High efficiency of endogenous antigen presentation by MHC class II molecules

Transcription

1 International Immunology. Vol. 4, No. 10, pp Oxford University Press High efficiency of endogenous antigen presentation by MHC class II molecules Veronique Calin-Laurens, Frederique Forquet, Suzanne Lombard-Platet, Patrick Bertolino, Isabelle Chretien, Marie-Claude Trescol-Blemont, Denis Gerlier, and Chantal Rabourdin-Combe Immunobiologie Moleculaire, UMR 49, CNRS-ENS Lyon, Lyon Cedex 07, France Key words: antigen presentation, endogenous antigen, hemagglutinin, hen egg lysozyme, MHC class II molecules Abstract MHC class II molecules are Involved in the presentation of both exogenous and endogenous antigens to CD4 T cells. Using the trans-membrane hemagglutinin (HA) from measles virus and the secreted hen egg lysozyme (HEL) as antigen models, we have compared the efficiency of MHC class II presentation by naive antigen presenting cells (APCs) pulsed with exogenous antigen with that of their transfected counterparts synthesizing endogenous antigen. B cells expressing even a very low amount of trans-membrane HA were found to present endogenous HA to I-E d restricted T cell hybrldomas with a high efficiency whereas their naive counterparts required to be pulsed with a comparatively high amount of exogenous HA. Similarly, MHC class II presentation of endogenous secreted HEL was found to be much more efficient when compared with that of exogenous HEL. Biochemical studies did not reveal any enhanced intracellular degradation of endogenous HEL. As expected, HEL was released in the surrounding medium within <1 h. MHC class II presentation of endogenous HEL could not be explained by re-uptake by bystander APCs of HEL secreted In the surrounding medium. No sensitlzatlon of naive APCs could be observed either when co-cultured with HEL secreting cells or when cultured for 10 days with a sub-threshold amount of exogenous HEL. At the cell surface, I-E d molecules immunopreclpitated from HEL secreting cells were found to be slightly enriched In SDS-reslstant forms. These data raised the question of how peptides derived from endogenous transmembrane and secreted antigens can so efficiently reach an MHC class II loading compartment. Introduction T cells expressing an a0 TCR do not recognize an isolated nominal antigen but, rather, a complex associating MHC molecules with antigen derived peptides on an antigen presenting cell (APC) surface. Formation of peptide-mhc complexes in APCs thus requires antigen degradation into peptides and their association with MHC class I or class II molecules. Beside their distinct polypeptide structure, tissue expression, and respective CD8-TCR and CD4-TCR ligands, MHC class I and class II molecules differ by their intracellular trafficking. MHC class I molecules follow the 'default' secretory pathway (1,2) and MHC class II molecules, transiently associated with the invariant chain (li) (3), are routed to the endosomal compartment before expression at the cell surface (2,4-7). This difference if likely to be the key parameter which determines two functional sets of antigen derived peptides reaching the loading compartment of either MHC class I or MHC class II molecules. Indeed, accumulative data strongly support the endoplasmic reticulum (ER) as the functional loading compartment for MHC class I molecules (reviewed in 8) and the endosomal compartment has been postulated from very early experiments to be the major if not the only functional loading compartment for MHC class II molecules (9-12). Thus an important issue is to determine which antigen can give rise to peptides capable of reaching the functional MHC loading compartments and how they get there. We have undertaken a systematic approach using a given antigen to study the sensitization of APCs when the antigen is exogenousty supplied or endogenously synthesized and targeted to various APC compartments after transfection with the appropriate recombinant expression vector. We have previously reported that secreted and cytosolic endogenous hen egg lysozyme (HEL) but not exogenous HEL are presented by MHC class I molecules (13). In contrast, exogenous and secreted endogenous HEL but not cytosolic endogenous HEL are presented by MHC class II molecules (14). Accumulating reports Correspondence to: D. Gerlier Transmitting editor: B. Malissen Received 18 February 1992, accepted 22 June 1992

2 1114 Antigen presentation of MHC class II molecules have recently confirmed that MHC class II presentation occurs not only for exogenous antigens but also for endogenous antigens (14-21). This led us to ask what is the relative efficiency of MHC class II presentation of exogenous and endogenous antigens. We report here that, quite unexpectedly, both a trans-membrane and a secreted endogenous antigen appeared to be very efficiently presented by MHC class II molecules when compared with their exogenous counterparts. Methods Antigen and APCs HEL was purchased from Sigma Chemical Co. (St Louis, MO) and hemagglutinin (HA) was purified from measles virus infected cells and incorporated into disearoyl-phosphatidylcholine liposomes according to Gerlier et al. (22). The following cell lines maintained in Dulbecco's minimal essential medium (DMEM) supplemented with 6% FCS, 10 mm HEPES and 5x 10~ 5 M /Smercaptoethanol were used as naive APCs: H-^ M and H-2 k CH27 B lymphoma cells. From these cell lines, stable HEL secreting and HA expressing clones were derived after transfection with phmg-hels eukaryotic expression vector containing full-length HEL cdna (14) and with phmg-ha eukaryotic expression vector containing HA cdna (23) respectively. As markers of selection, pag475/2 coding for hygromycine resistance and prsv5 coding for mycophenolic acid resistance (24) were co-transfected. B cells were transfected by electroporation using the BioRad Gene Pulser (BioRad, Richmond, CA). Briefly, 10 7 M cells in 400 /il of DMEM supplemented with 10 mm HEPES containing 20 ^g of phmg-hels and 20 p.q of prsv5 were given an electric pulse of 250 V with the capacitance set at 960 /*F in a 4 mm width Gene Pulser cuvette. Electroporation of CH27 cells was performed using a 270 V electric pulse and pag475/2 as selection marker. One day after transfection, cells were washed once then cultured for 3 days before cloning in the presence of the relevant antibiotic. Stable transfectants secreting HEL or expression HA were selected either by detecting the protein [secreted HEL (HELs) or cell surface HA] or by deteoting the corresponding mrna using either a RNA cytodot or a Northern blot assay previously described (13,14). The following clones were used: M12.HELs3, CH27.HELs10, M12.HA.5.1, and M12.HA.25. In one experiment, M cells transfected with phmg-helc vector containing a signal sequence deleted HEL cdna (13,14) and expressing a cytosol targeted HEL (clone M12.HELc4) was also used as a control. The cell lines were regularly checked for the absence of mycoplasma contamination by growth on microbial medium. Before their use in the T cell stimulation assay, fibroblastic L cells were grown at high density. Radiolabelling and immunoprecipitation Cells were metabolically radidabelled and solubilized according to Calin-Laurens et al. (14). Briefly, 10 6 cells were washed twice in serum-free medium without methionine then labelled at 37 C for 3 h with 125 /tci of [^Jmethionine (Amersham, Les Ulis, France) in 0.5 ml of medium. The cells were then washed in PBS and lysed in buffer containing 2% Triton X100. After ultracentrifugation, supernatants were pre-cleared with normal mouse serum for 1.5hat4 C, then with 0.1 mg of Protein A - Sepharose (Pharmacia, Uppsala, Sweden) previously saturated with 50 ng of polyclonal rabbit anti-mouse antibodies (Biosys, Compiegne, France) for 1.5 h at 4 C. After centrifugation and elimination of Sepharose beads, immunoprecipitation was then performed with 50 ng of monoclonal mouse anti-ha antibody (cl55) (25), coupled with rabbit anti-mouse and Sepharose beads as above. After 2 h incubation at 4 C, Sepharose beads were recovered, washed twice in lysis buffer and processed for SDS - PAGE analysis. For pulse-chase studies, 15X10 6 cells were washed, incubated in methionine and cysteine free medium for 1 h and pelleted. The cell pellet was re-suspended in 50 /d (37 MBq) of psjmethionine and psjcysteine (Expre 3^3^, NEN-Dupont, Paris, France). After 15 min labelling, cells were quickly washed in medium, re-suspended in complete culture medium supplemented with 5 mm of cold methionine and cysteine, and incubated for up to 24 h. The cells were lysed as indicated above, and the extracts and cell-free supernatants were immunoprecipitated with a pool of antibodies directed against native and denatured HEL (14) in the presence of Protein A-Sepharose beads previously saturated with rabbit anti-mouse Ig antibodies. After extensive washes, the material was eluted from Protein A -Sepharose beads and loaded onto SDS polyacrylamide gel for electrophoresis. The material immunoprecipitated was revealed by autoradiography To study the SDS-resistant forms of the MHC class II expressed at the cell surface, 5x 10 6 M and M12.HELs3 viable cells were lodinated at 4 C with 0.5 mci 125 l using the lactoperoxydase procedure, washed, lysed in 2% NP-40, 6 mm CHAPS buffer (26), and then centrifuged. The supernatants were immunoprecipitated with I-E d specific S mabs and Protein A - Sepharose according to the procedure described by Germain and Hendrix (26). After washes, the immunoprecipitates were eluted in Laemli sample buffer without /3-mercaptoethanol for 30 min at room temperature. Half of the immunoprecipitates were boiled at 100 C for 2 min before analysis onto 12.5% SDS - PAGE and autoradiography. The autoradiographies from unboiled samples were scanned using a densitometer to determine the percentage of I-E d molecules in SDS-resistant forms, the total amount of I-E d molecules immunoprecipitated being determined by adding the signal of SDS-resistant forms (C forms) to that of free MHC class II chains (U forms). T cell hybndomas The two HEL-specific I-A k restricted 3A9 (27) and I-E d restncted B10D24.42 (kindly provided by E. Sercarz) were used in this study. The HEL epitope has been determined, HEL for 3A9 (27) and HEL for B10D24.42 T cell hybridomas (E. Sercarz, personal communication). The HA-specific T cell clone TH5.124 was obtained as previously described (23). 7" cell functional assay Specific antigen stimulation of the T cell hybridomas was performed by co-cultivating 10 s hybridoma cells and various amounts of APCs with or without various amounts of antigen in a final volume of 200 y\. After incubation for 20 h at 37 C in 96-well microplates, IL-2 production in supernatants was measured in a biological assay using the growth of IL-2- dependent CTL-L2 cell line revealed by using the MTT assay (14). Standard deviation of replicates was usually below 5%. In one experiment, M cells were grown in tissue culture

3 Antigen presentation of MHC class II molecules medium containing 10 ^g HEL/ml with re-seeding every 3-4 days at 5 x 10" cells/ml before their use in the T cell functional assay. Co-cultures were performed by seeding 2 x H-2d M12-HELs cells with 3 x 10* H-2* CH27 cells in 200 p\ of culture medium. After 24 h of incubation, 105 I-Ak restricted 3A9 T cells were added. As controls, 3A9 T cells were stimulated with 3 x 10" CH27 cells together with various amounts of HEL and various numbers of M12-HELs were used to stimulate I-Ed restricted B10D24.42 T cells. To quantify the relative efficiency of MHC class II presentation of exogenous and secreted endogenous antigen, the following calculations were done. The amount of HEL produced within 24 h by one HELs cell was determined by measuring the amount of HEL produced within 24 h by 105 cells in 200 pi of culture medium using an HEL-specific ELISA. In addition, HEL contents in HELs cells was estimated after running cell extracts onto SDS-PAGE and analysis by Western blot as previously described (14). Results APCs expressing a very low amount of cell-surface HA can efficiently stimulate MHC class ll-restricted T cells M B cells were used to present exogenous HA or endogenous transmembrane HA after transfection with phmg.ha recombinant expression vector to HA-specific T cell 1115 hybridomas. Both naive M B cells pulsed with an optimum concentration of HA and M12.HA.24 cells synthesizing a high amount of HA were able to stimulate the HA-specific TH5.124 T cell hybridoma (Fig. 1) as well as four other independent T cell hybridomas (data not shown). Moreover, MHC class II presentation of endogenous HA by B cells was found to be very efficient since only a few hundred M12.HA.24 cells stimulated the B10D24.42 T cells (Fig. 1B). In contrast, to induce a similar level of T cell activation, 3 x 10" naive M cells had to be pulsed with 3 /xg/ml of exogenous HA (Fig. 1A). In addition, a similar efficiency of MHC class II presentation of endogenous HA was observed when M12.HA.5.1, another HA transfectant expressing very little HA, was used to stimulate TH5.124 T cells (Fig. 1B). This was not due to the isolation of an M12 sub-clone particularly potent as APCs, since when M12.HA.24 and M12.HA.5.1 were used to present exogenous HEL to an I-Ed restricted and HELspecific T cell hybridoma, they were as potent APCs as the parental M cells (data not shown). The amount of HA synthesized by M12HA.24 and M12HA.5.1 cell clones, as evidenced after metabolic radiolabelling immunoprecipitation, and SDS - PAGE analysis (Fig. 2B), was found to correlate with the amount of HA mrna as determined in a Northern blot analysis (Fig. 2A). This indicates that the very low amount of HA (see the faint band in lane 2 of Fig 2B) detected in M12.HA.5.1 cell extracts is not due to synthesis of an unstable HA overdegraded in this particular sub-clone. The amount of HA in M12.HA.24 cells could not be directly determined due to the high cross-reactivity on the parental M cells of every anti-ha antibodies we have tested. These data indicated not only that endogenous trans- 0.4 o d z 1 03 cell number/well 10' 3 4 M 10" Fig. 1. Efficient presentation of endogenous HA to the MHC class llrestricted T cells TH5.124 hybridoma T cells (10s) were stimulated with 3X10 4 M B cells pulsed with exogenous HA (A) or with various numbers ol M12.HA (B) expressing low (M12HA5.1, open squares) or high(m12.ha.24, closed squares) levels of endogenous HA. After 24 h of stimulation, IL-2 was measured in the supernatants using the CTL-L2 bio-assay revealed by MTT assay Fig. 2. Synthesis of endogenous HA in cells transfected with phmg.ha expression vector. (A and B) The amount of HA synthesized by M12.HA.5.1 and M12.HA.24 cells clones correlates with the amount of HA mrna. (A) Northern blot of 20 pq of total RNA extracted from M (lane 1), M12.HA.5.1 (lane 2), and M12.HA.24 (lane 3) cells analysed using HA cdna probe. (B) Immunoprecipitation after metabolic radiolabelling of M12.HA.5.1 (lanes 1 and 2) and M.12.HA.24 (lanes 3 and 4) cells using normal mouse serum (lanes 1 and 3) or HA specific monoclonal antibody (lanes 2 and 4). M: 69 kda molecular weight marker.

4 1116 Antigen presentation of MHC class II molecules l l O ' l i g HEL 3.3 > 10 4 ng HEL 1.1 i IO 4 ng HEL Fig. 3. Synthesis and fate of endogenous HEL in cells transfected with phmg.hel expression vectors. (A) HEL mrna expression in M (lane 1), M12.HELc4 (lane 2), M12.HELs3 (lane 3). CH27 (lanes 4 and 40, and CH27 HELsiO (lanes 5 and 50 as detected by RNA cytodot. Rapid RNA extraction was performed for 108 cells, and each sample (two dilutions) was dotted on nitrocellulose then hybridized with ^P-radiolabelled BamH\ HELc cdna probe. HEL mrna was detected by autoradiography after 16 h (lanes 1-5 ) or 3 week (lanes 4' and 50 exposure. (B) Pulse-chase study of HEL secreted by M12 HELs3. Cells were pulsed with ["^Jmethionine and [^Icysteine for 15 min followed by a chase with an excess of cold methionine and cysteine for 0 (lanes 1 and 6), 1 (lanes 2 and 7), 2 (lanes 3 and 8), 3 (lanes 4 and 9), and 6 (lanes 5 and 10) h. Cell extracts (lanes 1-5) and cell free supernatants (lanes 6-10) were immunopreciprtated with HEL specific antibodies and analysed by SDS-PAGE. Molecular weight markers: 30, 21 5,14.3,6.5, and 3.4 kda. membrane HA was presented by MHC class II molecules but also that presentation of endogenous HA might be even more efficient than presentation of its exogenous counterpart. This prompted us to investigate the efficiency of MHC class II presentation of another endogenous antigen which would not be retained within the APCs but secreted. Endogenous HELs is very efficiently presented to MHC class IIresthcted T cells H-2* M and H-2k CH27 B cells were transfected with phmg.hels recombinant expression vector coding for a secreted form of HEL and, as control, M cells were also transfected with phmg.helc recombinant expression vector coding for an HEL form lacking the signal peptide and thus targeted to the cytoso) (14). The M12.HELs3 clone secreting up to 300 ng HEL/106 cells/24 h and the CH27.HELs10 secreting < 1 ng HEL/106 cells/24 h were used in this study. The amount of HEL mrna expressed in these transfectants correlates with the amount of HELs since in M12.HELs3 cells, HEL mrna was detected after only few hours of autoradiographic exposure of the RNA cytodot (Fig. 3A, lane 2) whereas a 3 week exposure was necessary to detect HEL mrna in CH27.HELs10 cells (Fig. 3A, lanes 5 and 50. As a control, M12.HELc4 expressing the cytosolic form of HEL contains as much HEL mrna as M12.HELs3 cells (Fig. 3A, lane 3). We have previously reported 0.34 slo 4 Dg HEl 10* I 0.3 call* ag HCLi In c t l t i 30 ng crttad HEL / 2 4 boars Fig. 4. Quantitative comparison between the I-Ed restricted presentation of exogenous HEL and that of endogenous HELs B10D24.42 T cells (105) were stimulated with vanous numbers of (A) naive H-2? M B cells and various amount of exogenous HEL or (B) M12.HELs3 cells secreting endogenous HEL (closed squares) in a final volume of 200 pi. As a control, various numbers of M12.HELc4 cells synthesizing cytosoj targeted HEL (open squares) were used as APCs. To allow a quantitative comparison between the efficiency of MHC class II presentation of exogenous versus endogenous HEL, the amounts of HEL and the amount of HEL secreted in the surrounding medium for 24 h (i.e. the length of the T cell stimulation assay) are indicated as two additional scales on the abscissa. These scales are valid only for M12 HELs3 cells (closed squares) and not for M12 HELc4 cells (open squares). (14) that HEL targeted to the cytosol is very short-lived (ty, < 5 mm) and is not secreted. The fate of HEL synthesized by M12.HELs3 cells was studied by pulse - chase analysis (Fig. 3B). After 15 min of radiolabelling, neo-synthesized HEL was found in cell extract and not in the supernatant. After a 1 h chase with cold amino acids, most of the radiolabelled HEL had already been released into the supernatant, and the release was completed after 3 h. If the intracellular HEL contents (10 ng/106 M12.HELs3 cells) are released every hour, we would expect an accumulation of 240 ng of HEL in the surrounding medium after 24 h, which is in agreement with the amount of HEL found to be secreted per 24 h (300 ng/106 M12.HELs3 cells). Further incubation of the cells never resulted in detectable re-uptake of radiolabelled HEL even after 24 h (data not shown) and no detectable degradation of radiolabelled HEL was observed in the supernatant A pulse-chase study using CH27.HELs10 cells showed a similar result (data not shown). M12.HELs3 cells were used as APCs to stimulate the I-Edrestricted HEL-speciflc B10D24.42 T cell hybridoma. As an

5 Antigen presentation of MHC class II molecules J 10' SO 0Ong KEL «7ngHB. Fig. 5. Quantitative comparison between the I-A k restricted presentation of exogenous HEL and that of endogenous HELs. 3A9 T 10 s cells were stimulated with various numbers of (A) naive H-2 k CH27 B cells and various amount of exogenous HEL or (B) CH27.HELs10 cells secreting endogenous HEL (closed squares) in a final volume of 200 y\ approach to compare the efficiency of MHC class II presentation of endogenous HELs and that of exogenous HEL, 10 5 B10D24.42 cells were either stimulated with various numbers of M cells in the presence of various amounts of exogenous HEL or with various numbers of M12.HELs3 cells (Fig. 4). Endogenous HELs was very efficiently presented by MHC class II molecules since as few as 100 transfected cells were able to significantly stimulate B10D24.42 T cells (Fig. 4B). We then tried to compare the efficiency of MHC class II presentation of endogenous HELs with that of exogenous HEL. A similar T cell activation signal was observed when B10D24.42 T cells were stimulated (in a final volume of 200 #d) either with 2X10 3 M cells fed with 10 5 ng of exogenous HEL for 24 h (i.e. the duration of the T cell stimulation) (Fig. 4A) or with 2x 10 3 M12.HELS3 cells secreting only 0.6 ng of HEU24 h and containing 0.02 ng of HEL (on its way to being secreted) (Fig. 4B). In contrast, cells expressing HEL targeted to the cytosol (M12.HELc.4) failed to stimulate the B10D24.42 T cell hybridoma (Fig. 4B). This was not due to an inherent inability of M12.HELc4 cells to process and present HEL, since they were as efficient as their parental counterpart in presenting exogenous HEL (data not shown). This high efficiency of MHC class II presentation of endogenous HELs was confirmed using the CH27 transfectant secreting a barely detectable amount of HEL (Fig. 5). CH27.HELs10, 5 x 10 4, (secreting <0.05 ng HEL/24 h) were found to stimulate the l-a k -restricted 3A9 T cell hybridoma as efficiently as 5 x 10 4 parental CH27 cells fed with 610 ng of exogenous HEL for 24 h. Re-uptake by bystander cells cannot account for MHC class II presentation of HELs Since HELs expressing APCs release the antgen into the surrounding medium, MHC class II presentation of endogenous HELs may be due to re-uptake of HELs by bystander cells. Co-culture between naive APCs and HEL secreting cells expressing different MHC class II haplotypes were performed to test this hypothesis. As shown in Rg. 6(A) co-cufturing naive H-2* CH27 B cells with a large amount of H-2«M12.HELs3 cells cannot lead to the presentation of HEL to l-a k -restricted 3A9 T cells. It should be stressed that 3A9 is a high avidity T cell hybridoma (27) and that the amount of HEL released into the supernatants of 2x10* M12.HELs3 cells (up to 60 ng/24 h) remains below the minimal amount of exogenous HEL (600 ng in 200 /d, see Fig. 5) required to sensitize CH27 cells for recognition by 3A9 cells at a detectable level. As control, M12.HELs3 cells (H-^ were potent stimulators of the rather low avidity l-e d -restricted B10D24.42 T cell hybridoma (Fig. 6B). Similar results were obtained using various co-culture combinations (data not shown). Thus, re-uptake of released HEL by bystander cells cannot account for the MHC class II presentation of endogenous HEL. Continuous pre-loading of APCs with exogenous HEL cannot result in MHC class II presentation as efficient as that of endogenous HELs Since short-term co-culture experiments between HEL secreting cells and naive APCs did not result in MHC class II presentation of any HEL after re-uptake, we undertook to mimic the growing conditions of HEL secreting APCs. M cells (5X10 4 cells/ml) were cultured and subcultured every 2-3 days for 10 days in the presence of 10 4 ng/ml of HEL (i.e to 1000-fold the amount of HEL which an equivalent number of M12.HELs cells could release in their surrounding medium). This amount of HEL was, however, well below the amount required to sensitize naive M cells during the T cell stimulation assay. When tested for their ability to stimulate B10D24.42 T cells, HEL pre-loaded M failed to present the HEL these cells could have engulfed during their growth. Unexpectedly, HEL preloaded M cells were found to present exogenous HEL added on the day of the assay no more efficiently than their naive counterparts (Rg. 7), and this suggests that these cells did not even express a sub-threshold number of HEL peptide-mhc class II complexes. In contrast, a similar number of M12.HELs3 cells could stimulate the hybridoma T cells without addition of any exogenous HEL. Addition of a large amount of exogenous HEL to M12.HELs cells further increased their ability to stimulate T cells as if they still had MHC class II molecules available for presentation of exogenous antigen. Increase in cell surface expression of SDS-resistant forms of I-E d molecules in cells secreting HEL Since it has been recently suggested that the SDS-resistant form of MHC class II molecules may represent the functional peptide-mhc class II complexes expressed at the cell surface (26,28-30) and could be up-regulated when APCs are fed with antigens such as HEL (26), the level of SDS-resistant forms of I-E d molecules at the cell surface was determined in unpulsed naive M cells and in M12.HELs3 cells. As shown in Rg. 8, two bands were observed by SDS - PAGE in the absence of

6 1118 Antigen presentation of MHC class II molecules A CH nM HEL CH27 + M12XEL8 M12J1EL8 CH Z of 3A9 T call reipom* S of BI0D2442 T call ntpoma 60 Fig. 6. HEL released into the supernatant of HELs transjected cells cannot sensitize naive APCs. Left panel. 3 x 10 4 H-2 k naive CH27 B cells were cultured for 24 h with 1 ^g/ml HEL or co-cultured with 2X10 5 H-2* 1 M12.HELs3 or then used to stimulate 10 s I-A K restricted 3A9 T cells for 24 h. As control, 1.3X10 3 M12 HELs3 cells and 10 5 naive M cells were used to stimulate 10* I-E d restricted B10D24.42 T cells (right panel). Results are expressed as per cent of the maximal T cell response observed using 5 x 10 4 naive CH27 cells in the presence of 500 ^g/ml HEL Oeft panel) or 5 x 10 4 M12 HELs3 cells (right panel). /3-mercaptoethanol when the I-E d specific immune-precipitates were not boiled, one migrating with an apparent mol. wt of ~70kDa (C arrow) and corresponding to the compact heterodimer forms, and one migrating with an apparent mol wt of 30 kda (U arrow) and corresponding to the /3 chain of unstable dimers. a chains were not detected since they were not efficiently lodinated. When the samples are boiled, the C dimer forms (or SDS-resistant forms) were converted into U forms. In two separate experiments, M12.HELs3 cells exhibited as slight increase in the proportion of SDS-resistant forms. From densitometry analysis, 44% of the total cell surface I-E d molecules immunoprecipitated from M12.HELs3 cells and 33% of I-E d from M cells were found to be in SDS-resistant forms. 800 Discussion We demonstrate here that MHC class II presentation of two endogenous antigens, the trans-membrane HA and the secreted HEL, is very efficient since APCs need to produce very little endogenous antigen to be able to activate T cells. In addition, in order to induce a similar T cell activation signal using a given number of APCs, naive APCs have to be fed with quite a high amount of exogenous antigen in comparison with the amount of endogenous antigen produced by transfected APCs. From the data shown on Fig. 4, the efficiency of the presentation of endogenous HELs can be estimated to be 160,000-fold more efficient than that of exogenous HEL (since 2x^(fi APCs secreting 0.6 ng HEL and 2x 10 3 naive APCs fed with 10 s ng of exogenous HEL gave the same T cell stimulation signal). The MHC class II presentation of endogenous antigens produced in a very low amount is most likety observed because recognition of only a few hundred peptide-mhc class II complexes can result in T cell activation (31). The highly efficient MHC class II presentation of endogenous trans-membrane HA can be explained because every HA molecule remains associated with the APCs and as such is potentially fully available for processing. Indeed, we could not Fig. 7. Pre-loading of naive M APCs with 10 nglm\ HEL for several days does not result in MHC class II presentation of HEL. I-E d restricted B10D24.42 T cells (10 s ) were stimulated with either 6 25X10 3 M12.HELs3 (closed squares) or naive M (open circles) or M cultured for 10 days with 10 pg/ml HEL prior to the T cell assay (closed circles) in the absence or presence of various amounts of exogenous HEL detect any HA shed in the surrounding medium and we have obtained preliminary data showing that HA expressed at the cell surface is spontaneously internalized by endocytosis. The highly efficient MHC class II presentation of endogenous secreted HEL is more difficult to understand. Most if not all endogenous HEL will ultimately leave the APCs within < 1 h after synthesis translocation into the ER as shown by the pulse - chase analysis. Thus, only a very small fraction of endogenous HEL could remain associated with the APCs and be available for antigen processing. It is unlikely that MHC class II presentation of HELs is due to a cytosolic degradation of HEL prior to the ER translocation because: (i) the amount of HELs synthesized correlates with the amount of HEL mrna expressed by the cells, (ii) the HELs is stable early on after synthesis and remains stable thereafter even after secretion in the surrounding medium, and

7 Antigen presentation of MHC class II molecules M M Fig. 8. Invnunopreciprtalion of 12sl-labelled cell surface I-Ed molecules from M Oanes 1 and 2) and M12.HELs3 Oanes 3 and 4). The immunoprectpitates were eluted with the Laemmli sample buffer at room temperature then run on SDS - PAGE untreated (lanes 1 and 3) or after boiling (lanes 2 and 4). M: 92.5, 69, 46, 30, 21.5, and 14.3 kda. (in) furthermore, cell transfectants expressing a high amount of mrna coding for HEL targeted to the cytosol and synthesizing a very unstable cytosolic HEL (14) cannot present up to now any detectable HEL derived peptide by MHC class II molecules. So far, two cell compartments have been considered to be the site of MHC class II loading with antigenic peptides, the endosomeyiysosome (9-12) and putatively the ER (32,33). MHC class II loading in the ER with peptides derived from endogenous secreted HEL would imply proteolytic cleavage within the ER. However, pulse-chase analyses did not reveal any degradation of HEL undergoing maturation. Using several APCs, and several I-Ak, I-Ek, I-Ad, and I-Ed restricted T cells (this paper and data not shown), we have found so far that all available anti-hel T cell hybridoma recognizing different HEL determinants are stimulated by every APC secreting HEL (and not by APCs synthesizing cytosolic HEL). These T cells have been derived and selected for their ability to recognize exogenous HEL, i.e. peptides derived from antigen cleavage by endosomal proteolytic enzymes. This in agreement with the report of Moreno et al. (21) indicating that in 'professional' APCs (i.e. B cells) endogenous and exogenous HEL can give rise to a similar set of peptidesalthough this may not be true when using certain 'non-professional' APCs. Therefore, the degradation of secreted endogenous HEL must given rise to a panel of peptides similar or identical to those derived from endocytosed exogenous HEL. This would imply that, if HEL degradation occurs in the ER, ER contains a proteolytic machinery able to generate such peptides. However, so far, ER and endosome/lysosome protein degradations involve different proteolytic enzymes which work in different physicochemical conditions and which can be distinguished by their sensitivity to various inhibitors (34,35). Moreover, ER loading of MHC class II molecules with HEL-derived peptides would occur only on aft heterodimers not associated with li which has been 1119 shown to compete in vivo (36) and in vitro (37) with peptide binding. However, since we have used B cells expressing an excess of invariant chain li, free or/? heterodimers are unlikely to be available. In addition, MHC class II presentation of both exogenous HA and HEL is enhanced when li is expressed (23,38), and we have some evidence that presentation of their endogenous counterparts is also increased when li is expressed (23 and manuscript in preparation). Taken together these data strongly suggest that formation of MHC class II-peptide complexes in the ER from endogenous HEL is unlikely. Thus, both endogenous trans-membrane HA and HELs are likely to have the capability to reach the endosomal compartment relevant for peptide - MHC class II complex formation with a high efficiency. Delivery of HA to the endosome/lysosome compartment is likely to occur since membrane expressed HA undergoes spontaneous internalization (data not shown). In addition, immunofluorescence studies suggest that HA may also localize partly in some MHC class II containing cytoplasmic vesicles. This is currently under investigation using confocal immunofluorescence microscopy. The high efficiency of MHC class II presentation of secreted endogenous HEL indicates that an HEL fraction large enough is carried over into the endosomal compartment whereas only a low fraction of exogenous antigen would have access to this compartment. If the high endosomal loading occurs through endocytosis, it implies the re-uptake by the same cell of some secreted HEL locally concentrated within a very short distance from the plasma membrane since re-uptake by bystander cells cannot play a major role (see co-culture experiments). In addition, attempts to mimic the efficiency of the presentation of endogenous HELs by culturing naive APCs in the presence of a HEL concentration far exceeding the amount of HEL which secreting cells can accumulate in their surrounding medium failed. This indicates that a particularly efficient routing of some endogenous HELs to the endosome/lysosome compartment does exist. Endogenous HEL may be efficiently taken back at the surface of the same cell through direct HEL 'pouring' from exocytic vesicles to adjacent endocytic vesicles on-going internalization. Since HEL is an highly basic protein, its re-uptake may also be facilitated by its non-specific binding onto negatively charged cell surface proteins undergoing internalization. Indeed, a very efficient antigen uptake through a receptor mediated process such as the one observed when antigen specific B cells are used as APCs can result in a comparable to 100,000-fold increase in the efficiency of MHC class II presentation of an exogenous soluble antigen (39). Alternatively, some HELs may leak from the secretory pathway to the endosomal compartment through, for example, passive transport by vesicles involved in the transport of targeted molecules (MHC class II molecules, mannose-6-phosphate receptor, etc.) from the trans-golgi network to the late endosomal/pre-lysosomal compartment. We thus favor that MHC class II loading with peptides derived from endogenous HELs (and trans-membrane HA) occurs in the endosome/lysosome compartment as documented for the loading with peptides derived from exogenous antigen (9-12). Accordingly, Adorini et al. (40) have recently shown that exogenous peptides compete for MHC class II presentation of both exogenous and endogenous antigens, suggesting a lack of strict compartimentalization between endogenous and exogenous pathways of antigen presentation. Moreover, at the cell surface of HEL secreting cells, the proportion

8 1120 Antigen presentation of MHC class II molecules of SDS-resistant I-E d molecules was found to be slightly increased when compared with that observed in their naive counterparts and this is in agreement with the strong increase in SDS-resistant I-A k when B cells have been fed with exogenous HEL observed by Germain and Hendrix (26). The SDS-resistant form appears simultaneously to the li dissociation from MHC class II molecules (26,30) and is acquired in endosomes (30). Although we cannot ascertain that the increase in SDS-resistant forms of I-E d molecules observed in HEL secreting cells corresponds to the true accumulation of stable complexes made of HEL derived peptides and I-E d molecules, this further strengthens the hypothesis that the MHC class II processing pathway of endogenous HELs may not be different from that of exogenous HEL. MHC class II presentation of secreted endogenous antigen was found to be very efficient when compared with that of exogenous antigen, since APCs synthesizing a very low amount of antigen express a number of peptide- MHC class II complexes sufficient to activate specific T cells. This indicates that every MHC class II expressing cell of an organism must express a large number of complexes associating MHC class II molecules with peptides derived from self-proteins synthesized within the cell and, accordingly, the peptides eluted from MHC class II molecules have been recently identified to be derived mostly from cell components (41). Therefore, not only MHC class I (reviewed in 8) but also MHC class II molecules are loaded with peptides derived mostly from the self as previously predicted (42) This phenomenon must be physiologically relevant and play a role in shaping the T cell repertoire during thymic education and maintenance of self-tolerance. Trans-membrane and secreted self proteins are likely to play a role in positive and/or negative selection of CD4 T cells and/or in anergizing CD4 T cells in peripheral tissue. As another consequence and in agreement with very old observations made on the induction of the primary response, MHC class II presentation of exogenous soluble foreign proteins in vivo is likely to be of a rather limited efficiency and this could explain the usual requirement for an adjuvant. Acknowledgements This work was supported in part by grant from INSERM (D G. no ) ARC (C.R.-C. no. 6108), and MNES (C R.-C ). We thank N. Koch, B. Malissen, P. Allen, E Sercarz, E. Long, and A. Garapin for providing us with cell lines or plasmids; J. Salamero, J. Davoust, and C. Lethias for microscopic analysis; and F. Cretin and D. Naniche for helpful discussions and critical reading of the manuscript. Abbreviations APC DMEM ER HA HEL HELs References antigen presenting cell Dulbecco's minimal essential medium endoplasmic reticulum hemagglutinin hen egg lysozyme secreted HEL 1 Krangel, M. S., Orr, H. T., and Strominger, J. L Assembly and maturation of HLA-A and HLA-B antigens in vivo. Cell 18:979 2 Neefjes, J. J., Stoflorz, V., Peters, P. J., Geuze, H. J., and Ploegh, H. L The biosynthetic pathway of MHC class II but not class I molecules intersects the endocytic route Cell 61: Kvist, S., Wiman, K., Claesson, L, Peterson, A., and Dobberstein, B Membrane insertion and oligomeric assembly of HLA-DR histocompatibility antigens. Cell 29:61. 4 Cresswell, P Intracellular class II HLA antigens are accessible to transferrin - neuroaminidase conjugates internalized by receptormediated endocytosis Proc. NatlAcad. Sci. USA Lotteau, V., Teyton, L., Peleraux, A., Nilsson, T., Karisson, L, Schmid, S., Quaranta, V., and Peterson, P A Intracellular transport of class II MHC molecules directed by invariant chain. Nature Bakke, O. and Dobberstein, B MHC class ll-associated invariant chain contains a sorting signal for endosomal compartments Cell 63: Peters, P J., Neefjes, J. J., Oorschot, V., Ploegh, H. L., and Geuze, H. J Segregation of MHC class II molecules from MHC class I molecules in the Golgi complex for transport to lysosomal compartments. Nature Geriier, D., Calin-Laurens, V, Bertdino, P., and Rabourdin-Combe, C Molecular biology of antigen presentation. In Fornusek, L., ed., Immune System Accessory Cells, p CRC Press, Boca Raton, FL. 9 Ziegler, H. K. and Unanue, E. R Decrease in macrophage antigen catabolism caused by ammonia and chloroquine is associated with inhibition of antigen presentation to T cells. Proc. Natl Acad. Sci. USA 79: McCoy, K. L, Miller, J., Jenkins, M, Ronchese, F., Germain, R. N., and Schwartz, R H Diminished antigen processing by endosomal acidification mutant antigen-presenting cells. J. Immunol Guagliardi, L E, Koppelman, B., Blum, J. S., Marks, M. S., Cresswell, P., and Brodsky, F. M Co-localization of molecules involved in antigen processing and presentation in an early endocytic compartment. Nature Harding, C. V., Unanue, E. R., Slot, J. W., Schwartz, A. L, and Gueze, H. J Functional and ultrastructural evidence for intracellular formation of major histocompatibility complex class II-peptide complexes during antigen processing. Proc. Natl Acad. Sci. USA 87: Forquet, F., Calin, V, Trescol-Biemont, M. C, Kanellopoulos, J., Mottez E, Kourilsky, P., Rabourdin-Combe, C, and Geriier, D Generation of hen egg lysozyme-specific and major histocompatibility complex class l-restricted cytolytic T lymphocytes: recognition of cytosolic and secreted antigen expressed by transfected cells. Eur. J. Immunol. 20: Calin-Laurens, V., Forquet, F., Mottez, E., Kanellopoulos, J, Godeau, F., Kourtsky, P., Geriier, D., and Rabourdin-Combe, C Cytosolic targeting of hen egg lysozyme gives rise to a short-lived protein presented by class I but not by class II major histocompatibility complex molecules. Eur. J. Immunol Jacobson, S., Sekaly, R. P., Jacobson, C. L., McFariand, H. F., and Long E. O MHC class ll-restricted presentation of cytoplasmc measles virus antigens to cytotoxic cells J. Virol Jaraquemada, D., Marti, M., and Long, E. O An endogenous processing pathway in vaccinia virus-infected cells for presentation of cytoplasmic antigens to class ll-restricted T cells. J. Exp. Med. 172: Yurin.V. L, Rudensky, A. Y.,Mazel,S. M., and Blechman, J. M Immunoglobulin-specific T-B cell interaction II T cell clones recognize the processed form of B cell own surface immunoglobulin in the context of the major histocompatibility complex class II molecule. Eur. J. Immunol. 19: Weiss, S and Bogen, B MHC Class ll-restricted presentation of intracellular antigen. Cell 64: Bikoff, E Class ll-restricted lgg2a b -specific T cells recognize a signal-minus form of the V-CH3 b antigen. Eur J. Immunol. 21: Brooks, A., Hartley, S., Kjer-Nielsen, L, Perera, J., Goodnow, C. C, Basten, A, and McCluskey, J Class ll-restncted presentation of an endogenously derived immundominant T-cell determinant of hen egg lysozyme. Proc. Natl Acad. Set. USA 88: Moreno, J., Vignali, D. A. A., Nadimi, F., Fuchs, S., Adorini, L., and Hammerting, G. J Processing of an endogenous protein can generate MHC class ll-restncted T cell determinants distinct from those derived from exogenous antigen. J. Immunol. 147:3306.

9 Antigen presentation of MHC class II molecules Geriier, D., Gamier, F.. and Forquet, F Haemaggkrtinin of measles virus: purification and storage with preservation of biological and immunotogical properties. J. Gen. Virol. 69: Bertolino, P., Forquet, F., Pont, S., Koch, N., Gerlier, D., and Rabourdin-Combe, C Correlation between invariant chain expression level and capability to present antigen to MHC class II restricted T cells. Int. Immunol. 3: Long, E. O., Rosen-Bronson, S., Karp, D. R., Maltani, M., Sekaiy, R. P., and Jaraquemada, D Efficient cdna expression vectors for stable and transient expression of HLA-DR in transfected fibroblast and lymphoid cells. Hum. Immunol. 31: Giraudon, P and Wild, T. F Correlation between epitopes on hemagglutinin of measles virus and biological activities: passive protection by monoclonal antibodies is related to their hemagglutination inhibiting activity. Virology 144: Germain, R N. and Hendnx, L. R MHC class II structure, occupancy, and surface expression determined by post-endoplasmic reticulum antigen binding. Nature 353: Allen, P. M., Matsueda, G. R., Haber, E., and Unanue, E. R Specificity of the T cell receptor: two differents determinants are generated by the same peptide and the I-A k molecule. J. Immunol. 135: Dornmair, K., Rothenhausler, B, and McConnell, H. M Structural Intermediates in the reactions of antigenic peptides with MHC molecules. Cold Spring Harbor Symp. Quant. Biol Davidson, H W.,Reid, P. A., Lanzavecchia, A, and Watts, C Processed antigen binds to newly synthesized MHC class II molecules in antigen-specific B lymphocytes. Cell Neefjes, J. J. and Ploegh, H. L Inhibition of endosomal proteolytic activity by leupeptin blocks surface expression of MHC class II molecules and their conversion to SDS resistant a/3 heterodimers in endosomes. EMBO J Harding, C. V. and Unanue, E R Quantrtation of antigenpresenting cell MHC class ll/peptide complexes necessary for T-cell stimulation. Nature Sekaiy, R. P., Jacobson, S., Richert, J R, Tonnelle, C, McFarland, H F., and Long, E Antigen presentation to MHC class ll-restricted measles virus-specific T-ceB clones can occur in the absence of the invariant chain. Proc. Natl Acad. Set. USA Nuchtern, J. G.. Biddison, W. E., and Klausner. R. D Class II MHC molecules can use the endogenous pathway of antigen presentation. Nature 343: Wikstrom, L. and Lodish, H. F Nonlysosomal, pre-gcrfgi degradation of unassembled asialoglycoprotein receptor subunits: a TLCK- and TPCK-sensitive cleavage within the ER. J. Cell Biol. 113: Stafford, F. and Bonifacio, J. S A permeabilized cell system identifies the endoplasmic reticulum as a site of protein degradation J. Cell Biol. 115: Teyton, L, O'SulHvan, D., Dickson, P., Lotteau, V., Sette, A., Fink, P, and Peterson, P Invariant chain distinguishes between the exogenous and endogenous antigen presentation pathways. Nature Roche, P. A and Cresswell, P Invariant chain association with HLA-DR molecules inhibits immunogenic peptide binding. Nature Nadimi, F., Moreno, J., Momburg, F., Heuser, A., Fuchs, S., Adorini, L., and Hammerling, G J 1991 Antigen presentation of hen egg-white lysozyme but not of ribonuclease A is augmented by the major histocompatibility complex class ll-associated invariant chain. Eur. J. Immunol. 21: Lanzavecchia, A Antigen uptake and accumulation in antigenspecific B cells Immunol Rev. 99:39 40 Adorini, L., Moreno, J., Mombourg, F., Hammerling, G J., Guery, J. C, Valli, A., and Fuchs, S Exogenous peptides compete for the presentation of endogenous antigens to major histocompatibility complex class-ll restricted T cells. J. Exp. Med. 174: Rudensky, A. Y., Preston-Hurlburt, P., Hong, S -C, Barlow, A, and Janeway, C. A Jr Sequences analysis of peptides bound to MHC class II molecules. Nature Kourilsky, P. and Clavene, J. M The peptidic self model, a hypothesis on the molecular nature of the immunological self. Ann. Inst. Pasteur Immunol 137D:3.

10