International Immunology, Vol. 8, No. 12, pp. 1857-1865 © 1996 Oxford University Press

Functional consequences of the binding ofMHC class ll-derived peptides to MHCclass IIMaryam Feili-Hariri1-3, Henry Kao2, Timothy A. Mietzner2 and Penelope A. Morel1"3

Departments of 1Medicine, and 2Molecular Genetics and Biochemistry, University of Pittsburgh, and3University of Pittsburgh Cancer Institute, Pittsburgh, PA 15213, USA

Keywords: MHC class II, non-obese diabetic mouse, peptide antagonists

Abstract

Three MHC class ll-derived synthetic peptides (l->»p»71-16, Mp9752-77 and l-4a9763-82YC) were

analyzed for their ability to bind to syngeneic and allogeneic MHC class II molecules using a wholecell, competitive peptide binding assay. These studies demonstrated that the 4pfl71-16 peptide wasable to specifically bind to syngeneic as well as to four allogeneic MHC class II molecules. The4a

9763-82YC peptide bound to self MHC class II molecules with a lower relative affinity and wasable to bind to three out of the four allogeneic cells tested. The binding of the three l-A^7-derivedpeptides to the self MHC class II was functionally significant. The 4p

971-16 and Af,g752-77 peptidesinhibited the proliferation of a heat shock protein 60 peptide-specific Th1 clone by MHC blockade.Interestingly, the 4a

9763-82YC peptide appeared to interact directly with T cells as pretreatment ofthe Th1 clone with this peptide resulted in inhibition of antigen-induced proliferation. Thisphenomenon was analyzed in more detail and it was found that this peptide could behave as apartial agonist. Incubation of T cells with the y*a

ff763-82YC peptide resulted in up-regulation ofIL-2R a chain expression and induction of IFN-y secretion. In addition T cells pretreated with thispeptide were rendered hyporesponsive to further antigenic stimulation. Thus, a peptide derivedfrom MHC class II may be used in an immunoregulatory capacity.

Introduction

It has been known for some time that peptides derived frompolymorphic MHC class II regions are capable of binding toboth syngeneic and allogeneic MHC class II molecules(1-3). T cell responses to donor-derived MHC class II peptidescan be detected during allograft rejection (4) and theseresponses are capable of inducing donor-specific cytotoxic Tlymphocytes in vivo (5). In a murine syngeneic setting it hasbeen shown for I-Ak that T cell responses to two self l-A-derivedpeptides could be induced following peptide immunization (1).These peptides, /4a

k1-18 and /\pk1-16, could inhibit the prolif-eration of an antigen-specific l-Ak-restricted T cell hybridoma,demonstrating the ability of these peptides to bind self I-Ak

molecules. These experiments, and others, have demonstratedthe importance of indirect recognition of allogeneic class IImolecules in the T cell response to an allograft.

An interesting feature of MHC-derived peptides is their abilityto exert an immunoregulatory role. In particular, it has beenshown that a human MHC class I peptide can induce allografttolerance in a rat cardiac allograft model (6). MHC class ll-

derived peptides have also been used to induce oral tolerancein the rat (7). The use of such peptides in transplantation modelshas demonstrated that the tolerance induction is specific andlong-lived (8).

Another situation for which such peptides might'be usefulis in autoimmunity where, in many cases, target organs aredestroyed by autoreactive T cells. Several therapeuticapproaches have been successfully tried including anti-CD4antibodies, anti-MHC class II antibodies and analogs of pep-tides known to induce disease (9-12). These approaches allhave potential disadvantages in either inducing widespreadimmune suppression or in requiring the identification of thedisease-causing peptide. Immunoregulatory peptides derivedfrom MHC class I or class II, such as those described byClayberger et al. (13), seem to have the property to induceanergy only in those T cells that are activated at the time ofadministration. Thus, it is possible that peptides derived fromMHC class II or class I could have a therapeutic role in auto-immune disease.

Correspondence tor. P. A. Morel, Pittsburgh Cancer Institute, Biomedical Science Tower W1057, 200 Lothrop Street, Pittsburgh, PA 15213, USA

Transmitting editor. E. Sercarz Received 5 September 1995, accepted 12 August 1996

1858 Interaction of MHC class ll-derived peptides with MHC class II

We have been studying the non-obese diabetic (NOD)murine model of human type I diabetes (14). This is character-ized by an early infiltration of CD4+ T cells at 3-4 weeks ofage, followed by the development of overt diabetes at -30weeks of age. NOD mice have a unique MHC class IImolecule, I-A9?, which is thought to play an important role insusceptibility to disease. In this report, studies were initiatedto determine whether self class II peptides, derived fromI-A97, would behave the same way in an autoimmune mouseversus a healthy mouse. We performed peptide binding andT cell inhibition experiments to determine whether threepeptides derived from I-A97 could specifically bind to syng-eneic MHC class II molecules. These peptides were alsoassessed for their ability to bind allogeneic class II molecules,and alanine analogs of one of the peptides were preparedand analyzed in order to determine the important contactresidues. Of potential interest is that one peptide derivedfrom l-/4as

7, /4a9763-82YC, appeared to function in an immuno-

regulatory role, as it was found to behave as a partial agonist.

Methods

Peptides

Synthetic peptides representing residues 437-460 of the heatshock protein 60 (hsp60) (VLGGGCALLRCIPALDSLTPANED),residues 1-16 of the \-A^7 (GDSERHFVHQFKGECY), res-idues 63-82YC of the \-Aa9

7 (GLQNIAAEKHNLGILTKRSVYC),residues 52-77 of the \-A^7 (ELGRHSAEYYNKQYLERTRA-ELDTAC) and Tap (HPGTVTALVGPNGSGKSTC), and sixsingle alanine-substituted analogs of Mp071-16 peptide atpositions 2, 4, 6, 9, 12 and 14 were prepared using standardmethods (15) by the Peptide Facility of the University ofPittsburgh Cancer Institute. All peptides were HPLC purifiedfollowing synthesis.

The N-terminus of the hsp60 peptide was biotinylated aspreviously described (16). Briefly, 5 mg of immuno-pure N-hydroxy succinimidyl biotin (NHS-LC-Biotin; Pierce, Rockford,IL) was added to a 5 mg/ml suspension of peptide in dimethylformamide (DMF) diluted 1:1 in PBS, pH 7.4, and incubatedwith stirring at room temperature for 2 h. Biotinylated peptidewas purified by HPLC, dried under vacuum and resuspendedat 1 mg/ml in PBS.

MiceFemale NOD/MrkTacfBR (NOD) mice were obtained fromTaconic (Germantown, NY) and were used at 6-10 weeks ofage. The following strains were obtained from the JacksonLaboratory (Bar Harbor, ME): 129/J (H-2b), SWR (1-1-29), SJL(H-2S) and A.CA/Sn (H-2f), and were used at 8-10 weeksof age.

Peptide binding assay

T cell-depleted NOD spleen cells at 2X105 cells per 50 (il ofDMEM (Gibco, Grand Island, NY), supplemented with 10%FBS, 2 mM L-glutamine, 50 lU/ml penicillin, 50 ng/ml strepto-mycin, 5X10"5 M 2-mercaptoethanol and 10 mM HEPES,were incubated for 1.5 h at 37°C with increasing concentra-tions (0.78-25 nM) of biotinylated hsp60 peptide. Followingincubation, the cells were washed in PBS containing 2% FBS,

0.02% NaN3 (FACS medium) and stained with FITC-avidin(Sigma, St Louis, MO) for 30 min at 4°C. After staining, cellswere washed twice, with the last wash in FACS mediumcontaining 1 (xg/ml propidium iodide (PI; Sigma) and resus-pended in 200 (il of FACS medium. Cells were analyzed forFITC and PI on a FACStar flow cytometer. Cells were excitedat 488 nm with a 3 W argon laser and emission spectrawere collected with 530/30 and 585/42 nm band pass filtersrespectively. A live gate was used in all FACS analyses.

To determine the specificity of binding, T cell-depletedNOD spleen cells were preincubated for 3 h at 37°C with100-fold molar excess of unlabeled hsp60 peptide, Tap orwith 10 ng/ml of the mAb 10.2.16 (anti-l-Ak) (9) or MKD6 (anti-I-Ad) (ATCC, Rockville, MD). At this point the biotinylatedpeptide was added to a final concentration of 12.5 ̂ M. Thecells were incubated for a further 1.5 h at 37°C, washed andincubated for 30 min at 4°C with FITC-avidin as describedabove. Similarly, competition with class ll-derived competitorpeptides (6.25- up to 100-fold molar excess) was accomp-lished as described above.

Generation of T cell clones

T cell clones specific for a hsp60-derived peptide (437-460)were generated as previously described (17). Briefly, a NODmouse was immunized at the base of the tail with 200 |ag ofhsp60 peptide in complete Freund's adjuvant (Difco Detroit,Ml) and a T cell line established. Clone C3.5 was generatedby limiting dilution of 3 cells/well and further subcloned at1 cell/well by placing them in 96-well plates in the presenceof 5xiO5/well irradiated NOD spleen cells, antigenic peptideand 20 U/ml IL-2. The specificity of clones and line weredetermined by proliferation assays carried out with eitherantigenic or irrelevant peptides. This clone was characterizedas Th1 type and was maintained by re-stimulation every 10-12 days with 1-2x106/ml irradiated (2500 rad) NOD spleencells and antigenic peptide (10 |xg/ml).

Inhibition of T cell proliferation by MHC class ll-derived peptide

C3.5 T cells (2xiO4/well) were placed in 96 well plates in thepresence of 2x105 irradiated antigen-presenting cells (APC)and 15 ng/ml hsp60 peptide in the presence or absence ofcompetitor peptides or mAb. The cells were incubated at37°C for 42 h and were pulsed with 0.5 (iCi/well [3H]thymidinefor the last 18 h. The cells were harvested onto glass fiberstrips and [3H]thymidine incorporation was counted in a betascintillation counter.

In the APC pretreatment assays, T cell-depleted and irradi-ated spleen cells were preincubated for 3 h at 37°C, inmedium alone or with competitor peptides at the indicatedconcentrations. Following this 15 ng/ml of hsp60 peptide wasadded. The cells were incubated further for 1.5 h, washedand placed (2x105/well) in wells with 2X104 T cells. Thewells were pulsed and harvested as described above.

In the T cell pretreatment assays, C3.5 cells were incubatedin medium or with the competitor peptides at the indicatedconcentrations for 3 h at 37°C. After washing, the cells wereincubated with irradiated APC and 15 ng/ml hsp60 peptide.The cells were then incubated and harvested as describedabove.

In the anergy induction assays C3.5 cells (5X105 cells/ml)

Interaction of MHC class ll-derived peptides with MHC class II 1859

were incubated with irradiated APC (4X106 cells/ml) in thepresence of the indicated peptides in a six-well plate (finalvolume of 5 ml). The cells were incubated for 48 h, afterwhich dead cells and APC were removed by centrifugationover Ficoll-Hypaque. These cells were rested for an additional24 h, counted and placed in a standard proliferation assaywith fresh APC and antigenic peptide as described above.In some experiments, cells were removed at 48 h for FACSanalysis of cell size and IL-2R a chain expression using therat mAb PC61.5 (18). Supematants were also collected after48 h in order to determine the level of IFN-y production byELISA assay as previously described (19).

Results

Specificity of binding of the biotinylated hsp60 peptide to I-A&7

A binding assay in which I-A9? expressing cells were incub-ated with biotinylated hsp60 peptide was used to establishMHC class ll-peptide interactions. The binding of the biotinyl-ated hsp60 peptide to T cell-depleted NOD spleen cells wasdose-dependent (Fig. 1a). The specificity of binding can beseen in Fig. 1(b) as the binding was nearly completely inhib-ited by a 100-fold molar excess of unlabeled hsp60 peptideor the mAb 10.2.16 which recognizes the I-A97 molecule. Anon-specific mAb, MKD6 (anti-l-Ad) and the Tap peptide wereunable to block the peptide binding.

Competition of binding of the biotinylated hsp60 by class ll-derived synthetic peptides

This assay was used to determine whether self peptidesderived from I-A^7 could inhibit the binding of the biotinylatedhsp60 peptide. In these experiments, cells were preincubatedwith increasing concentrations of the unlabeled hsp60 peptideor l-A-derived peptides and analyzed by FACS. The datademonstrate that hsp60 and /4p071-16 peptides were able toblock the binding of labeled peptide in a dose-dependentfashion (Fig. 2). /4a

9763-82YC peptide inhibited the biotinyl-ated hsp60 peptide binding at higher concentrations. Theamount of peptide required to achieve a 50% decrease ofhsp60 peptide binding to I-A97 can be seen in Fig. 2. Fromthese data it would appear that hsp60 peptide has thehighest affinity for I-A97, whereas /y?71-16 and ^ 7 6 3 - 8 2 Y Cpeptides were less able to compete for hsp60 peptide binding.Similar experiments were performed with the A$9752-17 pep-tide. We observed a 40% inhibition in binding using a 10-foldmolar excess of A$g752-77 peptide (data not shown). Theresults indicate that this peptide can also bind to the I-Ag7

molecule. The binding of one of the peptides (/4ps71—16) wasanalyzed using a set of six alanine analog peptides and itwas found that positions 2 and 4 were essential for peptidebinding as well as for inhibition of T cell proliferation (datanot shown).

We have pursued these studies in different strains ofmice to determine whether this was a specific or generalphenomenon. From Fig. 3 it can be seen that the hsp60peptide showed the highest level of binding to I-A97 (NOD)and I-Aq (SWR) spleen cells. This binding is inhibited byunlabeled hsp60 peptide in all cases. The ,4p971-16 peptideis able to compete for binding in all the strains, but the

oaEca

10000 -i

1000-

100-

100.1 1 10

[hsp65-bio] (iM

100

1000 -i

800-

600-

400-

200-

Fig. 1. Direct binding of the biotinylated hsp60 peptide 437-460 toT cell-depleted NOD spleen cells, (a) Cells were incubated withincreasing concentrations of (0.78-25 nM) labeled peptide for 1.5 hat 37°C; (b) cells were first preincubated for 3 h at 37°C with 100-fold molar excess of unlabeled peptides or mAb 10.2.16 (anti-l-Ak),MKD6 (anti-l-Ad), then with 12.5 \iM biotinylated hsp60 peptide for1.5 h at 37°C. Cells were washed and stained for 30 min at 4°C withFITC-avidin, after washing they were analyzed by flow cytometry.Results are representatives of three to eight experiments and areexpressed as the mean log fluorescence observed with FITC-avidin.Background FITC-avidin staining was assessed using cells incubatedwithout peptide.

Aa9763-82YC peptide is unable to inhibit the binding to I-As.

These data demonstrate that both the hsp60 and the A^7\-16 peptides are promiscuous in their binding to different MHCclass II molecules.

Inhibition of T cell proliferation by MHC class ll-derivedpeptides

Experiments were performed to determine whether the MHCclass ll-derived peptides were able to inhibit antigen-induced

1860 Interaction of MHC class ll-derived peptides with MHC class II

lOO-i

80-

60-

40-

20

A

500 1000

[Peptide] HM

- D — hsp65 (437-460)

1500

—O—

-A—

P (1-16)

a (63-82YC)

Fig. 2. Inhibition of binding of the biotinylated hsp60 peptide AA 437-460 by unlabeled hsp60 peptide, MHC class ll-derivedI-Ap9

71-16 and l-/y763-82YC peptides to T cell-depleted NODspleen cells. Cells were pretreated with the indicated concentrationsof competitor peptides for 3 h at 37°C and incubated further with12.5 |iM labeled hsp60 peptide for 1.5 h at 37°C. Cells were washedand stained with FITC-avidin. Data shown are the mean of fourdifferent experiments which are expressed as percent inhibition ofbinding of labeled peptide by competitor peptides. This is calculatedby: [(mean fluorescence with competitor peptide - background/meanfluorescence without competitor peptide - backgroundxiOO) - 100].

proliferation of the C3.5 clone. The results (Fig. 4a) indicatedthat all three peptides were able to inhibit the proliferation ofthis clone when the MHC class ll-derived peptides wereadded simultaneously with the antigenic peptide, whereas anirrelevant peptide (Tap) was unable to do so. Other controlsincluded the inhibition by an anti I-Ak mAb (10.2.16) and thelack of inhibition by an anti I-Ad mAb (MKD6) (Fig. 4a). Inaddition, the preincubation of the irradiated APC (Fig. 4b) withthe l-A97-derived peptides resulted in reduced proliferation ofthe hsp60-specific T cell line suggesting that these peptidesinhibit T cell proliferation by MHC blockade.

The next question we asked was whether any of thesepeptides could interact directly with the C3.5 clone. In theseassays T cells were preincubated, in the absence of anyAPC, with the MHC class ll-derived peptides for 3 h, washedand then placed in a standard proliferation assay. Thesestudies revealed that only the /4a9

763-82YC peptide was ableto inhibit the antigen-induced proliferation of T cells (Fig. 5a).Repeated experiments with the other two peptides failed todemonstrate any direct interaction with the T cells, furthersupporting the suggestion that the inhibition observed previ-ously was due to competition at the level of MHC peptidebinding. Figure 5(b) demonstrates that the effects of A^^Q^82YC peptide were most pronounced at low antigen concen-trations.

Aa9763-82YC is a partial agonist

In order to further characterize the effect of the /Aas763-82YC

on the T cell clones experiments were carried out to determinewhether this peptide could function as a partial agonist of theC3.5 clone. In these assays C3.5 cells were incubated withthe hsp60 peptide (15 ng/ml), A*9763-82YC (450 ng/ml) and/ y 7 1 - 1 6 (450 ng/ml) peptides in the presence of APC. As apositive control for anergy induction the cells were alsoincubated on plates that had been coated with anti-TCR mAb(10 n.g/ml). Cells were analyzed after 48 h for an increase incell size and up-regulation of IL-2R a chain expression. Cellsthat had been incubated with hsp60, /\a0

763-82YC or anti-TCR mAb were found to have increased in size and to haveup-regulated expression of the IL-2R a chain (Fig. 6). Thiswas not observed in cells incubated with the /Apg71-16 peptide,which were indistinguishable from cells that had been incub-ated with medium alone. In addition supernatants collectedat 48 h were found to contain significant amounts of IFN-yfollowing incubation with hsp60, /\a

9763-82YC peptides oranti-TCR coated plates but not with the /4p

971-16 peptide(Fig. 7). In view of the fact that the /V763-82YC peptide wasunable to stimulate proliferation of the C3.5 clone (data notshown) these data suggested that the /4a9

763-82YC peptidecould act as a partial agonist for the C3.5 clone.

A feature of some partial agonist peptides has been theirability to induce T cell anergy (20, 21) and thus experimentswere performed in order to determine whether the Aa

9763-82YC peptide could induce anergy in the C3.5 clone. C3.5cells were incubated with the /Aa9

763-82YC peptide in thepresence of APC for 48 h, removed from the APC and thenrested for a further 24 h, after which time they were assayedfor their ability to respond to fresh APC and hsp60 peptide.Cells treated in this way were found to be profoundly hypores-ponsive to the second stimulation (Fig. 8). Cells incubatedon anti-TCR-coated plated were also unable to respond tothe second antigenic stimulus, whereas cells incubated withthe /4p971-16 peptide were responsive in the subsequentantigenic stimulation. These data suggest therefore that the/4as

763-82YC peptide does act as a partial agonist and caninduce a state of hyporesponsiveness in a hsp60-specific Tcell clone.

Discussion

In this study, we synthesized three l-A97-derived peptides,two from the AQ37 chain (1-16 and 52-77), and one from theAa9

7 chain (63-82YC), and we examined the ability of thesepeptides to compete with a biotinylated peptide derived fromhsp60 (437-460) for binding to MHC class II molecules.These studies clearly demonstrated that all three l-A97-derivedpeptides bind self I-A molecules. This binding is functionallysignificant as these peptides were able to inhibit the prolifera-tion of an antigen-specific T cell clone. For two of the peptides(Ap971-16 and A$7b2-ll) T cell inhibition appeared to bemediated via MHC blockade as preincubation of APC withthese peptides resulted in the inhibition of the T cell prolifera-tion. In contrast, the Aa3763-82YC peptide appeared tointeract with T cell as pretreatment of the Th1 clone with thispeptide resulted in reduced proliferative responses to the

Interaction of MHC class ll-derived peptides with MHC class II 1861

1200-

1000-

800-

NOD1200-

1000-

800-

600-

400-

200-

129/J

1 1 H1

1

C

1200-

1000-

800-

600-

400-

200-

SWR

•

il J 1

A.CA/Sn

3

1200-

1000-

800-

600-

400-

200-

SJL

1 1 s

D

G3

m

•

PBS only

Bio-hsp65 (437-460)

+ hsp65 (437-460)

+ 0(1-16)

+ a(63-82YC)

S £ 8 2 8 2 8 R2 + + - -

Fig. 3. Binding of 12.5 nM biotinylated hsp60 peptide 437-460 to T cell-depleted l-E-negative MHC class II spleen cells. Data are presentedas the mean log fluorescence binding of labeled peptide to different I-A molecules preincubating with/without the indicated competitorpeptides as described in Fig. 2. Data shown are representative of three to five experiments carried out with each strain. NOD (H-297), 129/J(H-2b), SJL (H-2S), SWR (H-29) a n d A.CA/Sn (H-2f).

antigenic peptide. This effect was most pronounced at lowerantigen concentrations. This inhibition was not due to thephenomenon of TCR antagonism, as described by De Mag-istris et al. (22), since no effect on the T cell proliferativeresponse was observed when /4a0

763-82YC peptide wasadded to APC preincubated with the antigenic peptide (datanot shown). Rather this peptide appears to act as a partialagonist as demonstrated by an increase in cell size, up-regulation of IL-2R a chain expression and IFN-y secretion.In addition, the -4a

9763-82YC peptide was able to induce a

state of hyporesponsiveness in the hsp60 (437-460)-specificT cell clone.

The peptides /\p»71-16 and /y?763-82YC were also foundto bind several different I-A alleles. The mouse strains chosenfor these assays were all l-E-negative in order to ensurethat the binding observed was only involving I-A. Our datademonstrate that the binding of /\p971-16 is not specific forself MHC as this peptide binds to all five MHC class IImolecules tested. The /V763-82YC peptide bound to selfMHC class II molecule and was able to bind to three of the

1862 Interaction of MHC class ll-derived peptides with MHC class II

AlOOOO-i

T

20000

15000-

10000

5000

0

n

T

T

1

T

x x x x x x x x x x x x *

- * 9- - "> 2

Fig. 4. Inhibition of hsp60 peptide-specific T cell proliferation byMHC class ll-derived I-A97 peptides. (a) C3.5 T cell clone (2x107well) was incubated in the presence of 2xiO5/well irradiated APCand 15 ng/ml of hsp60 peptide 437-460 with/without the indicatedconcentrations of class ll-derived peptides or mAb for 24 h, pulsedwith [3H]thymidine and harvested after an additional 18 h ofincubation, (b) Irradiated APC (4xiO6/ml) were first preincubatedwith/without the indicated concentrations of class ll-derived peptidesfor 3 h at 37°C and were further incubated with 15 ng/ml of hsp60peptide for 1.5 h at 37°C. Cells were washed and 2xiO5/well wereincubated with 2xiO4/well T cell line for 24 h and pulsed asdescribed above. Data shown are the mean (±SD) of triplicate wellsrepresentative of three to five different experiments.

four other I-A molecules tested. The promiscuous binding ofthese two I-A derived peptides is of relevance in the contextof allorecognition. These data suggest that many MHC classll-derived peptides could bind to the recipient MHC class IIto induce indirect allorecognition. This phenomenon would

40000-1

CL 30000-

at

20000-

10000-

•••o—Med

a (63-82YC)

0-9"o.i 10 100

Fig. 5. Inhibition of proliferation of the C3.5 T cell clone pretreatedwith MHC class ll-derived I-A9? peptides. (a) T cells (4xiO5/ml) werepreincubated with/without the indicated concentrations of class ll-derived peptides for 3 h at 37°C. After washing, 2xio4/well peptide-pretreated T cells were placed in a proliferation assay with irradiatedAPC and hsp60 peptide (15 ng/ml). Cells were incubated, pulsedand harvested as described in Fig. 4. (b) T cells (4xiO5/ml werepreincubated with 375 ng/ml of class ll-derived /4a9763-82YC peptideor in medium for 3 h. After washing T cells were placed in the wellsas described above with indicated antigen concentrations. Resultsare the mean (±SD) of triplicate wells representative of at least threedifferent experiments.

depend on the ability of these peptides to be generated bythe natural processing of MHC class II molecules whichremains to be resolved.

The binding of /4p^71-16 to I-A97 appears to follow the samerules as those that govern the binding of antigenic peptidesto MHC. In data not shown in this report, alanine analogs of,4p971-16 were made and, using these analogs in bindingassays and T cell proliferation assays, we were able to identify

3

~~ 6O

Interaction of MHC class ll-derived peptides with MHC class II 1863

M A I -

"iboRelative Huorescence

[ Num

ber

Cel

!

tea

14O

120

100

so

GO

40

2O

r

A/

J |(%so ida 1S0 2OO 2SO

Relative Size (FSC)

Fig. 6. The /V763-82YC peptide induces an increase in size andup-regulation of IL-2R a chain expression. C3.5 cells (5xiO5/ml)were incubated with /V763-82YC peptide (thick dotted line), hsp60peptide (thin dotted line), l-/4p971-16 peptide (thin solid line) in thepresence of APC (4x iO6/ml) or on anti-TCR-coated plates (thick solidline) for 48 h. APC and dead cells were removed by centrifugationover Ficoll-Hypaque and the cells were stained with an antibodyspecific for the IL-2R a chain (PC61.5). After washing the cells wereincubated with a FITC-conjugated goat anti-rat IgG antibody, washedand analyzed by flow cytometry as described in Methods. The upperpanel represents the results obtained for the IL-2R a chain expressionand the lower panel is the forward angle scatter on the same cellsindicating an increase in size in those cells that were treated withhsp60 peptide, /4as

763-82YC or with anti-TCR-coated plates. Theseresults are representative of four similar experiments.

important contact residues. Analog peptides with substitutionsof Ala for Asp or Glu at positions 2 or 4 did not compete withthe biotinylated hsp60 peptide for the MHC binding, indicatingthat the presence of Asp and Glu in these two positions wereimportant for the binding of ,4p971-16 peptide to the selfmolecule. These data demonstrate the presence of importantanchor residues in the binding of -4p971-16 to I-A97. A peptidebinding motif for I-A97 has been proposed (23) based onpeptides eluted from I-A^7. Not all of the peptides eluted inthese studies contained this motif and neither do the hsp60(437-460) or A$g7'\-'\6 peptides. In common with these elutedpeptides, however, contact residues exist along the length ofthe /4p071-16 peptide. It has been also suggested that selfpeptides with high affinity of binding to self MHC may have

cx-TCR hsp65 P (l-16)<x (63-82YC)

Fig. 7. The A*9763-82YC peptide induces secretion of IFN-y. C3.5cells were treated as described in the legend to Fig. 6 andsupernatants were collected at the end of the 48 h incubation. Thesupernatants were tested for the secretion of IFN-y using an ELISAassay. The results of three independent experiments are shownindicating that the /Act

9763-82YC peptide can induce IFN-y secretionto a level similar to that obtained with antigen (hsp60 peptide).

20000n

0- 15000-

o. 10000-

5000-

-D— hsp 65

"O~~ p(l-16) T

' ° ' "" a (63-82YC)

-A— anti-TCR / /

1 0 100

[hsp65 437-460] |ig/ml

Fig. 8. Anergy induction of C3.5 T cell clone by MHC class ll-derivedpeptides. T cells were incubated, in the presence of APC and theindicated peptides for 48 h at 37°C. The live cells were then purifiedby centrifugation over Ficoll and the cells were allowed to rest for afurther 24 h. The T cells were counted and placed ( 2 X 1 0 4 cells/well)in a proliferation assay in which fresh irradiated spleen cells (5x10°cells/well) were added in the presence of increasing concentrationsof antigen (hsp60 437-460 peptide). The cells were pulsed after24 h, and harvested and counted after an additional 18 h of culture.The results shown are representative of four similar experiments.

a role in modulation of immune response (24,25). In particular,Chicz et al. (24) have proposed that MHC class II moleculesmay contain many peptides derived from self molecules,including MHC itself, and that these peptides serve as compet-itors to ensure that only high-affinity foreign peptides bind.

It has recently been shown that the I-Ag7 molecule isintrinsically unstable and that it is a poor peptide binder (26).In the assays that we performed we did not detect any

1864 Interaction of MHC class ll-derived peptides with MHC class II

difference in the ability of I-A^7 to bind the various peptidesthat we tested, when compared with four other murine strainstested. This is likely to be due to the less stringent nature ofthe whole cell assay that we have performed as comparedto the immunoprecipitations performed by Carrasco-Marinet al. (26). The whole cell binding assay was specific as thepeptide derived from the Tap protein was unable to bind toI-A97. The assay described here may therefore prove a usefulway of identifying peptides that are capable of binding to I-A97.

The case of the /V763-82YC peptide is interesting as itappears that this peptide may be able to interact directly withT cells. Whether this peptide is actually interacting with TCRor another cell surface structure remains to be determined. Ithas been shown, however, that a peptide derived from the I-A$k chain (amino acids 68-83) was able to bind directly tothe TCR as determined using a direct binding assay (27). Itwas also shown that this peptide could inhibit T cell prolifera-tion and inhibit the binding of an anti-clonotype antibody tothe T cell clone (27). This peptide was derived from the pchain of the I-Ak molecule whereas the one that we havestudied is derived from the a chain. Both peptides are derivedfrom the a-helical region of the respective I-A chain. Anothercommon feature between these two peptides is that bothpeptides were engineered with an additional cysteine at oneend which may allow the peptides to dimerize and favor moreefficient binding to the TCR.

In these studies we have also been able to demonstratethat the /\a

9763-82YC peptide can act as a partial agonist tothe hsp60 (437-460)-specific T cell clone and that it caninduce hyporesponsiveness to subsequent antigenic stimula-tion. In these studies it was found that the presence of thecysteine did have an influence on the response as the Aa9

763-82 peptide was synthesized and it was found to cause anincrease in size and to up-regulate IL-2R a chain expressionbut it was unable to induce IFN-y secretion (data not shown).These data therefore suggest that the a chain sequence isresponsible for binding to the TCR. In addition, it is possiblethat the presence of the cysteine allows the peptide todimerize and thus cross-link more effectively.

These data are similar to those that have been describedfor altered peptide ligands in which antigenic peptides aresubstituted at individual positions and the resulting mutantsare unable to induce proliferation but can stimulate IL-4secretion and induce anergy (28,29). This is the first time thata peptide derived from MHC class II has been shown to actas a partial agonist. Nisco et al. (6) have recently shown thata MHC class l-derived peptide can be used to inducedallograft tolerance in a rat heart transplant model. This classl-derived peptide was previously shown to inhibit T cellactivation by direct interaction with the T cell (30). It is possiblethe /A(X£

f763-82YC peptide may have a similar effect and sucha peptide may prove useful in the therapy of autoimmunedisease, and these experiments are underway.

Peptides derived from MHC class II have been used ascompetitors and antagonist peptides in T cell proliferationassays. Rosloniec et al. (3) showed that the l -^-der ivedpeptides that competed for binding was also able to inhibitT cell proliferation. It has been reported in the collagen-induced arthritis model of human rheumatoid arthritis, whichcan be induced in susceptible H-2q strains, a peptide derived

from the \-Aaq chain can be used to inhibit the proliferation

of collagen-specific T cell clones (31). The peptide used inthese studies was derived from the same region as the onethat we report in this paper. The fact that such a peptide mayinduce hyporesponsiveness/anergy has important implica-tions for potential immunotherapy in autoimmunity. Thus in anautoimmune disease in which the precipitating antigen isunknown, it may be possible to design an inhibitory peptide(s)derived from the predisposing MHC class II molecules thatcould modulate disease initiation and progression.

Acknowledgements

We would like to thank Mr Joseph Campbell for expert flow cytometry.This work was funded by an NIH grant Al 25251.

AbbreviationsAPCDMFhsp60NODPI

antigen-presenting celldimethyl formamideheat shock protein 65non-obese diabeticpropidium iodide

References

1 Benichou, G., Takizawa, P. A., Ho, P. T, Killion, C. C, Olson, C.A., McMillan, M. and Sercarz, E. E. 1990. Immunogenicity andtolerogenicity of self-major histocompatibility complex peptides.J. Exp. Med. 172:1341.

2 Harris, P. E., Liu, Z. and Suciu-Foca, N. 1992. MHC class IIbinding of peptides derived from HLA-DR 1. J. Immunol. 148:2169.

3 Rosloniec, E. F, Vitez, L. J., Buus, S. and Freed, J. H. 1990. MHCclass ll-derived peptides can bind to class II molecules, includingself molecules, and prevent antigen presentation. J. Exp. Med.171:1419.

4 Benichou, G., Takizawa, P. A., Olson, C. A., McMillan, M. andSercarz, E. E. 1992. Donor major histocompatibility complex(MHC) peptides are presented by recipient MHC moleculesduring graft rejection. J. Exp. Med. 175:305.

5 Lee, R. S., Grusby, M. J., Glimcher, L H., Winn, H. andAuchincloss, H. 1994. Indirect recognition by helper cells caninduce donor-specific cytotoxic T lymphocytes in vivo. J. Exp.Med. 179:865.

6 Nisco, S., Vriens, P., Hoyt, G., Lyu, S-C, Farfan, F, Pouletty, P.,Krensky, A. M. and Clayberger, C. 1994. Induction of allografttolerance in rats by an HLA class l-derived peptide andcyclosporine A. J. Immunol. 152:3786.

7 Sayegh, M. H., Khoury, S. J., Hancock, W. W., Weiner, H. L. andCarpenter, C. B. 1992. Induction of immunity and oral tolerancewith polymorphic class II major histocompatibility complexallopeptides in the rat. Proc. Natl Acad. Sci. USA 89:7762.

8 Krensky, A. M. and Clayberger, C. 1994. The induction of toleranceto alloantigens using HLA-based synthetic peptides. Curr. Opin.Immunol. 6:791.

9 Boitard, C, Bendelac, A., Richard, M. F, Carnaud, C. and Bach,J. F. 1988. Prevention of diabetes in nonobese diabetic mice byanti I-A monoclonal antibodies: transfer of protection by splenicT cells. Proc. Natl Acad. Sci. USA 85:9719.

10 Shizuru, J. A., Taylor-Edwards, C, Banks, B. A., Gregory, A. K.and Fathman, C. G. 1988. Immunotherapy of the nonobesediabetic mouse: treatment with an antibody to T-helperlymphocytes. Science 240:659.

11 Wauben, M. H. M., Boog, C. J. P., Van der Zee, R., Joosten, I.,Schlief, A. and Van Eden, W. 1992. Disease inhibition by MHCbinding peptide analogues of disease associated epitopes: morethan blocking alone. J. Exp. Med. 176:667.

12 Lamont, A. G., Sette, A., Fujinami, R., Colon, S. M., Miles, C.and Grey, H. M. 1990. Inhibition of experimental autoimmune

Interaction of MHC class ll-derived peptides with MHC class II 1865

encephalomyelitis induction in SJL/J mice by using a peptidewith high affinity for lAs molecules. J. Immunol. 145:1687.

13 Clayberger, C, Lyu, S.-C DeKruyff, R., Parham, P. and Krensky,A. 1994. Peptides corresponding to the CD8 and CD4 bindingdomains of HLA molecules block T lymphocyte immune responsesin vitro. J. Immunol. 153:946.

14 Leiter, E. L, Prochazka, M. and Coleman, D. L. 1987. Animalmodel of human disease, the nonobese diabetic (NOD) mouse.Am. J. Pathol. 128:380.

15 Carpino, L. A. and Han, G. Y. 1972. 9-Fluorenylmethoxy-carbonylamino protecting group. J. Org. Chem. 37:3404.

16 Sinha, A. A., Mietzner, T. A., Lock, C. B. and McDevitt, H. O.1993. Fluorescence activated cell sorter analysis of peptide majorhistocompatibility complex. In Alt, F. W. and \fogel, H. J., eds,Molecular Mechanisms of Immunological Self Recognition, p.187. Academic Press, San Diego, CA.

17 Morel, P. A., Livingstone, A. M. and Fathman, C. G. 1987.Correlation of T cell receptor Vp gene family with MHC restriction.J. Exp. Med. 166:583.

18 Ceredig, R. J., Lowenthal, J. W., Nabholz, M. and MacDonald, H.R. 1985. Expression of interleukin-2 receptors as a differentiationmarker on intrathymic stem cells. Nature. 314:89.

19 Mosmann, T. R., Cherwinski, H., Bond, M. W., Giedlin, M. A. andCoffman, R. L. 1986. Two types of murine helper T cell clones. I.Definition according to profiles of lymphokine activities andsecreted proteins. J. Immunol. 136:2348.

20 Evavold, B. D., Sloan-Lancaster, J. and Allen, P. M. 1993. Ticklingthe TCR: Selective T-cell functions stimulated by altered peptideligands. Immunol. Today 14:602.

21 Sette, A., Alexander, J., Ruppert, J., Snoke, K., Franco, A., Ishioka,G. and Grey, H. M. 1994. Antigen analogs/MHC complexes asspecific T cell receptor antagonists. Annu. Rev. Immunol. 12:413.

22 De Magistris, M. T., Alexander, J., Coggeshall, M., Altman, A.,

Gaeta, F. C. A., Grey, H. M. and Sette, A. 1992. Antigen analog-major histocompatibility complexes act as antagonists of the Tcell receptor. Cell. 68:625.

23 Reich, E.-P, Grafenstein, H. V., Barlow, A., Swenson, K. E.,Williams, K. and Janeway, C. A. 1994. Self peptides isolated fromMHC glycoproteins of non-obese diabetic mice. J. Immunol.152:2279.

24 Chicz, R. M., Urban, R. G., Gorga, J. C, Vignali, D. A., Lane, W.S. and Strominger, J. L. 1993. Specificity and promiscuity amongnaturally processed peptides bound to HLA-DR alleles. J. Exp.Med. 178:27.

25 Rotzschke, O. and Falk, K. 1994. Origin, structure and motifs ofnaturally processed MHC class II ligands. Curr. Opin. Immunol.6:45.

26 Carrasco-Marin, E., Shimuzu, J., Kanagawa, O. and Unanue, E.R. 1996. The class II MHC I-A97 molecules from non-obesediabetic mice are poor peptide binders. J. Immunol. 156:450.

27 Williams, W. V., Weiner, D. B., Borofsky, M. A., Rubin, D. H.,Katsuyuki, Y. and Greene, M. I. 1992. Modulation of T cellresponses with MHC-derived peptides. Immunol. Res. 11:11.

28 Sloan-Lancaster, J., Evavold, B.D. and Allen, P.M. 1993. Inductionof T cell anergy by altered T cell-receptor ligand on live antigen-presenting cells. Nature 363:156.

29 Evavold, B. D. and Allen, P. M. 1991. Separation of IL-4 productionfrom Th cell proliferation by an altered T cell receptor ligand.Science 252:1308.

30 Clayberger, C , Parham, P., Rothbard, J., Ludwig, D. S., Schoolnik,G. K. and Krensky, A. M. 1987. HLA-A2 peptides can regulatecytolysis by human allogeneic T lymphocytes. Nature 330:763.

31 Banerjee, S. and David, C. S. 1991. Immunogenetics of Type IIcollagen-induced arthritis in mice. In Farid, N. R., eds, TheImmunogenetics of Autoimmune Diseases, p. 119. CRC Press,Boca Raton, FL.