MHC Class II Presentation Is Controlled by the Lysosomal Small GTPase, Arl8b

Dendritic cells (DCs) are the most potent APCs found in mammals. They are critical regulators of immunity that continuously survey their microenvironment by taking up antigenic materials, processing them, and presenting Ags to Ag-specific T cells. Peptides and lipids Ags are displayed for immune recognition by MHC and CD1 Ag-presenting molecules, respectively (1, 2). In general, MHC class I (MHC I) and II (MHC II) molecules present peptides that originate from different sources. Peptide Ags degraded in the cytosol by the proteasome are translocated into the endoplasmic reticulum (ER) through the transporter associated with Ag presentation and loaded onto MHC I molecules (3). Once MHC I molecules acquire their peptides in the ER, the loaded MHC I molecules then traffic along the secretory route to the cell surface for presentation to CD8⁺ T cells (4).

In contrast, MHC II molecules sample peptides of exogenous origin. MHC II assembly occurs in the ER where the two polymorphic α- and β-chains are associated with the invariant chain (Ii), a type 2 transmembrane protein. This trimer assembles in the ER into a nonameric complex (αβIi)₃ that exits the ER and passes through the Golgi apparatus to the trans-Golgi network (TGN). MHC II molecules are directed to the late endolysosomal system either from the TGN or by endocytosis at the plasma membrane. Two di-leucine sorting motifs located in the cytosolic tail of Ii are recognized by the sorting adaptors AP1 for sorting directly from TGN and AP2 for sorting from the plasma membrane (5–9). MHC II molecules are finally delivered to late endolysosomal MHC II compartments. Once in the lysosome, Ii is sequentially degraded, leaving the small CLIP fragment still bound to the Ag-binding groove. The remnant CLIP fragment is then exchanged for an antigenic peptide in a reaction catalyzed by H2-M/HLA-DM molecules (10, 11). Finally, the MHC II–peptide complex is delivered to the plasma membrane for recognition by CD4⁺ T cells (12–14).

A critical aspect of these orchestrated Ag presentation systems is the trafficking of Ag-presenting molecules and Ags to the same compartment for peptide loading. Vesicle trafficking in APCs requires cargo selection and motoring to the destination compartment in a process controlled by small GTPases of the RAS superfamily (15). The delivery to lysosomes of endocytosed cargo occurs through an endosomal maturation process. This is notably controlled by RAB-5 that localizes to early endosomes and RAB-7 that localizes to late endosomes. During endosomal maturation, RAB-5 endosomes are converted to RAB-7 late endosomes that finally fuse to lysosomes. Fusion to lysosomes is mediated by the recruitment of the tether complex homotypic fusion and protein sorting (HOPS) (16, 17). The transport of MHC II to the late endolysosomal system seems to occur through endosomal maturation and seems to require dynein motors and ESCRT complexes (3, 18–20). Yet, the GTPase regulating the delivery of Ag to MHC II compartments for generating MHC II–peptide complexes remains unknown.

We recently identified a crucial role for the Arf-like GTPase ADP-ribosylation factor-like 8b (Arl8b) in directing CD1d and other cargo to lysosomes for lipid Ag presentation (21). To date, Arl8b is the only GTPase known to localize specifically to lysosomes. Interestingly, depending on the effectors it recruits, Arl8b can be involved in both cargo delivery to and motility of lysosomes. For example, Arl8b was shown to control the outward movement of lysosomes and lysosome-related organelles such as lytic granules in NK cells by recruiting SifA and kinesin-interacting protein that, in turn, binds to the kinesin motor Kif5b (22, 23). Recent studies also suggested that Arl8b plays a role in lysosomal tubulation in macrophages and in Salmonella-infected cells, a process that requires the recruitment of kinesin motors (24, 25). Separately, we found that Arl8b directs certain cargo into lysosomes in mammals by recruiting the VPS41 subunit of the tethering HOPS complex that mediates the fusion of late endosomes to lysosomes, whereas others defined a similar role for Arl8b in Caenorhabditis elegans (21, 26, 27).

Because the trafficking regulators for delivery of Ags to lysosomes and its subsequent loading onto MHC II molecules are not known, we sought to determine whether the lysosomal GTPase Arl8b controlled MHC II presentation in DCs. We found that Arl8b silencing in DCs results in a decrease of MHC II presentation to cognate CD4⁺ T cells. This impaired presentation is due to a delay in the formation of MHC II–peptide complexes in lysosomes. Accordingly, Arl8b-silenced DCs display a reduced amount of MHC II–peptide complexes at the cell surface that directly correlates with a defect in activating Ag-specific CD4⁺ T cells.

D1 cells are a long-term cultured growth factor–dependent immature splenic DC line derived from C57BL/6 mice maintained as previously described (28). Bone marrow–derived DCs (BMDCs) were generated from C57BL/6 mice. BMDCs were cultured for 6–7 d in RPMI 1640 (Life Technologies), 10% heat-inactivated FBS (Gemini), HEPES (Life Technologies), l-glutamine, penicillin/streptomycin (Life Technologies), and 2-ME (Life Technologies) supplemented with 10 ng/ml recombinant mouse GM-CSF (Miltenyi Biotec). BMDCs were then purified with CD11c⁺ magnetic beads (Miltenyi Biotec). Primary mouse splenic DCs were isolated from spleens of C57BL/6 mice, cut into small pieces, and digested with RPMI 1640, collagenase IV 1 mg/ml (Worthington), DNase 0.1 mg/ml (Roche) at 37°C for 1 h. Every 15 min, cells released during the digestion were collected and placed at 4°C, and fresh digestion buffer was added to the remaining splenocytes. CD8⁺ OT-I and CD4⁺ OT-II T cells were freshly isolated from mouse spleen with CD4 or CD8 magnetic beads T cell isolation kits (Miltenyi Biotec). C1R and SPF3 T cells were maintained as previously described (29).

Abs used in this study were rabbit anti-ARL8 (21), mouse anti–lysosome-associated membrane protein 1 (anti–LAMP-2; Developmental Studies Hybridoma Bank), rabbit anti–EEA-1 (Calbiochem), mouse anti–I-Ab–Eα_(52–68), clone YAe (eBioscience), mouse anti–β-actin, clone AC-74 (Sigma-Aldrich), rat anti–LAMP-1 and directly conjugated to Alexa 488, clone 1D4B, rat anti–I-A/I-E, clone M5/114.15.2 (Biolegend), and rat anti–H2-M (BD Biosciences). Secondary Abs goat anti-mouse, -rabbit, -rat coupled to Alexa dyes 488, 568 or 647 were purchased from Life Technologies.

For Western blotting, cells were lysed in HEPES pH 7.5, 50 mM, NaCl 150 mM, Nonidet P-40 1%, glycerol 5%, SDS 0.1%, and proteases inhibitors (Bio-Rad). Protein extracts were separated on 15% acrylamide gels, transferred to polyvinylidene difluoride membranes (Bio-Rad) and blotted with ARL8 Ab diluted in TBST, 5% nonfat milk (Santa Cruz), at 1 μg/ml followed by goat anti-rabbit HRP at 1:5000 dilution. Membranes were reprobed with mouse anti–β-actin (Sigma-Aldrich) 1:50,000 followed by goat anti-mouse HRP at 1:10,000. All blots were developed with SuperSignal West Pico Chemiluminescent Substrate (Pierce) and exposed to film (Santa Cruz). Goat anti-mouse HRP and goat anti-rabbit HRP Abs were purchased from Jackson Immunoresearch Laboratories.

D1 cells and BMDCs were electroporated with ON-TARGETplus nontargeting small interfering RNA (siRNA) #1 (Dharmacon) as a control or Arl8b siRNA (CGAGGAGTCAATGCAATTGTT) (21) (Dharmacon) using an Amaxa system with the mouse immature DC kit (Lonza). In brief, 1.5 × 10⁶ cells were mixed with 800 pmol siRNA and electroporated using the Y-001 Amaxa program. Media were changed 24 h after electroporation and assays were performed 72 h after transfection. C1R cells were transduced with control and Arl8b short hairpin RNA as previously described (21).

For D1 and BMDCs assay, 25,000 cells/well were plated on a 96-well, flat-bottom plate (Corning) and allowed to adhere for 2 h or overnight. Cells were incubated with OVA at the indicated concentrations (Biovendor Research and Diagnostic Products). After 3 h, OVA was removed and the cells were washed in PBS, and 100,000 T cells/well were added. Supernatants were collected after 24 h and assayed by sandwich ELISA on EIA/RIA plates (Costar). Murine IL-2 ELISA was performed with Ab pairs 554424 and 554426 (BD Biosciences). Detection was carried out using streptavidin coupled to HRP 554066 (BD Biosciences) and then with TMB substrate kit (Pierce), and absorbance was read at 405 nm. To measure T cell activation, we harvested cells after 20 h of coculture with T cells/APCs, incubated with Fc Block (clone 93; Biolegend) to block FcRs, then stained for the surface markers CD25, CD69, CD44, TCRβ, and CD4, and analyzed by flow cytometry. Freshly isolated CD4⁺ T cells were extensively washed with PBS, labeled with 2 μM CFSE (Biolegend), and then cocultured with APCs to assess T cell proliferation. Cells were harvested 72 h later, stained for the surface markers CD4 and TCRβ, and then analyzed by flow cytometry.

Cells were stained with Abs against cell-surface molecules and analyzed on a BD FACSCanto II cytometer (BD Biosciences). Abs used were PE rat anti-CD86, allophycocyanin rat anti-CD40 (BD Biosciences), allophycocyanin Armenian hamster anti-CD69 (clone H1.2F3; Biolegend), PE rat anti-CD25 (clone PC61; BD Biosciences), FITC rat anti-CD44 (clone IM7; BD Biosciences), PE-Cy7 rat anti-CD4 (clone GK1.5; eBioscience), BV421 Armenian hamster anti-TCRβ (clone H57-597; Biolegend), Alexa Fluor 488 rat anti–I-A/I-E (clone M5/114.15.2; Biolegend), nonconjugated anti–I-Ab–Eα_(52–68) (clone YAe; eBioscience) followed by PE-conjugated goat anti-mouse. Live/dead cells discrimination was done by staining with Zombie Aqua (Biolegend). Flow cytometry data were analyzed with FlowJo software (Tree Star).

D1 cells were seeded on 12-mm coverslips (VWR) precoated with rat fibronectin (100 mg/ml) for 1 h at 37°C (Sigma). Cells were then incubated with OVA (200 μg/ml; Life Technologies), Eα_(52–68) (10 μM; Anaspec), recombinant Eα_(52–68)-GFP (200 μg/ml), or dextran 1 mg/ml (Life Technologies) for the time indicated. Cells were then washed in cold HBSS, fixed, and processed for immunostaining. When indicated, cells were preincubated during 1 h with 100 nM LysoTracker Red DND-99 (Life Technologies). For flow-cytometry analysis, cells were incubated with 10-kDa dextran Alexa Fluor 546 during the indicated time, then placed in cold FACS buffer (PBS, 1% serum, 2 mM EDTA) and analyzed by FACS. Recombinant protein Eα_(52–68)-GFP was expressed in BL21 bacteria transformed with pTrcHis2-TOPO-Eα_(52–68)-GFP plasmid, kindly provided by Marc Jenkins (30). The chimeric protein expression was induced for 4 h with IPTG (Calbiochem) and then purified from bacterial lysate using His60 Ni Superflow Resin (Clontech) followed with fast protein liquid chromatography using a gel-filtration column.

Transfected D1 cells were seeded on coverslips (VWR), then fixed for 15 min in HBSS (Life Technologies), 3.7% paraformaldehyde (PFA, Electron Microscopy Sciences), washed, blocked, and permeabilized for at least 30 min in PBS, 10% FBS, and 0.2% saponin. Cells were then stained with primary Abs overnight at 4°C. Cells were washed three times and incubated with secondary Abs for 1 h at room temperature. Cells were washed another three times and then mounted with the antibleaching mounting medium Fluoro Gel with DABCO (Electron Microscopy Sciences). When costaining I-A/I-E and LAMP-1 or H2-M and LAMP-1, cells were first stained with I-A/I-E or H2-M and then incubated with the secondary Ab. After an extra step of blocking with 1% rat serum for 1 h, the cells were incubated with the anti–LAMP-1 directly conjugated to Alexa Fluor 488. Cells were analyzed on a Nikon TE2000-U inverted microscope equipped with the laser-scanning C1 confocal system using a Plan Apochromat 60×/1.40 NA oil objective. Image analysis and colocalization quantifications were done using National Institutes of Health ImageJ software (http://rsbweb.nih.gov/ij) and MetaMorph v7.6.4 (MDS Analytical Technologies), respectively. For colocalization quantification, each channel (R, G, B) was considered separately with each pixel assigned an intensity value 0–255 with care taken to assure that no pixel was oversaturated in any image used for quantification. A lower threshold is set whereby areas of the image not containing cells or staining are excluded. The software then calculates the area of I-A/I-E, H2-M, OVA, or dextran overlapping with LAMP-1 or LysoTracker-positive areas and reports it in terms of percentage colocalized. Each quantification was done on 15–30 cells for each time point, and representative images are shown from at least 2 different experiments. For Fig. 4, the total integrated signal density/cell was measured using National Institutes of Health ImageJ software after setting up a threshold as previously described. For each time point, 30–40 cells were analyzed. All images were processed with Adobe Photoshop Software and Adobe Illustrator.

Bacteria expressing Eα_(52–68)-GFP, kindly provided by Marc Jenkins (30), were grown in LB media at 37°C until OD₆₀₀ = 0.3–0.4. E. coli bacteria were incubated with 1 mM IPTG (Calbiochem) for 16 h to induce chimeric protein expression. Bacteria were washed and then resuspended in D1 media. A total of 100,000–200,000 D1 cells/well were seeded in 96-well plates, and bacteria were added at various multiplicities of infection (MOI) 10:1, 20:1, or 50:1. Cells and bacteria were incubated for 1–2 h; then cells were washed three times and finally incubated with fresh media containing gentamicin 10 μg/ml (Life Technologies) for another 1–2 h at 37°C.

To test whether the loss of Arl8b impacts MHC II presentation, we used murine D1 cells, a previously described splenic DC line (14, 28). D1 cells were transfected with Arl8b siRNA or control siRNA, pulsed with OVA for 3 h, washed, and then incubated with MHC II–restricted OT-II T cells. For comparison, cells were incubated in parallel with OVA and MHC I–restricted OT-I T cells. The formation of MHC II–OVA_323–337 and MHC I–OVA_257–264 peptide complexes was determined by assessing T cell stimulation via IL-2 cytokine ELISA. Arl8b-silenced D1 cells displayed a marked decrease in MHC II Ag presentation compared with control siRNA cells given the same concentration of protein Ag (Fig. 1A, compare 830 pg/ml IL-2 production in control cells at 100 μg/ml OVA with 370 pg/ml IL-2 production in Arl8b-silenced cells). In contrast, Arl8b-silenced and control cells were able to induce secretion of the same amount of IL-2 by OT-I T cells (Fig. 1B). Moreover, Arl8b-silenced and control siRNA cells loaded with the MHC I peptide OVA_257–264 (SIINFEKL) do not show any difference in the level of T cell activation (Supplemental Fig. 1C). Concurrently, Arl8b-silenced D1 cells loaded with the MHC II peptide OVA_329–337 do not display a reduction of IL-2 secretion compared with control siRNA D1 cells (Supplemental Fig. 1C). Next, we tested the role of Arl8b in MHC II presentation with Arl8b-depleted primary BMDCs. Arl8b-silenced BMDCs showed a decrease in MHC II presentation of OVA protein as measured by OT-II T cell IL-2 production (Fig. 1C) when compared with control siRNA BMDCs. To extend these findings, we tested whether silencing of Arl8b in the human B cell line C1R impacted MHC II presentation. Similar to our findings for murine DCs, Arl8b silencing reduced MHC II presentation of tetanus toxoid protein to human T cells as measured by IFN-γ cytokine ELISA (Fig. 1D). To complement MHC II–restricted OT-II T cell activation measured by IL-2 secretion, we also measured T cell proliferation and induction of surface activation markers such as CD25, CD69, and CD44. In agreement with our previous experiments, Arl8b-silenced BMDCs (Fig. 1E, 1F) or D1 cells (Fig. 1G) pulsed with OVA are significantly diminished in their ability to activate T cells as measured by the percentage of OT-II T cells that are CD25⁺, CD69⁺, and CD44^high+. Arl8b siRNA and control siRNA-treated BMDCs or D1 cells were pulsed with OVA and then cocultured with CFSE-labeled OT-II T cells for 72 h to measure T cell proliferation. As expected, the percentage of T cells that had divided was significantly reduced when using Arl8b-silenced BMDCs (Fig. 1H, 1I) and D1 (Fig. 1J) compared with control cells. The efficiency of Arl8b silencing was confirmed by quantitative PCR and Western blots showing a marked decrease of Arl8b mRNA and protein levels, without any effect on the closely related Arl8a small GTPase (Fig. 1K and Supplemental Fig. 1A, 1B).

To determine whether the defect in MHC II presentation after Arl8b silencing could be explained by a defect in MHC II transcription or a general defect in trafficking of molecules to the cell surface, we measured the level of transcription for the MHC II β-chain (H2-ab) by quantitative PCR and cell-surface levels of MHC II (I-A/I-E), and the costimulatory molecules CD40 and CD86 by flow cytometry. Control and Arl8b-silenced D1 cells displayed the same transcription rate of MHC II molecules (Supplemental Fig. 1A, 1B). Moreover, control and Arl8b-silenced D1 cells all displayed similar levels of cell-surface MHC II, CD86, and CD40 (Fig. 1L). Together, these results suggest that Arl8b-silencing reduces MHC II Ag presentation without altering costimulatory molecule surface expression, impacting MHC I presentation or MHC II cell-surface expression levels.

Arl8b has previously been shown to localize to lysosomes and lysosome-related organelles in mouse and human cells (21, 23). Because MHC II is enriched in lysosomes, we asked whether Arl8b localized in lysosomes and MHC II⁺ compartments in DCs. Three types of DCs were costained for Arl8b, LAMP-1, and MHC II. We confirmed that Arl8b localized in LAMP-1⁺ and LAMP-2⁺ compartments in D1 cells (Fig. 2A and Supplemental Fig. 2), BMDCs (Fig. 2C), and primary splenic DCs (Fig. 2E). Moreover, costaining of DCs with anti-Arl8b and the MHC II Ab, anti-IA/I-E, revealed that Arl8b localized to MHC II⁺ compartments (Fig. 2B, 2D, 2E).

To form MHC II–peptide complexes, exogenous Ags must traffic through the early endocytic system before being delivered to lysosomes. In this compartment, the chaperone protein H2-M (human homolog HLA-DM) facilitates peptide exchange to load processed peptides onto MHC II molecules. Therefore, we tested whether Arl8b depletion impacted the steady-state lysosomal localization of MHC II and H2-M. First, we stained control and Arl8b-silenced D1 cells for MHC II and LAMP-1 (Fig. 3A, 3B). Colocalization analysis revealed a similar percentage of MHC II colocalization to LAMP-1⁺ compartments (control ± SEM: 78.9 ± 3.737% versus Arl8b D1–silenced cells ± SEM: 79.7 ± 3.667%; Fig. 3C). Second, we stained control siRNA and Arl8b-silenced D1 cells for H2-M and LAMP-1 (Fig. 3D, 3E). H2-M localized to LAMP-1⁺ lysosomes in both Arl8b-silenced and control cells to the same degree (control ± SEM: 71.5 ± 6.073% versus Arl8b D1–silenced cells ± SEM: 77.9 ± 6.724%; Fig. 3F). These results indicate that the Arl8b depletion did not result in loss of MHC II or H2-M, or their failure to colocalize in lysosomes at steady-state.

Because the decrease in T cell activation observed with the Arl8b-silenced DCs is not due to a defect in the cell-surface expression of MHC II molecules or in the localization of MHC II and H2-M to lysosomes, we next asked whether delivery of exogenous Ag to MHC II compartments was impaired in Arl8b-silenced DCs. To assess whether the delivery of protein Ags to the lysosomes is impacted by Arl8b silencing, we monitored the trafficking of endocytosed fluorescently tagged OVA by confocal microscopy. D1 cells were pulsed with OVA for various times and stained for EEA-1 to label early endosomes (Fig. 4C, 4D) and LAMP-1 to label lysosomes (Fig. 4A, 4B). Strikingly, Arl8b-silenced D1 cells displayed less colocalization of OVA in LAMP-1⁺ compartments over time, compared with control siRNA-treated D1 cells. Two-fold less colocalization of OVA with LAMP-1 after 60-min incubation was noted (control ± SEM: 64.53 ± 3.228% versus Arl8b D1–silenced cells ± SEM: 33.68 ± 3.173%; Fig. 4A, 4B, quantification 4E). In contrast, delivery of OVA to early endosomes after 10 min of incubation was not impacted by Arl8b silencing (control ± SEM: 57.13 ± 4.478% versus Arl8b-silenced D1 cells ± SEM: 54.96 ± 4.786%; Fig. 4C, 4D, quantification 4F). To extend these findings, we used LysoTracker red to identify acidified lysosomes and fluorophore conjugates of the carbohydrate dextran, which is taken up into cells through pinocytosis and traffics to and accumulates in lysosomes. Similar to the results observed with OVA, Arl8b-silenced D1 cells have a lower colocalization rate of endocytosed dextran to LysoTracker red⁺ compartments after 60-min incubation (control ± SEM: 44.93 ± 4.941% versus Arl8b D1–silenced cells ± SEM: 15 ± 1.629%; Supplemental Fig. 3A, 3B, quantification, Supplemental Fig. 3C). These results clearly indicate that Arl8b-silenced cells have a pronounced delay in the delivery of endocytosed dextran to acidic compartments such as lysosomes. Finally, we asked whether Arl8b silencing impacted the rate of Ag uptake into D1 cells. Thus, control and Arl8b siRNA D1 cells were incubated with fluorescent OVA or dextran for various times, and the amount of Ag endocytosed was analyzed by flow cytometry. Both Arl8b-silenced and control siRNA D1 cells acquired similar amounts of OVA (Fig. 4G, 4H) and dextran over time (Supplemental Fig. 3D). Together, these results demonstrate that Arl8b-silenced D1 cells endocytose similar amounts of Ag but have a kinetic delay in the trafficking of endocytosed Ag to lysosomes. This alteration likely contributes to the defect in Ag presentation observed in Arl8b-silenced cells.

Forming Ag-loaded MHC II complexes is a critical intermediate step for successful Ag presentation. To determine whether the delay in cargo delivery could contribute to the defect in formation of MHC II–peptide complexes, we directly assessed complex formation by confocal microscopy, using an Ab that specifically recognizes the YAe epitope formed by the MHC II molecule I-Ab associated with the I-E α-chain peptide, Eα_(52–68) (31–33). Control siRNA and Arl8b-silenced D1 cells were pulsed with the peptide Eα_(52–68) for various times, costained for the YAe epitope and LAMP-1, and then the mean integrated signal density of YAe was assessed (Fig. 5A). Because surface-loaded complexes could confound the results when loading cells with peptide, flow cytometry was not performed in this study. Instead, we used confocal microscopy and intentionally excluded the plasma membrane from the analysis. To identify complex formation, we focused only on lysosomal I-A_b–Eα_(52–68). Compared with control cells, Arl8b-silenced D1 cells showed a >2-fold decrease in I-A_b–Eα_(52–68) complex signal intensity in LAMP1⁺ lysosomes after a 60-min pulse with Ag (control ± SEM: 776.4 ± 120.7 versus Arl8b D1–silenced cells ± SEM: 372.4 ± 75.39; Fig. 5B). However, using this peptide system, we cannot exclude the possibility that some MHC II complexes were loaded with the peptide at the surface and then trafficked to the lysosome. To complement this experiment, we produced and purified a recombinant chimeric protein, Eα_(52-68)-GFP, that requires processing within the lysosome to generate free Eα_(52–68) peptide that can load onto the MHC II molecules. Control and Arl8b siRNA D1 cells were incubated with 200 μg/ml Eα_(52–68)-GFP for 90 min, fixed, and stained as previously described, and the mean integrated signal density of YAe epitope was assayed by confocal microscopy (Fig. 5C). As observed with free peptide Eα_(52–68), Arl8b-silenced D1 cells showed a significantly reduced I-A_b–Eα_(52–68) complex signal intensity (control ± SEM: 1554 ± 177.7 versus Arl8b D1–silenced cells ± SEM: 804.7 ± 72.75; Fig. 5C). All together, these data strongly indicate that Arl8b-silenced cells have a defect in the formation of MHC II–peptide complexes in lysosomal compartments.

Given the delay in delivery of Ag to lysosomes and the delay in MHC II–peptide complex formation in Arl8b-silenced DCs, we hypothesized this might result in a reduced amount of YAe epitope presented at the cell surface. To assess the role of Arl8b also under conditions relevant to microbial infection, we set up an infection assay of D1 cells using an E. coli strain that expresses the peptide Eα_(52–68) fused to the fluorescent protein GFP. Importantly, presentation of YAe in this system requires lysosomal destruction of the bacteria and processing of the Eα_(52–68)-GFP to allow the loading of Eα_(52–68) onto MHC II molecules. D1 cells were incubated for 1–2 h with bacteria at different MOI, after which extracellular bacteria were removed by successive washes and cells were then incubated for 1–2 h in normal medium supplemented with gentamicin to kill any remaining nonphagocytosed bacteria. Finally, cells were placed at 4°C and stained for the YAe epitope and total MHC II using an anti–I-A/I-E Ab and analyzed by flow cytometry. Infected cells were gated as GFP⁺ cells (Fig. 6A), and YAe and MHC II fluorescence intensity were measured. Importantly, control siRNA and Arl8b-silenced D1 cells were equally infected at each MOI tested (Fig. 6B). Moreover, the GFP mean fluorescence intensity was equivalent between control siRNA and Arl8b D1-silenced cells indicating that infected cells contained the same amount of bacteria per cell (Supplemental Fig. 4A, 4B). However, Arl8b-depleted D1 cells displayed a lower amount of surface I-A_b–Eα_(52–68) complex at the MOI tested (Fig. 6C). Indeed, the quantification of the mean fluorescence intensity of the YAe epitope revealed significantly reduced amounts of surface I-A_b–Eα_(52–68) presented by D1 cells when infected at MOI 20:1 and 50:1. In parallel, we controlled for the total surface MHC II levels on control siRNA and Arl8b-silenced DCs using an Ab that binds to all MHC II molecules. No significant difference in the global levels of surface MHC II was noted between control siRNA and Arl8b-depleted cells (Supplemental Fig. 4C, 4D).

To extend these findings in primary cells, we performed an infection assay in BMDCs at MOI 10:1. As described earlier, the percentage of control siRNA and Arl8b-silenced cells infected was equivalent (Supplemental Fig. 4G), but Arl8b-depleted cells displayed significantly reduced levels of surface I-A_b–Eα_(52–68) complex (Supplemental Fig. 4E, 4F). Taken together, these results confirm that Arl8b-silenced DCs have a defect in MHC II Ag presentation because of delayed processing/loading of internalized Ags within lysosomes that results in lower amounts of MHC II–Ag complexes presented at the cell surface.

In this article, we report a role for the small GTPase Arl8b in controlling MHC II Ag presentation in DCs. The depletion of Arl8b in DCs leads to a delay in the delivery of Ags to lysosomes and in the formation of MHC II–peptide complexes in MHC II compartments, as well as reduced display of complexes at the cell surface. The T cell activation defect in Arl8b-silenced DCs is correlated with reduced amounts MHC II–peptide complexes in the lysosome and a subsequent decrease in surface MHC II–peptide complexes. We previously showed that the presentation of lipid Ags by CD1d molecules was dependent on Arl8b directing traffic to lysosomes and facilitating the formation of CD1d–lipid complexes in lysosomes (21). Moreover, Garg et al. (21) showed that Arl8b interacts with the VPS41 subunit of the HOPS complex to mediate the delivery of endocytic vesicles to lysosomes. Interestingly, Arl8b interaction with VPS-41 has also been implicated in the clearance of apoptotic cells in C. elegans (27). HOPS is a tethering complex that facilitates the SNARE assembly on lysosomes and promotes the fusion of late endosomes to lysosomes (17, 34, 35). In this study, we extend those results to MHC II Ag presentation in DCs. Indeed, MHC II presentation is controlled by the GTPase Arl8b, which regulates the delivery of Ags to lysosomes and, therefore, the formation of MHC II–peptide complexes. The observed defect in delivery of Ags to lysosomes in Arl8b-depleted DCs is likely due to a defect in the recruitment of the HOPS complex to lysosomes by Arl8b, resulting in a defect in fusion of late endosomes to lysosomes.

These descriptions of the role of Arl8b in CD1d and MHC II Ag presentation highlight the crucial role of Arl8b in vesicle trafficking and immunological host defense. Many pathogens including Mycobacterium tuberculosis, Salmonella typhimurium, and Listeria monocytogenes escape cellular trafficking pathways to delay or avoid their delivery to lysosomes where they may be degraded (36, 37). Small GTPases are key regulators of vesicle trafficking, and Arl8b seems to be an excellent target for pathogens endeavoring to escape lysosomal degradation and the subsequent adaptive immune response (38). Indeed, it was demonstrated recently that the M. tuberculosis protein PtpA inhibits phagosome acidification and phagosome-lysosome fusion by subversion of the HOPS complex subunit VPS33b (39, 40). Consequently, M. tuberculosis may survive in the host cell and prevent MHC II and/or CD1d Ag presentation. Interestingly, another study revealed that Arl8b is required for the cell-to-cell transfer of S. typhimurium in HeLa cells. Here, the bacterium takes advantage of the second known effector of Arl8b, the kinesin-1 motor (25). These studies underscore the importance of Arl8b and its effectors as potential targets of immune evasion.

Newly synthesized MHC II molecules are transported to the lysosomes directly from the Golgi or the plasma membrane (6, 8, 9, 41). Interestingly, at steady-state we did not detect a defect in the localization of MHC II or H2-M molecules as both localize normally to lysosomes. This result may not have been expected because both molecules must traffic through endosomal routes to reach lysosomes. Several possibilities may explain this observation. First, RNA interference silencing is not complete, and under steady-state conditions normal localization levels may eventually be achieved, even if there is a delay in trafficking. Second, mammalian cells express a close Arl8b homolog, Arl8a, that although usually expressed at lower levels may partially compensate for loss of Arl8b. Finally, the GTPase Rab7 localizes to late endosomes and has been implicated in recruiting the HOPS complex, and thus contributes to the fusion of endosomes to lysosomes (17, 34, 42). A previous study reported that inward transport of the MHC II compartment along microtubules depends on the Rab7 interacting protein that is controlled by the cholesterol content within the cell (19).

Arl8b has also been described as a regulator of lysosome motility. Arl8b-depleted cells display lysosome clustering at the microtubule organizing center, whereas its overexpression causes a mislocalization of lysosomes to the cell periphery (21–23). Arl8b-regulated motility appears to be dependent upon its interaction with SifA and kinesin-interacting protein that binds the kinesin motor Kif5b. Moreover, Arl8b and these partners have been implicated in the control of lysosome tubulation in macrophages (24). Arl8b-depleted macrophages were unable to generate lysosomal tubules in response to LPS stimulation. Recently, we showed that Arl8b also controls the secretion of lysosome-related organelles such as lytic granules in NK cells (23). This raises the question of the role of Arl8b in the retrograde transport of MHC II molecules from the lysosome to the plasma membrane to enable T cell recognition. Interestingly, retrograde transport of peptide-loaded MHC II molecules in DCs requires lysosomal tubulation (14, 43–45). As the only GTPase known to be specific for lysosomes and in light of its roles in lysosome transport, Arl8b is an excellent candidate to promote MHC II transport to the plasma membrane during DC maturation. In our experimental approach, we cannot exclude that a component of the defect in MHC II presentation is partially due to a defect on the export of MHC II–peptide complexes from lysosomes to the cell surface. However, we did not detect any defect in MHC II cell-surface levels at steady-state or of surface MHC II levels post bacterial infection. Because MHC II is massively transported to the plasma membrane from the lysosome in DCs upon DC maturation (43), we had expected the possibility that Arl8b silencing might lead to a reduced amount of MHC II at the cell surface post bacterial infection. However, it is also possible that a part of the cell-surface MHC II is from a newly synthesized pool resulting from an increase of the transcription induced by maturation signals of bacterial origin. The exact role of Arl8b in the retrograde transport of MHC II will be interesting to explore in more detail during DC maturation. Given its emerging roles in both delivery to and movement of lysosomes, Arl8b now appears to be a centrally important GTPase regulating lysosomal biology and Ag presentation.

We thank M. Jenkins (University of Minnesota, Minneapolis, MN) for Eα_(52–68)-GFP construct. We thank L. Lynch, V. Hsu, and members of the Brenner laboratory for helpful discussions and critical reading of the manuscript. We also thank S.Y. Park for assistance with fast protein liquid chromatography purification.

The authors have no financial conflicts of interest.