Endogenous MHC Class II Processing of a Viral Nuclear Antigen After Autophagy

See allHide authors and affiliations

Science  28 Jan 2005:
Vol. 307, Issue 5709, pp. 593-596
DOI: 10.1126/science.1104904


CD4+ T cells classically recognize antigens that are endocytosed and processed in lysosomes for presentation on major histocompatibility complex (MHC) class II molecules. Here, endogenous Epstein-Barr virus nuclear antigen 1 (EBNA1) was found to gain access to this pathway by autophagy. On inhibition of lysosomal acidification, EBNA1, the dominant CD4+ T cell antigen of latent Epstein-Barr virus infection, slowly accumulated in cytosolic autophagosomes. In addition, inhibition of autophagy decreased recognition by EBNA1-specific CD4+ T cell clones. Thus, lysosomal processing after autophagy may contribute to MHC class II–restricted surveillance of long-lived endogenous antigens including nuclear proteins relevant to disease.

The main protein degradation machineries in eukaryotic cells are the proteasome and lysosomal proteases. The immune system monitors the products of these catabolic processes for pathogenic determinants. For this purpose, peptides generated by the proteasome are presented on MHC class I products, whereas products of lysosomal degradation are displayed on MHC class II (1, 2). CD8+ and CD4+ T cells survey these MHC class I and II complexes, respectively. Classically, CD8+ T cell epitopes are of endogenous origin, synthesized in the antigen presenting cell (APC), whereas CD4+ T cell epitopes are of exogenous origin, endocytosed by the APC. However, analysis of natural ligands eluted from MHC class II molecules has revealed that a large proportion of MHC II–bound peptides are derived from endogenous antigens (3).

In addition, endogenous MHC class II processing has been described for self-[GAD65 (4), complement C5 (5), and immunoglobulin light chain λ2 (6)], viral [influenza matrix protein (7), influenza nucleoprotein (8), and influenza hemagglutin (9)], and model antigens [hen egg lysozyme (10) and neomycin phosphotransferase II (11)]. However, these studies all involved overexpression and often ectopic expression of the antigens by transfection or infection with recombinant viral vectors. To date, the underlying endogenous MHC class II processing pathways are poorly characterized at the cell biology level.

We investigated endogenous MHC class II processing of the nuclear antigen 1 (EBNA1) of the Epstein-Barr virus (EBV). EBNA1 is the dominant EBV-latent antigen for CD4+ T cells and can be detected by CD4+ T cells after endogenous MHC class II processing in EBV-positive lymphoma cells (12). In order to address whether EBNA1 could enter MHC class II processing via lysosomal degradation, we incubated EBV-transformed lymphoblastoid cells (LCLs) with inhibitors of lysosomal acidification, ammonium chloride (Fig. 1) or chloroquine (13), for 2 days (14). On blocking of lysosomal acidification, EBNA1 accumulated in cytosolic vesicles, partially co-staining with the lysosome-resident protein LAMP1 (Fig. 1A). Subcellular fractionation confirmed that EBNA1 was enriched in microsomes on blocking of lysosomal acidification (Fig. 1B). In EBNA1 transfectants of the EBV-negative Hodgkin's lymphoma cell line L428 (L428E1PC5), EBNA1 was confined to whole-cell lysate and the nuclear fraction. After chloroquine treatment, EBNA1 could be found in the postnuclear supernatant and the high-speed pellet derived thereof. Discontinuous sucrose gradient centrifugation revealed that EBNA1 had accumulated in microsomes after chloroquine treatment. EBNA1 fractionated together with the lysosomal marker LAMP1. Accumulation of EBNA1 in microsomes after chloroquine inhibition was confirmed in the EBV-transformed B cell line LG2 (Fig. 1C). EBNA1, but not another nuclear EBV antigen and prominent CD8+ T cell antigen, EBNA3A, was enriched in the microsomal fraction after chloroquine treatment. These results demonstrate that EBNA1 accumulates in cytosolic, partially lysosomal vesicles upon inhibition of acidification.

Fig. 1.

EBNA1 accumulates in lysosomes after inhibition of lysosomal acidification. (A) The EBV-transformed lymphoblastoid cell line SL-LCL was stained for EBNA1, LAMP1, and DNA content [4′,6′-diamidino-2-phenylindole (DAPI)] without (first row) or with (second and third row as well as large field) 20 mM NH4Cl treatment for 2 days. Arrows indicate cells with EBNA1/LAMP1 colocalization. One of three experiments is shown. (B and C) The EBNA1 transfectant of the EBV Hodgkin's lymphoma cell line L428 [L428E1PC5 (B)] or the EBV-transformed lymphoblastoid cell line LG2 (C) were treated with 50 μM chloroquine for 3 days (+CQ). After homogenization (H), intact cells and nuclei (N) were pelleted (twice at 3000 g for 10 min), and the postnuclear supernatant (S1) was centrifuged at 100,000 g for 1 hour to obtain a microsomal pellet (P) and a postmicrosomal supernatant (S2). Mirosomes were further purified from the interphase of a two-step sucrose gradient (2 M/0.5 M sucrose; M). EBNA1, EBNA3A, and the lysosomal marker LAMP1 were visualized by Western blot analysis (B and C). One of nine experiments is shown.

Next, we investigated whether EBNA1 follows an autophagic route to lysosomes. Autophagy is involved in the steady-state turnover of long-lived proteins and damaged organelles (15). During this process, cytoplasmic material is enveloped into double-membrane vesicles, which then fuse with lysosomes (16). This fusion can be prevented by inhibition of lysosomal acidification (17). Thus, autophagosomes accumulated in chloroquine-treated LCLs (Fig. 2A) and could be specifically visualized with the fluorescent dye monodansylcadaverine (MDC) (18). The majority of cytosolic EBNA1 was found in these large, usually less than 10 MDC+ vesicles (Fig. 2A), of which only a part co-stained for LAMP1 (13). To assess whether the ultrastructural features of EBNA1-containing vesicles conform to previously described autophagic vacuoles, we performed immuno-electron microscopy (IEM) on sections of EBNA1-transfected L428 Hodgkin's lymphoma cells. On inhibition of lysosomal acidification, EBNA1 could be observed in cytosolic vesicles, surrounded by double membranes (Fig. 2C, black arrows). Some of these structures had the cup-shaped appearance of forming autophagosomes (Fig. 2C, middle). An isotype control antibody produced no staining (13). Thus, autophagy participates in the delivery of EBNA1 to lysosomes.

Fig. 2.

EBNA1 localizes to autophagosomes upon inhibition of lysosomal acidification. (A) The lymphoblastoid cell line SL-LCL was stained for EBNA1, autophagic vesicles (MDC), and DNA content (DAPI) in the absence (–CQ) or presence (+CQ) of 50 to 100 μM chloroquine. One of 10 experiments is shown. (B) Electron micrographs of cell organelle morphology in chloroquine-treated (+CQ) or untreated (–CQ) EBNA1-transfected L428 Hodgkin's lymphoma cells. Black arrows indicate electron-dense autophagosomes or lysosomes. (C) Electron micrographs of chloroquine-treated EBNA1-transfected L428 Hodgkin's lymphoma cells stained for EBNA1 with the monoclonal antibody 5F12. Blue arrows indicate nuclear EBNA1; black arrows, EBNA1 in forming or mature autophagosomes. One of three experiments is shown.

To assess the kinetics of EBNA1 degradation via autophagosomes, we followed cytosolic EBNA1 after removal of lysosomotrophic reagents. After 2 days of chloroquine or ammonium chloride treatment, a substantial amount of EBNA1 accumulated in cytosolic vesicles of LCLs (fig. S1A) (13). After removal of the lysosomotrophic reagents, cytosolic EBNA1 was degraded with a half-life of around 30 min (fig. S1B). Thus, EBNA1 accumulation in autophagosomes and lysosomes represents a short-lived intermediate state of EBNA1 degradation, which can be stabilized by inhibition of lysosomal acidification.

Many long-lived proteins are substrates for autophagy (15). One well-characterized example of autophagic protein degradation is the proteasome (1921). In chloroquine- or ammonium chloride–treated cells, the proteasome 20S core particle accumulated in cytoplasmic vesicles, most of which also co-stained for EBNA1 (fig. S2) (13). Thus, EBNA1 and a known autophagy substrate, the proteasome, accumulate in the same vesicles on inhibition of lysosomal acidification.

In order to test whether autophagy leads to endogenous MHC class II processing of EBNA1, we determined whether inhibition of this pathway would block antigen presentation to EBNA1-specific CD4+ T cell clones (22, 23). Treatment with the autophagy inhibitor 3-methyladenine (24) for 2 to 4 days decreased EBNA1-specific CD4+ T cell recognition of EBV-transformed B cells and EBNA1-transfected Hodgkin's lymphoma cells by 30 to 70% in interferon-γ (IFN-γ) enzyme-linked immunospot (ELISPOT) assays, whereas the proteasome inhibitor lactacystin (25) had little effect (Fig. 3A). Furthermore, we used RNA interference to silence the essential autophagy gene Atg12. Atg12 is a small ubiquitin-like protein, and its complex with Atg5 is essential for autophagosome formation (26, 27). Small interfering RNAs (siRNAs) against two different regions of Atg12 (Atg12.1 and Atg12.2), but not a green fluorescent protein (GFP)–specific siRNA, down-regulated efficiently the Atg12 mRNA level (Fig. 3B). Knockdown of Atg12 with both siRNAs decreased the IFN-γ response of EBNA1-specific CD4+ T cells by 40 to 60%, whereas treatment with the GFP siRNA had no effect (P = 0.0002 for Atg12.1 versus mock, P = 0.0030 for Atg12.1 versus GFP, P = 0.0004 for Atg12.2 versus mock, and P = 0.0014 for Atg12.2 versus GFP). In contrast, EBNA3A recognition by the CD8+ T cell clone MS.B11 was not affected by the Atg12 knockdown (Fig. 3C). Treatment with 3-methyladenine or Atg12-specific siRNAs did not affect the MHC class II levels on the B cell lines (Fig. 3D). In addition, recognition of inhibitor- or siRNA-treated target cells could be completely restored by pulsing the cells with the cognate T cell epitopes before ELISPOT or enzyme-linked immunosorbent assay (ELISA) (13). Furthermore, exogenous MHC class II processing of EBNA1 could be excluded, because LCLs were unable to sensitize bystander B cells for EBNA1-specific CD4+ T cell recognition (12) and EBNA1 was undetectable in the supernatants of LCL- and EBNA1-transfected L428 cells with the use of ELISAs (detection limit was 0.1 μg/ml), whereas 1 to 10 μg/ml were necessary to increase CD4+ T cell detection of EBV+ and EBV B cell lines (13). Thus, blocking of autophagy specifically inhibits endogenous MHC class II processing of EBNA1 but not the presentation of the MHC class I–restricted antigen EBNA3A.

Fig. 3.

Blocking of autophagy leads to down-regulation of MHC class II–restricted CD4+ T cell recognition of EBNA1. (A) The EBNA1-specific CD4+ T cell clones A4.E116 and RJD.79 were stimulated with cognate target cells [A4.E116 with human histocompatibility leukocyte antigen (HLA)–DR1+EBV+ LCL721.221, RJD.79 with HLA-DQ2/3+EBNA1+ L428E1PC5, and HLA-DQ2/3+EBNA1 L428 as a negative control] in IFN-γ ELISPOT assays. Where indicated, targets were pretreated with 10 mM 3-methyladenine or 1 μM lactacystin for 2 to 4 days. One of four experiments is shown. Error bars indicate standard deviations. (B) mRNA levels of glyceraldehyde-3-phosphate dehydrogenase and Atg12 were analyzed on day 4 after knockdown with GFP- or Atg12-siRNAs (Atg12.1 and Atg12.2) with use of semiquantitative reverse transcription polymerase chain reaction. The results are representative of five experiments and correspond to the experiment shown in (C). (C) IFN-γ ELISAs on supernatants of the EBNA1-specific CD4+ T cell clone P3-B7 and the EBNA3A-specific CD8+ T cell clone MS.B11, co-cultured with cognate target cells (P3-B7 with HLA-DP3+ Ag876 and MS.B11 with HLA-B8+ BM-LCL). Where indicated, targets were electroporated with 10 μM siRNA twice in 4 days before overnight culture with T cells. One of three experiments is shown. Error bars indicate standard deviations. (D) HLA-DR surface levels after treatment with 10 mM 3-methyladenine (3-MA) and 1 μM lactacystin for EBNA1-transfected L428 Hodgkin's lymphoma cells (L428E1PC5, left) or with GFP- and Atg12-siRNAs (Atg12.1 and Atg12.2) for the EBV-transformed B cell line Ag876 (right). Mean fluorescence values are given in the flow cytometer blots. One of two experiments is shown.

Autophagy has been suggested as the main degradation pathway for long-lived proteins (15), whereas short-lived proteins are probably mainly regulated by proteasomal processing (28). A long half-life is also characteristic for EBNA1 (29) and the proteasome (20, 21). Whereas we find autophagy and lysosomal processing of full-length EBNA1 for MHC class II presentation, glycine-alamine repeat-deleted EBNA1 and defective ribosomal products (DRiPs) of EBNA1 were described to be short-lived and processed by the proteasome for efficient presentation by MHC class I (3033). Similarly, influenza's matrix protein follows in part a proteasome and transporter associated with antigen processing–independent, chloroquine-sensitive but brefeldin A–insensitive endogenous MHC class II processing pathway. Long-lived forms of this antigen were primarily processed onto MHC class II, but short-lived forms mainly onto MHC class I (7, 34). Lastly, the most prominent sources of natural MHC class II ligands are long-lived cytosolic or nuclear antigens, and natural MHC class I ligands primarily originate from short-lived proteins (table S1). Thus, MHC class II processing after autophagy might contribute to CD4+ T cell recognition of long-lived viral and tumor antigens.

Apart from APCs like EBV-transformed B cells, MHC class II presentation of endogenous antigens might be especially relevant in MHC class II–positive tissues with low phagocytic ability. At sites of inflammation, many cell types like epithelial cells up-regulated MHC class II and were found to be surrounded by cytolytic CD4+ Tcells(3537). Thus, CD4+ T cells may play a role in immune surveillance of APCs and inflamed tissues after endogenous MHC class II processing involving autophagy, and EBNA1 is the first pathogen-derived antigen found to follow this pathway.

Supporting Online Material

SOM Text

Figs S1 and S2

Table S1

References and Notes

References and Notes

View Abstract

Navigate This Article