Ž .Journal of Neuroimmunology 95 1999 73–84

Interleukin-4 deficiency facilitates development of experimentalmyasthenia gravis and precludes its prevention by nasal administration

of CD4q epitope sequences of the acetylcholine receptor

Peter I. Karachunski a, Norma S. Ostlie a, David K. Okita a, Bianca M. Conti-Fine a,b,) ,1

a Department of Biochemistry, College of Biological Sciences, UniÕersity of Minnesota, St. Paul, MN 55108, USAb Department of Pharmacology, School of Medicine, UniÕersity of Minnesota, Minneapolis, MN 55455, USA

Received 5 October 1998; revised 25 November 1998; accepted 25 November 1998

Abstract

Ž . Ž .Immunization with acetylcholine receptor AChR causes experimental myasthenia gravis EMG . We investigated EMG in interleukinŽ . Ž .IL -4 knock out B6 KO mice, that lack Th2 cells. EMG was more frequent in KO than in wild type B6 mice. KO and B6 micedeveloped similar amounts of anti-AChR antibodies. They were IgG2a and IgG2b in KO mice, IgG1 and IgG2b in B6 mice. CD4q cellsfrom KO and B6 mice recognized the same AChR epitopes. Nasal administration of synthetic AChR CD4q epitopes reduced antibodysynthesis and prevented EMG in B6, not in KO mice. Thus, Th2 cells may have protective functions in EMG. q 1999 Elsevier ScienceB.V. All rights reserved.

Keywords: Acetylcholine receptor; Experimental myasthenia gravis; Th2 cells; Nasal tolerance; Autoimmunity

1. Introduction

CD4q cells comprise Th1 and Th2 cells, that differ inŽtheir function and in the cytokines they secrete Abbas et

.al., 1996; Romagnani, 1997; Weigle and Romball, 1997 .Th1 cells mediate effector functions of the immune re-sponse. They secrete pro-inflammatory cytokines, such as

Ž .IFN-g, TNF and interleukin IL -2, can be cytotoxic andhelp the synthesis of IgG subclasses that bind complement.Th2 cells help the synthesis of antibodies that do not bindcomplement, such as IgE, IgG4 in humans and its homo-logue in mice, IgG1. Also, they modulate immune re-sponses by secreting anti-inflammatory cytokines, like IL-4and IL-10, that down regulate the function of antigen

Ž .presenting cells APC and Th1 cells.Th1 cells have been implicated in the pathogenesis of T

cell mediated autoimmune diseases, Th2 cells in theirŽdown regulation Miller and Karpus, 1994; Racke et al.,

) Corresponding author. Department of Biochemistry, Universityof Minnesota, 1479 Gortner Avenue, St. Paul, MN 55108, USA.Tel.: q 1-612-6246796; Fax: q 1-612-6255780; E-m ail:conti@biosci.cbs.umn.edu

1 Previously known as Bianca M. Conti-Tronconi.

1994; Cua et al., 1995; Liblau et al., 1995; Abbas et al.,1996; Cong-Qui and Londei, 1996; Tian et al., 1996;Mueller et al., 1996; Falcone and Bloom, 1997; PrabhuDas et al., 1997; Romagnani, 1997; Shaw et al., 1997; vonHerrath and Oldstone, 1997; Weigle and Romball, 1997;

.O’Garra, 1998 . However, also Th2 cells may cause TŽcells mediated autoimmune diseases Anderson et al., 1993;

Ferber et al., 1996; Hultgren et al., 1996; Lee et al., 1996;Willenborg et al., 1996; Lafaille et al., 1997; Manoury-Schwartz et al., 1997; Pakala et al., 1997; Vermeire et al.,

.1997 , and TGF-b, a cytokine secreted by modulatory Tcells that may represent a distinct T cell lineage, may

Žprotect from experimental autoimmune responses Hafler.et al., 1997; Liblau et al., 1997; O’Garra et al., 1997 .

Thus, in T cell mediated autoimmune diseases both Th2and Th1 cells, and the cytokines they secrete, may haveeffector or down regulatory functions, or both.

Both Th1 and Th2 cells may be implicated in antibody-mediated autoimmune disease. Th1 cells help synthesis of

Žantibodies able to fix complement Abbas et al., 1996;Romagnani, 1997; Weigle and Romball, 1997; O’Garra,

.1998 , that would be especially effective in causing tissuedamage. For example, in MRlrlpr mice deposits of IgGand complement activation cause autoimmune glomerulo-

0165-5728r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0165-5728 98 00262-8

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8474

nephritis, and expression of IFN-g is necessary for theŽdevelopment of autoimmune glomerulonephritis Hass et

.al., 1997 . Because of their effective T helper function,also Th2 cells may be involved in antibody-mediated

Žautoimmune diseases Erb et al., 1997; Fuss et al., 1997;.Nakajima et al., 1997; Peng et al., 1997 . However, Th2-

induced antibodies do not fix complement or bind theŽphagocyte Fc receptor Abbas et al., 1996; Romagnani,

.1997; Weigle and Romball, 1997; O’Garra, 1998 , and areunlikely to cause severe tissue injury.

Antibodies against the muscle acetylcholine receptorŽ . Ž .AChR cause myasthenia gravis MG and its animal

Ž .model, experimental MG EMG , induced by immuniza-Žtion with purified AChR reviewed in Conti-Fine et al.,

.1997 . Th1 cells may have a role in the pathogenesis ofMG and EMG. MG patients had AChR-specific Th1 cellsŽ .Moiola et al., 1994a,b; Wang et al., 1997, 1998 . Trans-genic mice that produced IFN-g at the neuromuscularjunction developed functional disruption of the junction

Ž .and clinical weakness reminiscent of MG Gu et al., 1995 ,and mice deficient in IFN-g appeared to be resistant to

Ž .induction of EMG Balasa et al., 1997 . Mice deficient inIL-12—a cytokine necessary for the development of Th1responses—were resistant to EMG induction, whereas ad-ministration of IL-12 at the time of the immunization with

Ž .AChR facilitated EMG development Moiola et al., 1998 .On the other hand, Th2 cells may have a modulatory rolein EMG. Nasal or subcutaneous administration to C57Blr6Ž . qB6 mice of synthetic AChR peptides forming CD4epitopes activated Th2 cells specific for the peptides ad-ministered, caused reduced synthesis of anti-AChR anti-

Žbodies, and prevented EMG Karachunski et al., 1997,.1999; Wu et al., 1997 .

Induction of EMG in B6 mice requires multiple AChRinjections, and the frequency of EMG is 20–70% even

Žafter prolonged AChR immunization reviewed in Conti-. Ž .Fine et al., 1997 . IL-4 knock out KO mice have an

effective Th1 function, but a severely impaired Th2 func-Ž . qtion Kuhn et al., 1991 . Their CD4 cells do not express¨

other Th2 cytokines, and they have very low amounts ofŽTh2-driven IgG1 in the serum Kuhn et al., 1991; Kopf et¨.al., 1993; Lawrence et al., 1995 . The defective Th2 cell

activity in IL-4 KO mice agrees with the demonstrationobtained in studies in vitro, that IL-4 is essential for the

Žgeneration of Th2 responses Le Gros et al., 1990; Swain.et al., 1990 . IL-4 KO mice of B6 background are an

excellent model system to investigate the role of IL-4, andindirectly that of Th1 and Th2 cells, in the pathogenesisand prevention of EMG. In this study we investigated theappearance of anti-AChR antibodies and of EMG in wildtype B6 and in IL-4 KO B6 mice, immunized with Tor-

Ž .pedo AChR TAChR . Also, we determined whether nasaladministration to the IL-4 KO mutants of immunodomi-nant CD4q epitopes of the TAChR affected the anti-TAChR antibody synthesis and prevented EMG, as it does

Ž .in wild type B6 mice Karachunski et al., 1997 .

2. Materials and methods

2.1. Mice

ŽB6 and IL-4 KO mice of B6 background Jackson.Laboratory, Bar Harbor, ME were housed at the animal

facility of the University of Minnesota.

2.2. Purification of Torpedo AChR

We purified TAChR from Torpedo californica electricorgan as alkali-stripped TAChR-rich membrane fragmentsŽ .Bellone et al., 1991a . We measured the protein concen-

Ž .tration by the Lowry assay Lowry et al., 1981 and theŽ .TAChR concentration as a-bungarotoxin a-BTX binding

Ž .sites Bellone et al., 1991a . The TAChR preparations weused contained 3.8–5.8 nmol of sitesrmg protein. Weassessed the protein composition by sodium dodecyl sul-

Ž . Žfate SDS polyacrylamide gel electrophoresis Laemmli,.1970 : the TAChR preparations we used contained only

the four TAChR subunits as the main protein bands. Foruse in cell cultures, we diluted the TAChR-rich membranefragments in RPMI-1640 as needed, and sterilized them byultraviolet irradiation. For immunization and antibody as-say, we solubilized the membranes in 1% Triton X-100Ž .Bellone et al., 1991a , diluted them to 0.5 mgrml in PBSand stored them at y80 C8.

2.3. Peptide synthesis and characterization

We used four panels of overlapping peptides, ;20residue long, spanning the sequences of the TAChR a , b,g and d subunits, and synthesized as described inŽ .Houghten, 1985 . They overlapped by approximately 5residues. We reported their characterization previouslyŽ .Bellone et al., 1991a, 1993 . For proliferation assays, weused the peptides as roughly equimolar pools of all thepeptides spanning the sequence of one TAChR subunit.We also used the a subunit peptides individually. Forproliferation assays, we used peptide solutions in PBS,sterilized by ultraviolet irradiation and stored frozen.

Three a subunit sequence regions, corresponding toresidues 150–169, 181–200 and 360–378 are dominant forsensitization of the anti-TAChR CD4q cells in B6 miceŽ .Bellone et al., 1991a, 1993; Karachunski et al., 1995 . Weindicate these peptides with codes that include Ta forTAChR a subunit and two numbers, referring to theposition on the a subunit sequence of the first and lastresidues of the peptide. We used these peptides for nasaltolerization procedures. We routinely characterized themas follows. We carried out reverse-phase high pressureliquid chromatographic analysis of the peptides on a C18

Ž .column Ultrasphere ODS, Beckman, Fullerton, CA and agradient of acetonitrile in 0.1% trifluoroacetic acid inwater. We consistently found one main peak of optical

Ž .density not shown . We verified the amino acid composi-

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–84 75

tion of the peptides by derivatization of the amino acidresidues released by acid hydrolysis with phenylisothio-cyanate, followed by separation on a reverse-phase high

Žpressure liquid chromatography column Heinrikson and.Meredith, 1984 . The results of the amino acid composi-

tion corresponded closely to the expected theoretical val-Ž .ues not shown . We verified the sequence and purity of

some randomly selected batches of peptides by mass spec-trometry. For all peptides, we found a major peak of the

Ž .expected molecular weight not shown .

2.4. Immunizations

We immunized eight 10-week-old mice by subcuta-neous injections, along the back and at the base of the tail,

Ž .of solubilized TAChR 25 mg in 100 ml PBS emulsifiedwith an equal volume of complete Freund adjuvant. Weboosted them twice at 4 week intervals with the sameamount of TAChR emulsified in incomplete Freund adju-vant.

2.5. Nasal administration of synthetic TAChR CD4q epi-topes

We anesthetized the mice by intra peritoneum injectionŽ .of Ketaset 100 mgrkg; Alveco, Fort Dodge, IA , and

instilled into the mouse nostrils 25 ml of phosphate bufferedŽsaline solution PBS: 10 mM Na phosphate buffer, pH 7.4,

.2.7 mM KCl, 137 mM NaCl containing 50 mg of each ofŽpeptides Ta150–169, Ta181–200 and Ta360–378 re-

.ferred to as ‘a epitope pool’ . Control mice received cleanPBS. We administered the a epitope pool or clean PBSweekly, starting 2 weeks before beginning of the TAChRimmunization, for a total of 12 treatments. We investigatedpreviously the distribution in the respiratory tract of solu-

Ž .tions instilled nasally Karachunski et al., 1997 . We foundthe solutions in the mouse nostrils, larynx and trachea, and

Žto a much lesser extent in the lung parenchyma Karachun-.ski et al., 1997 .

2.6. EÕaluation of clinical symptoms of EMG

We quantified the symptoms of EMG using a forcedexercise by the inverted hang technique, sensitized by a

Žminute amount of pancuronium bromide 0.03 mgrkg. Žintra peritoneum , given just before the test Karachunski

.et al., 1995 . The mice hang from a grid, and we measuredthe time it took for the mouse to release its hold and fall

Ž .three times ‘holding time’ . We tested the mice on the dayof the first nasal administration, on the day before eachimmunization, and approximately 14 days after the thirdimmunization, just before sacrificing the mice. We per-formed the test blindly, i.e., without knowledge of thetreatment that the mouse had received. This test is para-metric, and gives a quantitative assessment of the severityof the mouse weakness. To verify the myasthenic nature of

Žthe weakness we injected edrophonium chloride Reversol,.Organon, West Orange, NJ intra peritoneum. Reversol is a

cholinesterase inhibitor, and it immediately increased thestrength of the mice.

The holding time of normal mice is 10.4"2.1 minŽ . Ž .ns99 Karachunski et al., 1995, 1997 . We considered

Žmyasthenic the mice with holding times of 6.2 min the.holding time of normal mice minus 2 SD or less. Normal

Žmice never have holding times shorter than 6.2 min Fig..1; Karachunski et al., 1995, 1997 . Paralyzed mice or mice

that died of respiratory paralysis are represented in thefigures as having holding time of zero.

2.7. Anti-AChR antibody assay

We obtained sera after each clinical testing. We mea-sured the serum concentration of anti-TAChR antibody byradioimmunoprecipitation assay, using TAChR solubilized

Fig. 1. IL-4 KO mice are more susceptible to EMG than wild type B6mice, and are not protected from EMG by nasal administration ofsynthetic TAChR CD4q epitopes. Strength of IL-4 KO and B6 mice, asindicated at the right side of the panels, measured by the pancuronium-sensitized hanging test. The mice were sham-treated nasally with clean

Ž . ŽPBS top two panels or treated nasally with the a epitope pool bottom.two panels , and immunized with TAChR. The tests were carried out just

Ž .before the first TAChR immunizing injection panels ‘0 weeks’ , and 4, 8and 10 weeks after beginning of the immunization, as indicated below the

Žpanels. The horizontal lines indicate a holding time of 6.2 min the.holding time of normal mice minus 2 SD . Mice with holding time of 6.2

min or less were considered to have EMG. The average holding time ofthe different groups of mice is indicated by black diamonds. See text forexperimental details.

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8476

in Triton X-100 and labeled by the binding of 125I-a-BTXŽ .Bellone et al., 1993 . We express the antibody concentra-tion as mM precipitated 125I-a-BTX.

2.8. Assay of anti-TAChR and anti-peptide IgG subclasses

We measured by ELISA the relative concentration ofIgG subclasses of anti-TAChR and anti-peptide antibodiesin the sera of peptide- and sham-treated mice, after immu-nization with TAChR. We used pooled sera from three–four mice that had received identical treatments. We usedfor this assay the sera obtained at the end of the observa-

Žtion period 10 weeks after beginning of the anti-TAChR.immunization .

Ž .Ninety-six well plates Nunc, Karstrup, Denmark werewashed extensively and incubated as follows: 4 h with a10 mgrml solution of TAChR or of individual peptidesTa150–169, Ta181–200 and Ta360–378 in 0.1 M Na

Ž .phosphate buffer, pH 9.5 two wells for each antigen , 1 hŽ .with PBS plus 3% bovine serum albumin BSA , 2 h with

Žmouse serum serial dilutions from 1:100 to 1:40,000 in.PBS plus 3% BSA , 1 h with a dilution 1:1000 in PBS plus

3% BSA of goat antibody specific for the total mouse IgG,Žor IgG1, or IgG2a, or IgG2b or IgG3 Mouse Monoclonal.Isotyping Kit, Sigma, St. Louis, MO , and 30 min with a

1:3000 dilution of peroxidase-conjugate rabbit anti-goatŽ .IgG Sigma in PBS plus 3% BSA. We washed the plates,

and developed them for 20–60 min with ABTS peroxidaseŽsubstrate system Kirkegaard and Perry Laboratories,.Gaithersburg, MD . We stopped the reaction with a 1%

solution of SDS in PBS. We read the OD at 405 nm.This assay is qualitative, but it allows comparison of the

relative amount of IgG subclasses within the same sample,and in different samples run simultaneously. We did thisby comparing the OD values obtained within the linearrange of the dose dependence curve.

2.9. Detection of anti-AChR IgG and mouse complementbound to muscle AChR by immunofluorescence microscopy

The hind limb muscle of naive, peptide- and sham-treated IL-4 KO mice were frozen in liquid nitrogen andstored at y708C. We embedded the frozen tissue in O.C.T.

Ž .Compound Tissue-TEK Miles Laboratories, Elkhart, INand sectioned it in the transverse direction into 10 mm

Žsections using a Jung Frigout 2800E Kryostat Leica,.Nublach, Germany . We incubated the sections at room

temperature in PBS for 10 min, and for 1 h with a 1:200dilution of goat anti-mouse IgG conjugated with biotinŽ .Sigma in PBS containing 3% BSA. We washed thesections with PBS for 15 min three times, and stained themfor 1 h at room temperature with Texas Red labeled

Ž .a-BTX Molecular Probes, Eugene, OR , FITC labeledŽgoat anti-mouse complement C3 antibody Nordic Im-

.munological Laboratories, Capistrano Beach, CA , biotineŽ .conjugated anti mouse IgG Sigma in PBS containing 3%

BSA at 1:4000, 1:100 and 1:200 dilutions respectively, andŽ .AMCA-S labeled streptavidine Molecular Probes diluted

in PBS at a 1:500 dilution. We washed the sections threetimes for 15 min with PBS and viewed them in a fluores-

Ž .cence microscopy Nikon eclipse E 800, Japan . We col-Žlected digital images using Image Pro Plus Media Cyber-

.netics, L.P., Silver Spring, MD .

2.10. Lymphocyte proliferation assay

Two weeks after the last immunization, we obtainedŽ .spleen cells Bellone et al., 1991a from three identically

treated IL-4 KO mice. We pooled the cells and depletedthem in CD8q cells using paramagnetic beads and rat

q Ž .anti-mouse CD8 antibody Pharmingen, San Diego, CA .ŽWe suspended the cells in RPMI-1640 Gibco, Grand

.Island, NY supplemented with 10% heat inactivated fetalŽ .calf serum Gibco , 50 mM 2-mercaptoethanol, 1 mM

L-glutamine, 10 mM Hepes, 1 mM sodium pyruvate, 100Ž 6Urml penicillin and 100 mgrml streptomycin 1=10

.cellsrml . We seeded the cells in triplicate in 96 flat-bot-Ž .tom well plates 200 mlrwell . We added one of the

following antigens or stimulants: 10 mgrml phytohemoag-Ž . Ž .glutinine PHA Sigma ; 10 mgrml TAChR; the pools of

overlapping peptides spanning the a , b, g or d subunitŽ .sequences 5 mgrml of each peptide ; and 10 mgrml of

the individual a subunit peptides. Controls were triplicatewells cultivated without any antigen, or with or a 20-re-sidue control peptide synthesized by the same method,

Ž .unrelated to the TAChR sequence 10 mgrml . After 43 Ždays we labeled the cells for 16 h with H-thymidine 1

mCi per well, specific activity 6.7 Cirmmol, Dupont,. ŽBoston, MA , and harvested them Titertek, Skatron, Ster-

. 3ling, VA . We measured the H-thymidine incorporationby liquid scintillation.

2.11. Cytokine secretion by CD8q depleted spleen cells inresponse to stimulation with TAChR

Two weeks after the last TAChR immunization weprepared CD8q depleted spleen cells from three identicallytreated mice, using the procedure described above. Weresuspended the CD8q depleted spleen cells at 5=106

cellsrml, and cultured them with and without 10 mgrmlŽ .TAChR or a epitope pool 10 mgrml of each peptide in

24 well plates. In some experiment we set up two indepen-dent cultures for each antigen. Cells cultivated without anyantigen served as controls for spontaneous secretion ofcytokines. We harvested the culture supernatants after 72and 96 h. We measured the concentrations of IFN-g, IL-2,IL-4 and IL-10 by capture ELISA, using duplicate sam-ples. We used anti-IFN-g, anti-IL-2, anti-IL-4 and anti-IL-

Ž .10 monoclonal and polyclonal antibodies Pharmingen ,Žand recombinant IFN-g, IL-2, IL-4 and IL-10 Pharmin-

.gen as standards, and followed the manufacturer’s instruc-tions.

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–84 77

2.12. Statistical analysis

We determined the significance of the differences of theaverage responses of two groups using a two-tailed Stu-dents’ t-test.

3. Results

3.1. IL-4 KO mice had increased susceptibility to EMG

The top panels of Fig. 1 report the results we obtainedtesting the strength of B6 and IL-4 KO mice sham-treatednasally with clean PBS, and immunized with TAChR. Wereport the results obtained 4, 8 and 10 weeks after begin-ning of the immunization. For each group we report alsothe results obtained for the same mice before the TAChR

Ž .immunization panels ‘0 weeks’ . The results we obtainedat 8 and 10 weeks were consistent. They reflect themaximum frequency of EMG symptoms that we observed.All IL-4 KO mice, and 58% of the wild type B6 micedeveloped EMG. Fig. 1 reports also the average holdingtime of the groups of mice used in these experimentsŽ .black symbols . Wild type B6 mice had significantlylonger holding time than the IL-4 KO mice at 8 weeksŽ . Ž .P)0.001 and 10 weeks P)0.03 .

3.2. Nasal instillation of the a epitope pool did notpreÕent EMG in IL-4 KO mice

The bottom panels of Fig. 1 report the results weobtained testing the strength of B6 and IL-4 KO micetreated nasally with the TAChR peptide epitopes, andimmunized with TAChR. We report the results obtained 4,8 and 10 weeks after beginning of the immunization. Thepanels ‘0 weeks’ report the results obtained for the samemice before the TAChR immunization. Among the 24

Ž .peptide-treated B6 mice, two mice 8% had EMG symp-Ž .toms at 8 weeks and one 4% at 10 weeks, as compared to

58% of the sham-treated mice. Seven of the eight peptide-Ž .treated IL-4 KO mice 88% had EMG weakness at 8 and

10 weeks.Fig. 1 reports also the average holding time of the

Ž .groups of mice used in these experiments black symbols .B6 mice treated nasally with the a epitope pool had

Ž .significantly P)0.001 longer holding time than thesham-treated B6 mice. The average holding time of pep-tide-treated IL-4 KO mice was the same as the group thathad inhaled clean PBS.

3.3. Nasal treatment of IL-4 KO mice with the a epitopepool did not affect the synthesis of anti-TAChR antibody

We measured by radioimmmunoprecipitation assays theanti-TAChR antibody concentration in the sera of B6 and

IL-4 KO mice treated nasally with the a epitope pool orsham-treated with clean PBS, and immunized with TAChR.We used sera obtained 4, 8 and 10 weeks after thebeginning of the TAChR immunization.

In wild type B6 mice the nasal treatment with thesynthetic CD4q epitopes reduced substantially the synthe-

Ž .sis of anti-TAChR antibodies Fig. 2A, top panel . Peptide-and sham-treated IL-4 KO mice had similar serum concen-trations of anti-AChR antibodies, which were comparable

Žto those of sham-treated wild type B6 mice Fig. 2A,.bottom panel .

3.4. Nasal treatment of IL-4 KO mice with the a epitopepool did not affect the synthesis of Th1-induced anti-TAChRantibodies

We investigated the effect of nasal treatment with the a

epitope pool on the synthesis of Th1-dependent antibodiesagainst the TAChR. We assessed by ELISA the relativeconcentrations of total anti-TAChR IgG and of differentanti-TAChR IgG subclasses, in the sera of peptide-treatedand sham-treated IL-4 KO and B6 mice immunized withTAChR. We determined the relative concentration of IgG

Žsubclasses synthesized with the help of Th1 IgG2a and. Ž .IgG2b, IgG3 or Th2 IgG1 cells. We used sera obtained

10 weeks after beginning of the TAChR immunization.

Fig. 2. Nasal treatment of IL-4 KO mice with the a epitope pool does notcause reduced synthesis of anti-TAChR antibodies or affect the synthesis

Ž .of Th1-induced anti-TAChR antibodies. A Average anti-TAChR anti-Ž .body concentration "SD , measured by radioimmmunoprecipitation as-

Ž .say, in sera of B6 mice treated nasally with the a epitope pool ns9 orŽ .sham-treated with clean PBS ns10 , and of IL-4 KO mice treatedŽ .nasally with the a epitope pool ns8 or sham-treated with clean PBS

Ž .ns7 . The sera were obtained 4, 8 and 10 weeks after the beginning ofthe immunization with TAChR, as indicated along the abscissa of the

Ž .plots. B Relative amounts of anti-TAChR IgG of different subclasses, asindicated below the plots, expressed as percent of total anti-TAChR IgG,in sera from B6 and IL-4 KO mice, as indicated inside the plots. The

Ž .mice had been treated nasally with the a epitope pool black columns , orŽ .had been sham-treated with clean PBS white columns . The sera were

obtained 10 weeks after the beginning of the immunization with TAChR.See text for experimental details.

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8478

In agreement with the results of the radioimmmunopre-cipitation assays, the amount of total anti-AChR IgG wasidentical in peptide- and sham-treated IL-4 KO mice,whereas it was significantly and substantially reduced inthe peptide-treated B6 mice as compared to the sham-

Ž .treated B6 mice data not shown . Fig. 2B reports theresults of one of three consistent experiments that assessedthe relative concentrations of anti-TAChR IgG1, IgG2a,IgG2b and IgG3, using pooled sera from three–four micethat had received identical treatments. In B6 mice therelative concentration of Th1-dependent IgG subclasseswere similar in peptide- and sham-treated mice, whereasthe relative concentration of the Th2-driven anti-TAChR

Ž .IgG1 was significantly P-0.001 increased in the pep-tide-treated mice. This is likely related to synthesis of

Ž .antibodies to the peptides administered see below . Anti-peptide antibodies may cross-react with TAChR in ELISAbecause of the partial denaturation of the TAChR when

Ž .absorbed onto the plastic plates Conti-Fine et al., 1996 .In IL-4 KO mice the relative concentrations of all IgGsubclasses were the same in the peptide-treated and in thesham-treated mice. The IL-4 KO mice had substantialamounts of anti-TAChR IgG2a and IgG2b, that were lowerthan but comparable to the amount of total anti-TAChRIgG. This suggests that in IL-4 KO mice the anti-TAChRIgG are mostly IgG2a and IgG2b. The mice had minimalamounts of anti-TAChR IgG3. IL-4 KO mice had signifi-

Ž .cantly less P-0.0001 anti-TAChR IgG1 than B6 mice.Synthesis of IgG1 is both IL-4 and IL-2 dependent, andtherefore IL-4 KO mice may still synthesize IgG1.

3.5. Synthesis of Th1-induced anti-peptide antibodies inpeptide-treated IL-4 KO mice

In B6 mice nasal instillation of the CD4q epitopepeptides used here stimulated the synthesis of Th2-induced

Ž .anti-peptide antibodies Karachunski et al., 1997 . Weinvestigated whether nasal treatment of IL-4 KO mice with

the a epitope pool stimulated the synthesis of anti-peptideantibodies. We determined the relative concentration oftotal IgG and of IgG subclasses against the three epitopepeptides used for the nasal treatment. We used pooled seraof three–four mice that had been peptide- or sham-treated,and immunized with TAChR. We obtained the sera 10weeks after beginning of the TAChR immunization. Fig. 3reports the results of one of two consistent experiments.Sham-treated mice had very little IgG antibodies reactivewith the peptides. Nasal treatment with the a epitope poolcaused synthesis of Th1-induced antibodies against each ofthe peptides administered. The antibodies were primarilyor exclusively IgG2b. Peptides Ta150–169 and Ta360–378 elicited also a moderate synthesis of IgG2a. We couldnot detect Th2-induced anti-peptide IgG1 in IL-4 KOmice.

3.6. Presence of mouse IgG and complement at the neuro-muscular junction of IL-4 KO mice immunized with TAChR

Rodents immunized with TAChR, as well as MG pa-tients, have complement at the neuromuscular junction,and complement-induced destruction of the neuromuscularjunction is believed to be an important pathogenic mecha-

Žnism in both EMG and MG reviewed in Conti-Fine et al.,.1997 . We investigated the presence of IgG and of the C3

component of complement at the neuromuscular synapsesof sham-treated and peptide-treated IL-4 KO mice, immu-nized with TAChR. Fig. 4 reports the results of one ofseveral consistent experiments. Both peptide-treated andsham-treated mice had IgG and C3 bound to the musclesynapses.

3.7. Nasal treatment of IL-4 KO mice with the a epitopepool caused reduced proliferatiÕe responses of CD4q

cells to the TAChR and to the administered peptides

We assessed the effect of nasal treatment of IL-4 KOmice with the epitope peptides, on sensitization of CD4q

Fig. 3. Nasal treatment of IL-4 KO mice with the a epitope pool causes synthesis of Th1-induced IgG isotypes against the peptides administered.Concentrations of total IgG and of IgG of different subclasses, as indicated below the plots, against the sequences Ta150–169, Ta181–200 and

Ž .Ta360–378, as indicate above each panel, in sera of IL-4 KO mice treated nasally with the a epitope pool black columns , or sham-treated with cleanŽ .PBS white columns . The columns represent the average"SD of triplicate ELISA determinations, using pooled sera from three–four identically treated

mice from each group. The sera were obtained 10 weeks after beginning of the anti-TAChR immunization. See text for experimental details.

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–84 79

Fig. 4. Presence of mouse IgG and complement at the neuromuscular junction of IL-4 KO mice. Muscle sections from IL-4 KO mice treated nasally withthe a epitope pool or sham-treated nasally with clean PBS and immunized with TAChR, and from naive IL-4 KO mice, as indicated at the right of the

Ž . Ž .panels. The sections were stained for the presence of mouse IgG blue fluorescence and of the C3 component of complement green fluorescence asŽ .indicated below the panels. We localized the neuromuscular synapses using triple immunofluorescence staining with a-BTX red fluorescence , as

indicated below the panels. Magnification: 1000= . See text for experimental details.

cells to the TAChR and to the administered epitopes. Wereported previously the results of similar studies using wild

Ž .type B6 mice Karachunski et al., 1997 . We sacrificedIL-4 KO mice treated nasally with the a epitope pool orsham-treated with clean PBS and immunized with TAChR,10 weeks after beginning of the TAChR immunization. Wetested the proliferative responses of their CD8q depletedspleen cells to the TAChR, to the different TAChR sub-

Ž .units a , b, g and d , and the individual peptides screen-ing the a subunit sequence. As representative antigens ofthe TAChR subunits, we use equimolar pools of overlap-ping synthetic peptides spanning the sequence of eachsubunit. We demonstrated previously that the proliferativeresponses to those peptide pools of CD4q cells fromTAChR-immunized mice compared well with the re-

Ž .sponses to purified TAChR subunits Bellone et al., 1991a .We used for each mouse group pooled CD8q depleted

spleen cells from three mice that had received identicaltreatments. This was necessary to obtain sufficient CD8q

depleted cells to test all the antigens. Fig. 5 reports theresults of one of two experiments, that yielded consistent

Ž .results. The cells from sham-treated mice white columns

responded vigorously to the TAChR and to the a subunit,whereas they responded minimally or not at all to the other

Ž .subunits Fig. 5, inset . This agrees with the results of ourŽprevious studies Bellone et al., 1991a, 1993; Karachunski

.et al., 1995 , that demonstrated that the a subunit domi-nated the sensitization of anti-TAChR CD4q cells in B6

Ž .mice. The cells from peptide treated mice black columnsŽresponded well to the TAChR and to the a subunit Fig. 5,

.inset . However, their responses were significantly lowerŽthan those observed for cells of sham-treated mice P-

10y6 for the response to the TAChR, and P-0.002 for.the response to the a subunit .

ŽSimilar to the wild type B6 mice Bellone et al.,1991a,b, 1993; Wall et al., 1994; Karachunski et al.,

. q1995 , CD8 depleted spleen cells from sham-treated IL-4KO mice recognized strongly peptides Ta150–169,Ta181–200, Ta360–378 and Ta146–162, that overlaps

Ž . qTa150–169 Fig. 5, white columns . The CD8 depletedspleen cells from peptide-treated mice responded mini-mally or not at all to the peptides administered nasally, but

Žthey recognized Ta146–162 strongly Fig. 5, black.columns .

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8480

3.8. Cytokine secreted by CD8q depleted spleen cellsfrom IL-4 KO mice immunized with TAChR, after chal-lenge with TAChR or a epitope pool

We tested the secretion of IL-2, IFN-g, IL-4 and IL-10by CD8q depleted spleen T cells from peptide-treated andsham-treated IL-4 KO mice immunized with TAChR, afterchallenge in vitro with TAChR or with the a epitope pool.We carried out two independent experiments. For eachexperiment we used pooled CD8q depleted spleen cellsfrom three identically treated mice, obtained 10 weeksafter the beginning of the TAChR immunization. We setup two independent cultures for the TAChR, and one forthe a epitope pool. Cells cultivated without any antigenserved as controls for spontaneous cytokine secretion. Weobtained consistent results in the two independent experi-ments, and in the duplicate cultures of each experiment.We assessed the concentration of secreted cytokines after72 and 96 h of culture, with comparable results.

Fig. 6 reports the results we obtained in one experiment,that assessed the secretion of IL-2 and IFN-g after 72 h ofculture with TAChR or the a epitope pool, as indicated ineach panel. The symbols represent the average of duplicate

Ž . ŽELISA assays of IL-2 top panels and IFN-g bottom. qpanels in the supernatant of cultures of CD8 depleted

Ž .spleen cells from sham-treated white symbols or a epi-Ž .tope pool treated black symbols IL-4 KO mice, immu-

nized with TAChR. We have subtracted from the data thespontaneous secretion of IL-2 and IFN-g by cells culti-vated without any stimulus. The cells from peptide- andsham-treated mice, after challenge with TAChR or the a

epitope pool, secreted similar levels of IL-2. The cellsŽfrom peptide-treated mice secreted significantly less P-

Fig. 5. CD8q depleted spleen cells from IL-4 KO mice treated nasallywith the a epitope pool and immunized with TAChR have a reducedproliferative response to the to the TAChR, and to the peptides adminis-tered nasally. Responses of CD8q depleted spleen cells from mice

Ž .treated nasally with the a epitope pool black columns or sham-treatedŽ .with clean PBS white columns to the TAChR and to pools of overlap-

ping synthetic peptides spanning the sequence of each TAChR subunitsŽ .inset , and to individual overlapping synthetic peptides screening theTAChR a subunit sequence, as indicated at the bottom of the panel. Thecolumns represent average cpm"SD of triplicate cultures. The columnsindicated as ‘Cont. pep.’ are average cpm"SD of triplicate culturescultivated in the presence of a 20-residue peptide synthesized by the samemethod, unrelated to the TAChR sequence. See text for experimentaldetails.

Fig. 6. Secretion of Th1 cytokines by CD8q depleted spleen cells fromIL-4 KO mice immunized with TAChR, after challenge in vitro with

Ž . Ž .TAChR or a epitopes pool. IL-2 top panels and IFN-g bottom panelsin the supernatant of cultures of CD8q depleted spleen cells from IL-4

Ž .KO mice treated nasally with the a epitope pool black symbols orŽ .sham-treated with clean PBS white symbols , and immunized with

TAChR. Two independent cell cultures were cultivated in the presence ofTAChR, and one culture in the presence of the a epitope pool, asindicated inside the plots. The symbols represent the average"SD ofduplicate ELISA determinations, using increasing amounts of the super-natant of each culture, as indicates below the plots, after 72 h ofincubation with the antigen. The spontaneous secretion of IL-2 and IFN-gby cells cultivated in the absence of any stimulus has been subtractedfrom these data. See text for experimental details.

.0.005 IFN-g than the cells from sham-treated mice, bothafter challenge with TAChR or with the a epitope pool.We could not detect secretion of IL-4 and IL-10 by CD8q

Ž .depleted spleen cells from IL-4 KO mice not shown .

4. Discussion

This study demonstrates that IL-4 is not necessary fordevelopment of EMG, and suggest that this cytokine, andtherefore Th2 cells, may have a down-regulatory functionon the development of anti-TAChR antibodies, and of theresulting EMG symptoms. Absence of IL-4 and of aneffective Th2 function facilitated EMG development, andanti-TAChR Th2 cells appeared to have an important rolein the protection from EMG resulting from nasal ‘toleriza-tion’ procedures.

ŽIL-4 KO mice have functional Th1 cells Kuhn et al.,¨.1991 . Consequently, the present results indicate that sensi-

tization of Th1 cells is sufficient to cause EMG. Thisagrees with the conclusion of a recent study, that demon-

Žstrated that IL-12 is involved in induction of EMG Moiola.et al., 1998 . Another study also suggested that Th1 cells

are indispensable for EMG induction, because IFN-g KOŽmice immunized with TAChR did not develop EMG Bal-

.asa et al., 1997 . However, in that study IFN-g KO miceimmunized with TAChR had very low levels of serumanti-TAChR antibodies of all subclasses, including Th2-

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–84 81

dependent subclasses, as compared to wild type miceŽ .Balasa et al., 1997 . IFN-g enhances the expression ofMHC proteins and antigen presentation, and thereforefacilitates priming of all CD4q cells. Also, it may facilitatesensitization of Th2 cells by up regulating IL-4 production

Žand down regulating IL-2 production, or both Vermeire et.al., 1997 . The reported absence of EMG in IFN-g KO

Ž .mice Balasa et al., 1997 might have been caused byineffective sensitization of anti-TAChR Th1 or Th2 cellsor both, and inadequate help for anti-TAChR antibodysynthesis.

Tolerance induced by nasal, oral or systemic adminis-tration of an antigen or of epitope sequences can resultfrom several mechanisms. They include: anergy or deletionof antigen-specific T cells and induction of antigen spe-

q Žcific regulatory CD4 Th2 cells Weiner et al., 1994;.Chen et al., 1995, 1996 . The present results suggest that

modulatory Th2 cells have an important role in the protec-tive effect of nasal administration of synthetic TAChRepitopes, because nasal treatment of IL-4 KO mice withpeptide epitopes did not affect the development of EMG.Anergy or deletion of Th1 cells had occurred in thepeptide-treated IL-4 KO mice, because their CD8q de-pleted spleen cells had a reduced proliferative responses tothe TAChR, to the TAChR a subunit and to the individual

Ž . qpeptides administered nasally Fig. 5 . Also, their CD8depleted spleen cells synthesized less IFN-g after chal-

Ž .lenge in vitro with TAChR or the a epitope pool Fig. 6 .The reduced anti-TAChR Th1 response did not suffice to

Ž .protect from EMG Fig. 1 or to reduce the synthesis ofŽ .anti-TAChR antibodies Fig. 2 . Consequently, the strong

reduction in synthesis of anti-TAChR antibodies observedin B6 mice after nasal or subcutaneous administration of

Žthe TAChR peptide epitopes Fig. 2; Karachunski et al.,.1997, 1999; Wu et al., 1997 must be mediated by activa-

tion of regulatory Th2 cells. Anergy of anti-TAChR Th1cells might protect from EMG if one could use a fullcomplement of the TAChR epitope sequences recognizedby the pathogenic Th1 clones, in doses adequate to causeanergy or deletion of all the specific Th1 clones.

When we used proliferative assays, that detect primarilythe response of Th1 cells, the pattern of recognition of theTAChR subunits and the epitope repertoire on the a

subunit of CD8q depleted spleen cells from IL-4 KO miceŽ . ŽFig. 5 was identical to that of wild type B6 mice Bellone

.et al., 1991a,b, 1993; Karachunski et al., 1995 . Thus, theabsence of IL-4 does not appear to affect the sensitizationof Th1 cells to the TAChR. Th1 and Th2 cells may

Žrecognize different epitopes on the same antigen Julia etal., 1996; Mikszta and Kim, 1996; Das et al., 1997; Fresno

.et al., 1997 . B6 mice must have anti-TAChR Th2 cells,because they have a strong anti-TAChR IgG1 responseŽ .Fig. 2 . Anti-TAChR Th2 cells in B6 mice may recognizethe same epitope sequences as the Th1 cells, or otherepitopes, whose detection might require assays other thanthe proliferative assay.

Spleen cells from IL-4 KO mice that sniffed the a

epitope pool responded to peptide Ta150–169 much lessthan the sham-treated mice, but they still responded

Ž .strongly to the overlapping peptide Ta146–162 Fig. 5 .This indicates that the sequence region a146–169 containsoverlapping epitopes that sensitize different pathogenicCD4q clones. The residual population of Th1 cells thatrecognized epitopes within the amino terminal part of the

Žsequence region a146–169 i.e., those that responded to.Ta146–162, not to Ta150–169 sufficed to cause EMG

in the peptide-treated IL-4 KO mice. This verifies theimportance of this sequence for sensitization of pathogenic

ŽTh1 cells Bellone et al., 1991b; Infante et al., 1991;.Asthana et al., 1993; Karachunski et al., 1995 .

Nasal administration of a epitope pool to IL-4 KOmice caused synthesis of Th1-induced anti-peptide anti-

Ž .bodies Fig. 3 . The anti-peptide antibodies did not cross-react significantly with native TAChR, because the serumconcentration of anti-TAChR antibodies measured in ra-dioimmunoprecipitation assays was identical in peptide- or

Ž .sham-treated mice Fig. 2 . Anti-peptide antibodies cross-Žreact seldom with the cognate native antigen reviewed in

.Conti-Fine et al., 1996 . The Th2-induced anti-peptideantibodies synthesized by wild type B6 mice after nasalpeptide treatment may have recognized partially denaturedTAChR in the ELISA, which revealed a significantlyhigher relative concentration of anti-TAChR IgG1 antibod-ies—an IgG subclass that is primarily Th2-induced—in

Ž .peptide-treated B6 mice Fig. 2 .Sensitization of modulatory Th2 cells that protect from

EMG explains the previous findings, that nasal or subcuta-neous administration of a single TAChR CD4q epitope

Ž .peptide, Ta150–169 Karachunski et al., 1997, 1999 orŽ .Ta146–162 Wu et al., 1997 , protected from EMG.

Regulatory Th2 cells against one epitope can down regu-late the Th1 response to the whole cognate antigen throughsecretion of cytokine, such as IL-4 and IL-10, that act onall Th1 cells in topographic proximity, irrespective of their

Ž . Žspecificity antigen-driven bystander suppression Weiner.et al., 1994 . Furthermore, Th2 determinant spreading may

Ž .occur Tian et al., 1997 , that also explains how sensitiza-tion of Th2 cells to an individual epitope protects from anautoimmune disease that involves the whole cognate au-toantigen.

A recent study has also investigated EMG in IL-4 KOŽ .mice Balasa et al., 1998 . In agreement with our results,

that study also concluded that IL-4 was nor required forthe induction and progression of EMG. However, at differ-ence with the data we report here, that study did not find adifferent susceptibility to EMG of IL-4 KO and wild typeB6 mice, and concluded that IL-4 does not influence thesusceptibility to EMG. EMG is difficult to assess in mice

Žbecause the symptoms are frequently subclinical reviewed.in Conti-Fine et al., 1997 , and their presence and quantifi-

cation requires sensitized, parametric tests of musclestrength, like the one we used here. The study of Balasa et

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8482

Ž .al. 1998 may have missed the higher susceptibility toEMG of IL-4 KO mice because the test of muscle strengthused was not sensitive enough, and it was qualitative.

The main pathogenic mechanisms of anti-AChR anti-bodies in MG and EMG are activation of complement andaccelerated degradation of AChR molecules cross linked

Ž .by the antibodies reviewed in Conti-Fine et al., 1997 .Complement components are consistently present at the

Žneuromuscular junction in both MG and EMG Lennon et.al., 1978; Nakano and Engel, 1993 . The present results

underline the importance of complement in the pathogene-sis of EMG, since Th1-dependent antibodies are mosteffective at binding complement. The present results raisethe possibility that one of the protective mechanisms ofIL-4 and of anti-TAChR Th2 cells is stimulation of synthe-sis of anti-TAChR IgG that do not activate complement,and may compete for AChR binding with the Th1-induced,complement activating antibodies. The finding that IL-4KO and B6 mice develop similar levels of anti-AChRantibodies, but in B6 mice the anti-TAChR IgG include a

Ž .substantial fraction of IgG1 Fig. 2 , is consistent with thispossibility.

The present results do not exclude the possibility thatTh2 cells may have also a pathogenic role, at least incertain circumstances. Anti-AChR antibodies may causeaccelerated degradation of the AChR without the need of

Ž .complement reviewed in Drachman, 1994 . Th2-inducedanti-AChR antibodies may be pathogenic in muscle groupsthat are especially susceptible to myasthenic weaknesswhen the concentration of AChR at the synapses is re-duced, even in the absence of complement-mediated dam-age of the neuromuscular junction. The functional proper-

Ž .ties of extrinsic ocular muscles EOM make them espe-cially susceptible to weakness when a reduction of the

ŽAChR function occurs Kaminski et al., 1991; Kaminski.and Ruff, 1997 . Thus, Th2 cells might be pathogenic in

ocular MG, given the extreme susceptibility of the EOM todevelop myasthenic weakness when the concentration ofAChR at the synapses is reduced. Further studies will benecessary to exclude a pathogenic role of Th2 cells in thepathogenesis of EMG, and especially MG.

Acknowledgements

This work was partially supported by the NINDS grantŽ .NS 23919 to B.M.C.-F. .

References

Abbas, A.K., Murphy, K.M., Sher, A., 1996. Functional diversity ofhelper T lymphocytes. Nature 383, 787–793.

Anderson, T.T., Cornelius, J.G., Jarpe, A.J., Winter, W.E., Peck, A.B.,1993. Insulin-dependent diabetes in the NOD mouse model: II. Betacell destruction in autoimmune diabetes is a Th2 and not a Th1mediated event. Autoimmunity 15, 113–122.

Asthana, D., Fujii, Y., Huston, G.E., Lindstrom, J., 1993. Regulation of

antibody production by helper T cell clones in experimental autoim-mune myasthenia gravis is mediated by IL-4 and antigen-specific Tcell factors. Clin. Immunol. Immunopathol. 67, 240–248.

Balasa, B., Deng, C., Lee, J., Bradleyu, L.M., Dalton, D.K., Christadoss,Ž .P., Sarvetnick, N., 1997. Interferon g IFN-g is necessary for the

genesis of acetylcholine receptor-induced clinical experimental au-toimmune myasthenia gravis in mice. J. Exp. Med. 186, 385–391.

Balasa, B., Deng, C., Lee, J., Christadoss, P., Sarvetnick, N., 1998. TheTh2 cytokine IL-4 is not required for progression of antibody-depen-dent autoimmune myasthenia gravis. J. Immunol. 161, 2856–2862.

Bellone, M., Ostlie, N., Lei, S., Conti-Tronconi, B.M., 1991a. Experi-mental myasthenia gravis in congenic mice. Sequence mapping andH-2 restriction of T helper epitopes on the a subunits of Torpedocalifornica and murine acetylcholine receptors. Eur. J. Immunol. 21,2303–2310.

Bellone, M., Ostlie, N., Lei, S., Wu, X.-D., Conti-Tronconi, B.M., 1991b.The I-Abm12 mutation, which confers resistance to ExperimentalMyasthenia Gravis, drastically affects the epitope repertoire of murineCD4q cells sensitized to nicotinic acetylcholine receptor. J. Immunol.146, 2253–2261.

Bellone, M., Ostlie, N., Karachunski, P., Conti-Tronconi, B.M., 1993.Cryptic epitopes on nicotinic acetylcholine receptor are recognized byautoreactive CD4q cells. J. Immunol. 151, 1025–1038.

Chen, Y.J., Inobe, R., Marks, P., Gonnella, P., Kuchroo, V., Weiner, H.,1995. Peripheral deletion of antigen-reactive T cells in oral tolerance.Nature 376, 177–180.

Chen, Y.J., Inobe, R., Kuchroo, V., Baron, J., Janeway, C.J., Weiner, H.,1996. Oral tolerance in myelin basic protein T-cell receptor transgenicmice: suppression of autoimmune encephalomyelitis and dose-depen-dent induction of regulatory cells. Proc. Natl. Acad. Sci. U.S.A. 93,388–391.

Cong-Qui, C., Londei, M., 1996. Induction of Th2 cytokines and controlof collagen-induced arthritis by nondepleting anti-CD4q antibodies.J. Immunol. 157, 2685–2689.

Conti-Fine, B.M., McLane, K.E., Lei, S., 1996. Antibodies as tools tostudy the structure of membrane proteins. The case of the nicotinicacetylcholine receptor. Ann. Rev. Biophys. Biomol. Struct. 25, 197–229.

Conti-Fine, B.M., Bellone, M., Howard, J.F. Jr., Procti, M.P., 1997.Myasthenia Gravis. The Immunobiology of an Autoimmune Disease.Landes RG., Chapman & Hall, Springer, p. 230.

Cua, D.J., Hinton, D.R., Stohlman, S.A., 1995. Self-antigen-induced Th2Ž .responses in experimental allergic encephalomyelitis EAE -resistant

mice. Th2-mediated suppression of autoimmune disease. J. Immunol.155, 4052–4059.

Das, M.P., Nicholson, L.B., Greer, J.M., 1997. Autopathogenic T helperŽ .cell type 1 Th1 and protective Th2 clones differ in their recognition

of the autoantigenic peptide of myelin proteolipid protein. J. Exp.Med. 186, 867–876.

Drachman, D.B., 1994. Myasthenia gravis. New Engl. J. Med. 330,1797–1810.

Erb, K.J., Ruger, B., von Brevern, M., Ryffel, B., Schimpl, A., Rivett, K.,Ž .1997. Constitutive expression of Interleukin IL-4 in vivo causes

autoimmune-type disorders in mice. J. Exp. Med. 185, 329–339.Ž .Falcone, M., Bloom, B.R., 1997. A T helper cell 2 Th2 immune

response against non-self antigens modifies the cytokine profile ofautoimmune T cells and protects against experimental allergic en-cephalomyelitis. J. Exp. Med. 185, 901–907.

Ferber, I.A., Brocke, S., Taylor-Edwards, C., Ridgway, W., Dinisco, C.,Steinman, L., Dalton, D., Fathman, C.G., 1996. Mice with a disruptedIFN-g gene are susceptible to the induction of experimental autoim-

Ž .mune encephalomyelitis EAE . J. Immunol. 156, 5–7.Fresno, M., Kopf, M., Rivas, L., 1997. Cytokines and infectious diseases.

Immunol. Today 18, 56–58.Fuss, I.J., Strober, W.J., Dale, K., Fritz, S., Pearlstein, G.R., Puck, J.M.,

Lenardo, M.J., Strauss, S.E., 1997. Characteristic T helper 2 T cellcytokine abnormalities in autoimmune lymphoproliferative syndrome,

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–84 83

a syndrome marked by defective apoptosis and humoral autoimmu-nity. J. Immunol. 158, 1912–1918.

Gu, D., Wogensen, L., Calcutt, N.A., Xia, C., Zhu, S., Merlie, J.P., Fox,H.S., Lindstrom, J., Powell, H.C., Sarvetnick, N., 1995. Myastheniagravis-like syndrome induced expression of interferon-g in neuromus-cular junction. J. Exp. Med. 181, 547–557.

Hafler, D.A., Kent, S.C., Pietrusewicz, M.J., Khoury, S.J., Weiner, H.L.,Fukaura, H., 1997. Oral administration of myelin induces antigen-specific TGF-beta 1 secreting T cells in patients with multiple sclero-sis. Ann. New York Acad. Sci. 835, 120–131.

Hass, C., Ryffel, B., Le Hir, M., 1997. IFN-g is essential for thedevelopment of autoimmune glomerulonephritis in MRLr lpr mice. J.Immunol. 158, 5484–5491.

Heinrikson, R.L., Meredith, S.C., 1984. Amino acid analysis by reverse-phase high-performance liquid chromatography: precolumn derivati-zation with phenylisothiocyanate. Anal. Biochem. 136, 65–74.

Houghten, R., 1985. General method for the rapid solid phase synthesisof large numbers of peptides: specificity of antigen–antibody interac-tion at the level of individual amino acids. Proc. Natl. Acad. Sci.U.S.A. 82, 5131–5135.

Hultgren, B., Huang, X., Dybdal, N., Stewart, T.A., 1996. Geneticabsence of gamma interferon delays but does not prevent diabetes inNOD mice. Diabetes 45, 812–817.

Infante, A.J., Thompson, P.A., Krolick, K.A., Wall, K.A., 1991. Determi-nant selection in murine experimental autoimmune myasthenia gravis:effect of the bm12 mutation on T cell recognition of acetylcholinereceptor epitopes. J. Immunol. 146, 2977–2982.

Julia, V., Rassoulzadegan, M., Glaichenhaus, N., 1996. Resistance toLeishmania major induced by tolerance to a single antigen. Science274, 421–423.

Kaminski, H.J., Ruff, R.L., 1997. Ocular muscle involvement in myasthe-nia gravis. Ann. Neurol. 41, 419–420.

Kaminski, H.J., Maas, E., Spiegel, P., Ruff, R.L., 1991. Why are eyemuscles frequently involved in myasthenia gravis? Neurology 40,1663–1669.

Karachunski, P.I., Ostlie, N., Bellone, M., Infante, A.J., Conti-Fine,B.M., 1995. Mechanisms by which the I-Abm12 mutation influencessusceptibility to Experimental Myasthenia Gravis. A study in ho-mozygous and heterozygous mice. Scand. J. Immunol. 42, 215–225.

Karachunski, P.I., Ostlie, N., Okita, D., Conti-Fine, B.M., 1997. Protec-tion from experimental myasthenia gravis in C57Blr6 mice by sniff-ing of synthetic CD4q T cells epitopes. J. Clin. Invest. 100, 3027–3035.

Karachunski, P.I., Ostlie, N., Okita, D., Garman, R., Conti-Fine, B.M.,1999. Subcutaneous administration of T epitope sequences of theacetylcholine receptor prevents experimental myasthenia gravis. J.Neuroimmunol. 93, 108–121.

Kopf, M., Le Gros, G., Bachmann, M., Lamers, M.C., Bluethmann, H.,Kohler, G., 1993. Disruption of the murine IL-4 gene blocks Th2cytokine responses. Nature 362, 245–248.

Kuhn, R., Rajewsky, K., Muller, W., 1991. Generation and analysis of¨ ¨interleukin-4 deficient mice. Science 254, 707–710.

Laemmli, U., 1970. Cleavage of structural proteins during the assemblyof the head of bacteriophage T4. Nature 227, 680.

Lafaille, J., Van de Keere, F., Hsu, A.L., Baron, J.L., Haas, W., Raine,C.S., Tonegawa, S., 1997. Myelin basic protein-specific T helper 2Ž .Th2 cells cause experimental autoimmune encephalomyelitis inimmunodeficient hosts rather than protect them from the disease. J.Exp. Med. 186, 307–312.

Lawrence, R.A., Allen, J.E., Gregory, W.F., Kopf, M., Maizels, R.M.,1995. Infection of IL-4-deficient mice with the parasitic nematodeBrugia malayi demonstrates that host resistance is not dependent on aT helper 2-dominated immune response. J. Immunol. 154, 5995–6001.

Lee, M.S., Mueller, R., Wicker, L.S., Peterson, L.B., Sarvetnick, N.,1996. IL-10 is necessary and sufficient for autoimmune diabetes inconjunction with NOD MHC homozygocity. J. Exp. Med. 183,2663–2668.

Le Gros, G., Ben-Sasson, S.Z., Seder, R., Finkelman, F.D., Paul, W.E.,Ž .1990. Generation of interleukin 4 IL-4 -producing cells in vivo and

in vitro: IL-2 and IL-4 are required for in vitro generation ofIL-4-producing cells. J. Exp. Med. 172, 921–929.

Lennon, V.A., Seybold, M.E., Lindstrom, J.M., Cochrane, C., Ulevitch,R., 1978. Role of complement in the pathogenesis of experimentalautoimmune myasthenia gravis. J. Exp. Med. 147, 973–983.

Liblau, R.S., Singer, S.M., McDevitt, H.O., 1995. Th1 and Th2 CD4q Tcells in the pathogenesis of organ-specific autoimmune disease. Im-munol. Today 1, 34–38.

Liblau, R., Steinman, L., Brocke, S., 1997. Experimental autoimmuneencephalomyelitis in IL-4-deficient mice. Int. Immunol. 9, 799–803.

Lowry, O., Rosebrough, N., Farr, A., Randall, R., 1981. Protein measure-ment with Folin phenol reagent. J. Biol. Chem. 193, 256.

Manoury-Schwartz, B., Chiocchia, G., Bessis, N., Abehsira-Amar, O.,Batteux, F., Muller, S., Huang, S., Boissier, M.C., Fournier, C., 1997.High susceptibility to collagen-induced arthritis in mice lacking IFN-greceptors. J. Immunol. 158, 5501–5506.

Mikszta, J.A., Kim, B.S., 1996. Conversion of low antibody respondersinto high responders by up-regulating the T cell response to aselective epitope. J. Immunol. 157, 2883–2890.

Miller, S.D., Karpus, W.J., 1994. The immunopathogenesis and regula-tion of T cell mediated demyelinating diseases. Immunol. Today 8,356–361.

Moiola, L., Karachunski, P., Protti, M.P., Howard, J.F., Conti-Tronconi,B.M., 1994a. Epitopes on the b subunit of human muscle acetyl-choline receptor recognized by CD4q cells of Myasthenia Gravispatients and healthy subjects. J. Clin. Invest. 93, 1020–1028.

Moiola, L., Protti, M.P., Howard, J.F., Conti-Tronconi, B.M., 1994b.Myasthenia Gravis: residues of the a and g subunits of muscleacetylcholine receptor involved in formation of immunodominantCD4q epitopes. J. Immunol. 152, 4686–4698.

Moiola, L., Galbiati, F., Martino, G., Amadio, S., Brambilla, E., Comi,G., Vincent, A., Grimaldi, L.M.E., Adorini, L., 1998. IL-12 isinvolved in the induction of experimental autoimmune myastheniagravis, an antibody-mediated diseases. Eur. J. Immunol. 28, 2487–2497.

Mueller, R., Krahl, T., Sarvetnick, N., 1996. Pancreatic expression ofinterleukin-4 abrogates insulitis and autoimmune diabetes in nonobese

Ž .diabetic NOD mice. J. Exp. Med. 184, 1093–1099.Nakajima, A., Hirose, S., Yagita, H., Okumura, K., 1997. Roles of IL-4

and IL-12 in the development of lupus in NZBrW F mice. J.1

Immunol. 158, 1466–1472.Nakano, S., Engel, A.G., 1993. Myasthenia gravis: quantitative immuno-

cytochemical analysis of inflammatory cells and detection of comple-ment membrane attack complex at the end-plate in 30 patients.Neurology 43, 1167–1172.

O’Garra, A., 1998. Cytokine induce the development of functionallyheterogenous T helper cell subsets. Immunity 8, 275–283.

O’Garra, A., Steinman, L., Gijbes, K., 1997. CD4q T-cell subsets inautoimmunity. Curr. Opin. Immunol. 9, 872–883.

Ž .Pakala, S.V., Kurrer, M.O., Katz, J.D., 1997. T helper 2 Th2 T cellsinduce acute pancreatitis and diabetes in immune-compromised

Ž .nonobeses diabetic NOD mice. J. Exp. Med. 186, 299–306.Peng, S.L., Mosiehi, J., Craft, J., 1997. Roles of interferon-g and

interleukin-4 in murine lupus. J. Clin. Invest. 99, 1936–1946.Prabhu Das, M., Nicholson, L.B., Greer, J.M., Kuchroo, V.K., 1997.

Ž .Autopathogenic T helper cell type 1 Th1 and protective Th2 clonesdiffer in their recognition of the autoantigenic peptide of myelinproteolipid protein. J. Exp. Med. 186, 867–876.

Racke, M.K., Bonomo, A., Scott, D.E., Cannella, B., Levine, B., Raine,C.S., Shevach, E.M., Rocken, M., 1994. Cytokine-induced immunedeviation as a therapy for inflammatory autoimmune disease. J. Exp.Med. 180, 1961–1966.

Romagnani, S., 1997. The Th1rTh2 paradigm. Immunol. Today 18,263–266.

Shaw, M.K., Lorens, J.B., Dhawan, A., Dal Canto, M., Tse, A.B., Tran,

( )P.I. Karachunski et al.rJournal of Neuroimmunology 95 1999 73–8484

C., Bonpane, S.L., Eswaran, S., Brocke, N., Sarvetnick, N., Steinman,L., Nolan, G.P., Fathman, C.G., 1997. Local delivery of interleukin 4by retrovirus-transduced T lymphocytes ameliorates experimental au-toimmune encephalomyelitis. J. Exp. Med. 185, 1711–1714.

Swain, S.L., Weinberg, A.D., English, M., Huston, G., 1990. IL-4 directsthe development of different subsets of helper T cells. J. Immunol.145, 3796–3806.

Tian, J., Atkinson, M.A., Clare-Salzler, M., Herschenfeld, A., Forsthuber,T., Lehmann, P.V., Kaufman, D.L., 1996. Nasal administration of

Ž .glutamate decarboxylase GAD65 peptides induces Th2 responsesand prevents murine insulin-dependent diabetes. J. Exp. Med. 183,1561–1567.

Tian, J., Lehmann, P.V., Kaufman, D., 1997. Determinant spreading of TŽ .helper cell 2 Th2 responses to pancreatic islet autoantigens. J. Exp.

Med. 186, 2039–2043.Vermeire, K., Heremans, H., Vandeputte, M., Huang, S., Billiau, A.,

Matthys, P., 1997. Accelerated collagen-induced arthritis in IFN-greceptor-deficient mice. J. Immunol. 158, 5507–5513.

von Herrath, M.G., Oldstone, M., 1997. Interferon-g is essential fordestruction of b cells and development of insulin-dependent diabetesmellitus. J. Exp. Med. 185, 531–539.

Wall, K.A., Hu, J.-Y., Currier, P., Southwood, S., Sette, A., Infante, A.J.,1994. A disease-related epitope of Torpedo acetylcholine receptor.

Residues involved in I-Ab binding, self–nonself discrimination, andTCR antagonism. J. Immunol. 152, 4526–4646.

Wang, Z.-Y., Okita, D., Howard, J.F. Jr., Conti-Fine, B.M., 1997. Th1epitope repertoire on the a subunit of human muscle acetylcholinereceptor in myasthenia gravis. Neurology 48, 1643–1653.

Wang, Z.-Y., Okita, D., Howard, J.F. Jr., Conti-Fine, B.M., 1998. T-cellrecognition of muscle acetylcholine receptor subunits in generalizedand ocular myasthenia gravis. Neurology 50, 1045–1054.

Weigle, W.O., Romball, C.G., 1997. CD4q T-cell subsets and cytokinesinvolved in peripheral tolerance. Immunol. Today 18, 533–538.

Weiner, H.L., Friedman, A., Miller, A., Khoury, S.J., Al-Sabbagh, A.,Santos, L., Sayegh, M., Nussenblatt, R.B., Trenthan, D.E., Hafler,A.D., 1994. Oral tolerance: immunologic mechanisms and treatmentof animal and human organ-specific autoimmune diseases by oraladministration of autoantigens. Ann. Rev. Immunol. 12, 809–837.

Willenborg, D.O., Fordham, S., Bernard, C.A., Cowden, W.B., Ramshaw,I.A., 1996. IFN-g plays a critical down-regulatory role in the induc-tion and effector phase of myelin oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis. J. Immunol. 157, 3223–3227.

Wu, B., Deng, C., Goluszko, E., Christadoss, P., 1997. Tolerance to adominant T cell epitope in the acetylcholine receptor molecule in-duces epitope spread and suppresses murine myasthenia gravis. J.Immunol. 159, 3016–3023.