Autophagic Compartments Gain Access to the MHC Class II Compartments in Thymic Epithelium

The recognition of self-peptides in the context of MHC molecules on cortical thymic epithelial cells (TECs)² is critical for double-positive naive T cells to acquire MHC-restriction by positive selection, whereas the recognition of such complexes on medullary TECs is crucial for CD4⁻ or CD8-single positive T cells to acquire central tolerance by negative selection (1). The formation of the complexes composed of nominal antigenic peptides and MHC molecules in professional APCs, such as macrophages and dendritic cells, has been extensively studied (2, 3). However, the mechanisms that underlie the formation of the complexes composed of self-peptides and MHC molecules in both cortical and medullary TECs are poorly understood.

Concerning the MHC class II-restricted Ag presentation, professional APCs mainly use three types of intracellular delivery pathways for nominal Ags into the MHC class II compartments. These pathways are the endocytic pathway, in which endocytosed Ags are delivered from endosomes or phagosomes into the MHC class II compartments (4); the constitutive secretory pathway, in which Ags endogenously synthesized in endoplasmic reticulum are delivered into the MHC class II compartments through the trans-Golgi network (4); and the autophagic pathway, in which cytoplasmic and nuclear Ags are sequestered in autophagosomes and delivered into the MHC class II compartments (5, 6, 7). After the Ags are digested into peptides, complexes of MHC class II αβ heterodimers and the antigenic peptides (referred to as MHC class II plus peptide complexes) are formed in the MHC class II compartments.

In in vitro-established TEC lines, endocytosed Ags are delivered into the MHC class II compartments via the endocytic pathway and endogenous Ags in the endoplasmic reticulum are delivered into the MHC class II compartments via the constitutive secretory pathway (8, 9, 10, 11, 12). In addition, autophagy in TECs is critically involved in shaping the CD4-single positive T cell repertoire and establishing central tolerance (13). However, the intersection of autophagy with the MHC class II-restricted Ag-presentation pathway in TECs has not been verified. In general, autophagy is the bulk degradation process of cytoplasmic components, including organelles, to generate recycled amino acids during stress and starvation (14). However, thymic autophagy constitutively occurs irrespective of nutrient conditions (14). In the autophagy process, some of the cytoplasmic components are sequestrated in autophagosomes and then delivered into autolysosomes, which are formed after the fusion of autophagosomes with lysosomes. Therefore, it appears that thymic autophagy may be extensively involved in lysosomal delivery and the degradation of cytoplasmic Ags in thymic stromal cells, and may intersect the MHC class II presentation pathway, in which the complexes of thymic peptides with MHC class II molecules are formed. To investigate this possibility, we examined the expression and the localization profiles of LC3 by Western blot analysis and confocal microscopy in the in vitro-established TEC lines, in which both the autophagy process and the formation of MHC class II plus peptide complexes were arrested by pepstatin A, an aspartyl protease inhibitor, and E64d, a cysteine protease inhibitor. In the presence of these protease inhibitors, LC3 was detected in cell lysates from TEC lines and was colocalized in compartments that contained LAMP-1, H2-DM, and MHC class II plus class II-associated invariant chain peptides (CLIP) complexes. To address whether the autophagic process had access to the MHC class II presentation pathway in thymus, we examined the expression and the localization profiles of LC3 on thymic cryosections of 1-day-old mice by confocal microscopy. We found that LC3 colocalized with the compartments in which MHC class II plus CLIP complexes as well as H2-DM molecules resided. These results suggest that the autophagic process may intersect the MHC class II presentation pathway not only in the in vitro established TEC lines but also in TECs in situ. We discuss the role of the autophagic process in the MHC class II-restricted presentation of thymic self-peptides in light of the thymic selection of T cell repertoires.

Cortical and medullary TEC lines were established and cultured as previously described (15).

We obtained C57BL/6 mice from Japan SLC and performed all animal experiments according to the guidelines of the Institutional Animal Care and Use Committee of Nagasaki University.

The following Abs were purchased and used in this study: rabbit anti-pan cytokeratin Ab (A0575; DakoCytomation), rabbit anti-mouse cytokeratin 5 Ab (Covance), mouse anti-cytokeratin 8 mAb (PROGEN Biotecnik), mouse mAb to LC3 (clone 4E12; MBL International), and FITC-conjugated rat anti-mouse CD107a (LAMP-1) mAb (BD Pharmingen). Rabbit antisera against LC3 were raised against thioredoxin-human LC3 and purified by affinity chromatography on a glutathione-S-transferase-immobilized human LC3-Sepharose column (16). We obtained Y3P mouse mAb against mouse I-A^b (17) from Dr. Saizawa (Nihon Medical College, Tokyo, Japan), M5–114 rat mAb against mouse I-A^b,d (18) from Dr. Uchida (National Institute of Infectious Diseases, Tokyo, Japan), In-1.1 mAb against mouse Ii (19) from Dr. G. J. Hämmerling (German Cancer Research Center, Heidelberg, Germany), and 30-2 mAb, which reacts with the I-A^b molecules associated with CLIP (20), from Dr. A. Rudensky (Howard Hughes Medical Institute, University of Washington School of Medicine, Seattle, WA). Rabbit antisera to H2-DMβ2 chain was raised against a keyhole limpet hemocyanin-coupled synthetic peptide derived from the cytoplasmic tail of the H2-DMβ2 chain (RKSHSSSYTPLPGSTYPEGRH) (21). We purified the antisera with the synthetic peptide-conjugated, epoxy-activated Sepharose 6B and designated it as anti-H2-DM Ab after purification. Abs were labeled with biotin or Alexa Fluor according to the manufacture’s instructions (Molecular Probes).

TEC lines were cultured on eight-hole heavy Teflon-printed glass slides (Thermo Fisher Scientific) in 10% FCS-DMEM and then incubated in DMEM containing 10% FCS and 10³ IU/ml IFN-γ (PeproTech) for 72 h to express MHC class II molecules and their related molecules. The culture medium was replaced with fresh medium every 24 h. After incubation with IFN-γ, the TEC lines were incubated with DMEM containing 10% FCS or DMEM containing 10% FCS and 6 μg/ml E64d and 6 μg/ml pepstatin A for 4 h. The TEC lines were incubated for 10 min in serum-free medium at room temperature and then fixed with PBS containing 4% paraformaldehyde (pH 7.0) for 10 min at room temperature. Slides were treated with 0.1 M glycine (adjusted to pH 7.0 with 1 M Tris) to block the remaining paraformaldehyde activity and then washed three times with TBS for 10 min. To stain the TEC lines with Abs against intracellular compartments, the cells were permeabilized with TBS containing 0.05% saponin for 10 min after fixation and then washed three times with TBS. The TEC lines were incubated overnight with 1–3 μg/ml of the primary Ab in 1% BSA-TBS in a refrigerator. After washing three times with TBS, the samples were incubated with 1 μg/ml Alexa Fluor 546-conjugated anti-IgG Ab in 1% BSA-TBS for 3 h at room temperature. After washing three times with TBS, the samples were incubated overnight in a refrigerator with 1–3 μg/ml biotin-conjugated secondary Ab in 1% BSA-TBS or 3 μg Alexa Fluor 488-conjugated secondary Ab or 1 μg of secondary Ab labeled by Zenon Alexa Fluor 488 IgG labeling reagent (Molecular Probes). After incubation with biotin-conjugated Ab, the samples were washed with TBS three times and then incubated with 1 μg/ml Alexa Fluor 488-conjugated streptavidin for 3 h at room temperature. After washing three times with TBS, the slides were sealed with PBS(−)-glycerol (for fluorescence microscopy; Merck) solution (1:9).

Thymuses from 1-day-old mice were removed, embedded in optimal cutting temperature (O.C.T) compound, and frozen in liquid nitrogen. To stain with Abs against cytokeratins, the 5-μm cryosections from an O.T.C-embedded thymus were fixed with cold acetone (kept at 4°C) for 10 min and washed three times with TBS. The cryosections were incubated overnight with 3 μg/ml Alexa Fluor 546-conjugated primary Ab in 1% BSA-TBS in a refrigerator. After washing three times with TBS, the samples were incubated overnight in a refrigerator with 3 μg/ml Alexa Fluor 488-conjugated secondary Ab or 1 μg of secondary Ab labeled by Zenon Alexa Fluor 488 rabbit IgG labeling reagent in 1% BSA-TBS. After the slides were washed three times with TBS, they were sealed with PBS(−)-glycerin solution (1:9).

The sealed samples were imaged with a Zeiss LSM 510 confocal microscope equipped with a ×64 1.4 NA plan-apochromat oil-immersion lens or a ×40 1.2 NA C- apochromat water-immersion lens (Carl Zeiss). A 488-nm Ar laser line and a 545-nm He-Ne laser line were used for excitation of Alexa Fluor 488 and Alexa Fluor 543, respectively. Emission wavelengths were separated by band pass (505–530 nm) and long pass (560 nm) filters, respectively.

Pictures were taken with the LSM510 confocal software and were assembled into RGB images with Photoshop (Adobe Systems). Quantification of red, green, or yellow fluorescing dots in the pictures was performed with the image analysis software MetaMorph (Molecular Devices).

Magnetic beads (M-450 Dynabeads conjugated with anti-rabbit IgG Ab; Dynal Biotech) were conjugated with rabbit anti-H2-DM Ab according to the manufacturer’s protocol. The TEC lines were cultured in 100-mm dishes in DMEM with 10% FCS and then incubated in DMEM with 7.5% FCS and without IFN-γ or with 10³ IU/ml IFN-γ (PeproTech) for 72 h to express MHC class II complexes and their related molecules. The culture medium was replaced by fresh medium every 24 h. The TEC lines were incubated in 7.5% FCS-DMEM or 7.5% FCS-DMEM containing 6 μg/ml E64d and 6 μg/ml pepstatin A for 4 h. After incubation, these TEC lines were rinsed once with 5 ml of PBS and once with 2.0 ml of 0.25M sucrose solution containing 0.02% EDTA, 1 mM PMSF, and 0.1 mM tosylphenylalanyl chloromethyl ketone (TPCK). Then the cells were scraped with 1.0 ml of 0.25M sucrose solution containing 0.02% EDTA and protease inhibitors (1 mM PMSF, 0.1 mM TPCK, 10 μg/ml aprotinin, and 0.1 mM leupeptin). The cell suspension was placed into a 15-ml plastic tube and homogenized by passing the 0.25M sucrose solution 20 times through a 23-gauge needle attached to a disposable 1-ml plastic syringe. The postnuclear supernatant (PNS) was obtained by centrifugation at 1000 × g for 10 min. Rabbit anti-H2-DM Ab-conjugated Dynabeads M450 (2 × 10⁷ beads) were added to 1 ml of PNS. The mixture was rotated gently for 24 h in a refrigerator. The magnetic bead fraction was washed twice with PBS with the aid of a magnet. The fraction was stored at −80°C until used for SDS-PAGE.

To analyze the H2-DM-positive vesicles bound to the magnet beads by SDS-PAGE and Western blotting, 50 μl of SDS-sample buffer without 2-ME was added to the magnetic bead-fraction, and the proteins included in the H2-DM-positive vesicles were eluted. We diluted the eluate 1/4 with H₂O and determined its protein concentration by the Bradford method (Bio-Rad) according to the manufacturer’s protocol. We added an equal volume of sample buffer containing 10% (v/v) of 2-ME to each sample. Proteins (1 μg), with or without boiling, were separated on a 12.5% polyacrylamide gel. For Western blotting, proteins were transferred from polyacrylamide gels to polyvinylidene difluoride membranes (Immobilon-P transfer membrane; Millipore). The membranes were blocked with TBS containing 5% BSA and probed with 1–2 μg/ml the primary Ab in TBS containing 1% BSA and 0.02% NaN₃. The membranes were washed once with TBS containing 0.02% Tween 20 and twice with TBS. Then primary Abs were probed with HRP-conjugated anti-Ig Ab in TBS containing 1% BSA, followed by washing once with TBS containing 0.02% Tween 20 and twice with TBS. Signals were detected by using super signal chemiluminescent substrate (Pierce).

Values of p were calculated by the Student’s t test (Microsoft Excel software) with two-tailed distribution and two-sample unequal variance parameters.

Both the cortical TEC (cTEC) and medullary TEC (mTEC) lines express MHC class II as well as MHC class II-related molecules, such as Ii chains and H2-DM molecules, upon stimulation with IFN-γ (15). In both TEC lines, MHC class II molecules are localized in H2-DM-positive compartments and on the plasma membrane. Ii chains and the MHC class II plus CLIP complexes are also localized in the H2-DM-compartments (10, 11, 12). To determine whether the autophagic process in both TEC lines is involved in the MHC class II-Ag presentation pathway, we examined the expression levels and localization profiles of LC3 by Western blot analysis and immunofluorescence staining. In the cell lysates from both TEC lines stimulated with IFN-γ, the rabbit anti-LC3 Ab detected LC3 as a membrane-bound form, i.e., LC3-II, which is ∼16 kDa in molecular mass (Fig. 1 A) (16). This Ab hardly detected a cytosolic form of LC3, i.e., LC3-I, which is ∼18 kDa in molecular mass. In fact, this Ab may be more likely to react to LC3-II than to LC3-I in HeLa cell and A431, which are originated from epithelium; however, this Ab reacts to LC3-I as well as LC3-II in HEK 293 cell (16). The expression level of LC3-II was apparently increased when the TEC lines were precultured in the presence of both 6 μg/ml pepstatin A (an aspartyl protease inhibitor) and 6 μg/ml E64d for 4 h. The combination of pepstatin A and E64d arrests the autophagic process at the stage of autophagosomes or autolysosomes; thus, these vesicles accumulate in the cytoplasm of the treated cells (16). Therefore, the increased levels of LC3-II in lysates of both TEC lines may be due to the accumulation of the autophagic compartments in the cytoplasm.

Immunostaining analyses revealed that the intracellular compartments bound with LC3-II (referred to as LC3-positve compartments) were clearly probed with the anti-LC3 Ab in the cytoplasm of both TEC lines stimulated with IFN-γ (Fig. 1,B, b and e). Furthermore, when both TEC lines were precultured in medium containing the protease inhibitors, pepstatin A and E64d (referred to as the protease inhibitors hereafter), the amounts of LC3-positve compartments were increased in the cytoplasm of both the cTEC and the mTEC lines (Fig. 1,B, c and f). When the amounts of LC3-positve compartments shown in Fig. 1,B were quantified, they were increased with statistical significance in the cytoplasm of the both TEC lines after incubation with the protease inhibitors (Fig. 1 C). These results suggested that, in the presence of protease inhibitors, the increased amount of LC3-II in the cell lysates is attributed to the accumulation of the LC3-positive compartments in the cytoplasm of both TEC lines.

In the same culture condition, ∼50% of the LC3-positive compartments in the cTEC line and 32% of the LC3-positive compartments in the mTEC line were costained with anti-LAMP-1 Ab (Table I and Fig. 1, D–E), suggesting that the autophagic process was arrested at the autolysosome stage. Furthermore, ∼40% of the LC3-positive compartments in both TEC lines were costained with both anti-H2-DM Ab (Table I and Fig. 1, F–G) and with anti-MHC class II Ab, respectively (Table I and Fig. 1, H–I). Taken together, in both TEC lines, LC3-II seemed to be colocalized in the lysosomal compartments that contain both MHC class II and H2-DM molecules. However, there was a difference in the ratio of the colocalization of LC3-positive compartments with LAMP-1 between in the cTEC line and the mTEC line, suggesting that the nature of the autolysosome in the cTEC line may be subtly different from that in the mTEC line (10, 11, 12).

The combination of pepstatin A and E64d not only arrests the autophagic process by accumulation of autophagosomes/autolysosomes in the cytoplasm, but also affects both the generation of CLIP from Ii chains and the dissociation of CLIP from MHC class II αβ heterodimers (22, 23, 24). Thus, the MHC class II plus CLIP complexes accumulate in the H2-DM-positive compartments. In the next experiment, TEC lines were cultured with or without the protease inhibitors and then stained with anti-LC3 Ab followed by staining with 30-2 mAb that reacts to the MHC class II plus CLIP complexes (20). In the cultures without the protease inhibitors, some of the LC3-positive and the 30-2-positive compartments were detected in the cytoplasm of both TEC lines (Figure 2, A and C); however, the distribution of the 30-2-positive compartments were uneven. Both TEC lines have been established via several cloning procedures, but when both TEC lines are placed on slide glass, they may adhere to the glass and start growing in its own manner. Then the cell cycle of both the TEC lines may not synchronize. In addition, when the TEC lines are incubated in the medium containing IFN-γ, the expression of MHC class II and H2-DM molecules in both the TEC lines may not synchronize. The above un-synchronizing cell cycle and expression may cause the uneven distribution of the 30-2-positive compartments. When the staining profiles with anti-LC3 Ab were merged to those with 30-2 mAb, a few of the yellow compartments, in which the LC3-positive compartments were colocalized with the 30-2-positive compartments, were found in the cytoplasm of both TEC lines (Fig. 2, A and C). Approximately 10% of LC3-positive compartments in the cTEC line and 7% of those in the mTEC line were colocalized with the 30-2-positive compartments (Fig. 2,F). In the cultures with the protease inhibitors, the amounts of the 30-2- and LC3-positive compartments were increased in the cytoplasm of both TEC lines (Fig. 2, B and D), although the distribution of the 30-2-positive compartments was uneven. After the addition of the protease inhibitors, quantification analysis of red and green color-compartments in the pictures of both the TEC lines clearly showed that not only the amount of LC3-positive compartments but also the amount of the 30-2-positive compartments were increased in the cytoplasm of both TEC lines after culture with the protease inhibitors (Fig. 2,E). When the staining profiles with 30-2 mAb were merged to those with anti-LC3 Ab, it was observed that the amount of the yellow compartments increased in the cytoplasm of both TEC lines (Fig. 2, B and D). When the amounts of yellow color-compartments in the merged-pictures were quantified, the amount of yellow color-compartments increased (Fig. 2,E). In this condition, ∼20% of LC3-positive compartments were colocalized with the 30-2-positive compartments in the cTEC lines as well as in mTEC lines (Fig. 2 F). These results indicated that the considerable amounts of the 30-2-positive compartments in both TEC lines were colocalized with the LC3-positive compartments. These results also suggested that the LC3-positive compartments could fuse to the H2-DM-positive compartments in which MHC class II complexes were formed.

To confirm that LC3-positive compartments were docked to the compartments containing H2-DM molecules, we isolated H2-DM-positive vesicles from the PNS of both TEC lines after incubation with the protease inhibitors. The isolated vesicles were then subjected to SDS-PAGE and Western blotting analyses. We detected H2-DM molecules in the isolated vesicles from both TEC lines (Fig. 3,A). LAMP-1, a typical lysosomal marker, and a series of soluble endosomal/lysosomal proteases, i.e., cathepsins-B, -D, -L, and -S, were detected in the same manner as in our previous reports (Ref. 15 and data not shown). Both MHC class II heterodimers and P70 complexes that were composed of the heterodimers and partially digested Ii chains (10, 11, 12) were probed with M5–114 (Fig. 3,B), and two intact isoforms of Ii chains (35 kDa and 31 kDa) were probed with anti-Ii chain mAb (In-1.1; Fig. 3,C). It is noteworthy that LC3-II was detected on this blotting (Fig. 3,D). These results, together with the results in Figs. 1 and 2, may imply that the combined protease inhibitors not only affect the degradation of Ii chains but also accumulate the LC3-II molecules in the H2-DM-positive vesicles.

To determine whether the autophagic process participates in the presentation of cytoplasmic self-Ags not only in the in vitro-established TEC lines but also in the thymus in situ, we analyzed mouse thymic cryosections that were immunostained. Thymic autophagy occurs even without starvation. In addition, the autophagy is immediately induced in various tissues after birth and is then involved in the degradation of self-proteins and the production of amino acids (14). Thus, it is plausible that thymic autophagy might occur with a greater frequency shortly after birth than in the constitutive condition. We therefore analyzed thymuses taken from 1-day-old mice.

LC3-II was detected across the cryosections of the 1-day-old thymus and colocalized with pan-cytokeratin-positive stromal cells (Fig. 4,A) and in a vesicle-like manner in the cytoplasm of the stromal cells (Fig. 4,B). The LC3-positive compartments in thymic cryosections resided in some of the medullary epithelial cells that were stained with anti-cytokeratin 5 Ab (Fig. 4,C) and in some of the cortical epithelial cells that were stained with anti-cytokeratin 8 Ab (Fig. 4,D). Quantification analyses (Table II) indicated that ∼40% of the LC3-positive compartments in the thymic cryosections were costained with anti-LAMP-1 Ab (Fig. 4,E) and anti-H2-DM Ab (Fig. 4,F), suggesting that, similar to the TEC lines as shown above, most of these compartments would be autolysosomes. The LC3-positive compartments were costained with anti-MHC class II Ab (Fig. 4,G); however, the ratio of colocalization with the LC3 was around 20%. The anti-MHC class II Ab used in this staining reacts to MHC class II plus peptide complexes generated from MHC class II plus Ii complexes. Therefore, the lower colocalization ratio suggests that autophagosomes fused with compartments transporting MHC class II plus Ii complexes would progress via autolysosomes to the LC3-positive compartments colocalized with MHC class II plus peptide complexes. In contrast, ∼60% of the LC3-positive compartments were stained with 30-2 mAb (Fig. 4,H and Table II), suggesting that the LC3-positive compartments colocalized with MHC class II plus CLIP complexes would be generated immediately after the fusion of autophagosomes with compartments transporting MHC class II plus Ii complexes. Taken together, these results indicate that the autophagic compartments in thymus in situ could access the H2-DM positive compartments in which the MHC class II plus peptide complexes are formed.

The presentation of cytoplasmic self-Ags by TECs to developing/immature T cells is critical for the generation of T cell repertoires (1). Although the presentation process of exogenous Ags to mature T cells performed by professional APCs has been extensively investigated, mechanisms underlying the presentation of endogenous Ags, including self-derived cytoplasmic Ags, by TECs are almost unknown. The present study introduces the intriguing possibility that the autophagic process contributes to the MHC class II presentation pathway for cytoplasmic self-Ags in TECs.

We demonstrated that LC3-II, a membrane marker of autophagosomes, is localized in the H2-DM-positive compartments in which MHC class II plus CLIP complexes are formed (Fig. 1 G). The combination of two lysosomal protease inhibitors, pepstatin A and E64d, arrests not only the autophagic process at the stage of autolysosomes (16), but also the formation of MHC class II plus peptide complexes at the formative stage of the MHC class II plus CLIP complex (21, 22, 23). E64d, a potent cysteine protease, may affect the proteolytic activity of cathepsins L and S to cleave Ii chains (4, 22, 23, 24). Therefore, the use of these protease inhibitors facilitates the intersection of the autophagic process in the MHC class II Ag-presentation pathway.

To determine whether the autophagic process could access the MHC class II compartments in the thymus in situ, we used immunofluorescent probes to analyze cryosections of thymuses from 1-day-old mice. In newborn mice, autophagy is actively induced and helps to rescue neonates from severe starvation (25). Therefore, thymic autophagy may also occur with an increased frequency during this period. As expected, LC3-II was expressed across the cryosection of the neonate thymus (Fig. 4,A), suggesting that the autophagic process occurs in both the cortical region and the medullary region. In fact, some LC3-II molecules were colocalized with the thymic stromal cells that are positive for cytokeratin 5 (Fig. 4,C) as well as cytokeratin 8 (Fig. 4,D). Moreover, in some thymic stromal cells, LC3-II was clearly colocalized with the compartments containing LAMP-1, H2-DM molecules, MHC class II molecules, and MHC class II plus CLIP complexes (Figs. 4, E–H). These results indicate that the autophagic process in the thymus may access the endosomal/lysosomal compartments in which MHC class II plus peptide complexes are formed.

We previously demonstrated that cTEC lines mainly present intracellular Ags via the constitutive secretory pathway but do not present the extracellular Ags via the endocytic pathway, whereas mTEC lines present the extracellular Ags via the endocytic pathway and the intracellular Ags via the constitutive secretory pathway (8, 9, 10, 11, 12). In the present study, we found that the autophagic process could intersect the MHC class II presentation pathway in both in vitro-established TEC lines and the thymus in situ. Therefore, we speculate that the thymic self-Ags in cTECs could be delivered into the MHC class II compartments via the autophagic pathway as well as the constitutive secretory pathway. In the mTECs, thymic self-Ags could be delivered into the MHC class II compartments via the following three pathways: the autophagic pathway, the endocytic pathway, and the constitutive secretory pathway.

Takahama et al. suggested that the unique characteristics of protein degradation in the cortical thymus are critical for positive selection (26). Nakagawa et al. demonstrated that cathepsin L, a lysosomal cysteine protease expressed by cTECs, is required for the generation of CD4-single positive T cells in the context of positive selection involved with MHC class II (22). We previously verified that cathepsin L is produced both in vitro in the cTEC line and in situ in the thymic cortex (27). Therefore, in cTECs, the autophagic process, in conjunction with the constitutive secretory pathway, could contribute to the efficient transport of thymic self-Ags into the MHC class II compartments in which cathepsin L could be involved in degradation of the thymic self-Ags and generation of the thymic self-peptides with the unique characteristics for positive selection. In contrast, promiscuous expression of the thymic self-Ags by autoimmune regulator activity in mTECs may be essential for negatively selecting autoreactive single-positive T cells in the medulla (28). Therefore, in mTECs, the autophagic pathway, in conjunction with the endocytic and constitutive secretory pathways, may contribute to the comprehensive collection of the thymic self-Ags promiscuously expressed by autoimmune regulator and transport them into the MHC class II compartments in which lysosomal enzymes may degrade the thymic self-Ags and generate the effective repertoire of thymic self-peptides for negative selection. In any case, it is suggested that the autophagic process in the thymus may participate in producing the various spectra of self-peptides for thymic selection.

In conclusion, we have shown that LC3-II molecules are in fact colocalized with the H2-DM-positive compartments in which MHC class II complexes are formed, not only in the in vitro-established TEC lines but also in thymic cryosections. This finding implies that autophagy could intersect the MHC class II presentation pathway and consequently play an important role in presenting self-Ags to developing/immature T cells in the thymus. Our conclusion is concordant with the report by Nedjic et al. in which they addressed the role of autophagy in T cell repertoire selection in the thymus by using Atg5-deficeint mice that are deprived of an essential component of autophagosome formation (13). In the present study, the intracellular localization profiles of the autophagic marker, i.e., LC3-II, in in vitro-established TECs as well as in the thymus in situ, confirms the involvement of autophagy in the presentation of the thymic self-Ags. The degradation of the thymic self-Ags via the proteosome processing system and the autophagy-mediated lysosomal processing system produced a variety of complexes composed of various thymic self-peptides and MHC molecules. These complexes may lead to efficient positive selection of double-positive T cells in the thymic cortex as well as negative selection of self-reactive, single-positive T cells in the thymic medulla. The autophagic process may play a supporting role or even a crucial role in the presentation of self-Ags in the thymus to shape the T cell repertoires.

We thank Drs. Y. Takahama and M. Matsumoto for valuable suggestions and for reviewing the manuscript.

The authors have no financial conflict of interest.