bacterial and viral superantigens: roles in autoimmunity? · caninteractwithaparticular...

Annals of the Rheumatic Diseases 1993; 52: S6-S16

Bacterial and viral superantigens: roles inautoimmunity?

Hans Acha-Orbea

AbstractSuperantigens are bacterial, viral, orretroviral proteins which can activatespecifically a large proportion of T cells.In contrast with classical peptide antigenrecognition, superantigens do not requireprocessing to small peptides but act ascomplete or partially processed proteins.They can bind to major histocompatibilitycomplex class II molecules and stimulateT cells expressing particular T cellreceptor Vp chains. The other poly-morphic parts of the T cell receptor,which are crucial for classical antigenrecognition, are not important for thisinteraction. When this strategy is used alarge proportion of the host immunesystem can be activated shortly afterinfection. The activated cells have a widevariety of antigen specificities. The abilityto stimulate polyclonal B (IgG) as well asT cell responses raises possibilities of arole for superantigens in the induction ofautoimmune diseases. Superantigenshave been a great tool in the hands ofimmunologists in unravelling some of thebasic mechanisms of tolerance andimmunity.

(Ann Rheum Dis 1993; 52: S6-S 16)

Ludwig Institute forCancer Research,Lausanne Branch,1066 Epalinges,SwitzerlandH Acha-Orbea

Classical antigen recognition by T cells occurs

by interaction of self major histocompatibilitycomplex (MHC) proteins, which bind smallfragments of proteins, with T cell receptor(TCR). These three components are brieflydescribed below. More detailed informationcan be found in excellent reviews and thereferences therein. 1-6The MHC encodes many different proteins.

Among these the MHC class I and class IImolecules are important in this context. Thesedimeric molecules are highly polymorphic inthe population. Of the order of 50 alleles havebeen defined for the different MHC class I andclass II molecules. The major function of thesemolecules is the presentation of antigens to theT cells by acting as receptors able to bindthousands of 8-9 (MHC class I) or 9-16(MHC class II) amino acid long processedpeptides. Such peptides are generated fromproteins in the cytoplasm (for MHC class Ipresentation) or in endosomes (for MHC classII presentation) by specialised proteases. Mostof the antigens entering the cell from theoutside will be associated with MHC class II

molecules, whereas proteins produced within

the cell are presented by MHC class Imolecules. Despite the high numbers ofpeptides which can be bound to a particularMHC molecule, these receptors show strikingspecificity for peptides they bind. Thisspecificity is due to differences in the aminoacid sequence in the N-terminal part ofMHCmolcules which forms the peptide bindingpocket that can accommodate two to fouramino acids of the peptide. Specific aminoacids could be implicated at specificlocalisations in the bound peptides for severaldifferent MHC class I as well as class IImolecules. Different MHC class I and class IIproteins bind different sets of peptides. Singleamino acid differences in this peptide bindingregion can heavily influence the peptides whichcan hind to a particular MHC class I or classII molecule.The formation of a trimolecular complex

between MHC, peptide, and TCR is highlyspecific in that a particular TCR can interactonly with specific MHC molecules which havebound a particular peptide. For this interactionTCR and MHC sequences as well as peptidesequences are important. As recognitionoccurs almost exclusively with self MHCmolecules, and the same peptides cannot bepresented in the context of another MHCmolecule, it is called MHC restricted. ThisMHC restricted antigen recognition by Tlymphocytes leads to formation of atrimolecular complex between short antigenicpeptides bound to MHC molecules andrecognition by a TCR (fig IA).

In addition to these molecules, a series ofother receptors and coreceptors intensify thisinteraction. Most of these intensifyingmolecules do not contribute to the specificityof the interaction and are not discussed here.One pair of these coreceptors intensifies theinteraction of the TCR with MHC class Imolecules (CD8) or MHC class II molecules(CD4). Cytotoxic T cells are found amongstthe former, and helper T cells amongst thelatter.The TCR is the end product of a series of

recombination events during the maturationand differentiation of T cells. This bringstogether the constant region which is sharedbetween all the receptor molecules with thetwo or three highly polymorphic parts: variable(V), diversity (D) (in the case of the I chainonly), and junctional (j) elements. Theseelements are encoded as V and J clusters,separated but in the same chromosomallocalisation as the constant region. In addition,a tremendous heterogeneity is introduced by

S6

on Decem

ber 20, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.52.S

uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Bacteial and viral superantigens

A BB cell

Peptide

V'

Tcell TcellFigure 1 Classical peptide and superantigen recognition. (A) In classical antigen recognition a processed peptide ispresented by major histocompatibility complex (MHC) class I or class II molecules to the T cell receptor (TCR). The CD4or CD8 molecule intensifies this interaction. (B) In superantigen recognition the superantigen crosslinks the TCR V, withMHC class II. It is not known whether antigenic peptides are bound toMHC and whether CD4 and CD8 are ofimportance.

random introduction and deletion of shortsequences in the V-D, V-J, and D-J regionsduring this recombination process. Thefunctional TCR is composed of an a and a ,chain (which are encoded on differentchromosomes and use different elements),and, as only one type of TCR molecule isexpressed on one T cell clone, is responsiblefor giving the T cell the high specificrecognition capability. In mice there are about25 Vp, 12 Jp, 75 Va, and 75 JL elementsencoded. With the introduced randomsequences of the order of 10'5 different TCRmolecules can theoretically be produced, by farexceeding the potential repertoire of a singlemouse. In humans about 100 Vp elementsexist, whereas the other TCR elements aresimilar in number in mice and humans. Inhumans the V8 families, which are defined by70% amino acid homology, are much largerthan in the mouse. The number of families iscomparable. In a classical MHC restrictedpeptide specific T cell response about one in100 000 T cells can interact with one particularimmunogenic peptide-selfMHC complex withsignificant affinity.

Biology ofsuperantigensThe term superantigen has been given toantigens which can activate of the order of onein 10 T cells, allowing the activation of about10 000 times more T lymphocytes with a singleantigen than in classical MHC-peptiderecognition.7-12 In contrast with T cellmitogens, superantigens have a high degree ofspecificity.

It became clear recently that superantigenshave a new way of interaction with MHCmolecules and the TCR. Instead of interactingwith the most polymorphic part ofthe receptor,they recognise one or several different TCR Vpregions which make only a small contributionto the overall heterogeneity. 1318 They caninteract with MHC molecules and cross linkthe TCR to form an aberrant trimolecularcomplex (fig lB). Neither the TCR nor thetrimolecular complex has yet been crystallised,but structural homology with previously

crystallised structures allows us to suggest atentative structure.' 4 These superantigens arenot processed to small peptides like theclassical antigens but function as entiremolecules'9-21 (or possibly partially cleavedmolcules in the case of retroviralsuperantigens, see below). Based on theanalysis of MHC class II and TCR mutantswhich lost or gained superantigen binding,as well as the recent crystallisation andmutagenesis analysis on the bacterial super-antigen staphylococcal enterotoxin B, it isnow thought that superantigens cross link theTCR with MHC class II molecules on thelateral side of the complex (fig lB).9-1 21-30

Recently, with the help of superantigens andtransgenic mice, direct evidence for two basictolerance mechanisms (clonal deletion of selfreactive cells in the thymus'3 14 31-36 andinduction of unresponsiveness (anergy) in theperiphery37 38) was produced. In this reviewonly the results obtained with superantigensare discussed.T cells mature in the thymus, where they

learn to distinguish between self and foreign.Early in life the immune system is plastic andcan adapt to learn what belongs to self andwhat is foreign. The thymus plays an importantpart in this decision. Cells which are stronglyself reactive are negatively selected (deleted),whereas T cells which can interact weakly withself MHC molecules are positively selected forexport into the periphery.33 3' Owing topositive selection the cells are trained torecognise antigens bound to self MHCmolecules, as otherwise a large percentage ofT cells which cannot interact with self MIHCmolecules would be wasted. Negative selectionin the thymus (deletion) of 'self reactive' Tcells has been found in mice expressingsuperantigens at birth.7 12-16 I2 With the help ofmonoclonal antibodies specific for Vp regionsof the TCR an almost complete elimination ofthe superantigen reactive V'-bearing T cellshas been noted in mature thymocytes as wellas in the peripheral T cell repertoire.A first encounter with a superantigen in

an adult mouse in vivo leads to an initialexpansion of the reactive T cells bearing

S7

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Acha-Orbea

one or several specific TCR VPregions.7 8 12-14 23 43 44 Shortly after thisencounter the T cells become unresponsive(anergic).37 38 454 In the case of anergy thecells are still present but have lost the abilityto respond a second time to the same antigen;this is in striking contrast with a secondchallenge with a conventional antigen where asecondary response is more intense. One of theimportant findings is that these anergic T cellslose the ability to produce one of the importantlymphokines for T cell survival and division-namely, interleukin 2. 38 45 48 49 In some butnot all cases this state of anergy can be reversedby addition of exogenous interleukin 2. Withthe availability of monoclonal antibodiesspecific for TCR Vp and V,,* regions it becameeasy to measure these changes in thefrequencies of the responding T cells. As theymake up about 10% of the total mouse T cellrepertoire they can easily be followed. Negativeselection in the thymus can eliminate 90-99%of T cells expressing a TCR Vp element thatcan interact with a particular superantigen.

Negative selection in the thymus can also beinduced with an exogenous antigen, such asa bacterial superantigen injected at birth.7Normally, however, negative selection in thethymus (as well as tolerance) is onlymaintained if the antigen is presentcontinuously and will fade out with thedisappearance of the antigen. For longlastingtolerance the antigen has to be presentcontinuously, as in the case of retroviralinfection (see below).

Nature ofsuperantigensSuperantigens have been found in differentinfectious agents. The recently describedsuperantigens originate from bacteria(staphylococci,7 1 1 44 50-52 streptococci,53 145-141

mycoplasma,54 148149 yersinia55), retroviruses(mouse mammary tumour virus(MMTV),9 56-66 possibly, human immuno-deficiency virus (HIV),67 68 possibly, murineleukaemia virus69), and other viruses (rabiesvirus70). The exclusive distribution of thesemolecules in microbes raises the questionof how important the expression of thesestructures is for microbes. Probably, thestrategy of activation of the host immuneresponse leads to an overstimulation of theimmune system which allows the pathogensmore efficient infection.71 72 Up to now,however, only few results show such anadvantage directly.

Bacterial superantigensOverall, the described bacterial superantigensare of similar size (20-30 kilodaltons) with highto very low amino acid sequence homology(20-80%) (for review see ref 11). Most likelymany of them developed by convergentevolution towards the same aim. Thesebacterial superantigens have been shown tointeract strongly with MHC class II but notMHC class I molecules.'9-21 73 74 Theiraffinities are of the same order of magnitude as

the binding affinities of antibodies to theirantigen (10-6-10-8 M). They can bind tomouse, rat, and human MHC class II antigens.Moreover, for T cell stimulation to occur,class II expressing cells are required. Table 1provides a summary of the differences betweensuperantigen and classical peptide recognition.Normally, T cells expressing the surfacemarker CD4 (amongst which helper T cells arefound) react with antigens presented by MHCclass II molecules, whereas cells expressingCD8 (which include cytotoxic T cells) reactwith antigens presented by MHC class Imolecules. Bacterial superantigens, however,break this rule; both T cell subpopulations areeasily stimulated by MHC class II expres-sing cells in the presence of bacterialsuperantigens.74 In addition, no MHCrestriction is seen. For example, mousesuperantigen presenting cells can stimulatehuman T cells. The only requirement is MHCclass II expression of the presenting cell andexpression of a specific TCR Vp chain by theT cell.

Different MHC class II isotypes havedifferent affinities for various bacterialsuperantigens.'9-21 73 In general, they interactbetter with human than mouse MHC class IIproteins. Superantigens are the most potent Tcell mitogens known. Their affinity constantsare in the picomolar to nanomolar range.Table 2 lists the currently known bacterialsuperantigens and their TCR Vp specificities inmice and humans. In mice many more TCRVp specific monoclonal antibodies are avail-able, which makes it much easier to definethese specificities. Therefore the TCRspecificities of bacterial superantigens in miceare more completely documented than inhumans.One of these superantigens, staphylococcal

enterotoxin B (SEB), has recently beencrystallised.30 In addition, introduction ofmutations in the SEB coding sequence allowslocalisation of the MHC and the TCRinteraction residues.29 For all the analysedsuperantigens the affinity for the TCR seemsto be much lower than for MHC molecules.In addition to contact residues on thesuperantigen, amino acid contact residues onthe TCR and the MHC class II molecules havebeen characterised for both retroviral (seebelow) and bacterial superantigens.2'-29The overall conclusions of these studiessuggest that unprocessed superantigenmolecules can interact with the lateral side ofthe TCR and the MHC class II molecules (seefig IB).As described above, injection of bacterial

superantigens into newborn mice leads tothymic clonal deletion of the reactive T cells.7Injection into adult mice leads to an expan-sion of the reactive T cells, which thereafterleads to induction of non-responsiveness(anergy).48 49 For a limited time period the Tcells expressing the responsive TCR Vp arereduced in number in vivo. Anergy can also beinduced in T cell clones with bacterial

75superantigens.paA recent report showed that injection of

S8

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Bacterial and viral superantigens

Table 1 Characteristics ofsuperantigens

Classical peptide antigen Superantigen

Origin Protein Bacterial ViralProcessing Yes (small peptides) No Partial or noPresentation MHC* class I, or class II MHC MHC class II, not MHC MHC class II on B cells, not

restricted restricted MHC restrictedFrequency of responding cells 1 in 104 to 10' 1 in 5 to 1 in 20 1 in 5 to 1 in 20TCR* elements V,J,,Vp,Dp,J8 V V

*MHC=major histocompatibility complex; TCR=T cell receptor.

Table 2 Bacterial superantigens

TCR*Vp In humans In mice Reference

1 - SEA* 134,135, Acha-Orbea, unpublished2 TSST-1*, streptococcal M protein - 52,1453 SEB SEA, SED, TSST-1, Yersinia 7,51,52,55,134,135,146, Acha-Orbea,

enterolytica extract, streptococcal unpublishedpyrogenic exotoxin A

4 Streptococcal M protein 1455 SEC3, SED, SEE 51,526 SEE MAM,* Yersinia enterolytica extract 51,52,55,130,1497 SEB, SEC3, SED 7,135,1368-1 SEE, streptococcal M protein SEB, SEC3, SED, MAM 7,51,52,130,135,136,145

8-2 SEE, streptococcal M protein SEB, SEC(1,2,3), SED, MAM, 7,51,52,130,135,136,145,146streptococcal pyrogenic exotoxin A

8-3 SEE, streptococcal M protein SEB, SEC1, SED, MAM 7,51,52,130,135,136,14510 - SEA, SEC2 23,13511 - SEA, SEC 1, SED, SEE, Yersinia 55,134,135,146, Acha-Orbea,

enterolytica extract, streptococcal unpublishedpyrogenic exotoxin A

12 SEB, SEC(1,2,3) SED SEA 51,52,134,135, Acha-Orbea,unpublished

13 SEC2 51,5214 SEB,SEC2 51,5215 SEB, SEC2 SEE, TSST-1 51,52,13517 SEB, SEC2 SEA, SEC(1,2), SED, TSST-1 51,52,13518 SEE 51,5220 SEB, SEC2 51,52

Note that the most homologous TCR V8 elements do not have the same numbers in mice and humans.137 The molecular structuresof Yersinia enterolytica, Mycoplasma arthritidis superantigens have not yet been characterised.*TCR=T cell receptor; TSST-1=toxic shock syndrome toxin; SEA, SEB, SEC, SED, SEE=staphylococcal enterotoxin A, B, C,D, E; MAM=Mycoplasma arthritidis supernatant.

superantigens can lead to a profound inhibitionof conventional antigen responses.7' It is stillnot clear, however, whether it is this effect thatforced bacteria to develop superantigens.

Well known effects of superantigens includefood poisoning after ingestion ofstaphylococcal exotoxins,76 77 the toxic shocksyndrome,78-80 arthritis after mycoplasmainfections,8' possibly Kawasaki disease,82 andpossibly rheumatoid fever after streptococcalinfections83-85 148 149 (see below). Table 3 liststhe diseases and the causing superantigens.

Retroviral superantigens in miceIn mice, genes have been mapped whichencode endogenous superantigens.9 56-63 66 86 87They were originally described as minorlymphocyte stimulating (Mls) antigens byFestenstein some 20 years ago.43 He discoveredthat a large proportion of cells responsive toMls antigens are stimulated in an MHC non-

Table 3 Bacterial superantigens and disease

Superantigen Disease

Staphylococcal enterotoxins Food poisoning, shock

Staphylococcal TSST-1* Toxic shock syndrome

Mycoplasma arthritidis supematant Arthritis, shock(MAS)

Streptococcal superantigens: M Rheumatic fever, shock,protein, pyrogenic exotoxins scarlet fever

*TSST-.1=toxic shock syndrome toxin.

restricted fashion when T cells from micelacking a particular Mls antigen, such as Mls-Ia, are mixed with B cells expressing the Mls-i1 antigen. It was clear that the genes encodingthese proteins are dominant and, surprisingly,always have an Mls allele and a null allelewhich are unable to stimulate such a vigorousT cell response.8 9 88-91 Many such biallelicsystems have been defined in laboratory mousestrains and map different chromosomallocalisations. It was discovered later that the Tcells reacting with Mls antigens shareexpression of particular TCR V~chains.9 13-18 42 65 92-99 Mice expressing such anMls gene delete most of the T cells expressingthe responsive TCR Vp chain during thymicmaturation.

After an intensive search for over 20 years itbecame clear recently that these Mls antigensare encoded by endogenous proviruses of theMMTV family.9 56-64 66 86 87 Figure 2 shows thetypical buildup of such a virus. In the 3' longterminal repeat, an open reading frame (ORF)was found which represented a puzzle toretrovirologists.9 100-102 They had difficultiesascribing a function to this potential protein,which was never found in normal or tumourcells. It was thought that this protein had aneffect on gene regulation and both positive andnegative gene regulation were described.'03 104It turned out that it is this protein whichrepresents a superantigen.9 59 60 As viral genesoften have more than one function it will beinteresting to see if other functions can be

S9

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Acha-Orbea

RNA of infectious virus5' LTR 3' LTR

I R JU5 gag pol env I U3 R IAAAAA

ted provirus5' LTR 3' LTR

3 IRIU5 I gag pol env IORF

otein

4$s\\\

t ATG, potential in frame initiation codons

4 RXRR sequences, potential protease cleavage sites

* Potential N-linked glycosylation sites

g\\\ Highly polymorphic C-terminal amino acid sequence(aa 290-320)

l J |Hydrophobic sequence (aa45-63)

Figure 2 Mouse mammary tumour virus. The superantigen encoded in the open readingframe (ORF) molecule has notyet been identified biochemically. Possibly, therefore, thepotential cleavage sites in the figure are cleaved before the superantigen fiunctions. Clearly,however, the superantigen is expressed on the B cell surface. 152 153 LTR=long terminalrepeat.

found for this ORF. Most laboratory as well aswild mice have two to 10 copies of differentMMTV proviruses integrated into theirgenome.'05 The different MMTVs are about95% homologous in their amino acidsequence. Table 4 lists the currently knownendogenous MMTVs and their TCR Vpspecificities.When ORF molecules interacting with the

different TCR Vp elements are aligned astriking correlation between the C-terminalamino acid sequence and the TCR Vpspecificity is observed.9 59 60 86 At present some15 ORF sequences are known and in all casesa striking correlation between C-terminalsequence and TCR Vp specificity exists. Thelast 10-30 amino acids seem to determine theVp specificity. Figure 2 shows the presumedstructure of an ORF molecule.These endogenous superantigens also lead

to thymic deletion of the reactive cells whenencountered at birth. Thymic deletion can alsobe induced after injection of cells bearing asuperantigen into adult mice.86 As describedabove for bacterial superantigens, immunestimulation of the reactive T cells occurs whenthe first encounter happens in adult life

Integra

ORF Pr

followed by induction of unresponsiveness(anergy). Whether the loss of the reactive Tcells after this initial stimulation is due toperipheral deletion or to dilution of the cellswhich can no longer divide is not clear at themoment.These endogenous superantigens originate

from infectious viruses which are still found inwild and laboratory mice (for reviews onMMTV see refs 102, 106-110). In contrastwith endogenous transmission, which followsthe principles of mendelian inheritance,infectious viruses are transmitted through milkfrom mother to offspring. The followingscenario for infection has been suggested byexperiments in newborn or adult mice: thevirus is taken up through milk from birth andcontacts lymphocytes in the Peyer's patches(Karapetian 0, Shakhov A N, KraehenbuhlJ P, et al, unpublished data). As soon as theimmune system matures sufficiently (aroundtwo to three days after birth), infection of Bcells is followed by superantigen (ORF)mediated T-B interaction with T cellsexpressing the responsive TCR Voelements.86 111 A strong augmentation of theinfected B cells which express the superantigenoccurs1ll; at the same time a continuousstimulation (which would be harmful for thehost) is prevented by anergisation of theresponsive T cells. The superantigenexpression allows about 1000 times moreefficient infection (Held W, Waanders G A,Shakhov AN et al, unpublished data). We haverecently shown that the virus infectslymphocytes exclusively in Peyer's patches andthe lymph nodes of the small intestine(Karapetian 0, Shakhov A N, KraehenbuhlJ P, et al, unpublished data). The infection oflymphocytes represents a crucial step in the lifecycle of the virus. Mice lacking responsive Tcells or mice with a defective immune systemcannot propagate the virus efficiently."12 Laterin life the virus jumps from lymphocytes to themammary gland epithelium where large scalevirus production occurs which is secreted intothe milk. This is the only place where free virusparticles have been found. When the motherstransmit the virus to their offspring the viciouscircle is closed. .Figure 3 shows a typical lifecycle ofMMTV.As the name implies, these viruses are the

major causative agent of mammary tumours inmice. Integration close to a mouse proto-oncogene can activate it and induce the firststeps towards carcinogenesis."13 114This virus has very cleverly profited from the

weaknesses of the mouse's immune system. Indoing so it has given us tools which can helpus to understand better its normal function.The finding of an 'infectious superantigen'(MMTV), especially, allows experimentswhich help towards understanding the immunesystem, the retroviral life cycle, and theinterplay between the two.64 86 111 5 Theseexperiments are still in progress. Oneimportant point for this discussion is the effecton B cells after local infection of adult micewith MMTV. Adult mice cannot eliminateMMTV; they remain infected for the rest of

Sl0

.** .1 I

I9

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/


Table 4 Retroviral superantigens

Mouse Human Retroviral superantigen References

2 ? Mtv-MA, MMTV(C4) Wei and Acha-Orbea, and Marche et al, inpreparation

3 ? Mtv-1, -3, -6, -13, -44, -MAI 58,59,63,138-1405 ? Mtv-6, MuLV? 57,695-1 ? Mtv-8, -9 56-585-2 ? Mtv-8, -9 56-586 12 Mtv-7, -43, -44, -SW, MMTV(SW) Shakhov and Acha-Orbea, in preparation7 ? Mtv-7, -29, -43, -SW, -29, MMTV(SW) 59,63,87,141, Shakhov and Acha-Orbea, in

preparation, Marrack, personal communication8-1 ? Mtv-7, -43, -SW, MMTV(SW) 59,63,87,141, Shakhov and Acha-Orbea, in

preparation9 Mtv-7, -43, -SW, MMTV(SW) 59,63,87,141, Shakhov and Acha-Orbea, in

preparation11 Mtv-8, -9, -11 6212 Mtv-8, -9, -11 97-9914 ? Mtv-2, MMTV(GR), MMTV(C3H) 27,59,14415 ? Mtv-2, MMTV(GR), MMTV(C3H) 2716 ? Mtv? 97-99,14317 ? Mtv-3, -8, -9

Mtv-6, -MAI 140, 15419 ? New Mtv? Hodes, personal communication20 ? Mtv-? 65

Note that the most homologous TCR V8 elements do not have the same numbers in mice and humans. Mtv-n stands for integratedproviruses, MMTV stands for infectious viruses. Most human TCR V8-MMTV ORF interactions have not yet been measured.TCR=T cell receptor; MMTV=mouse mammary tumour virus; ORF=open reading frame.

Virlac

Infection of babies through mothers' milk

Infection of Peyer's patchlymphocytes, amplification of

,us production in infected B cells through helper:tating females T cells expressing superantigen

reactive TCR, induction of T cellI anergy

Infection of mammary gland

Mammary tumours after integration(by chance) close to proto-oncogenes

Figure 3 Life cycle ofmouse mammary tumourvirus. TCR=T cell receptor.

their life and transmit the virus maternally totheir offspring. The first cells which seem to beinfected with MMTV are the B cells.Expression of the superantigen leads to strongactivation of helper T cells (but not cytotoxiccells in vivo) expressing the reactive TCR VPelement. At the same time this T-B interactionleads to terminal activation of the B cells.These B cells start secreting large amounts ofIgG antibody and increase in number."' Suchan induction of polyclonal IgG produc-tion might be part of autoimmune reactions(see below). Similar observations have beenmade with bacterial superantigens.54 71 116 117

Figure 4 provides a schematic summary ofthese observations.

In the case of MMTV superantigen theanswer to the question of why these viruseshave adopted such a strategy seems answered.The superantigen allows about 1000 timesmore efficient infection of the host lympho-cytes, which seems crucial for the survival ofthe virus.

There are still gaps in our knowledge aboutthe interplay between virus and host, but so far(and surely even more so in the near future)retroviral and bacterial superantigens havebeen great tools in the hands ofimmunologists,allowing consideration of questions about thenormal immune response as well as retro-virology.

Table 5 compares current knowledge onbacterial and retroviral superantigens.

Human viral superantigensDuring the last year two reports suggested thatHIV may encode a superantigen.67 68 The firstanalysed the expression of mRNA of thedifferent TCR V8 elements in the peripheralblood ofpatients at late stages ofAIDS. SeveralTCR Vp elements were found to be reducedstrongly in patients with late stage disease, afterthe drop in the CD4+ T cell population. Asemiquantitative polymerase chain reactionwas used to measure mRNA expression inunseparated blood isolates.67 Analyses onpurified CD4+ T cell populations are requiredto consider this question more directly as,possibly, the Vps which were reduced areunderrepresented in the CD8+ T cellpopulation.

In another study it was shown that fresh Tcell lines expressing TCR Vp6 7, 8, and 17 orTCR V012 could be infected similarly withHIV but, after addition of antigen presentingcells, TCR V 12 expressing cells had anapproximately 100-fold higher HIV titre thanthe other three, suggesting a superantigeneffect in HIV amplification.68 More needs to bedone to show convincingly that HIV encodesa superantigen. If it does, it opens newstrategies for vaccination and would make theMMTV system a model of choice for newvaccination studies directed at the super-antigen effects.

In another study it was clearly shown that therabies virus encodes a superantigen in itsnuclear protein.70 Recombinant proteinshowed strong binding to MHC class II

Sll

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Acha-Orbea

Anergyorcell deatlor both

9

I

Activationof the T cell

-A

B cell activation, differentiation, and division

Secretion of antibodies

Table 5 Comparison of retroviral and bacterial superantigens

Bacterial superantigens Retroviral superantigens

Stimulation Yes Yes or noDeletion Yes YesAnergy Yes YesCytotoxicity Yes NoAPCf: B* Yes YesAPC: class II expressing transfectants Yes NoAPC: macrophages Yes NoSequence homology to host proteins None NoneSequence homology between memberst 20-900/0 95%Binding to MHCI: class II Strong ? (Most likely)Binding to TCR4 Weak ? (Most likely)Processing required for presentation No No or partial cleavage

*B cells as antigen presenting cellstAmino acid sequence homology between different bacterial superantigens or between differentretroviral superantigens.tAPC=antigen presenting cell; MHC=major histocompatibility complex; TCR=T cell receptor.

molecules and stimulated T cells expressingV12 in humans and VP6 (the closest murinehomologue) in mice. It will be interesting to see

the effect of this superantigen on the efficiencyof infection with rabies virus. Once this virusreaches the nerve cells it is no longer accessibleto the immune system. In addition, it will beinteresting to see if and how this superantigenplays a part in the vaccines used to protectagainst rabies infection. The results of theseexperiments will potentially give insights intothe role of superantigens in vaccination.

Roles ofsuperantigens in autoimmunity?It is well known that normal subjects havesome autoreactive T and B cells in theircirculation. Polyclonal activation of B cellsinduces production of autoantibodies, andautoreactive CD4+ T cells can easily beisolated after stimulation in vitro withautoantigens. 118-125 The overall tolerancemechanisms can maintain a healthyequilibrium between the presence ofautoreactive cells and the prevention of theirpredominance. Several points in the above

discussion make it likely that superantigensmay play a part in the induction ofautoimmune diseases: (a) superantigens can

induce polyclonal antibody responses; (b)superantigens can activate T cells with many

different fine specificities, possibly includingautoreactive T cells.Mimicry by microbes has been proposed

previously for a trigger of autoimmunereactions.'26 In this hypothesis a sequence withsimilarity to host sequences is expressed by themicrobe. A strong immune reaction to theinvading microorganism leads to the breakingof tolerance to an autoantigen. Superantigensdo not even have to express such a cross

reactive epitope. By activation of a substantialproportion of the host immune cells expressingmillions of different antigen specificities,several autoreactive cells might be activated(fig 5). It is not clear at present what happensto these cells when they encounter an auto-

antigen after superantigen activation. Inexperimental allergic encephalomyelitis, a

model disease for human multiple sclerosis,few TCR molecules are found on theautoreactive T cells. Most share expression ofTCR V 8 2.'50 151 The bacterial superantigenSEB reactivates all the V08 bearing T cells.Preliminary experiments indicate that SEBinduces relapses of experimental allergicencephalomyelitis in mice, clearly suggestingthe possibility of a role in classical antigenresponses by potentially anergised cells(Steinman L A, personal communication).Most likely it depends on the cell type whichpresents the autoantigen after the superantigenencounter.Another elegant experiment gives support

for both the original mimicry as well as thesuperantigen hypothesis in the induction ofautoimmunity. 127 128 Mice were generatedwhich bear two types of transgene. One

Figure 4 Fate of T andBcells after mouse mammarytumour virus superantigenznteraction in vivo.TCR= T cell receptor.

S12

N

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/


sClass1 1 ,APC

I~~~~~~~~~

Superantigen T cell

Activation

Migration totarget organ

Antigen presenting cellor target cell

Figure S Potential role in autoimmunity ofsuperantigens. APC=antigen presenting cell; TCR=T ceU receptor.

transgene determines the expression of anepitope for cytotoxic T cells exclusively on thepancreatic , cells, the targets of autoimmunedestruction in insulin dependent diabetesmellitus. The antigen is the glycoprotein of theleucocytic choriomeningitis virus. The othertransgene controls the expression of the TCRa and 3 chains isolated from a cytotoxic T cellspecific for a peptide of this glycoprotein. Miceexpressing both transgenes are non-diabeticand remain normal. Infection with a controlleucocytic choriomeningitis virus which doesnot bear the T cell epitope does not inducediabetes, whereas infection with the leucocyticchoriomeningitis virus expressing the T cellepitope induces diabetes within a few days.

Several recent studies have begun toconsider questions about the role ofsuperantigens in autoimmunity. In rheumatoidarthritis, again using semiquantitative PCR, anoverrepresentation in the rheumatoid lesions ofa particular TCR Vp chain has been found,with reduced representation in theperiphery."29 These findings were made usingsemiquantitative PCR and have to beconfirmed by more quantitative methods. Assoon as the necessary monoclonal antibodiesare available, such analyses can be repeated ona large number of patients. In anotherautoimmune disease, Kawasaki disease, a highpercentage of T cells expressing V02 has beenfound in the peripheral blood of patients.82Kawasaki disease is an acute multisystemvasculitis affecting young children. It is oftenfound in Japanese patients. Mycoplasmaarthritidis antigens, which have not beenbiochemically characterised, might yet beinvolved in the induction of polyarthritisin mice.8' 130 148 149 Other candidates areyersinia superantigens and streptococcal

55 83-814 14antigens. 85 145-147 Reiter's disease is areactive arthritis, a term used for arthritisdeveloping after infection. Candidate in-fectious agents causing reactive arthritis inhumans are yersinia, shigella, salmonella, andothers.'31-"33 An involvement of superantigens

or classical mimicry in autoimmunity would becompatible as well with geographical variationsin the prevalence of autoimmunc diseases.

It is still too early clearly to correlate specificautoimmune diseases with specific bacterial orviral superantigens. Indeed, it is not yet clearwhether superantigens are important forautoimmunity or not. So far all the correlationsamount to little more than speculation. Evenif microbes can be clearly linked to theinduction of autoimmune diseases, mech-anisms other than superantigens are likely tocontribute. The near future will tell what theirrole is in the known autoimmune diseases.Whatever the outcome of these studies,pathogens compensate for the morbidity theyinflict by providing us with highly valuabletools for dissecting the immune system andperhaps will give us insight into the cause ofsome forms of autoimmunity and itsprevention.

HAO was supported by a START fellowship of the SwissNational Science Foundation.

1 Bjorkman P J, Parham P. Structure, function, and diversityof class I major histocompatibility complex molecules.Annu Rev Biochem 1990; 59: 253-88.

2 Nepom G T, Erlich H. MHC class II molecules andautoimmunity. Annu Rev Immunol 1991; 9: 493-526.

3 Rothbard J B, Gefter M L. Interactions betweenimmunogenic peptides and MIC. Annu Rev Immunol1991; 9: 527-66.

4 Davis M M, Bjorkman P J. T-cell antigen receptor genesand T-cell recognition. Nature 1988; 334: 395-402.

5 Jorgensen J L, Reay P A, Ehrlich E W, Davis M M.Molecular components of T-cell recognition. Annu RevImmunol 1992; 10: 835-73.

6 Janeway C A Jr. The T cell receptor as a multicomponentsignaling machine: CD4/CD8 coreceptors and CD45 inT cell activation. Annu Rev Immunol 1992; 10: 645-74.

7 White J, Herman A, Pullen A M, et al. The V-specificsuperantigen staphylococcal enterotoxin B: stimulation ofmature T cells and clonal deletion in neonatal mice. Cell1989; 56: 27-35.

8 Abe R, Hodes R. T cell recognition of minor lymphocytestimulating (Mis) gene products. Annu Rev Immunol1989; 7: 683-708.

9 Acha-Orbea H, Palmer E. Mls-a retrovirus exploits theimmune system. Immunol Today 1991; 12: 356-61.

10 Janeway C A Jr. Self-superantigens? Cell 1990; 63:659-61.

11 Marrack P, Kappler, J. The staphylococcal enterotoxinsand their relatives. Science 1990; 248: 705- 1.

12 MacDonald H R, Glasebrook A L, Schneider R, et al.T-cell reactivity and tolerance to Mls' encoded antigens.Immunol Rev 1989; 107: 89-108.

S13

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Acha-Orbea

13 MacDonald H R, Schneider R, Lees R L,et al. T-cellreceptorV,B use predicts reactivity and tolerance to Mls3-encoded antigens. Nature 1988; 332: 40-5.

14 Kappler J W, Staerz U D, WhiteJ, Marrack P C. Self-tolerance eliminates T cells specific for Mls-modifiedproducts of the major histocompatibility complex. Nature1988; 332: 35-40.

15 KapplerJ W,Roehm N, Marrack P. T cell tolerance byclonal elimination in the thymus. Cell 1987; 49: 273-80.

16 Kappler J W, Wade T,W hite J, et al. A T cell receptor V5segment that imparts reactivity to a class II majorhistocompatibility complex product. Cell 1987; 49:263-71.

17 Abe R, Vacchio MS, Fox B, Hodes R. Preferentialexpression of the T-cell receptor V,3 geneby Mlsc reactiveT cells. Nature 1988; 335: 827-30.

18 Fry A M, Matis L A. Self-tolerance alters T-cell receptorexpression in an antigen-specific MHC-restrictedimmune response.NVature 1988; 335: 830-2.

19 Mollick J A, Cook R G, Rich R R. ClassII MHC moleculesare specific receptors for staphvlococcal enterotoxin A.Scienlce 1989; 244: 817-20.

20 Fischer H, Dohlstein M, Lindwall M, Sjogren H-O,Carlsson R. Binding of staphylococcal enterotoxin A toHLA-DRon B cell lines.7 Immnuniol 1989; 142: 3151-7.

21 FraserJ D. High-affinity binding of staphylococcalenterotoxins A and B to HLA-DR. Nature 1989; 339:221-3.

22 Dellabonna P, Peccaud J, Kappler J, et al. Superantigensinteract with MHC class II molecules outside of theantigen groove. Cell 1990; 62: 1115-21.

23 Herman A, Kappler J, Marrack P, Pullen A M.Superantigens: mechanism of T-cell stimulation and rolein immune responses. Ann7lti Rev Inimtunol 1991; 9:745-72.

24 Pullen A, Wade T, Marrack P, Kappler J. Identification ofthe region of the T cell receptor P chain that interacts withthe self-superantigen Mls-l'. Cell 1990; 61: 1365-74.

25 Pullen A M, BillJ, Kubo R, Marrack P, Kappler J W.Analysis of the interaction site for the self superantigenMls-1l on T cell receptor V. _f Exp Med 1991; 173:1183-92.

26 Cazenave P A, Marche P, Jouvin-Marche E, et al. V517gene polymorphism in wvild-derived mouse strains: two-amino acid substitutions in the V 17 region alterdrastically T cell receptor specificity. Cell 1990; 63:717-28.

27 Choi Y, Herman A, DiGusto D, et al. Regions of thevariable region of the T-cell receptor,B-chain that interactwith S. aureus toxin superantigens. Nature 1990; 346:471-3.

28 Karp D R, Long E 0. Identification of HLA-DR1,B chainresidues critical for binding staphylococcal enterotoxins Aand E. JExp Med 1992; 175: 415-24.

29 Kappler J, Herman A, Clements J, Marrack P. Mutationsdefining functional regions of the superantigenenterotoxin B. ] Exp Med 1992; 175: 387-96.

30 Swaminathan S, Furev W, Pletcher J, Sax M. Crystalstructure of staphylococcal enterotoxin B, a superantigen.Nature 1992; 359: 801-6.

31 Kisielow P, Bluthmann H, Staerz U D, Steinmetz M, vonBoehmer H. Tolerance in T-cell-receptor transgenic miceinvolves deletion of nonmature CD4'8+ thymocvtes.Nature 1988; 333: 742-6.

32 von Boehmer H, Kisielow P. Self-nonself discrimination byT cells. Science 1990; 248: 1369-72.

33 Sha W C, Nelson C A, Newberrv R D, et al. Positive andnegative selection of an antigen receptor on T cells intransgenic mice. Nature 1988; 336: 73-6.

34 Uematsu Y, RyserS, Dembic Z, et al. In transgenic micethe introduced functional T cell receptor P chain geneprevents expression of endogenous P genes. Cell 1988; 52:831-41.

35 Pircher H P, Lang R, Hengartner H, Zinkernagel R M.Tolerance induction in double specific T cell receptortransgenic mice varies with antigen. NVature 1989; 342:559-61.

36 Berg L J, de St Groth B F, Pullen A M, Davis M M.Phenotypic differences between alpha beta versus betaT-cell receptor transgenic mice undergoing negativeselection. Nature 1989; 340: 559-62.

37 Webb S, Morris C, Sprent J. Extrathymic tolerance ofmature T cells: clonal elimination as a consequence ofimmunity. Cell 1990; 63: 1249-56.

38 Rammensee H-G, Kroschewsky R, Frangoulis B. Clonalanergy induced in mature V56 T lymphocytes onimmunizing Mls-lb mice with Mls-l' expressing cells.Nature 1989; 339: 541-4.

39 Kisielow P, Teh H S, Bluthmann H, von Boehmer H.Positive selection of antigen-specific T cells in thymus byrestricting MHC molecules. Nature 1988; 335: 730-3.

40 MacDonald H R, Lees R K, Schneider R, Zinkemagel RM, Hengartner H. Positive selection of CD4+ thymocytescontrolled hy MHC class II gene products. Nature 1988;336: 471-3.

41 Benoist C, Mathis D. Positive selection of the T cellrepertoire: where does it occur? Cell 1989; 58: 1027-33.

42 MacDonald H R, Pedrazzini T, Schneider R, et al.Intrathymic elimination of Mls3-reactive (VO6+) cellsduring neonatal tolerance induction to Mls-l' encodedantigens. _] Exp Med 1988; 167: 2005-10.

43 Festenstein H. Immunogenic and biological aspects of invitro allotransformation (MLR) in the mouse. TransplantRev 1973; 15: 62-88.

44 Janeway C A Jr, Yagi J, Conrad P J, et al. T-cell responses

to Mls and to bacterial proteins that mimic its behavior.Immunol Rev 1989; 107: 61-88.

45 Jones L A, Chin L T, Longo D L, Kruisbeek A M.Peripheral clonal elimination of functional T cells. Science1990; 250: 1726-9.

46 Fowlkes BJ, Schwartz P H, Pardoll D M, Deletion of self-reactive thymocytes occurs at a CD4'8+ precursor stage.Nature 1988; 334: 620-3.

47 Blackman M R, Gerhardt-Burgert H, Woodland D L, etal. A role for clonal inactivation in T cell tolerance to Mls-V'. Nature 1990; 345: 540-2.

48 Kawabe Y, Ochi A. Programmed cell death andextrathymic reduction of VA8 CD4+ T cells in micetolerant to Staphylococcus aureus enterotoxin B. Nature1991;349: 245-8.

49 MacDonald H R, BaschieriS, Lees R K. Clonal expansionprecedes anergy and death ofVp8+ peripheral T cellsresponding to staphylococcal enterotoxin B in vivo.EurIinniunol 1991; 21: 1963-6.

i0 Fleischer B, Schrezenmeier H. T cell stimulation bystaphylococcal enterotoxins. Clonally variable responseand requirements for major histocompatibility complexclass II molecules on accessory or target cells._7 Exp Med1988; 167: 1697-707.

51 KapplerJ, Kotzin B, Herron L, et al. V5-specificstimulation of human T cells by staphylococcal toxins.Science 1989; 244: 811-13.

52 Choi Y, Kotzin B, Herron L, et al. Interaction ofStaphylococcus aureus toxin superantigens with human Tcells. Proc Natl Acad Sci USA 1989; 86: 8941-5.

53 Tomai M, Kolb M, Majumdar G, BeacheyE H.Superantigenicity of streptococcal M protein._7 Exp Med1990; 172: 359-62.

54TumangJ R, Posnett D N, Cole B C, Crow M K,Friedman S M. Helper T cell-dependent human B celldifferentiation mediated by a mycoplasmal superantigenbridge. _Exp Med 1990; 171: 2153-8.

55 Stuart P M, Woodward J G. Yersinia enterolytica producessuperantigen activity. if Immunol 1992; 148: 225-33.

56 Woodland D, Happ M P, Bill J, Palmer E. Requirementfor cotolerogenic gene products in the clonal deletion ofI-E reactive T cells. Science 1990; 247: 964-7.

57 Woodland D L, Happ M P, Gollub K J, PalmerE. Anendogenous retrovirus mediating deletion of tijT cells?Ntature 1991; 349: 529-30.

58 Woodland D L, Lund FE, Happ M P, et al. Endogenoussuperantigen expression is controlled by mouse mammarytumor proviral loci. IExp Med 1991; 174: 1255-8.

59 Acha-Orbea H, Shakhov A N, Scarpellino L, et al. Clonaldeletion of V1 14 positive T cells in mammary tumor virustransgenic mice. Nature 199 1; 350: 207- 11.

60 Choi Y, KapplerJ W, Marrack P. A superantigen encodedin the open reading frame of the 3' long terminal repeatof mouse mammary tumor virus. Nature 1991; 350:203-7.

61 Choi Y, Marrack P, Kappler J. Structural analysis of amouse mammary tumor virus superantigen. if Exp Med1992; 175: 847-52.

62 Dyson PJ, Knight A M, FairchildS, SimpsonE, TomonariK. Genes encoding ligands for deletion of Vo11 T cellscosegregate with mammary tumor virus genomes. Nature1991;349: 531-2.

63 Frankel W N, Rudy C, CoffinJ M, Huber B T. Linkageof Mls genes to endogenous mammary tumor viruses ofinbred mice. Nature 199 1; 349: 526-8.

64 Marrack P, KushnirE, Kappler J. A maternally inheritedsuperantigen encoded by a mammary tumor virus. Nature1991;349: 524-6.

65 Six A, Jouvin-MarcheE, Loh D Y, Cazanave P A, MarcheP N. Identification of a T cell receptor P chain variableregion, V,20, that is differently expressed in variousstrains ofmice. Exp Med 1991; 174: 1263-6.

66 Pullen A M, Choi Y, KushnirE, Kappler J, Marrack P. Theopen reading frames in the 3' long terminal repeats ofseveral mouse mammary tumor virus integrants encodeV,3-specific superantigens. _f Exp Med. 1992; 175: 41-7.

67 Imberti L, Sottini A, Bettinardi A, Puoti M, Primi D.Selective depletion in HIV infection of T cells that bearspecific T cell receptor V5 sequences. Science 1992; 254:860-2.

68 Laurence J, Hodtsev A S, Posnett D N. Superantigenimplicated in dependence of HIV-1 replication in T cellson TCRV expression. Nature 1992; 358: 255-9.

69 Hugin AiV, Vacchio M S, Morse H C III. Science 1991;252: 424-7.

70 Lafon M, Lafage M, Martinez-Arends A, et al. Evidencefor a human superantigen in humans. Nature 1992; 358:507-10.

71 Mourad W, Scholl P, Diaz A, Geha R, Chatila T. Thestaphylococcal toxic shock syndrome toxin 1 triggers Bcell proliferation and differentiation via majorhistocompatibility complex-unrestricted cognate T/B cellinteraction. if Exp Med 1989; 170: 2011 22.

72 Miethke T, Wahl C, Heeg K, Echtenacher B, Krammer PH, Wagner H. T cell-mediated lethal shock triggered inmice by the superantigen staphylococcal enterotoxin B:critical role of tumor necrosis factor. if Exp Med 1992;175: 91-8.

73 Herrmann T, Accolla R S, MacDonald H R. Differentstaphylococcal enterotoxins bind preferentially to distinctmajor histocompatibility complex class II isotypes. Eur]Imnmunol 1989; 19: 2171-4.

74 Herrmann T, Mariansky J L, Romero P, Fleischer B,MacDonald H R. Activation of MHC class I-restrictedCD8+ CTL by microbial T cell mitogens. Dependence

S14

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/


upon MHC class II expression of the target cells and V5usage of the responder cells. Jf Immunol 1990; 144:1181-6.

75 O'Hehir R E, Lamb J R. Induction of specific clonal anergyin human T lymphocytes by Staphylococcus aureusenterotoxins. Proc Nad Acad Sci USA 1990; 87: 8884-8.

76 Bergdoll M S. Enterotoxins. In: Moutie T C, Kadis J, AjlS J, eds. Microbial toxins. III. Bacterial protein toxins. NewYork: Academic Press, 1970: 265-326.

77 Bergdoll M S. Staphylococcal intoxications. In: RiemannH, Bryan F L, eds. Food borne infections and intoxications.New York: Academic Press, 1979: 443-94.

78 Todd J K, Fishaut M, Kapral F, Welch T. Toxic shocksyndrome associated with phage-group I staphylococci.Lancet 1978; i: 1116-8.

79 Bergdoll M S, Crass B A, Reiser R F, Robbins R N, DavisJ P. A new staphylococcal enterotoxin F, associated withtoxic shock syndrome isolates. Lancet 1981; i: 1017-21.

80 Schlievert P M, Shands K N, Dan B B, Schmid G P,Nishimura R D. Identification and characterization of anexotoxin from Staphylococcus aureus associated withtoxic shock syndrome.7 InfectDis 1981; 143: 509-16.

81 Cole B C, Washburn L R, Taylor-Robinson D.Mycoplasma-induced arthritis. In: Razin S, Barile M F,eds. The mycoplasmas Vol IV. New York: Academic Press,1985: 105-60.

82 Abe J, Kotzin B L, Jujo K, et al. Selective expansion of Tcells expressing T-cell receptor variable regions Vp2 andV 8 in Kawasaki disease. Proc Natl Acad Sci USA 1992;89:4066-70.

83 Bartter T, Dascal A, Carroll K, Curley F J. Toxicstreptococcus syndrome. Arch Intern Med 1988; 148:1421-4.

84 Quinn DeJoy S, Ferguson K M, Sapp T M, et al.Streptococcal cell wall arthritis. Passive transfer with a Tcell line and crossreactivity of streptococcal cell wallantigens with Mycobacterium tuberculosis. Jf Exp Med1989; 170: 369-82.

85 Bisno A L. Group A streptococcal infections and acuterheumatic fever. NEngl7Med 1991; 325: 783-93.

86 Held W, Shakhov A N, Waanders G A, et al. An exogenousmouse mammary tumor virus with properties of Mls-lV(Mtv-7).7 Exp Med 1992; 175: 1623-33.

87 Beutner U, Frankel W N,.Cote M S, Coffin J M, HuberB T. Mls-l is tncoded by the long terminal repeat openreading frame of the mouse mammary tumor virus Mtv-7. Proc NatlAcad Sci USA 1992; 89: 5432-6.

88 Festenstein H, Berumen L. BALB.D2-Mlsa new congenicmouse strain. Transplantation 1983; 37: 322-4.

89 Abe R, Ryan J J, Hodes R J. Mls is not a single gene, allelicsystem. Different stimulatory Mls determinants are theproducts of at least two nonallelic, unlinked genes. Jf ExpMed 1987; 166: 1150-5.

90 Abe R, Ryan J J, Hodes R J. Clonal analysis of the Mlssystem. A reappraisal of polymorphism and allelismamong Mlsa, Mlsb, and Mlsd. Jf Exp Med 1987; 165:1113-29.

91 Pullen A M, Marrack P, Kappler J W. Evidence that Mls-2 antigens which delete V,3+ T cells are controlled bymultiple genes. Immunol 1989; 142: 3033-7.

92 Pullen A M, Marrack P, KapplerJ W. The T-cell repertoireis heavily influenced by tolerance to polymorphic self-antigens. Nature 1988; 335: 796-801.

93 Bill J, Appel V, Palmer E. An analysis of T cell receptorvariable region gene expression in majorhistocompatibility complex disparate mice. Proc NatlAcadSci USA 1988; 85: 9184-8.

94 Bill J, Kanagawa 0, Woodland D, Palmer E. The MHCmolecule I-E is necessary but not sufficient for the clonaldeletion of V 1 1-bearing T cells. Exp Med. 1989; 169:1405-19.

95 Happ M P, Woodland D C, Palmer E. A third T cellreceptor Vp gene encodes reactivity to Mls- geneproducts. Proc NatlAcad Sci USA 1989; 86: 6293-6.

96 Okada C Y, Holzmann B, Guidos S, Palmer E, WeissmanI L. Characterization of a rat antibody specific for adeterminant encoded by the Vp7 gene segment. J7Immunol1990; 144: 3473-7.

97 Vacchio M S, Hodes R. Selective decrease in T cellreceptor V5 expression: decreased expression of specificV5 families is associated with expression of multiple MHCand non-MHC gene products. Jf Exp Med 1989; 170:1335-46.

98 Vacchio M S, Ryan J J, Hodes R J. Characterization of theligand(s) responsible for negative selection of V1 1- andV 12-expressing T cells: effects of a new Mls-determinant.J Exp Med 1990; 172: 807-13.

99 Singer P A, Balderas R S, Theofilopoulos A N. Thymicselection defines multiple T cell receptor Vp "repertoirephenotypes" at the CD4/CD8 subtype level. EMBO Jf1990; 9: 3641-8.

100 Donehower L A, Huang A L, Hagler G L. Regulatory and

coding potential of the mouse mammary tumor virus longterminal repeat.J Virol 1981; 37: 226-37.

101 Fasel N, Pearson K, Buetti E, Diggelmann H.The regionofmouse mammary tumor virus DNA containing the longterminal repeat includes a long coding sequence andsignals for hormonally regulated transcription. EMBO1982; 1: 3-7.

102 Dickson C. Molecular aspects of mouse mammary tumorvirus biology. Int Rev Cytol 1985; 108: 119-47.

103 Salmons B, Erfle V, Brem G, Giinzburg W H. naf, a trans-regulating negative-acting factor encoded within themouse mammary tumor virus open reading frame region.

Virol 1990; 64: 6355-9.

104 van Klaveren P, Bentvelzen P. Transactivating potential ofthe 3' open reading frame of murine mammary tumorvirus. Virol 1988; 62: 4410-3.

105 Kozak C, Peters G. Pauley R, et al. A standardizednomenclature for endogenous mouse mammary tumorviruses.J7 Virol 1987; 61: 1651-4.

106 Acha-Orbea H, Held W, Waanders G A, et al. Exogenousand endogenous mouse mammary tumor superantigens.Immunol Rev. In press.

107 Nandi S, McGrath C M. Mammary neoplasia in mice. AdvCancer Res 1973; 17: 353-414.

108 Hageman P C, Calafat J, Hilgers J. The biology of themouse mammary tumor virus. In: Hilgers J, Sluyser M,eds. Mammary tumors of the mouse. Amsterdam, NewYork, Oxford: Elsevier, North Holland Biomedical Press,1981: 391-463.

109 Bentvelzen P, Hilgers J. Murine mammary tumor virus. In:Klein G, ed. Viral oncology. New York: Raven Press, 1980:311-55.

110 Dickson C, Peters G. Proteins encoded by mousemammary tumor virus. Curr Top Microbiol Immunol 1983;106: 1-34.

111 Held W, Shakhov A N, Izui S, et al. Superantigen-reactiveCD4+ T cells are required to stimulate B cells afterinfection with mouse mammary tumor virus. Exp Med.In press.

112 Tsubura A, Inaba M, Imai S, et al. Intervention of T-cellsin transportation of mouse mammary tumor virus (milkfactor) to mammary gland cells in vivo. Cancer Res 1988;48: 6555-8.

113 Nusse R, Varmus H E. Many tumors induced by themouse mammary tumor virus contain a provirusintegrated in the same region of the host genome. Cell1982;31:99-109.

114 Peters G, Brookes S, Smith R, Dickson C. Tumorigenesisby mouse mammary tumor virus: evidence for a commonregion for provirus integration in mammary tumors. Cell1983; 33: 369-77.

115 Ignatowicz L, Kappler J, Marrack P. The effects of chronicinfection with a superantigen-producing virus. Exp Med1992; 175: 917-23.

116 He X, Goronzy J, Weyand C. Selective induction ofrheumatoid factors by superantigens and human helper Tcells. 7 Clin Invest 1992; 89: 673-80.

117 Tumang J R, Cherniack P, Gietl D M, et al. T helper celldependent, microbial superantigen-induced murine B cellactivation: polyclonal and antigen-specific antibodyresponses. Immunol 199 1; 147: 432.

118 Hohlfeld R, Toyka K V, Heininger K, et al. Autoimmunehuman T lymphocytes specific for acetylcholine receptor.Nature 1984; 310: 244-6.

119 Martin R, Jaraquemada D, Flerlage M, et al. Finespecificity and HLA restriction of myelin basic protein-specific cytotoxic T cell lines from multiple sclerosispatients and healthy individuals. 7 Immunol 1990; 145:540-8.

120 Pette M, Fujita K, Wilkinson D, et al. Myelinautoreactivity in multiple sclerosis: recognition of myelinbasic protein in the context of HLA-DR2 productsby T lymphocytes of multiple-sclerosis patients andhealthy donors. Proc Natl Acad Sci USA 1990; 87:7968-72.

121 Wucherpfennig K W, Ota K, Endo N, et al. Shared humanT cell receptor Vp usage to immunodominant regions ofmyelin basic protein. Science 1990; 248: 1016-9.

122 Sun J-B, Olsson T, Wang W-Z, et al. Autoreactive T andB cells responding to myelin proteolipid protein inmultiple sclerosis and controls. Eur Immunol 199 1; 2 1:1461-8.

123 Ben-Nun A, Liblau R S, Cohen L, et al. Restricted T-cellreceptor V5 gene usage by myelin basic protein-specificT-cell clones in multiple sclerosis: predominant genesvary in indivuduals. Proc Natl Acad Sci USA 1991; 88:2466-70.

124 Liblau R, Tournier-Lasserve E, Maciazek J, et al. T cellresponse to myelin basic protein epitopes in multiplesclerosis patients and healthy subjects. Eur Immunol1991;21: 1391-5.

125 Shirai A, Cosentino M, Leitman-Klinman S F, KlinmanD M. Human immunodeficiency virus infection inducesboth polyclonal and virus-specific B cell activation. ClinInvest 1992; 89: 561-6.

126 Oldstone M B A. Molecular mimicry and autoimmunedisease. Cell 1987; 50: 819-20.

127 Ohashi P S, Oehen S, Burki K, et al. Ablation of"tolerance" and induction of diabetes by virus infectionin viral antigen transgenic mice. Cell 1991; 65: 305-17.

128 OldstoneM B A, Nerenberg M, Lewicki H. Virus infectiontriggers insulin-dependent diabetes mellitus in atransgenic model: role of anti-self (virus) immuneresponses. Cell 1991; 65: 319-31.

129 Paliard X, West S G, Lafferty J A, et al. Evidence for theeffects of a superantigen in rheumatoid arthritis. Science1991; 253: 325-9.

130 Cole BC, Kartchner D R, Wells D J. Stimulation ofmouselymphocytes by a mitogen derived from mycoplasmaarthritides (MAM). VIII. Selective activation of T cellsexpressing distinct V5 T cell receptors from various strainsof mice by the superantigen MAM. Immunol 1990; 144:425-31.

131 Ahvonen P, Sievers K, Aho K. Arthritis associated withYersinia enterolytica infection. Acta Rheum Scand 1969;15: 232-53.

132 Keat A. Reiter's syndrome and reactive arthritis inperspective. NEngl3rMed 1983; 309: 1606-15.

S15

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

Acha-Orbea

133 Inman R D. The interplay of microbe and MHC in thepathogenesis of Reiter's syndrome. Clin Exp Rheumatol1986;4:75-82.

134 Takimoto H, Yoshikai Y, Kishihara K, et al. Stimulationof all T cells bearing V51, V3, V,11 and Vp12 bystaphylococcal enterotoxin A. Eur 7 Immunol. 1990; 20:617-21.

135 Callahan J E, Herman A, Kappler J, Marrack P.Stimulation of B10.Br T cells with superantigenicstaphylococcal enterotoxins. _7 Inmmunol 1989; 144:2473-9.

136 Yagi J, Baron J, Buxser S, Janeway J A Jr. Bacterial proteinsthat mediate the association of a defined subset of T cellreceptor: CD4 complexes with class II MHC. _7 Immunol1990; 144: 892-901.

137 Wilson R K, Lai E, Concannon P, Barth R K, Hood L E.Structure, organization and polymorphism of murine andhuman T-cell receptor and P chain gene families.Immunol Rev 1988; 101: 149-72.

138 Fairchild S, Knight A M, Dyson P J, Tomonari K. Co-segregation of a gene encoding a deletion ligand for Tcrb-V3+ T cells with Mtv-3. Immunogenetics 1991; 34:227-30.

139 Fairchild S, Rosenwasser 0 A, Dyson P J, Tomonari K.Tcrb-V3+ T-cell deletion and a new mouse mammarytumor provirus, Mtv-44. Immunogenetics 1992; 36:189-94.

140 Jouvin-Marche E, Cazenave P-A, Voegtle D, Marche P.V517 T-cell deletion by endogenous mammary tumorvirus in wild-type-derived mouse strains. Proc Natl AcadSci USA 1992; 89: 3232-5.

141 Rudy C K, Kraus E, Palmer E, Huber B T. Mls-1-likesuperantigen in the MA/MyJ mouse is encoded by a newmammary tumor provirus. _7 Exp Med 1992; 175:1613-22.

142 Tomonari K, Fairchild S. Positive and negative selectionof Tcrb-V6+ T cells. Immunogenetics 1992; 36: 230-7.

143 Vacchio M S, Kanagawa 0, Tomonari K, Hodes R J.Influence of T cell receptor V,c expression on Mlsasuperantigen-specific T cells. _7 Exp Med 1992; 175:1405-8.

144 Ferrick D A, Cho K, Gemmell-Hori L, Morris D W.Genetic analysis of the effects of Mtv-2 on the T cell

repertoire in the WXG-2 mouse strain. InternationalImmunology 1992; 4: 805-10.

145 Tomai M A, Aelion J A, Dockter M E, et al. T cell receptorV gene usage by human T cells stimulated with thesuperantigen streptococcal M protein. _7 Exp Med 1991;174: 285-8.

146 Imanishi K, Igarashi H, Uchiyama T. Activation of murineT cells by streptococcal pyrogenic exotoxin type A. 7

Immunol 1990; 145: 3170-6.147 Nelson K, Schlievert P M, Selander R K, Musser J M.

Characterization and clonal distribution of speA geneencoding pyrogenic exotoxin A (scarlet fever toxin) in

Streptococcus pyrogenes. 7 Exp Med 1991; 174: 1271 -4.148 Cole B C, Kartchner D R, Wells D J. Stimulation of mouse

lymphocytes by a mitogen derived from Mycoplasmaarthritidis. VII. Responsiveness is associated withexpression of a product(s) of the V58 gene family presenton the T cell receptor (x/c for antigen. _7 Imtlmunol 1989;142: 4131-7.

149 Cole B C, Kartchner D R, Wells D J. Stimulation of mouselymphocytes by a mitogen derived from Mycoplasmaarthritidis (MAM). VIII. Selective activation of T cellsexpressing distinct V5 T cell receptors from various strainsof mice by the "superantigen" MAM. _7 Imnmunol 1990;144: 425-31.

150 Acha-Orbea H, Mitchell D J, Timmerman L, et al. Limitedheterogeneity of T cell receptors in experimental allergicencephalomyelitis. Cell 1988; 54: 263-73.

151 Acha-Orbea H, Steinman L, McDevitt H 0. T cellreceptors in murine autoimmune diseases. Annu RevImmunol 1989; 7: 371-405.

152 Acha-Orbea H, Scarpellino L, Shakhov A N, et al.Inhibition of mammary tumor virus-induced T cellresponses in vivo by monoclonal antibodies to an openreading frame protein. Exp Med 1992; 176: 1769-72.

153 Winslow G M, Scherer M T, Kappler J W, Marrack P.Detection and characterization of the mouse mammarytumor virus 7 superantigen (Mls-l'). Cell 1992; 71:719-30.

154 McDuffie M, Schweiger D, Reits B, et al. I-E-independentdeletion of V,17a+ T cells by Mtv-3 from the nonobesediabetic mouse. _7 Immunol 1992; 148: 2097-102.

S16

on Decem


http://ard.bmj.com

/A

nn Rheum


uppl_1.S6 on 1 M

arch 1993. Dow

nloaded from

http://ard.bmj.com/

bacterial and viral superantigens: roles in autoimmunity? · caninteractwithaparticular...

Documents