of November 17, 2017.This information is current as

Class IIHLAMyelin Autoantigen Presentation by the

Inducible Heat Shock Protein 70 Promotes

Odyniec, Celia F. Brosnan and Krzysztof W. SelmajArturBozena Szymanska, Grzegorz Kudla, Lukasz Kilianek,

Marcin P. Mycko, Hanna Cwiklinska, Jacek Szymanski,

http://www.jimmunol.org/content/172/1/202doi: 10.4049/jimmunol.172.1.202

2004; 172:202-213; ;J Immunol 

        average*  

4 weeks from acceptance to publicationSpeedy Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

?The JIWhy

Referenceshttp://www.jimmunol.org/content/172/1/202.full#ref-list-1

, 24 of which you can access for free at: cites 50 articlesThis article

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2004 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on N

ovember 17, 2017

http://ww

w.jim

munol.org/

Dow

nloaded from

Inducible Heat Shock Protein 70 Promotes Myelin AutoantigenPresentation by the HLA Class II1

Marcin P. Mycko,* Hanna Cwiklinska,* Jacek Szymanski,† Bozena Szymanska,*Grzegorz Kudla,‡ Lukasz Kilianek,§ Artur Odyniec,* Celia F. Brosnan,¶ andKrzysztof W. Selmaj2*¶

In this study, we investigated the role of the inducible form of heat shock protein 70 (hsp70) in the presentation of the majorputative autoantigen in multiple sclerosis, myelin basic protein (MBP), in the context of appropriate MHC class II. By coimmu-noprecipitation, we found that MBP is associated with hsp70 in APC in an ATP/ADP-dependent manner. Additionally, usingconfocal microscopy, hsp70 was detected in the endocytic pathway of APC, where it colocalized with MBP and HLA-DR. Theimmunodominant epitopes of MBP 85–99 and 80–99 were shown to bind selectively and specifically to hsp70 by surface plasmonresonance. The functional significance of MBP interaction with hsp70 was demonstrated by the detection of enhanced responsesof an MBP-specific T cell hybridoma to MBP and MBP 80–99 with increasing levels of hsp70 and reduced responses when hsp70expression was diminished within APC-expressing DRA*0101, DRB1*1501 (DR1501). However, when MBP 85–99 was used as thestimulus, T cell hybridoma responses were not enhanced by hsp70 overexpression within APC, suggesting that hsp70 contributesto Ag processing rather than Ag presentation. The importance of a direct association between MBP and hsp70 in the presentationpathways was demonstrated by enhanced efficacy of MBP presentation by APC transfected with a plasmid vector encoding afusion hsp70-MBP protein. This is the first report on the involvement of self-inducible hsp70 in MHC class II-dependent autoan-tigen processing by APC. It implicates that aberrant self hsp expression may lead to the enhancement/modulation of autoimmuneresponses. The Journal of Immunology, 2004, 172: 202–213.

A role for heat shock proteins (hsp)3 as facilitators of im-mune responses to proteins and peptides has now beenwidely documented both in vivo and in vitro (1–5). The

suggestion that hsp-associated peptides are immunogenic, whereasthe peptides alone are not, was first provided by Udono and Sriv-astava (6). They showed that an hsp70 preparation derived fromcancer cells and containing cancer-specific peptides induced strongCTL response against the cancer cells after autologous vaccina-tion. Subsequently, several studies using tumor models or virallyinfected cells confirmed that hsp-peptide complexes are capable ofevoking immune response against peptides chaperoned by hsp (7–11). Most of these studies demonstrated the generation of a CTLresponse that was restricted by MHC class I molecules. Althoughthe mechanisms involved in the enhanced immunogenicity of hsp-

peptide complexes have not been fully delineated, recent studiessuggest that hsp facilitate uptake of small amounts of peptide byAPC via the common receptor CD91 that interacts with mostclasses of hsp (12, 13). Additional receptors for hsp that haverecently been characterized on APC include Toll receptor 4 forhsp60 (14) and CD36 for gp96 (15).

Several studies supported the notion that hsp are also integrallyinvolved in intracellular Ag presentation pathways. Transfection ofhsp70 was shown to be able to rescue presentation of tumor Ags byMHC class I molecules in tumor cell lines (16). Hsp90 and hsp70also have been shown to associate with TAP and other moleculesof the MHC class I presentation pathway (17, 18). Treatment ofAPC with deoxyspergualin, an inhibitor of hsp70 and hsp90 (19,20), significantly reduced their potency to present MHC class I-associated Ags to T cells (21). Although the majority of reports onthe enhancing effect of hsp on Ag presentation address the MHCclass I Ag presentation pathway, more recent studies have alsoimplicated the hsp70 family in MHC class II presentation. Thus,hsp70 cognate form (hsc70) was recently identified as a cytosolicpartner capable of interacting with the MHC class II invariantchain (Ii), as well as being responsible for the enlargement of theendocytic compartments (22). Furthermore, a dominant-negativeversion of hsc70 counteracted the ability of Ii to modify the en-docytic pathway, demonstrating an interaction in vivo of Ii withhsc70 (22). However, the exact role of hsp70, the inducible formof hsp70 in the pathway of the MHC class II Ag presentation andas a potential chaperonin for MHC class II-associated Ags, re-mains to be addressed.

In this study, we investigated the involvement of hsp70 expres-sion in the presentation of myelin basic protein (MBP) and MBP-derived peptides in an MHC class II-dependent manner. MBP rep-resents one of the most immunogenic proteins of the CNS and is

*Department of Neurology, Laboratory of Neuroimmunology, and†Department ofMolecular and Medical Biophysics, Medical University of Lodz, Lodz, Poland;‡In-ternational Institute of Molecular and Cell Biology, Warsaw, Poland;§Nencki Insti-tute of Experimental Biology, Warsaw, Poland; and¶Departments of Pathology andNeuroscience, Albert Einstein College of Medicine, Bronx, NY 10461

Received for publication September 11, 2002. Accepted for publication October6, 2003.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by State Committee for Scientific Research Grant 4 P05A005 19 to M.P.M., and U.S. Public Health Service Grants NS31919 and NS11920.G.K. is the recipient of a scholarship from the Postgraduate School of MolecularMedicine.2 Address correspondence and reprint requests to Dr. Krzysztof W. Selmaj, Depart-ment of Neurology, Medical University of Lodz, Kopcinskiego Str. 22, 90-153 Lodz,Poland. E-mail address: kselmaj@afazja.am.lodz.pl3 Abbreviations used in this paper: hsp, heat shock protein; EEA1, early endosomalautoantigen-1; GFP, green fluorescence protein; hsc, hsp cognate form; Ii, MHC classII invariant chain; MBP, myelin basic protein; MS, multiple sclerosis; PKC, proteinkinase C.

The Journal of Immunology

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

the putative autoantigen in multiple sclerosis (MS) (23, 24). Abun-dant expression of hsp has been noted in the pathologic lesions ofMS (25–27). We now show that hsp70 overexpression signifi-cantly contributes to T cell recognition of MBP or an elongatedMBP peptide that requires endosomal processing. We propose thataberrant hsp70 expression might contribute to recognition of au-toantigens and modulation of autoimmune responses.

Materials and MethodsCells and reagents

MHC class II-restricted MBP presentation was tested using MBP-specificCD4� T cell hybridomas that recognize MBP peptide 85–99 in the contextof HLA class II DRA*0101, DRB1*1501 (DR1501). The hybridomas weregenerated from a human autoreactive CD4� T cell clone, derived from aDR1501 homozygous MS patient, recognizing MBP peptide 85–99. TheTCR � and � full-length cDNA were generated from RNA extracted fromthe clone. The TCR � and � expression vectors (pRep7-TCR� and pRep8-TCR�) were generated by incorporation of TCR � and � cDNAs intopRep7 and pRep8 vectors (Invitrogen, Carlsbad, CA). TCR-negative JurkatJ.RT3-T3.5 hybridoma cells (American Type Culture Collection, Manas-sas, VA) were cotransfected with pRep7-TCR� and pRep8-TCR� usingLipofectamine Plus reagents (Life Technologies, Gaithersburg, MD), ac-cording to the manufacturer’s instructions. Stable transfectant lines weregenerated by selection with hygromycin B (Invitrogen), and positivelytransfected hybridomas were selected by single cell cloning. Several lineswere confirmed to express both TCR � and � by flow cytometry usingFITC-labeled anti-human TCR �� T10B9.1A-31 Ab (BD PharMingen,San Diego, CA). Because all selected lines expressed equal levels of TCR,as well as identical MBP recognition profiles, one of them was selected forsubsequent experiments (data not shown).

For the presentation assays, a native murine fibroblast cell line (L cells)transfected with DR1501 (L1501) was used as the APC. L cells wereshown to express functional HLA-DM and Ii and to be able to process andpresent MHC class II-associated Ags (28). Hsp70-inducible form, hsp70.1,hsp40, and MBP full-length cDNAs were generated from a murine spleencDNA library using specific primers. Hsp70-MBP and hsp40-MBP fusionprotein cDNA were generated by ligation of hsp70 cDNA or hsp40 cDNAand MBP cDNA, respectively, such that MBP was linked directly to the 3�end of hsp70 or hsp40. Protein expression vectors were generated by theincorporation of the full-length cDNA into the expression vector pVax(Invitrogen), according to the manufacturer’s instructions. Green fluores-cence protein (GFP) full-length cDNA (Clontech Laboratories, Palo Alto,CA) was incorporated into the pVax vector to generate pVax-GFP expres-sion vector. Empty linearized pVax vector was used as a control for themock transfection experiments.

Hybridoma MBP presentation and DR inhibition assays

To exclude the possibility of autopresentation of MBP in the context of selfMHC class II molecules by the human T cell clones, the MBP-reactivehuman TCR-derived T cell hybridomas were used for the presentationassays. T hybridoma cells (5 � 104/well) were cocultured with 5 � 104/well L1501 or native L cells in round-bottom 96-well plates. Ags weregiven in serial dilutions, and the total well volume was 200 �l. Presentationwas allowed to occur for 24 h in culture, following which 100 �l of su-pernatant was harvested and transferred to a fresh plate for measurement ofIL-2. The concentration of IL-2 in the supernatant was measured using theIL-2-specific HT.2 cell lines by coincubation of supernatants with 5 �103/well HT.2 cells, followed by measurement of HT.2 proliferation using[3H]thymidine incorporation over 18 h (29). The data represent �cpm cal-culated by subtracting the background proliferation of unstimulated HT.2cells. All presented values are means of triplicate cultures, and all exper-iments were performed four to six times. The requirement for MHC classII in the presentation assays was demonstrated using an anti-DR inhibitionassay. For this assay, DR-blocking Ab (clone L243; BD PharMingen) wasadded at a concentration of 10 �g/ml at the initiation of the hybridomas andAPC cocultures (30). Presentation was allowed to occur for 24 h in culture,following which 100 �l of supernatant was harvested and assayed for thepresence of IL-2, as described above.

Hsp70 overexpression

Hsp70 overexpression was achieved by transfection of the fibroblast mu-rine cell line expressing DR1501 (L1501) with full-length murine hsp70.1cDNA expression vector (pVax-hsp70). The transfection was performedusing Lipofectamine Plus reagents (Life Technologies), according to themanufacturer’s instructions. The kinetics and efficacy of hsp70 overexpres-

sion were tested by Western blot analysis of L1501 cell lysates using anti-hsp70 mAb sc24 (Santa Cruz Biotechnology, Santa Cruz, CA). Because thesc24 Ab recognizes both hsp70 and hsc70, this allowed us to simulta-neously control expression of both forms of hsp70. As a control, cells weretransfected with empty pVax expression vector. No difference in HLAclass II expression levels was detected between hsp70 and mock-trans-fected cell lines, as determined by flow cytometry, using FITC-labeledanti-HLA-DR mAb TU36 (BD PharMingen). The efficiency of the pVax-hsp70 transient transfection was assessed in the parallel pVax-GFP trans-fection. The percentage of FL1-positive cells was �75%, as assayed byflow cytometry (data not shown).

Hsp70 down-regulation

Hsp70 down-regulation was achieved in L1501 cells by infection withhsp70 antisense sequence containing Ad.ashsp70 adenovirus. Subconfluentcells plated 24 h earlier were infected with adenovirus for 15 min at 25°Cwith continuous rocking, followed by 60 min at 37°C in RPMI 1640. Theconcentrations of virus used were optimized to cause 100% infection ineach cell line without visible toxic effects according to the published pro-tocol (31). Ad.ashsp70 adenovirus was kindly provided by M. Jaattela(University of Copenhagen, Copenhagen, Denmark). The efficacy of hsp70down-regulation was tested by Western blot analysis of L1501 cell lysatesusing anti-hsp70 mAb sc24 (Santa Cruz Biotechnology) after several timepoints from 24 to 96 h postinitial infection, and the 60-h time point wasselected for additional experiments.

Hsp70-MBP and hsp40-MBP fusion protein expression

Hsp70-MBP and hsp40-MBP fusion protein expression were achieved bytransfection of the fibroblast murine cell line expressing DR1501 (L1501)with full-length murine MBP cDNA sequence linked with either full-lengthmurine hsp70.1 cDNA sequence or full-length murine hsp40 cDNA se-quence in the expression vectors (pVax-hsp70-MBP or pVax-hsp40-MBP,respectively). Hsp40 is another member of the hsp family that was shownto be an important hsp70 cochaperone (32). The efficacy and kinetics of thefusion protein hsp70-MBP expression were tested by Western blot analysisof L1501 cell lysates using anti-hsp70 mAb sc24 (Santa Cruz Biotechnol-ogy), as described above. Control cells were transfected with empty pVaxexpression vector. For the in vitro expression of MBP protein, L1501 cellswere transfected with full-length murine MBP cDNA expression vector(pVax-MBP). The efficiency of MBP, fusion hsp70-MBP, and fusionhsp40-MBP protein expression was assessed by Western blot analysis us-ing anti-MBP mAb (Chemicon, Temacula, CA). The amounts of MBP andMBP fusion protein expression vectors used for transfections were opti-mized to assure similar levels of MBP and MBP fusion protein synthesis inL1501 cells (see Fig. 6B, confirmed by densitometry; data not shown). Nodifference in HLA class II expression levels was detected between hsp70-MBP, hsp40-MBP, MBP, and mock-transfected cell lines, as determinedby flow cytometry using FITC-labeled anti-HLA-DR mAb TU36 (BDPharMingen). The efficiency of the pVax-MBP, pVax-hsp70-MBP, andpVax-hsp40-MBP transient transfections was assessed in the parallelpVax-GFP transfection. The percentages of FL1-positive cells were�75%, as assayed by flow cytometry (data not shown).

Hsp70 peptide-binding assay by surface plasmon resonance

Bacterial hsp70 analog (DnaK; kind gift from M. Zylicz, InternationalInstitute of Molecular and Cell Biology) was immobilized on the surface ofa CM5 sensorchip (Biacore AB, Uppsala, Sweden) at 40 �g/ml in 100 mMsodium acetate (pH 4.0). Between 5000 and 8000 resonance units werecoupled on each occasion in flow cell 2, while BSA was coupled (at thesame level) to flow cell 1, which served as a reference. MBP peptidesAc1–11, 38–59, 61–82, 80–99, 85–99, and 148–162, and two controlpeptides, PLEKQHEKERKQEEGES (corresponding to �1-acid glycopro-tein peptide 185–201) and SKEQREPELEGEHQKEK (�1-acid glycopro-tein peptide 185–201 scrambled peptide; both peptides were a kind giftfrom C. Cierniewski, Medical University of Lodz), at the concentration of200 �g/ml, were resuspended in HBS buffer, and binding of fluid-phasepeptides was determined at 37°C by surface plasmon resonance using theBIAcore X (Biacore AB) in neutral pH (7.0) and acidic pH (5.5). To per-form a competitive analysis of binding of MBP and MBP peptides to thesurface coupled with hsp70, the sensorchip was washed with HBS supple-mented with either of the MBP peptides at a concentration of 300 �g/ml(Sigma-Aldrich, St. Louis, MO) for 45 min, at a flow rate of 5 �l/min, atacidic (5.5) or neutral pH (7.0), before assaying the hsp70 MBP-bindingcapacity. MBP at concentration of 10 �g/ml in HBS was injected onto thesensor. To test ATP/ADP dependence of hsp70 interactions with peptides,the hsp70-coupled sensorchip surface was washed with HBS supplementedwith either 1 mM ADP or 1 mM ATP (Sigma-Aldrich) for 30 min at a flow

203The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

rate of 5 �l/min, before assaying the peptide-binding capacity of hsp70.HBS was used as buffer throughout, flow rate was 2 �l/min, and injectionvolume was 10 �l. When needed, the sensorchip surface was regeneratedwith a pulse of 100 mM HCl.

Analysis of coimmunoprecipitation of MBP-hsp70 complexesfrom APC

L1501 cells were pulsed with MBP at 37°C for 1 h. After washing (threetimes, 10 ml PBS), cells were incubated in lysis buffer (TBS, 1% TritonX-100, 1 mg/ml BSA, 0.2 U/ml aprotinin, 1 mM PMSF) for 1 h on ice. Thesupernatants from the last wash were collected and checked for the pres-ence of soluble MBP by Western blot. To exclude the possibility of non-specific interactions between hsp70 and other proteins, we included anexcess of BSA in the lysis buffer (see above). The lysate was subsequentlyclarified by centrifugation, and the aqueous phase was collected. To ex-clude the possibility of cell surface interaction between hsp70 and MBP,the cells were trypsinized before lysis. Cells were incubated in 0.25% tryp-sin (Sigma-Aldrich)/PBS buffer for 15 min at 22°C and washed extensivelythree times in PBS with 0.2% BSA. The trypsinization conditions wereselected in order not to disrupt cell membrane integrity (33). For immu-noprecipitation, primary Ab against hsp70, goat polyclonal anti-hsp70 IgG,sc1060 (Santa Cruz Biotechnology), was added at a concentration of 5�g/ml and incubated by shaking on ice for 1 h. Next, 50 �l of secondaryAbs anti-mouse IgG-agarose (Sigma-Aldrich) was added for an additional60-min incubation, shaking on ice. The homogenates were then washedfour times by centrifugation for 15 s at 14,000 rpm at 4oC. The supernatantswere discarded, and the immunoprecipitates were dissolved in SDS samplebuffer (0.1 Tris-Cl, 4% SDS, 20% glycerol, 0.05% bromphenol blue, 5%2-ME), resolved on a 12% polyacrylamide SDS gel, and subjected to elec-trophoresis (25 mA, 60 min, room temperature). Next, transfer to a poly-vinylidene difluoride membrane (Immobilon P) was performed, and theproteins were probed with anti-MBP mAb (Chemicon). The membraneswere stripped (2% SDS, 0.1 M 2-ME, 62.5 mM Tris-Cl, pH 6.8) for 30min, at 60°C, and probed with anti-hsp70 mAb sc24 (Santa Cruz Biotech-nology). As a control, the parallel immunoprecipitation was performedusing matched polyclonal goat IgG (anti-murine glial fibrillary acid pro-tein, sc6170; Santa Cruz Biotechnology). The hsp70 bacterial analog(DnaK; kind gift from M. Zylicz, International Institute of Molecular andCell Biology) and full-length bovine MBP protein (Sigma-Aldrich) wererun as a separate lane to serve as a reference for the immunodetectionexperiments. To test the ATP/ADP dependence of hsp70-MBP coimmu-noprecipitation, the cell lysate was incubated with either ADP or ATP(Sigma-Aldrich) in the concentration range 1–5 mM during the hsp70 im-munoprecipitation. The immunoprecipitates were subjected to the SDS-PAGE, as described above. To test the interactions of hsp70 with controlprotein, the L1501 cells were lysed and immunoprecipitated with anti-hsp70 Ab, as described above. The immunoprecipitates as well as L1501cell lysates after the separation on polyacrylamide SDS gel and transfer topolyvinylidene difluoride membrane were assayed with anti-protein kinaseC � (anti-PKC�) mAb sc216 (Santa Cruz Biotechnology).

Immunofluorescent staining of the cell lines and confocalmicroscopy

L1501 cells, induced for HLA-DR expression with IFN-� (400 U/ml for48 h), were grown on the glass coverslips to semiconfluence. Cells werefixed with 4% paraformaldehyde and Triton X-100 permeabilized for Abstaining. To visualize hsp70 subcellular localization, first staining was per-formed with anti-hsp70 SPA-812 (Stressgen Biotechnologies, San Diego,CA), followed by staining with secondary Cy5-coupled anti-rabbit IgG(Jackson ImmunoResearch Laboratories, West Grove, PA). To label theendosomal compartment, cells were incubated in serum-free medium for30 min, followed by incubation with FITC-coupled transferrin (MolecularProbes, Eugene, OR) for 30 min. To label early endosomal compartment,cells were stained with FITC-coupled anti-early endosomal autoantigen-1mAb (EEA1; BD Transduction Laboratories, San Diego, CA). To localizeintracellular HLA-DR, cells were stained with FITC-labeled anti-HLA-DRmAb TU39 (BD PharMingen). To visualize internalized MBP L1501, cellswere incubated in serum-free medium for 30 min, followed by incubationwith MBP (Sigma-Aldrich) at 100 �g/ml for 30 min and staining withprimary anti-MBP mAb (Chemicon) and secondary FITC-coupled anti-mouse IgG (Sigma-Aldrich). Confocal microscopy was performed usingthe Confocal Laser Scanning Microscopy TCS SP2 (Leica, Lasertechnik,Heidelberg, Germany). For each experiment, cells stained with each flu-orophore alone were imaged to ensure that fluorescence emission did notbleed between channels. Cells displayed for each experiment are represen-tative of many cells imaged over at least three independent experiments.

Heat shock-induced up-regulation of hsp70 expression in APCs

Splenocytes were isolated from naive, female SJL/J mice, 6–8 wk of age,and plated on 96-well round-bottom plates and used as murine APCs. Thecells were heat shocked for 1 h at either 40°C or 42°C, and the up-regu-lation of hsp70 was tested by Western blot analysis of splenocyte lysatesusing the anti-hsp70 mAb sc24 (Santa Cruz Biotechnology). The analysisof hsp70 expression was done at different time points after the start of heatshock (0, 10, 30, and 60 min).

MBP T cell line generation and proliferation assay

Female SJL/J mice, 6–8 wk of age, were immunized with MBP protein inCFA. Each mouse received 0.25 ml of a mixture of 0.8 mg of MBP proteindissolved in 0.1 ml of double-distilled H20 and 0.75 mg of Mycobacteriumtuberculosis, in 0.15 ml CFA, injected s.c. in two abdominal sites. On day12, cervical and inguinal lymph node cells were isolated from immunizedmice and placed on 24-well plates to generate MBP-specific T cell lines.The cultures were stimulated on the day of isolation with MBP protein at10 �g/ml, and on day 5 given murine IL-2 (10 U/ml; PeproTech, RockyHill, NJ) and Con A (2 �g/ml; Sigma-Aldrich). Lymph node cell cultureswere subsequently stimulated with MBP and IL-2 for another two roundsto generate MBP-specific T cell lines.

After the initial heat shock at 42°C for 1 h, splenocytes (see above) werepulsed with MBP at 1, 10, and 50 �g/ml and used for the proliferationassay as APCs. These APCs were added to MBP T cell lines in 96-wellround-bottom plates (2 � 105 APCs and 105 T cell line cells/well) for 24 hand cultured at 40°C. Heating conditions and duration of the cultures weredetermined in several experiments to achieve clear up-regulation of hsp70and to avoid nonspecific cell death due to high temperature, which wasobserved after 48 h of culture (data not shown). Cultures were pulsed with[3H]thymidine, 1 mCi/well, for the last 14 h of the assay. Cells were har-vested and [3H]thymidine incorporation was determined in a Wallac Be-taplate liquid scintillation counter (PerkinElmer Life Sciences, Wellesley,MA). As controls, we used cultures of nonheat-treated APCs. The datarepresent �cpm calculated by subtracting the background proliferation ofunpulsed cultures of APCs and T cell lines. All values shown represent themeans of triplicate cultures, and experiments were performed three to fourtimes.

Statistical analysis

Multifactor ANOVA test or Student’s t test was applied where appropriateto test significant differences in the IL-2 secretion assays and proliferationassay. Five percent (two tailed) was chosen as the level of significance.

ResultsHsp70 associates with MBP protein in the APC

To determine whether hsp70 associates with Ag in APC, we in-cubated the L1501 cells with MBP for 1 h and then immunopre-cipitated hsp70 from the cell lysates. To exclude the possibility ofnonspecific interactions, the samples were incubated during lysiswith an excess of BSA in the lysis buffer. Hsp70 and associatedproteins were separated by SDS-PAGE, transferred to ImmobilonP membrane, and probed with Abs specific for MBP (Fig. 1A).Immunoblot analysis with the MBP-specific Abs showed a 21-kDaband, which corresponded to full-length MBP protein, that coim-munoprecipitated with the hsp70 molecule (Fig. 1A, lane 3). Incontrast, no MBP was immunoprecipitated using a control poly-clonal goat IgG (Fig. 1A, lane 5) and no MBP was immunopre-cipitated from cells not pulsed with MBP (Fig. 1A, lane 2). Nodifference in MBP coimmunoprecipitation was found when the cellsurface was trypsinized before the immunoprecipitation, suggest-ing an intracellular origin of the MBP-hsp70 complexes (Fig. 1A,lane 4). The presence of MBP-hsp70 complexes was ATP depen-dent (Fig. 1B), which is characteristic for hsp70-chaperoning ac-tivity. Furthermore, the supernatants from the last wash before celllysis were assayed for the presence of soluble MBP. Because nosoluble MBP was found, it is unlikely that hsp70-MBP complexeswere formed extracellularly (Fig. 1C). Finally, the specificity ofhsp70-MBP complexes was tested by analysis of potential forma-tion of hsp70 complex with a control, intracellular protein. No

204 HSP70 IN MHC CLASS II PRESENTATION

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

coimmunoprecipitation of hsp70 and PKC�, an abundant cytoplas-mic protein, was detected (Fig. 1D). Thus, we have shown thatMBP-hsp70 complexes can be detected in the APCs, suggestingthe specific, selective self hsp70 involvement in the intracellularfate of MBP within the APCs.

Association of hsp70 with different MBP peptides

To investigate the significance of association between hsp70 andMBP, as well as MBP peptides, we used a BIAcore assay, whichis based on surface plasmon resonance. Surface plasmon resonancedetects molecular interactions, because there is a correspondingchange in refractive index when a macromolecule in solution bindsto a macromolecule immobilized on the sensorchip. Hsp70 and acontrol protein, BSA, were immobilized on the surface of the sen-sorchip, and binding of fluid-phase MBP peptides was measured(Fig. 2). We selected a panel of 11- to 22-aa-long peptides span-ning the entire MBP molecule: Ac1–11, 38–59, 61–82, 80–99,85–99, and 148–162. The BIAcore assays were run at two differentpH conditions: neutral (7.0) and acidic (5.5), to reflect possibleinteractions occurring in either intracellular or specifically endo-somal compartments. Sensorgrams of MBP peptide-hsp70 inter-actions showed a significant, specific binding of the 85–99 and80–99 MBP peptides to the hsp70 at neutral pH conditions (Fig.2A), while 85–99, 80–99, and 61–82 peptides bound at acidic pH(Fig. 2B). In contrast, none of the tested peptides showed anyinteractions with the control immobilized protein, BSA, for bothtested pH conditions. Furthermore, the control tested peptides,PLEKQHEKERKQEEGES and SKEQREPELEGEHQKEK, didnot show any significant binding to hsp70 (Fig. 2C). Thus, with thepanel of MBP peptides tested and two unrelated, control peptides,hsp70 was shown to be able to interact with MBP peptides 85–99,80–99, and 61–82, but not with any other tested peptides.

Because hsp70 is capable of binding MBP, as we have shownpreviously (34), as well as MBP peptides, we performed a com-petition assay to determine the specificity of the protein/peptidebinding site. We preincubated the sensorchip surface-immobilizedhsp70 with MBP peptides to saturate its binding capabilities and thentested whether the capacity to bind MBP was still present. Interest-ingly, incubation with either MBP peptide (85–99 or 80–99) at bothacidic and neutral conditions led to prevention of MBP binding tohsp70 (Fig. 2D). This result suggests that the MBP 80–99 amino aciddomain may be responsible for binding of MBP protein to hsp70.

The ATP/ADP dependence of the MBP peptide 85–99, 80–99,and 61–82 binding to hsp70 was tested by preincubating sensor-chip surface-immobilized hsp70 with either ADP or ATP and sub-sequent analysis of the MBP peptide binding for both acidic and

Lane 1, Represents the detection of reference hsp70 and MBP proteins. B,ATP/ADP dependence of MBP-hsp70 complexes. L1501 cells were incu-bated with MBP, lysed, and immunoprecipitated with anti-hsp70 goat IgG(lanes 2–4). The cell lysate was incubated with either ADP (lane 2) or ATP(lane 3 and 4) during the hsp70 immunoprecipitation (as described in Ma-terials and Methods). Lane 1, Represents the detection of reference hsp70and MBP proteins. C, Lack of MBP in extracellular supernatant of APC.L1501 cell supernatant collected before the cell lysis was probed for thepresence of MBP (lane 3). Lane 2, Represents the detection of referenceMBP proteins, line 1 weight marker. D, Lack of coimmunoprecipitation ofcontrol intracellular protein with hsp70. L1501 cells were lysed and im-munoprecipitated with anti-hsp70 goat IgG (lane 3) or with matched goatIgG (lane 2). All samples were probed with anti-PKC� mAb (lanes 2 and3, lower panel), and after Ab stripping, with anti-hsp70 (lanes 2 and 3,upper panel). Lane 1, Represents detection of the indicated proteins in theL1501 cell lysate. IgGh and IgGl bands represent H and L chains of Absused for immunoprecipitation experiments.

FIGURE 1. A, Demonstration of the presence of MBP-hsp70 complexesin APC. L1501 cells were lysed and immunoprecipitated with anti-hsp70goat IgG (lanes 2–4) or with matched goat IgG (lane 5). All samples wereprobed with anti-MBP mAb (lanes 1–5, lower panel), and after Ab strip-ping, with anti-hsp70 (lanes 1–5, upper panel). Lanes 3, 4, and 5, Repre-sent L1501 cells that were incubated with MBP, while lane 2 representsunpulsed L1501 cells. Lane 4, Represents L1501 cells that were treatedwith trypsin after the MBP pulse, before the immunoprecipitation step.

205The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

neutral pH conditions (Fig. 3). For all the tested peptides, ATPincubations abrogated their binding to hsp70 at both pH condi-tions, while ADP had no effect on these interactions. Thus, hsp70-specific, ATP-dependent peptide activity of the observed interac-tions was shown, confirming the specificity of the interaction ofMBP peptides 85–99, 80–99, and 61–82 with hsp70.

Overexpression of hsp70 leads to the increased presentation ofMBP protein and MBP peptide 80–99 by HLA class II

To analyze the role of hsp70 in Ag presentation in the context of HLAclass II, we overexpressed hsp70 in the APCs by transient transfectionwith pVax-hsp70 vector. The transfection with pVax-hsp70 resulted

FIGURE 2. A and B, Selective MBP peptides binding to hsp70 (bold lines), but not to BSA (thin lines), as measured by surface plasmon resonance.Several MBP peptides were passed independently over immobilized hsp70 or BSA protein at either neutral pH (7.0) (A) or acidic pH (5.5) (B). C, Twocontrol peptides, PLEKQHEKERKQEEGES and SKEQREPELEGEHQKEK, were passed independently over immobilized hsp70 protein in neutral pH(7.0). No significant binding to hsp70 was found. D, Competitive binding of MBP and MBP peptides by hsp70 was tested by preincubation of sensor-surface-immobilized hsp70 with either MBP peptide 85–99 or 80–99 at 300 �g/ml for 45 min, followed by the analysis of the MBP, in acidic (pH 5.5;left panel) and neutral (pH 7.0; right panel) conditions. Thin line represents hsp70 binding of MBP before incubation with MBP peptide; bold lines representthe binding after incubation with MBP peptide 85–99; shadow line after incubation with MBP peptide 80–99. Sensorgrams were obtained with BIAcoreX instrument equipped with a sensorchip, CM5, research grade. Hsp70 protein was immobilized on Fc2 channel, whereas Fc1 channel was used for thecontrol, BSA interactions. Flow rate, 2 �l/min; applied sample volume, 10 �l. The results of one representative of four independent experiments are shown.

206 HSP70 IN MHC CLASS II PRESENTATION

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

in a significant overexpression of hsp70, with peak expression oc-curring 24 h after transfection (Fig. 4A). In contrast, no differencefrom the control 0-h time point was noted in cells harvested 24 hposttransfection with the pVax control (Fig. 4A). Furthermore, nodifference was found with respect to the expression levels ofhsc70, the cognate form of hsp70 (Fig. 4A). To ensure maximumoverexpression of hsp70, L1501 cells were used for the MBP pro-tein/peptide presentation assay to MBP-specific T cell hybridomas24 h posttransfection with hsp70 or empty vector. We found nodifference in T cell hybridoma responses to MBP peptide 85–99(Fig. 4D) when presented by cells with hsp70 overexpression vsempty vector-transfected APCs. In contrast, the presentation ofeither the extended MBP peptide 80–99 (Fig. 4C) or full-lengthMBP (Fig. 4B) was significantly enhanced when hsp70 was over-expressed, p � 0.01 and p � 0.02, respectively. This result sug-gests that hsp70 plays a role in the Ag-processing pathway inAPCs. This effect was also critically dependent on the presence ofthe MHC class II Ags because native, DR-negative L cells wereunable to present either MBP protein (Fig. 4E) or MBP peptides(data not shown) to the T cell hybridomas, and a DR-blocking Ab(L243) abrogated the MBP protein recognition by hybridomas forboth L1501 and L1501 with hsp70 overexpression (Fig. 4E), p �0.02. Thus, the specificity of MHC class II presentation of MBPand MBP peptides associated with hsp70 was confirmed.

Down-regulation of hsp70 leads to a decrease in MBP protein/peptide presentation in the context of HLA class II

To confirm further the role of hsp70 in the presentation of HLAclass II-associated MBP Ags, we specifically down-regulated ex-pression of hsp70 in APCs by expression of hsp70 antisense usingan adenoviral delivery system (Fig. 5A). Interestingly, while thepresentation of both MBP peptides 85–99 and the N terminus ex-tended 80–99 was not modified by down-regulation of hsp70 (Fig.5, C and D), the presentation of the MBP protein by APCs toMBP-specific hybridomas was diminished (Fig. 5B), p � 0.02.These results correspond to the hsp70 overexpression data, sug-gesting the involvement of hsp70 in the MBP presentation, espe-cially in the processing phase.

Expression of hsp70-MBP fusion protein within the APCs leadsto increased HLA class II-associated Ag presentation

To analyze the importance of a direct association between hsp70 inHLA class II-associated Ag presentation, we transfected the APCs(L1501) with either an MBP expression vector, hsp70-MBP,hsp40-MBP fusion protein expression vectors, or a control emptyvector. The transfection of hsp70-MBP fusion protein resulted inmaximum levels of the fusion protein expression in 24 h posttrans-fection, with no influence on the hsp70 and hsc70 expression levels(Fig. 6A). Transfection conditions with MBP, hsp70-MBP, and

FIGURE 3. Hsp70 binds MBP peptides in an ATP/ADP-dependent manner, as measured by surface plas-mon resonance. MBP peptides 85–99, 80–99, and61–82 were passed independently over immobilizedhsp70 protein in either acidic pH (5.5) (left panel) orneutral pH (7.0) (right panel) in the presence of either1 mM ATP (bold lines) or 1 mM ADP (thin lines).Flow rate, 2 �l/min; applied sample volume, 10 �l.The results of one of four independent experiments areshown.

207The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

FIGURE 4. A, Western blot demonstration of hsp70 overexpression kinetics in L1501-transfected cells. L1501 cells were transfected with pVax-hsp70expression vector, as described in Materials and Methods. The cells were lysed at various time points posttransfection, and expression of hsp70 and hsc70 wasdetermined by Western blot with anti-hsp70 murine mAb (lanes 2–5). As a control, L1501 cells were transfected with empty pVax vector, and after 24 h cell lysateswere made and probed with anti-hsp70 (lane 1). Lane 6, Represents m.w. markers. B–D, The enhancing effect of hsp70 overexpression in APC on the presentationof MBP protein (B), p � 0.02; MBP-extended peptide 80–99 (C), p � 0.01; but not on MBP peptide 85–99 (D). As a control, L1501 cells were transfected withempty pVax vector (mock transfection). E, MHC class II dependence of the MBP protein presentation by hsp70-overexpressed L1501 cells was tested usingDR-negative L cells and anti-DR Ab. IL-2 production by T cell hybridomas is shown, as measured by incorporation of [3H]thymidine by the IL-2-dependent cellline HT.2 with background proliferation correction (�cpm, as described in Materials and Methods, �SD).

208 HSP70 IN MHC CLASS II PRESENTATION

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

hsp40-MBP fusion proteins were adjusted to yield similar kineticsand efficiency of protein synthesis in the APCs, as judged by West-ern blot analysis using anti-MBP Abs (Fig. 6B) and densitometry.This allowed us to make a direct comparison of the effects oftransfections of APCs with single or fusion proteins encoding vec-tors on HLA class II MBP Ag presentation. We found that MBPtransfection of APCs resulted in significantly higher responses ofMBP-specific T cell hybridomas in the absence of any additionalexogenous Ag, p � 0.04. Furthermore, expression of hsp70-MBPfusion protein led to an even stronger response of the T cell hy-bridomas, p � 0.04, whereas expression of hsp40-MBP fusionprotein elicited similar levels of hybridoma activation to MBPoverexpression, suggesting the enhancing effect of hsp70 fusion toMBP results from the direct association of both proteins ratherthan a nonspecific extension of the MBP transcript. These datafurther supported the role of hsp70 in the presentation of MBP Agsin the context of HLA class II, suggesting an important role ofdirect association of hsp70 with MBP during Ag processing.

Hsp70 subcellularly colocalizes with endosomal markers as wellas with HLA-DR and internalized MBP

To track the possible route of hsp70 action on the HLA-DR pre-sentation pathway, we localized intracellular hsp70 with respect to

the possible colocalization within the compartments critical forHLA-DR function (Fig. 7). Because the major site of Ag processingas well as loading of HLA-DR with peptides occurs in endosomes, welooked first for the presence of hsp70 within these structures. Trans-ferrin, which binds transferring receptor in clathrin-coated pits, wasused as a selective marker for endosomal trafficking. FITC-cou-pled transferrin was incubated with L1501 cells for 30 min andvisualized by confocal microscopy. A vast majority of transferrin-positive structures were also positive for the hsp70 (Fig. 7A). Also,staining with the fluorophore-labeled Ab against a marker for earlyendosomes, EEA1, also revealed significant colocalization withhsp70 (Fig. 7D). Thus, hsp70 is present in the endosomal com-partment as early as early endosome stage. Additional staining ofL1501 cells with Ab against HLA-DR (Fig. 7B) and, following apulse with MBP (Fig. 7C), with Ab against MBP revealed evi-dence for subcellular colocalization of hsp70 and both HLA-DRand internalized MBP. These data suggest that the presence ofhsp70 in the endosomes permits direct access to both HLA-DR andinternalized MBP. Colocalization of hsp70 and MBP also confirmsthe results of the MBP/hsp70 coimmunoprecipitation experiments,supporting the notion of subcellular association of hsp70 withMBP in APC.

FIGURE 5. A, Western blot analysis of hsp70 down-regulation using the Ad.ashsp70 vector. Lane 1, Molecular marker; lane 2, hsp70 expressionbefore Ad.ashsp70 infection; lane 3, 60 h postinfection. B–D, The inhibitory effect of hsp70 down-regulation in APC on the presentation of MBPprotein (B), p � 0.02, but not for MBP peptide 80 –99 (C) or MBP peptide 85–99 (D). IL-2 production by T cell hybridomas is shown, as measuredby incorporation of [3H]thymidine by the IL-2-dependent cell line HT.2 with background proliferation correction (�cpm, as described in Materialsand Methods, �SD).

209The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

Heat shock-induced up-regulation of hsp70 is associated withenhanced MBP-induced T cell proliferation

In attempt to relate the results of genetically induced hsp70 over-expression using the hybridoma system with more physiologic

hsp70 induction, we analyzed the potential influence of hsp up-regulation resulted from heat shock on the MBP presentation pro-cess. As a source of natural APCs, we used splenocytes of SJL/Jmice, and as responder cells an MBP-specific murine T cell line.Up-regulation of hsp70 was apparent in heat-shocked murinesplenocytes as early as 30-min incubation after temperature ele-vation (Fig. 8, A and B). Furthermore, the heat-induced up-regu-lation of hsp led to increased proliferative responses of MBP-spe-cific T cell line (Fig. 8C). These results demonstrate the potentialphysiological relevance of stress-induced expression of hsp70 andautoantigen recognition.

DiscussionIn this study, we have shown that the inducible form of hsp70enhances MHC class II-dependent presentation of the myelin-de-rived autoantigen, MBP. Hsp70 enhanced T cell recognition ofMBP, as well as an elongated MBP immunogenic peptide, but nota trimmed MBP peptide. We have also shown, using two indepen-dent techniques, coimmunoprecipitation and confocal microscopy,that hsp70 can associate with MBP within APC. Because hsp70was also present in the endosomal compartment and colocalizedwith HLA-DR, it suggests that hsp70 may modulate MBP pro-cessing within the MHC class II presentation pathway.

It is now well accepted that hsp act to chaperone peptidespresent in the cytosol for presentation and processing via the MHCclass I molecule-loading pathway (12, 35). It has also been shownthat hsp-associated peptides can enter an acidic compartment andbe loaded onto MHC class II molecules (36). Our results, usingMHC class II APC with overexpressed hsp70 and an MBP-specificTCR hybridoma as well as T cell lines, demonstrate a role forhsp70 in the immune responses dependent on MHC class II. Thedependence of MBP presentation on MHC class II expression wasconfirmed by the lack of presentation by APCs that had not beentransfected with DR1501, and by abrogation of the response withan Ab that blocked MHC class II on the cell surface. In addition,hsp70 has been localized in the endocytic compartment. This isalso the first demonstration to show that inducible self hsp70 en-hances Ag presentation in an MHC class II-dependent manner. Ithas also not been shown previously that hsp can enhance presen-tation of endogenous proteins that can serve as an autoantigen.Recently, Roth et al. (37) demonstrated an enhancing effect ofbacterial DnaK, the Escherichia coli analog of hsp70, in the MHCclass II-dependent presentation of recombinant human acetylcho-line receptor � subunit Ag. Panjwani et al. (38) demonstrated thatoverexpression of hsc73 in macrophages enhanced presentation ofthe exogenous Ags hen egg lysozyme and C5 protein to appropri-ate hybridoma T cells, and Matsutake and Srivastava (36) haveimplicated gp96 in MHC class II presentation. Inducible murinehsp70, although related to cognate forms of the protein as well asclosely related to its bacterial homologue, DnaK, is specificallytranscribed under different stress conditions and represents a typ-ical self stress-induced molecule involved in the majority of in-flammatory processes. Additionally, we have shown that inhibitionof hsp70 resulted in the diminished capability of APC to presentMBP. This observation underscores the important role of hsp70 inMBP presentation, but also indicates that hsp70 may not only en-hance Ag presentation in a stress-inducing environment, but mayalso be involved in regular mechanisms controlling Ag presentation.In this regard, the hypothesis of close evolutional similarities betweenhsp and MHC molecules serving as complementary mechanisms ofAg recognition seems to be of particular interest (1, 39).

We studied MBP, a putative autoantigen in the human autoim-mune neurologic disease, MS. It has been reported by us and oth-ers (25–27, 40, 41) that hsp, including hsp70, can accumulate in

FIGURE 6. A, Western blot analysis of kinetics of hsp70, hsc70, andhsp70-MBP fusion protein expression in APC transfected with hsp70-MBPexpression vectors detected with anti-hsp70 mAb, as described in Materi-als and Methods (lanes 2–6). Lane 1, Represents m.w. markers. B, Westernblot analysis of efficiency of MBP, hsp40-MBP, and hsp70-MBP fusionprotein expression detected with anti-MBP mAb in APC 24 h posttrans-fection with MBP, hsp70-MBP, and hsp40-MBP expression vectors, asdescribed in Materials and Methods (lanes 2–5). Lane 1, Represents m.w.markers. C, The enhancing effect of hsp70-MBP fusion protein expressionin APC on the HLA class II-dependent MBP protein presentation. L1501cells were transfected with either pVax-MBP, pVax-hsp70-MBP, or pVax-hsp40-MBP expression vectors or empty pVax vector as a control (mocktransfection) (�, pVax-MBP vs mock transfection, p � 0.04; ��, pVax-hsp70-MBP vs pVax-MBP, p � 0.04). IL-2 production by T cell hybrid-omas is shown, as measured by incorporation of [3H]thymidine by theIL-2-dependent cell line HT.2 with background proliferation correction(�cpm, as described in Materials and Methods, �SD).

210 HSP70 IN MHC CLASS II PRESENTATION

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

the pathologic lesions of MS within the CNS tissue. Using in situimmunocytochemical and molecular techniques, it has been shownthat accumulation of hsp70 occurs in glial cells implicated in Ag re-

stimulation within the CNS (42). We have shown in this study, usingsurface plasmon resonance technique, that hsp70 is capable of bindingboth MBP and MBP peptides 85–99 and 80–99. Furthermore, the

FIGURE 8. Western blot (A) and densito-metric (B) analysis of kinetics of hsp70 up-reg-ulation following heat shock at 40°C and 42°C.C, SJL/J splenocytes were incubated for 1 h at42°C, followed by pulsing with MBP, and usedas APC for MBP-reactive T cell lines. The pre-sentation was allowed to occur for 24 h in cul-ture at 40°C. Controls were cultures with APCsthat were nonheat treated (37°C). T cell prolif-erative responses are represented as �cpm (asdescribed in Materials and Methods, �SD),p � 0.04.

FIGURE 7. APC (L1501 cells) were formalde-hyde fixed and Triton X-100 permeabilized andstained for immunofluorescence analysis of intracel-lular colocalization of hsp70. Cells were labeled withanti-hsp70 and Cy5-coupled anti-rabbit IgG (red; leftpanel). A, To label the endosome compartment, cellswere allowed to internalize FITC-coupled transferrin(green; middle panel). B, To localize intracellulardistribution of HLA-DR molecules, cells werestained with FITC-coupled anti-DR (green; middlepanel). C, To visualize internalized MBP, cells wereincubated with MBP, followed by staining with anti-MBP and FITC-coupled anti-mouse Ig (green; mid-dle panel). D, Early endosome compartment was vi-sualized by staining with FITC-coupled anti-EEA1(green; middle panel). For all panels, superimposi-tions of Cy5 and FITC are represented in yellow(right panel).

211The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

competitive binding assay revealed the importance of MBP 80–99domain in MBP binding to hsp70. Such coincidence of hsp70 as-sociation with an immunodominant epitope of MBP makes feasi-ble a scenario in which hsp overexpressed by inflammatory medi-ators within the CNS lesions may be responsible for enhancedmyelin autoantigen presentation. This hypothesis would requirefurther confirmation, but implicates aberrant hsp expression in themechanisms involved in the propagation of inflammatory pro-cesses and autoimmunity.

Molecular mechanisms of hsp enhancement of Ag recognitionwithin the cell are currently not known. We have shown that over-expression of hsp70 only facilitates recognition of the whole MBPmolecule or the elongated peptide 80–99, but not the trimmedpeptide, which can directly be loaded onto MHC class II mole-cules. These results strongly suggest that hsp70 is involved inMBP processing within APC. The length of the principal humanHLA-DR2-restricted epitope, aa 85–99 of MBP that is recognizedin connection with HLA class II DRB1*1501 is 15 aa (43, 44). Ourresults indicate that presentation of such a peptide (MBP 85–99) isnot facilitated by hsp70, although it maintains hsp70-binding ca-pabilities assessed by surface plasmon resonance. However, MBPpeptide extended by 5 aa at the N-terminal side does require hsp70for efficient presentation. Recent findings by Menoret et al. (45)indicate that gp96 participates in trimming of N-terminal-extendedpeptides, which are then processed in the lumen of the endoplas-mic reticulum before MHC class I loading. They further showedthat gp96 is an aminopeptidase that can trim 19-mer precursors ofan octamer peptide derived from stomatitis virus to the ready-to-gooctamer for MHC class I presentation. Our results suggest thathsp70 might serve a similar function for N-terminal-elongatedpeptides processed for MHC class II presentation. Alternatively,hsp70 can cooperate with other aminopeptidases in the endosomalcompartments and protect the peptide from complete proteolysis,ensuring an appropriate peptide length for association with MHCclass II molecules, and chaperone them to the site of loading ontoMHC molecules (46). This possibility is supported by earlier find-ings that Abs to hsp70 inhibit MHC class II Ag presentation (47).Cooperation with other lysosomal enzymes seems feasible becausehsp70 down-regulation had a greater impact on the presentation ofthe whole molecule of MBP than for MBP 80–99 peptide.

We have also demonstrated a direct association between MBPand hsp70 within APC by coimmunoprecipitation and by colocal-ization of these two proteins using confocal microscopy. The im-portance of this association was demonstrated in transfectionexperiments with vectors encoding either MBP protein or hsp70-MBP and hsp40-MBP fusion protein. Although transfections withMBP and hsp70-MBP fusion protein yielded similar kinetics andefficiency of MBP synthesis in the APCs, a significantly higherresponse to APC-expressing fusion protein was detected with thehybridoma cell line. Furthermore, coupling of MBP to hsp40, an-other member of a closely related hsp family and an importanthsp70 cochaperone (32), by expression of hsp40-MBP fusion pro-tein expression in APCs did not lead to MBP recognition beyondthe point of MBP expression alone. These results might indicatethe importance of direct interactions between MBP and hsp70 pro-teins rather then indirect associations, through secondary cochap-erones, on entry to the endosomal/lysosomal compartment, repre-senting the early phase of the Ag presentation pathway. Lysosomallocalization is critical for proteolytic processing of human MBP(48, 49), and hsp70 has also been shown to be present in thiscompartment in a variety of cells (50).

Further studies will be required to provide definite evidence forthe role of hsp in autoimmune responses, but based on the currentfindings it is tempting to hypothesize that overexpression of hsp70

resulting from a noxious environment during inflammatory pro-cesses can lead to enhanced Ag recognition.

AcknowledgmentsWe thank Drs. M. Zylicz, C. Cierniewski, and L. Ignatowicz for helpfuldiscussions, and Dr. A. Wrzosek for the assistance during confocal mi-croscopy experiments.

References1. Srivastava, P. 2002. Roles of heat-shock proteins in innate and adaptive immu-

nity. Nat. Rev. Immunol. 2:185.2. Li, Z., A. Menoret, and P. Srivastava. 2002. Roles of heat-shock proteins in

antigen presentation and cross-presentation. Curr. Opin. Immunol. 14:45.3. Singh-Jasuja, H., N. Hilf, D. Arnold-Schild, and H. Schild. 2001. The role of heat

shock proteins and their receptors in the activation of the immune system. Biol.Chem. 382:629.

4. Srivastava, P. K. 2001. Immunotherapy of human cancer: lessons from mice. Nat.Immun. 1:363.

5. Wells, A. D., and M. Malkovsky. 2000. Heat shock proteins, tumor immunoge-nicity and antigen presentation: an integrated view. Immunol. Today 21:129.

6. Udono, H., and P. K. Srivastava. 1993. Heat shock protein 70-associated peptideselicit specific cancer immunity. J. Exp. Med. 178:1391.

7. Tamura, Y., P. Peng, L. Kang, M. Daou, and P. K. Srivastava. 1997. Immuno-therapy of tumors with autologous tumor-derived heat shock protein preparations.Science 278:117.

8. Yamazaki, K., T. Nguyen, and E. R. Podack. 1999. Cutting edge: tumor secretedheat shock-fusion protein elicits CD8 cells for rejection. J. Immunol. 163:5178.

9. Zheng, H., J. Dai, D. Stoilova, and Z. Li. 2001. Surface targeting of an intracel-lular heat shock protein gp96 on tumor cells induces dendritic cell maturation andanti-tumor immunity. J. Immunol. 167:6731.

10. Ciupitu, A. M., M. Petersson, C. L. O’Donnell, K. Williams, S. Jindal,R. Kiessling, and R. M. Welsh. 1998. Immunization with a lymphocytic chorio-meningitis virus peptide mixed with heat shock protein 70 results in protectiveantiviral immunity and specific cytotoxic T lymphocytes. J. Exp. Med. 187:685.

11. Moroi, Y., M. Mayhew, J. Trcka, M. H. Hoe, Y. Takechi, F. U. Hartl,J. E. Rothman, and A. N. Houghton. 2000. Induction of cellular immunity byimmunization with novel hybrid peptides complexed to heat shock protein 70.Proc. Natl. Acad. Sci. USA 97:3485.

12. Basu, S., R. J. Binder, T. Ramalingam, and P. K. Srivastava. 2001. CD91 is acommon receptor for heat shock proteins gp96, hsp90, hsp70, and calreticulin.Immunity 14:303.

13. Binder, R. J., D. K. Han, and P. K. Srivastava. 2000. CD91: a receptor for heatshock protein gp96. Nat. Immun. 1:151.

14. Ohashi, K., V. Burkart, S. Flohe, and H. Kolb. 2000. Cutting edge: heat shockprotein 60 is a putative endogenous ligand of the Toll-like receptor-4 complex.J. Immunol. 164:558.

15. Panjwani, N. N., L. Popova, M. Febbraio, and P. K. Srivastava. 2000. The CD36scavenger receptor as a receptor for gp96. Cell Stress Chaperones 5:391.

16. Wells, A. D., S. K. Rai, M. S. Salvato, H. Band, and M. Malkovsky. 1998.Hsp72-mediated augmentation of MHC class I surface expression and endoge-nous antigen presentation. Int. Immunol. 10:609.

17. Castellino, F., P. E. Boucher, K. Eichelberg, M. Mayhew, J. E. Rothman,A. N. Houghton, and R. N. Germain. 2000. Receptor-mediated uptake of antigen/heat shock protein complexes results in major histocompatibility complex class Iantigen presentation via two distinct processing pathways. J. Exp. Med.191:1957.

18. Chen, D., and M. J. Androlewicz. 2001. Heat shock protein 70 moderately en-hances peptide binding and transport by the transporter associated with antigenprocessing. Immunol. Lett. 75:1143.

19. Nadler, S. G., M. A. Tepper, B. Schacter, and C. E. Mazzucco. 1992. Interactionof the immunosuppressant deoxyspergualin with a member of the Hsp70 familyof heat shock proteins. Science 16:484.

20. Nadeau, K., S. G. Nadler, M. Saulnier, M. A. Tepper, and C. T. Walsh. 1994.Quantitation of the interaction of the immunosuppressant deoxyspergualin andanalogs with Hsc70 and Hsp90. Biochemistry 8:2561.

21. Binder, R. J., N. E. Blachere, and P. K. Srivastava. 2001. Heat shock protein-chaperoned peptides but not free peptides introduced into the cytosol are pre-sented efficiently by major histocompatibility complex I molecules. J. Biol.Chem. 276:17163.

22. Lagaudriere-Gesbert, C., S. L. Newmyer, T. F. Gregers, O. Bakke, andH. L. Ploegh. 2002. Uncoating ATPase Hsc70 is recruited by invariant chain andcontrols the size of endocytic compartments. Proc. Natl. Acad. Sci. USA 99:1515.

23. Steinman, L. 2001. Immunotherapy of multiple sclerosis: the end of the begin-ning. Curr. Opin. Immunol. 13:597.

24. Steinman, L. 1996. Multiple sclerosis: a coordinated immunological attackagainst myelin in the central nervous system. Cell 85:299.

25. Selmaj, K., C. F. Brosnan, and C. S. Raine. 1992. Expression of heat shockprotein-65 by oligodendrocytes in vivo and in vitro: implications for multiplesclerosis. Neurology 42:795.

26. Aquino, D. A., E. Capello, J. Weisstein, V. Sanders, C. Lopez,W. W. Tourtellotte, C. F. Brosnan, C. S. Raine, and W. T. Norton. 1997. Multiplesclerosis: altered expression of 70- and 27-kDa heat shock proteins in lesions andmyelin. J. Neuropathol. Exp. Neurol. 56:664.

212 HSP70 IN MHC CLASS II PRESENTATION

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

27. Bajramovic, J. J., H. Lassmann, and J. M. van Noort. 1997. Expression of �B-crystallin in glia cells during lesional development in multiple sclerosis. J. Neu-roimmunol. 78:143.

28. Lindstedt, R., N. Monk, G. Lombardi, and R. Lechler. 2001. Amino acid substi-tutions in the putative MHC class II “dimer of dimers” interface inhibit CD4� Tcell activation. J. Immunol. 166:800.

29. Ferreri, N. R., W. R. Herzog, and P. W. Askenase. 1993. Inhibition of IL-2-dependent proliferation by a prostaglandin-dependent suppressor factor. J. Im-munol. 150:2102.

30. Chauhan, B., A. Knutsen, P. S. Hutcheson, R. G. Slavin, and C. J. Bellone. 1996.T cell subsets, epitope mapping, and HLA-restriction in patients with allergicbronchopulmonary aspergillosis. J. Clin. Invest. 97:2324.

31. Nylandsted, J., M. Rohde, K. Brand, L. Bastholm, F. Elling, and M. Jaattela.2000. Selective depletion of heat shock protein 70 (Hsp70) activates a tumor-specific death program that is independent of caspases and bypasses Bcl-2. Proc.Natl. Acad. Sci. USA 97:7871.

32. Suh, W. C., C. Z. Lu, and C. A. Gross. 1999. Structural features required for theinteraction of the Hsp70 molecular chaperone DnaK with its cochaperone DnaJ.J. Biol. Chem. 274:30534.

33. Leung, K. C., M. J. Waters, I. Markus, W. R. Baumbach, and K. K. Ho. 1997.Insulin and insulin-like growth factor-I acutely inhibit surface translocation ofgrowth hormone receptors in osteoblasts: a novel mechanism of growth hormonereceptor regulation. Proc. Natl. Acad. Sci. USA 94:11381.

34. Cwiklinska, H., M. P. Mycko, O. Luvsannorov, B. Walkowiak, C. F. Brosnan,C. S. Raine, and K. W. Selmaj. 2003. Heat shock protein 70 associations withmyelin basic protein and proteolipid protein in multiple sclerosis brains. Int.Immunol. 15:241.

35. Singh-Jasuja, H., R. E. Toes, P. Spee, C. Munz, N. Hilf, S. P. Schoenberger,P. Ricciardi-Castagnoli, J. Neefjes, H. G. Rammensee, D. Arnold-Schild, andH. Schild. 2000. Cross-presentation of glycoprotein 96-associated antigens onmajor histocompatibility complex class I molecules requires receptor-mediatedendocytosis. J. Exp. Med. 191:1965.

36. Matsutake, T., and P. K. Srivastava. 2000. CD91 is involved in MHC class IIpresentation of gp96-chaperoned peptides. Cell Stress Chaperones 5:378.

37. Roth, S., N. Willcox, R. Rzepka, M. P. Mayer, and I. Melchers. 2002. Majordifferences in antigen-processing correlate with a single Arg71�–�Lys substi-tution in HLA-DR molecules predisposing to rheumatoid arthritis and with theirselective interactions with 70-kDa heat shock protein chaperones. J. Immunol.169:3015.

38. Panjwani, N., O. Akbari, S. Garcia, M. Brazil, and B. Stockinger. 1999. TheHSC73 molecular chaperone: involvement in MHC class II antigen presentation.J. Immunol. 163:1936.

39. Srivastava, P. K., A. Menoret, S. Basu, R. J. Binder, and K. L. McQuade. 1998.Heat shock proteins come of age: primitive functions acquire new roles in anadaptive world. Immunity 8:657.

40. Selmaj, K., C. F. Brosnan, and C. S. Raine. 1991. Colocalization of lymphocytesbearing �� T-cell receptor and heat shock protein hsp65� oligodendrocytes inmultiple sclerosis. Proc. Natl. Acad. Sci. USA 88:6452.

41. Birnbaum, G. 1995. Stress proteins: their role in the normal central nervoussystem and in disease states, especially multiple sclerosis. Springer Semin. Im-munopathol. 17:107.

42. Raine, C. S. 1994. Multiple sclerosis: immune system molecule expression in thecentral nervous system. J. Neuropathol. Exp. Neurol. 53:328.

43. Wucherpfennig, K. W., A. Sette, S. Southwood, C. Oseroff, M. Matsui,J. L. Strominger, and D. A. Hafler. 1994. Structural requirements for binding ofan immunodominant myelin basic protein peptide to DR2 isotypes and for itsrecognition by human T cell clones. J. Exp. Med. 179:279.

44. Vergelli, M., M. Kalbus, S. C. Rojo, B. Hemmer, H. Kalbacher, L. Tranquill,H. Beck, H. F. McFarland, R. De Mars, E. O. Long, and R. Martin. 1997. T cellresponse to myelin basic protein in the context of the multiple sclerosis-associ-ated HLA-DR15 haplotype: peptide binding, immunodominance and effectorfunctions of T cells. J. Neuroimmunol. 77:195.

45. Menoret, A., Z. Li, M. L. Niswonger, A. Altmeyer, and P. K. Srivastava. 2001.An endoplasmic reticulum protein implicated in chaperoning peptides to majorhistocompatibility of class I is an aminopeptidase. J. Biol. Chem. 276:33313.

46. Pepin, E., C. Villiers, F. Gabert, V. Serra, P. Marche, and M. Colomb. 1996. Heatshock increases antigenic peptide generation but decreases antigen presentation.Eur. J. Immunol. 26:2939.

47. Vanbuskirk, A., B. Crump, E. Margoliash, and S. Pierce. 1989. A peptide bindingprotein having a role in antigen presentation is a member of the hsp70 heat shockfamily. J. Exp. Med. 170:1799.

48. Beck, H., G. Schwarz, C. J. Schroter, M. Deeg, D. Baier, S. Stevanovic,E. Weber, C. Driessen, and H. Kalbacher. 2001. Cathepsin S and an asparagine-specific endoprotease dominate the proteolytic processing of human myelin basicprotein in vitro. Eur. J. Immunol. 31:3726.

49. Manoury, B., D. Mazzeo, L. Fugger, N. Viner, M. Ponsford, H. Streeter,G. Mazza, D. C. Wrath, and C. Watts. 2002. Destructive processing by asparagineendopeptidase limits presentation of a dominant T cell epitope in MBP. Nat.Immun. 3:169.

50. Agarraberes, F., S. Terlecky, and J. Dice. 1997. An intralysosomal hsp70 isrequired for a selective pathway of lysosomal protein degradation. J. Cell Biol.137:825.

213The Journal of Immunology

by guest on Novem

ber 17, 2017http://w

ww

.jimm

unol.org/D

ownloaded from