utilize what are
times lost at the lowest concentrations used. For example, at essentially only two binding sites (20). One, a high-affinity
the lowest concentrations of SPEA used, expansion of only interaction site, is focused on a highly conserved histadine
Vb14 and not Vb12 was observed. residue at position 81 of the class II b chain but also includes
Looking next at SEA responses, the Vb-specific response to sites which are polymorphic, for example, position 70 (23). The
SEA involves multiple Vb types; Vb1, 5, 8, 9, 16 and 22 have all other, lower affinity site on the class II a chain is polymorphic in
been suggested as targets (20). Since HLA-DR homozygous HLA-DQ but not in HLA-DR (24). SEA interacts at both these
donors are uncommon, it is very difficult to find multiple donors sites but the b chain site is most important (25). SPEA interacts
of comparable HLA-DR type to study the influence of SEA with the DQ a chain site and SEB interacts with the DR a chain
concentration on Vb-specific response in the way used for site (26). Our observations suggest that most, if not all,
studying SPEA and SEB responses in Fig. 3. For these interactions between Sags and HLA class II are likely to be
reasons, the relationship between SEA concentration and Vb subject to the influence of HLA class II polymorphism. Analysis
repertoire of the response to SEA is shown for each studied of the DR b chain sequence suggests that polymorphism
donor separately. A representative experiment is shown in at b70 may explain differential SEA binding. The DR15 and
Fig. 4. Exactly the same relationship between Vb response DR4 (DRB1*0401) b chains have glutamine at this position,
and Sag concentration was found as for SEB and SPEA. At the while DR11 has aspartate at position 70 (27). An alternative
lowest concentration at which a Vb-specific effect is apparent explanation for our observations relating to SEA, but not
(1 pg ml1), a response of only Vb22 is observed. At higher SPEA or SEB binding, is that differences in antigenic peptide
concentrations, Vb9, 8, 1 and 5.1 are drawn into the response. are responsible. An influence of peptide on binding of SEA
and TSST-1 has been clearly demonstrated (18, 28, 29). The
Influence of HLA class II on the Vb T cell response crystal structures of these Sags in complex with class II
We previously reported that Vb13.1 responses to SPEA could demonstrate interactions with peptide (30, 31). However, an
be detected in HLA-DQA1*01 homozygous donors but not in influence of peptide on SEB, and by inference SPEA, has beenTCR Vb signature of bacterial superantigens 1437
Percent of CD4 cells 30 A
25
20
15
10
5
0
13.1
13.1
13.1
13.1
13.1
13.1
13.1
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1
3
5
8
12
14
17
1000 100 10 1 0.1 0.01 PHA
Stimulus (SEB ng/ml or PHA)
50 B
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Percent of CD4 cells
40
30
20
10
0
13.1
13.1
13.1
13.1
13.1
13.1
3
5
8
12
14
17
3
5
8
12
14
17
3
5
8
12
14
17
3
5
8
12
14
17
3
5
8
12
14
17
3
5
8
12
14
17
500 50 5 0.5 0.05 PHA
Stimulus (SPEA ng/ml or PHA)
Fig. 3. Vb-specific changes in CD4 T cell repertoire following Sag stimulation. The response to SEB for eight Vb types is shown (A). Characteristic
skewing of the Vb repertoire towards Vb3, 17, 12 and 14 is seen at the lowest concentrations used. As concentration rises, the contribution of these
Vb types to the total CD4 response falls as other Vbs, notably Vb5, become drawn into the response. Vbs 8 and 13.1 remain absent at even the
highest concentrations used here. The response to SPEA for seven Vb types is shown (B). At the lowest concentration used, only a Vb14 response
is apparent. As concentration rises, Vb12, 13.1 and 3 are sequentially drawn into the response and the contribution of Vb14 T cells to the total
repertoire diminishes. Other Vb types remain absent from the response at the highest concentrations used. Control responses to PHA stimulation
are also shown in each case. Bars show means (n ¼ 4) 6 1SD.
sought and not demonstrated (32). This is in keeping with the used mutations of SEA targeted at the class II a- and b-chain-
crystal structure of the SEB–HLA-DR complex which shows no binding sites. Altered T cell mitogenicity and Vb-specific
interaction with peptide (26). In any event, since antigenic changes in T cell repertoire were observed (33).
peptide presentation is HLA class II restricted, the potential Whether or not a T cell responds to Sag stimulation is likely
role of antigenic peptide does not detract from the importance to be determined by the overall characteristics of the tri-
of the observation that HLA class II polymorphisms influence molecular TCR-Sag-class II interaction (34). This hypothesis
Sag presentation. is supported by our observation here that both higher Sag
Our observation, here and previously (13), that HLA class II concentration and higher Sag-class II-binding affinity are
polymorphisms associated with greater Sag binding are also associated with widening repertoire of Vb response. Although
associated with greater magnitude of T cell response leaves contamination of purified Sags with minute concentrations of
open the question, what is the mechanism behind differences other Sags could plausibly explain the observed widening of
in the magnitude of response? The observations presented Vb response at high concentrations, this is not the case for our
here provide a plausible mechanism. We have shown that observations concerning SPEA which was made with re-
a wider range of T cell Vb types are amenable to stimulation by combinant toxin purified from E. coli. Furthermore, this would
a Sag in the context of HLA class II polymorphisms associated not explain the marked narrowing of Vb repertoire at the lowest
with higher binding. Thus, the proportion of the total T cell concentrations used. Rather it seems to us that the T cell
repertoire amenable to stimulation will evidently be greater. response is determined by stimulus strength which will be
Our findings are in keeping with a previously reported determined by characteristics of the Vb region and the class II
relationship between the affinity of Sag-class II interactions involved and the concentration at which the Sag is acting. The
and the Vb-specific T cell response. In that study, Kotb et al. Vb type targeted by SPEA and SEB in the mouse is mVb8.21438 TCR Vb signature of bacterial superantigens
35
A
30
25
Percent of CD4 cells
20
15
10
5
0
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
100ng 10ng 1ng 100pg 10pg 1pg 0.1pg PHA
25
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
B
20
Percent of CD4 cells
15
10
5
0
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
17
20
22
23
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
1
2
3
8
9
100ng 10ng 1ng 100pg 10pg 1pg 0.1pg PHA
Stimulus (SEA concentration/ml or PHA)
Fig. 4. Influence of SEA dose on Vb-specific changes in T cell repertoire. Data from 10 Vb types are shown for two individuals in panels (A) and (B).
At the lowest concentration for which a Vb-specific change is apparent, a response of only Vb22 is observed. As concentration rises, other Vbs are
drawn into the response sequentially, Vb9, 8, 1 and 5.1. Other Vb types are absent from the response at the irrespective on SEA concentration.
and the sites at which SPEA and SEB interact with this Vb We are aware of six previous reports of the human Vb targets
region have been precisely defined (35, 36). The sites overlap of SPEA (14, 39–43). Many of these studies were conducted
extensively and this undoubtedly explains the overlapping Vb before recombinant SPEA was available and when only limited
specificities of these Sags. The SPEA interaction involves panels of mAbs to different Vb types were available. Vb12 and
more hydrogen bonds with more side-chain atoms than does Vb14 are consistently reported as targets of SPEA in these
the SEB interaction. Thus, the SPEA interaction with TCR is papers. Vb2 and Vb8 were identified in some early papers as
likely to be more dependent on TCR Vb amino acid sequence targets of SPEA probably because of contamination of purified
and this probably explains the narrower repertoire of the Vbs SPEA by other Sags. These Vb types share none of the Vb
response to SPEA (36). Comparison of the TCR Vb amino acid amino acid sequence features of mVb8.2 which are involved
sequence at sites of SPEA interaction identified by Mariuzza in SPEA binding. Of these six reports, only two looked at
et al. is very much in keeping with our observed hierarchy of Vb Vb13 and Vb3. Fleischer et al. (42) used T cell hybridomas-
responsiveness; Vb14, 12, 13, 3 (35–37). For example, E94 of expressing human Vbs of these types and found no responses
SPEA forms a hydrogen bond with the side chain of N28 in the to SPEA at up to 1 lg ml1 using RAJI cells (DQA1*0501) for
mVb8.2 TCR Vb region. The majority of human TCRs has antigen presentation. Their failure to observe response of
a glycine at this position and hydrogen bond with SPEA would these Vb types is therefore in keeping with our observation of
not form (38). Among human Vbs, Vb14, 12 and 13 have N28 Vb3 and Vb13 responses only above this concentration and in
while Vb3 uniquely has D28 which should allow hydrogen the presence of HLA-DQA1*01-expressing APCs. Yoshioka
bond formation with E94 of SPEA (38). et al. (14) reported Vb repertoire changes in PBMCs stimulatedTCR Vb signature of bacterial superantigens 1439
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Fig. 5. Influence of HLA class II on Vb-specific T cell response to SPEA or SEA. Selected Vbs are shown. For SPEA, presentation is by either
HLA-DQA1*01 homozygous APCs (PGF, TOK, WT46) (d) or HLA-DQA1*05 homozygous APCs (IDF, TISI) (s). For SEA, presentation is by either
HLA-DR4 homozygous APCs (WT51, BOLETH, PREISS) (n) or HLA-DR11 homozygous APCs (IDF, TISI, SWEIG) (h). Error bars show 61SD.
by SPEA at 1 lg ml1 and found Vb13 responses in six out of characterized by Vb2 and Vb6 changes and associated with
nine donors of unknown HLA type; Vb3 was not studied. streptococcal pyrogenic exotoxin C-producing strains of
A Vb-specific change in T cell repertoire is a sine qua non S. pyogenes is in keeping with the idea of HLA-associated
of superantigenicity. Such changes are often regarded as Sag responses (14). In toxic shock syndrome, attempts to
evidence that a disease has a Sag aetiology; Kawasaki’s define the Vb-specific changes seen in clinical samples have
disease is a prime example of this. The observation of variable produced conflicting results. Choi et al. found expansion of
changes in Vb repertoire in Kawasaki’s disease has been Vb2 T cells, while Mitchie et al. found reduction of Vb2 T cell
interpreted as meaning that the disease may be the common numbers in patients with toxic shock. Differences in these
end-point of exposure to one of several different Sags (9). studies could have arisen from differences in the timing of
Attempts to establish HLA associations with Kawasaki’s have sampling or from different Sags being present (46, 47). An
largely yielded negative results (44, 45). However, the recently alternative explanation for these observations is that Vb-
reported HLA association in patients with Kawasaki’s disease specific changes are not a fixed property of an individual Sag1440 TCR Vb signature of bacterial superantigens
but determined by factors including concentration at the site of comes of invasive group A streptococcal infections. Nat. Med.
Tcell activation and the HLA context in which the Sag is acting. 8:1398.
13 Llewelyn, M., Sriskandan, S., Peakman, M. et al. 2004. HLA class II
Infection by Sag toxin-producing strains of S. pyogenes and polymorphisms determine responses to bacterial superantigens.
S. aureus is followed by toxic shock syndrome in only a small J. Immunol. 172:1719.
proportion of cases. Furthermore, staphylococcal toxic shock 14 Yoshioka, T., Matsutani, T., Toyosaki-Maeda, T. et al. 2003. Relation
characteristically follows superficial infections while strepto- of streptococcal pyrogenic exotoxin C as a causative superantigen
coccal toxic shock only rarely follows superficial infections and for Kawasaki disease. Pediatr. Res. 53:403.
15 Sriskandan, S., Moyes, D., Buttery, L. K. et al. 1996. Streptococcal
Kawaskai’s disease appears to follow exposure to these same pyrogenic exotoxin A release, distribution, and role in a murine
toxins in the nasopharynx (22). The data presented here model of fasciitis and multiorgan failure due to Streptococcus
explain how both genetic factors and the nature of exposure pyogenes. J. Infect. Dis. 173:1399.
contribute to different outcomes following Sag exposure. 16 Hudson, K. R., Tiedemann, R. E., Urban, R. G., Lowe, S. C.,
Strominger, J. L. and Fraser, J. D. 1995. Staphylococcal
Additionally, these data show how we must be mindful of
enterotoxin A has two cooperative binding sites on major
HLA class II and the nature of exposure in searching for Vb- histocompatibility complex class II. J. Infect. Dis. 182:711.
specific changes in T cell repertoire of patients with putative 17 Hume, C. R., Accolla, R. S. and Lee, J. S. 1987. Defective HLA
Sag-mediated diseases. class II expression in a regulatory mutant is partially complemented
by activated ras oncogenes. Proc. Natl Acad. Sci. USA 84:8603.
18 Lavoie, P. M., McGrath, H., Shoukry, N. H., Cazenave, P. A., Sekaly,
R. P. and Thibodeau, J. 2001. Quantitative relationship between
Acknowledgements MHC class II-superantigen complexes and the balance of T cell
This work was funded through a Medical Research Council (UK) activation versus death. J. Immunol. 166:7229.
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Fellowship to Martin Llewelyn. 19 Gorga, J. C., Horejsi, V., Johnson, D. R., Raghupathy, R. and
Strominger, J. L. 1987. Purification and characterization of class II
histocompatibility antigens from a homozygous human B cell line.
J. Biol. Chem. 262:16087.
Abbreviations 20 Proft, T. and Fraser, J. 2003. Bacterial superantigens. Clin. Exp.
APC antigen-presenting cells Immunol. 133:209.
B-LCL B-lymphoblastoid cell line 21 Kotzin, B. L., Leung, D. Y., Kappler, J. and Marrack, P. 1993.
MFI mean fluorescence intensity Superantigens and their potential role in human disease. Adv.
Sag superantigen Immunol. 54:99.
SEA staphylococcal enterotoxin A 22 Llewelyn, M. and Cohen, J. 2002. Superantigens: microbial agents
SPEA streptococcal pyrogenic exotoxin A that corrupt immunity. Lancet Infect. Dis. 2:156.
TSST toxic shock syndrome toxin 23 Li, Y., Li, H., Dimasi, N. et al. 2001. Crystal structure of a
superantigen bound to the high-affinity, zinc-dependent site
on MHC class II. Immunity 14:93.
24 Sundberg, E. and Jardetzky, T. S. 1999. Structural basis for HLA-
References DQ binding by the streptococcal superantigen SSA. Nat. Struct.
1 Fraser, J. D. 1989. High-affinity binding of staphylococcal enter- Biol. 6:123.
otoxins A and B to HLA-DR. Nature 339:221. 25 Petersson, K., Thunnissen, M., Forsberg, G. and Walse, B. 2002.
2 Marrack, P. and Kappler, J. 1990. The staphylococcal enterotoxins Crystal structure of a SEA variant in complex with MHC class II
and their relatives. Science 248:705. reveals the ability of SEA to crosslink MHC molecules. Structure
3 Kappler, J., Kotzin, B., Herron, L. et al. 1989. V beta-specific (Camb) 10:1619.
stimulation of human T cells by staphylococcal toxins. Science 26 Jardetzky, T. S., Brown, J. H., Gorga, J. C. et al. 1994. Three-
244:811. dimensional structure of a human class II histocompatibility
4 Choi, Y. W., Kotzin, B., Herron, L., Callahan, J., Marrack, P. and molecule complexed with superantigen. Nature 368:711.
Kappler, J. 1989. Interaction of Staphylococcus aureus toxin 27 Robinson, J., Waller, M. J., Parham, P. et al. 2003. IMGT/HLA and
superantigens with human T cells. Proc. Natl Acad. Sci. USA IMGT/MHC: sequence databases for the study of the major
86:8941. histocompatibility complex. Nucleic Acids Res. 31:311.
5 White, J., Herman, A., Pullen, A. M., Kubo, R., Kappler, J. W. and 28 Wen, R., Broussard, D. R., Surman, S., Hogg, T. L., Blackman, M. A.
Marrack, P. 1989. The V beta-specific superantigen staphylococ- and Woodland, D. L. 1997. Carboxy-terminal residues of major
cal enterotoxin B: stimulation of mature T cells and clonal deletion histocompatibility complex class II-associated peptides control
in neonatal mice. Cell 56:27. the presentation of the bacterial superantigen toxic shock
6 Callahan, J. E., Herman, A., Kappler, J. W. and Marrack, P. 1990. syndrome toxin-1 to T cells. Eur. J. Immunol. 27:772.
Stimulation of B10.BR T cells with superantigenic staphylococcal 29 Wen, R., Cole, G. A., Surman, S., Blackman, M. A. and Woodland,
toxins. J. Immunol. 144:2473. D. L. 1996. Major histocompatibility complex class II-associated
7 McCormick, J. K., Yarwood, J. M. and Schlievert, P. M. 2001. Toxic peptides control the presentation of bacterial superantigens to
shock syndrome and bacterial superantigens: an update. Annu. T cells. J. Exp. Med. 183:1083.
Rev. Microbiol. 55:77. 30 Kim, J., Urban, R. G., Strominger, J. L. and Wiley, D. C. 1994. Toxic
8 Wucherpfennig, K. W. 2001. Mechanisms for the induction of shock syndrome toxin-1 complexed with a class II major histo-
autoimmunity by infectious agents. J. Clin. Invest. 108:1097. compatibility molecule HLA-DR1. Science 266:1870.
9 Meissner, H. C. and Leung, D. Y. 2000. Superantigens, conven- 31 Petersson, K., Hakansson, M., Nilsson, H. et al. 2001. Crystal
tional antigens and the etiology of Kawasaki syndrome. Pediatr. structure of a superantigen bound to MHC class II displays zinc
Infect. Dis. J. 19:91. and peptide dependence. EMBO J. 20:3306.
10 Yarwood, J. M., Leung, D. Y. and Schlievert, P. M. 2000. Evidence 32 von Bonin, A., Ehrlich, S., Malcherek, G. and Fleischer, B. 1995.
for the involvement of bacterial superantigens in psoriasis, atopic Major histocompatibility complex class II-associated peptides
dermatitis, and Kawasaki syndrome. FEMS Microbiol. Lett. 192:1. determine the binding of the superantigen toxic shock syndrome
11 Herman, A., Croteau, G., Sekaly, R., Kappler, J. and Marrack, P. toxin-1. Eur. J. Immunol. 25:2894.
1990. HLA-DR alleles differ in their ability to present staphylococ- 33 Newton, D. W., Dohlsten, M., Olsson, C. et al. 1996. Mutations in
cal enterotoxins to T cells. J. Exp. Med. 172:706. the MHC class II binding domains of staphylococcal enterotoxin A
12 Kotb, M., Norrby-Teglund, A., McGeer, A. and Low, D. 2002. differentially affect T cell receptor Vbeta specificity. J. Immunol.
An immunogenetic and molecular basis for differences in out- 157:3988.TCR Vb signature of bacterial superantigens 1441
34 Redpath, S., Alam, S. M., Lin, C. M., O’Rourke, A. M. and 41 Braun, M. A., Gerlach, D., Hartwig, U. F. et al. 1993. Stimulation of
Gascoigne, N. R. J. 1999. Trimolecular interaction of TCR with human T cells by streptococcal superantigen erythrogenic toxins
MHC class II and bacterial superantigen shows a similar affinity (scarlet fever toxins). J. Immunol. 150:2457.
to MHC: peptide ligands. J. Immunol. 163:6. 42 Fleischer, B., Necker, A., Leget, C., Malissen, B. and Romagne, F.
35 Li, H., Llera, A., Tsuchiya, D. et al. 1998. Three-dimensional 1996. Reactivity of mouse T-cell hybridomas expressing human
structure of the complex between a T cell receptor beta chain Vbeta gene segments with staphylococcal and streptococcal
and the superantigen staphylococcal enterotoxin B. Immunity superantigens. Infect. Immun. 64:987.
9:807. 43 Kline, J. B. and Collins, C. M. 1997. Analysis of the interaction be-
36 Sundberg, E. J., Li, H., Llera, A. S. et al. 2002. Structures of two tween the bacterial superantigen streptococcal pyrogenic exotoxin
streptococcal superantigens bound to TCR beta chains reveal A (SpeA) and the human T-cell receptor. Mol. Microbiol. 24:191.
diversity in the architecture of T cell signaling complexes. Structure 44 Barron, K. S. and Silverman, E. D. 1992. MHC class II alleles
(Camb) 10:687. in Kawasaki syndrome—lack of consistent correlation with disease
37 Folch, G. and Lefranc, M. P. 2000. The human T cell receptor beta or cardiac involvement. J. Rheumatol. 19:1790.
variable (TRBV) genes. Exp. Clin. Immunogenet. 17:42. 45 Fildes, N., Burns, J. C., Newburger, J. W., Klitz, W. and Begovich,
38 Berman, H. M., Westbrook, J., Feng, Z. et al. 2000. The Protein A. B. 1992. The HLA class II region and susceptibility to Kawasaki
data bank. Nucleic Acids Res. 28:235. disease. Tissue Antigens. 39:99.
39 Abe, J., Forrester, J., Nakahara, T., Lafferty, J. A., Kotzin, B. L. 46 Wantanabe-Ohnishi, R., Low, D., McGeer, A. et al. 1995. Selective
and Leung, D. Y. 1991. Selective stimulation of human T cells depletion of Vb-bearing T cells in patients with severe invasive
with streptococcal erythrogenic toxins A and B. J. Immunol. group A streptococcal infections and streptococcal toxic shock
146:3747. syndrome. J. Infect. Dis. 171:74.
40 Tomai, M. A., Schlievert, P. M. and Kotb, M. 1992. Distinct T-cell 47 Michie, C., Scott, A., Cheesbrough, J., Beverley, P. and Pasvol, G.
receptor V beta gene usage by human T lymphocytes stimulated 1994. Streptococcal toxic shock like syndrome; evidence of
with the streptococcal pyrogenic exotoxins and pep M5 protein. superantigen activity and its effects on T lymphocyte subsets
Downloaded from http://intimm.oxfordjournals.org/ by guest on October 12, 2015
Infect. Immun. 60:701. in vitro. Clin. Exp. Immunol. 98:140.You can also read
