The deubiquitinating enzyme ubiquitin C-terminal hydrolase-L1 (UCH-L1) is required for the maintenance of axonal integrity in neurons and is thought to regulate the intracellular pool of ubiquitin in the brain. In this study, we show that UCH-L1 has an immunological function in dendritic cell (DC) Ag cross-presentation. UCH-L1 is expressed in mouse kidney, spleen, and bone marrow–derived DCs, and its expression and activity are regulated by the immune stimuli LPS and IFN-γ. UCH-L1–deficient mice have significantly reduced ability to cross-prime CD8 T cells in vivo and in vitro because of a reduced ability of DCs to generate MHC class I (MHC I) peptide complexes for cross-presented Ags. Mechanistically, Ag uptake by phagocytosis and receptor-mediated endocytosis as well as phagosome maturation are unaffected by loss of UCH-L1 in DCs. Rather, MHC I recycling is reduced by loss of UCH-L1, which affects the colocalization of intracellular MHC I with late endosomal/lysosomal compartments necessary for cross-presentation of Ag. These results demonstrate a hitherto unrecognized role of the deubiquitinating enzyme UCH-L1 in DC Ag processing.
The CD8 T cell response is key in the elimination of infected or cancer cells. Naive CD8 T cells with their unique TCRs constantly survey and sample APCs within secondary lymphoid tissues in search of cognate peptide MHC molecules (1). Dendritic cells (DCs) are exceptional in their ability to stimulate naive T cells and are strategically located in both lymphoid and nonlymphoid sites where they constitutively sample the microenvironment and phagocytose microbial and apoptotic cells (2, 3). DCs have a specialized capacity to sample Ags from their environment by a process called cross-presentation to stimulate a CD8 T cell response. In this process, exogenous Ags are taken up by DCs and routed to the MHC class I (MHC I) rather than the MHC II pathway for presentation and activation of CD8 T cells (1). Mechanistically, there are two major cross-presentation pathways described to explain the acquisition of exogenous Ag on MHC I following internalization by endocytosis or phagocytosis. The phagosome-to-cytosol pathway describes the transfer of Ags from phagosomes or macropinosomes to the cytosol, where they access the proteasome and re-enter the phagosome for cross-presentation or follow the classical MHC I path of loading in the endoplasmic reticulum (ER). The vacuolar pathway, in contrast, does not require proteasomes or TAP and depends on gradual proteolysis of Ag within the endocytic system to generate appropriate peptides, which load on MHC I molecules within the same compartment (4).
Ubiquitination is a highly dynamic posttranslational modification necessary for a myriad of cellular processes, including protein degradation, trafficking, signaling, and epigenetic regulation (5). It begins with the covalent attachment of a single ubiquitin to a substrate Lys residue and can be modified further providing complex modifications. It provides an important mechanism for regulating immune responses (6). Indeed, polyubiquitination is a prerequisite for proteasome-dependent generation of peptide Ags in the MHC I classical and phagosome-to-cytosol cross-presentation pathways (4). Ubiquitination is, itself, a reversible process regulated by deubiquitinating enzymes (DUBs) that can fully deubiquitinate substrates to reverse function, edit polyubiquitin chains to direct a particular effect, or replenish monoubiquitin by cleaving polyubiquitin chains and recycling ubiquitin (7). UCH-L1 belongs to the ubiquitin C-terminal hydrolases (UCH) family of DUBs, which contain an unstructured loop in the active site that restricts access to short peptides conjugated to ubiquitin (8). UCH-L1 is thought to play an important role in the recycling of free ubiquitin whereby newly translated ubiquitin contains additional amino acids following the terminal glycine, which must be cleaved to expose the terminal glycine for conjugation (9). UCH-L1 is most highly expressed in mammalian brain (1–2% of total protein) and in the peripheral nervous system with low expression in testis, tumors, and kidney podocytes upon injury (10, 11). A major physiological role of UCH-L1 in neuronal function and axonal maintenance can be seen from UCH-L1–deficient animals that develop a dying-back type of axonal degeneration with paralysis and die prematurely (12–14). UCH-L1 deficiency affects synaptic plasticity with reduced acetylcholine release at the synaptic terminal at the neuromuscular junction (14). UCH-L1–mediated deficiencies at the synaptic level may be rescued by the introduction of monoubiquitin (15), suggesting a primary role for UCH-L1 in maintaining the monoubiquitin pool.
We recently described the expression of UCH-L1 in tubulointerstitial cells of the kidney of nonvascular origin, which phenotypically resemble DCs (16). We showed that UCH-L1–deficient mice exhibit an altered course of immune complex glomerulonephritis, and therefore, we wished to investigate whether UCH-L1 plays a role in the immune response. In the current study, we uncover a new role for UCH-L1 in DC cross-presentation. We describe the expression and regulation of UCH-L1 in DCs and show that loss of UCH-L1 impairs the CD8 T cell response to listeria and cell-associated Ag in vivo and in vitro. Mechanistically, Ag uptake and phagosome maturation are unaffected, whereas MHC I recycling and colocalization to late endosomal/lysosomal compartments show that UCH-L1 affects the ability of MHC I molecules to reach cross-presentation compartments competent for generating Ag-MHC I complexes.
Materials and Methods
Animals and Listeria monocytogenes infection model
UCH-L1 complete knockout animals on C57BL/6 background were generated (16) and bred at our animal facility. Both wild-type and heterozygous littermates were used as control group. OT-I animals transgenic for TCR-recognizing OVA257–264 peptide on H-2Kb and OT-II animals transgenic for TCR-recognizing OVA323–339 peptide on I-Ab were bred in-house. Animals were housed under pathogen-free conditions at the animal facility of the University Medical Center Hamburg-Eppendorf. All experiments were performed according to approved national and institutional ethical guidelines for animal care.
Age (>8 wk old)– and sex-matched animals were infected i.v. with the indicated doses of the wild-type Listeria monocytogenes strain LmEGD or of an L. monocytogenes strain expressing OVA (LmOVA) (17). Bacterial inocula were controlled by plating serial dilutions on tryptic soy broth agar plates. Bacterial load in spleen and liver was determined by mechanical disruption of the tissue, followed by 1:10 dilution series of the homogenate and plating on agar plates. LmEGD colonies were counted 24 h later.
Innate cell phenotyping of naive and 3-d infected spleens was performed by flow cytometry analysis of single-cell suspensions. Cells were Fc blocked with anti-CD16/32 (2.4G2; BD Biosciences) and stained with Abs against NK1.1 (PK136), CD11b (M1/70), CD11c (HL3), Ly-6GC (RB6-8C5), Ly-6C (AL-21), and MHC II (M5/114.15.2) (BioLegend and eBioscience).
CD8 T cell proliferation following LmOVA infection was determined by BrdU labeling. Briefly, mice were injected with 1 × 104 LmOVA i.v. On day 4 of infection, mice were injected with 1 mg BrdU (BD Biosciences) solution i.p. The following day, spleens were harvested, and cells were surface stained with Abs against CD3 (500A2) and CD8 (53-6.7) (BioLegend and eBioscience). Cells were then fixed, permeabilized, and treated with DNase to expose incorporated BrdU and stained with anti-BrdU Ab (BD Biosciences).
DC isolation, bone marrow–derived DC generation, and stimulation
Kidneys or spleens were excised, the kidney capsule was removed, and organs were incubated in 2 ml digest media/organ, injecting 1 ml directly into the organ (RPMI 1640 [Life Technologies], 0.1% FCS [Life Technologies], 1% HEPES [Invitrogen], 1% penicillin/streptomycin [Pen/Strep; Invitrogen], 1 mg/ml collagenase [C274; Sigma-Aldrich], 100 μg/ml DNase I [Roche]) for 30 min at 37°C. Organs were mashed and, in the case of kidney, incubated a further 30 min at 37°C. The cell suspension was filtered through a 100-μM filter. Red cell lysis was performed with 2 ml Ery-lysis buffer (NH4Cl 160 mM and Tris HCl 170 mM [9:1 ratio]). Cells were resuspended in MACS buffer (PBS without Ca2+Mg2+, 2 mM EDTA, and 1% FCS) and filtered through a 40-μM filter. Cells were blocked with normal mouse serum before incubation with anti-CD11c+ microbeads. DCs were isolated according to the manufacturer’s instructions (Miltenyi Biotec). Alternatively, DCs were depleted of B cells using Dynabeads (DYNAL) and FACS sorted with a BD FACSAria IIIu using the Abs against CD3 (500A2) and CD19 (1D3) (negative gating) and CD11c (HL3) and MHC II (M5/114.15.2) (positive gating) to isolate DCs.
Bone marrow–derived DCs (bmDCs) were prepared according to the method developed by Lutz et al. (18). Briefly, cells were flushed from femur and tibia leg bones, washed and counted, and 3 × 106 cells were plated in 10 ml media (RPMI 1640 [Life Technologies], 10% FCS [Life Technologies], 50 μM 2-ME (Invitrogen), 1 mM Na pyruvate [Invitrogen], 2 mM L-Glu [Invitrogen], 100 U/ml Pen/Strep [Invitrogen]) with 20 ng/ml murine GM-CSF (mGM-CSF) (PeproTech) on 100-mm plates. On day 3, 10 ml culture media with 20 ng/ml mGM-CSF was added. On day 6, 10 ml media with 10 ng/ml mGM-CSF was replaced. Day 8 nonadherent DCs were used for further experimentation. Day 8 bmDCs were characterized by Fc blocking with anti-CD16/32 (2.4G2; BD Biosciences) and staining with Abs against CD11c (HL3), MHC II (M5/114.15.2), CD40 (3/23), CD80 (16-10A1), CD86 (GL-1), and MHC I (AF6-88.5]). Cultured bmDCs were routinely over 80% positive for CD11c and MHC II and displayed low to intermediate levels of CD40, CD80, and CD86 and intermediate levels of MHC I and MHC II, therefore showing characteristics of immature DCs (18).
For stimulation, 1 × 106 per milliliter of cells were reseeded in 24-well plates with 1 μg/ml LPS (Escherichia coli 026:B6; Sigma-Aldrich) or 20 ng/ml IFN-γ (PeproTech).
Mouse spleen and human PBMC cell sorting for quantitative PCR
Mouse spleen was mashed and incubated with digestion medium (RPMI 1640, 10% FCS, 1% HEPES, 1% Pen/Strep, 0.4 mg/ml Liberase, and 100 μg/ml DNase) at 37°C with shaking for 30 min, followed by gentleMACS homogenization. Red cell lysis was performed with 2 ml Ery-lysis buffer (NH4Cl 160 mM and Tris HCl 170 mM [9:1 ratio]). Cells were resuspended in PBS and filtered through a 40-μM filter. Cells were stained with Near InfraRed Dead Cell Stain (Invitrogen), blocked with normal mouse serum (1:50), and incubated in MACS buffer with Abs (all at 1:200 dilution unless otherwise specified; BD Biosciences and BioLegend) against the following Ags to sort T cells and DCs: CD8 (53-6.7), CD45 (30-F11), MHC II (1:400 dilution, M5/114.15.2), CD11c (HL3), CD3 (17A2), and CD4 (RM4-5). Abs against CD19 (1D3), CD45 (30-F11), γδ (GL3), NK1.1 (PK136), and CD3 (17A2) were used to sort NK, NKT, γδ T, and B cells.
Human PBMCs were isolated by density gradient centrifugation from buffy coats obtained from the blood bank at the University Medical Center Hamburg-Eppendorf. For sorting, 150 × 106 PBMCs were resuspended in PBS and stained with Abs against CD45 (HI30, 1:500), CD3 (OKT3, 1:300), CD20 (2H7, 1:500), HLA-DR (L243, 1:200), CD14 (M5E2, 1:200), CD16 (3G8, 1:100), CD11c (3.9, 1:100), CD123 (6H6, 1:300), CD141 (M80, 1:500), and CD1c (L161, 1:300). DCs were defined as CD45+ lineage− HLA-DR+. From this population, plasmacytoid DCs were identified as CD11c− and CD123+ cells, and myeloid DCs (CD11c+, CD123lo) were further distinguished as conventional DC 1 (cDC1) (CD141hi, CD1c−) and cDC2 (CD1c+ CD141lo). All Abs were from BioLegend.
Cultured DCs were differentiated from sorted monocytes in a 24-well plate in full RPMI (RPMI containing 10% FCS, 1% Pen/Strep, and 1% l-glucose) with human GM-CSF (100 ng/ml; PeproTech) and human IL-4 (40 ng/ml; PeproTech) for 6 d.
In vitro restimulation and cytokine staining
Spleens were passed through a 70-μm filter into PBS. Erythrolysis was performed and the single-cell suspension was resuspended in complete RPMI 1640. Livers were perfused via the portal vein, followed by disruption through a 200-μm metal sieve in PBS and further through a 70-μm filter. Liver leukocytes were separated over a Percoll gradient, and erythrolysis was performed before resuspension in complete RPMI 1640. Cells were counted and restimulated in vitro with 1 × 10−6 M OVA257–264 and 1 × 10−5 M listeriolysin O (LLO)189–201 for 4.5 h. Ten micrograms per milliliter brefeldin A (eBioscience) was added for the last 3.5 h of stimulation. Cells were then washed with PBS and fixed in 2% paraformaldehyde (PFA) for 15 min at room temperature (RT) or in fixation buffer (eBioscience) for 1 h at 4°C. Cells were Fc blocked with anti-CD16/32 (2.4G2; eBioscience), surface stained with Abs against CD8 (53-6.7) and CD4 (RM4-5), and stained intracellularly with Abs against TNF-α (MP6-XT22) and IFN-γ (XMG1.2) in 0.3% saponin permeabilization buffer. Analysis was performed by flow cytometry.
Cell-associated OVA Ag
Ag for in vivo injection was prepared as previously described (19). In short, spleen and lymph node cells were prepared in serum-free medium from female BALB/c MHC-mismatched mice. Cells were incubated in hypertonic medium (0.25 M sucrose, 5% w/v polyethylene glycol 1000, and 10 mM HEPES in RPMI 1640 [pH 7.2]) with or without 10 mg/ml OVA for 10 min at 37°C. Cells were transferred into hypotonic medium (40% H2O and 60% RPMI 1640) and further incubated for 2 min at 37°C.
Cells were then washed with cold PBS and irradiated (1350 rad). A total of 20–35 × 106 mock or OVA-loaded cells were injected i.v. per mouse.
In vivo cross-priming assay
Lymph node and spleen cells from female OT-1 (mixed Thy1.1/1.2 genotype) mice were pooled. FITC-labeled Abs against CD11b (M1/70), CD19 (1D3), CD4 (RM4-4), Gr-1 (RB6-8C5), F4/80 (BM8), and MHC II (M5/114.15.2) were used to label APCs, neutrophils, and B and CD4 T cells. Labeled cells were depleted by negative selection using anti-FITC magnetic beads (Miltenyi Biotec). Eluted CD8 T cells were labeled with CFSE (5 μM, 15 min, 37°C; Molecular Probes) and injected i.v. (5 × 106 per mouse) into control littermate or UCH-L1–deficient mice. Next day, mice were immunized by i.v. injection of OVA- or mock-loaded BALB/c cells (20–35 × 106 per mouse). Two days later, spleens were collected, and cell suspensions were stained with Abs against Thy1.1 (HIS51), Thy1.2 (53-2.1), and CD8 (53-6.7) (eBioscience, BioLegend) and analyzed by flow cytometry.
In vivo cytotoxicity assay
Spleen cells were pulsed for 15 min at 37°C with OVA257–264 peptide (2 μg/ml), followed by labeling with 1 μM CFSE (CFSEhi cells) or left unpulsed and labeled with 0.1 μM CFSE (CFSElo cells). Both sets of targets (5 × 106 each) were injected i.v. After 4 h, the survival of target cells in the spleen was analyzed by flow cytometry. Specific lysis was calculated with the following formula: % specific cytotoxicity = 100 − (100 × (CFSEhi/CFSElo) primed/(CFSEhi/CFSElo) control).
In vitro cross-priming assay
A total of 1 × 106 day 8 bmDCs were seeded (1:1) with negatively selected CFSE-labeled CD8 OT-I or CD4 OT-II T cells in 24-well flat-bottom culture plates. Cell-associated Ag (5 × 106 cells)+/− OVA was added to the cultures and incubated for 1–3 d. Cells were collected and stained with Abs against Thy1.1 (HIS51), Thy1.2 (53-2.1), CD4 (RM4-5), CD8 (53-6.7), and CD69 (H1.2F3) (eBioscience, BioLegend) and analyzed by flow cytometry.
Phagocytosis and receptor-mediated endocytosis
For phagocytosis, 100 μg/ml OVA-A488 (Thermo Fisher Scientific) was adsorbed onto 3-μM latex beads (LB30-1 ml; Sigma-Aldrich) overnight according to the manufacturer’s instructions. A total of 5 × 105 bmDCs per 300 μl were incubated with OVA-A488–coated beads at 1:5 ratio with 0.1 μg/ml LPS for 5 h at 37°C. bmDCs incubated with OVA-A488–coated beads at 4°C served as negative controls. For receptor-mediated endocytosis, 5 × 105 bmDCs per 300 μl were incubated with 10 μg/ml soluble OVA-A488 and 0.1 μg/ml LPS for 20 min at 37°C. bmDCs incubated with OVA-A488 at 4°C served as negative controls. Cells were washed with ice-cold MACS buffer and phagocytosis or receptor-mediated endocytosis uptake was assessed by flow cytometry by gating on DCs.
Analyses were performed on a BD Biosciences LSR II system with Diva software, and data were analyzed using the FlowJo software, version 10 (Tree Star).
In vitro cross-presentation assay with beads
Two milligram per milliliter OVA (Sigma-Aldrich) was adsorbed onto 3-μM latex beads (LB30-1 ml; Sigma-Aldrich) overnight according to the manufacturer’s instructions. bmDCs were resuspended at 1 × 106/1 ml in 24-well plates on poly-d-lysine–coated coverslips (for immunofluorescence) or in 96-well plates (for FACS). Cells were pulsed with OVA-coated beads at 1:5 ratio plus 0.1 μg/ml LPS for 20 min, washed extensively with media, and chased for 5 h at 37°C.
Immunofluorescence and visualization of OVA-loaded MHC I complexes
For UCH-L1 staining of FACS-sorted splenic cells or bmDCs, cells were fixed with 4% PFA (Electron Microscopy Sciences) for 8 min at RT and washed with PBS three times for 5 min. Cells were blocked and permeabilized in blocking buffer (5% horse serum diluted in 0.05% Triton X-100 PBS) for 30 min at RT. Rat monoclonal anti–UCH-L1 Ab [U104; A. Grötzinger, Biochemistry University of Kiel (20)] was incubated at a 1:50 dilution over night at 4°C in blocking buffer. Staining was visualized using a Cy2-coupled donkey anti-rat Ab (1:400; Jackson ImmunoResearch Laboratories) for 30 min at RT in blocking buffer, and DNA was visualized using Draq5 (1:1000; Molecular Probes).
For visualization of OVA-loaded MHC I complexes, cells were fixed in 4% PFA, permeabilized in 2% Triton X-100, blocked in 5% horse serum, and Fc blocked with 1:100 rat anti-mouse Fc receptor (clone 2.4G2). Cells were stained with 1:25 unlabeled mouse 25-D1.16 Ab (eBiosciences) in 5% horse serum, an Ab that specifically binds OVA-loaded MHC I complexes. Secondary Ab binding was performed with an AF488-coupled donkey anti-mouse IgG (1:400; Jackson ImmunoResearch Laboratories). Specific staining was confirmed by lack of OVA complex signal in samples lacking OVA Ag and secondary Ab alone staining. The cytoskeleton was visualized using AF568 phalloidin (1:200; Molecular Probes), and DNA was visualized using Hoechst (1:1000; Molecular Probes) Quantitation of H-2Kb–OVA was performed by measuring the mean fluorescent intensity per cell area with Fiji in 10 random micrographs with two to three cells per micrograph per mouse in two independent experiments. Cell area was defined as the outer border of the actin cytoskeleton visualized by phalloidin immunofluorescence.
Staining of intracellular MHC I was performed with mouse H-2Kb Ab (AF6-18.104.22.168, 1:50; BioLegend) and visualized with Cy2 donkey anti-mouse secondary Ab (1:200; Jackson ImmunoResearch Laboratories). Staining of late endosomes was performed with rabbit Rab7 (1:200; Abcam), recycling endosomes with Rab11a (1:100; Aviva Systems Biology), and lysosomes with biotinylated LAMP-1 (1:100; eBioscience). Secondary Abs used were Cy5 donkey anti-rabbit Ab (1:100; Jackson ImmunoResearch Laboratories), or AF647 streptavidin (1:200; Vector). Staining was evaluated with an LSM 510 META microscope using the LSM software and an LSM 800 with Airyscan using the ZEN blue software (all ZEISS). Quantitation of colocalization of MHC I with the different vesicular compartments was performed using Pearson coefficient according to ZEISS instructions.
Single-phagosome flow cytometry
Analysis of DC phagosomes following phagocytosis of beads conjugated to OVA was analyzed by flow cytometry in isolated phagosomes as previously described (21). In brief, bmDCs were incubated with OVA-coated beads at 1:10 ratio plus 0.1 μg/ml LPS at 16°C for 30 min and washed and incubated for different chase times at 37°C to allow phagosome maturation. Cell-surface, noninternalized OVA beads were blocked with normal mouse serum (1:50) and Ab labeled with rabbit anti-OVA Ab (1:500; Sigma-Aldrich) and goat anti-rabbit Ab (1:1000; Jackson ImmunoResearch Laboratories) before cells were disrupted by passing them through a 23-gauge needle in homogenization buffer (250 mM sucrose, 3 mM imidazole [pH 7.4], and 1 × protease inhibitor mixture in PBS) to release/isolate phagosomes. Phagosomes were blocked (normal mouse serum; 1:50) and labeled with primary Abs against OVA (1:500; Sigma-Aldrich) and biotinylated LAMP-1 (1:100; eBioscience), followed by secondary anti-rabbit Cy5 and streptavidin BV421, respectively (1:1000; Jackson ImmunoResearch Laboratories, streptavidin, 1:1000; eBioscience). Phagosome pellets were resuspended in 200 μl PBS for acquisition by flow cytometry (LSR II; BD Biosciences).
Surface molecule recycling and internalization assays
Recycling of transferrin receptor (TfR) and MHC I was assayed by mAb labeling and direct capture of recycled mAb/protein complexes reaching the cell surface by fluorochrome-conjugated secondary Abs according to the quantitative method described by Blagojević Zagorac et al. (22). In brief, bmDCs were labeled with primary biotinylated Abs against TfR and MHC I (2 μg/ml in media; eBioscience) for 60 min to allow internalization. Uninternalized Abs were acid stripped in 0.2 M citric acid/0.2 M Na2HPO4 buffer (pH 3) on ice for 2 min. Cells were further incubated at 37°C over a time course to allow receptor recycling back to the surface, where secondary Ab labeling with streptavidin conjugate (1:100; BioLegend) of the recycled primary Abs occurred. Nonrecycled surface labeling (nonrecycle negative control) was determined following acid stripping by labeling protein/Ab complexes with streptavidin conjugate. The intracellular pool of TfR or MHC I before recycling (intracell t = 60) was assessed following acid stripping by fixing and permeabilizing cells and labeling intracellular protein/Ab complexes with streptavidin conjugate. Recycling was quantified by flow cytometry and percentage of recycling was calculated using the following equation:
Internalization of surface MHC I or TfR was performed by incubating bmDCs in media containing brefeldin A at 37°C over the time course 0, 60, 120, and 240 min. Cells were then washed, blocked, and stained with primary Abs against MHC I biotin or TfR biotin (1:250; eBioscience) and secondary Abs against biotin (streptavidin 1:200; BioLegend). Internalization was quantified by flow cytometry.
Immunoblots were performed from isolated kidney or spleen DCs or cultured bmDCs. Samples were lysed in T-PER (Pierce) (containing 1 mM sodium fluoride, 1 mM sodium vanadate, and 100 nM calyculin A, complete [Roche]) and denatured with 4 × LDS. SDS-PAGE and Western blotting procedures were carried out according to standard protocols. The following primary Abs were used: rat anti–UCH-L1 (1:250) (20), mouse anti–β-actin (1:3000; Sigma-Aldrich), mouse anti-GAPDH (1:1000; Sigma-Aldrich), rat anti-hemagglutinin (HA) (1:1000; Roche), mouse anti-ubiquitin (1:1000; Millipore), rabbit anti-K48 (1:1000, Apu2; MilliporeSigma), rabbit anti-β5 (1:1000; X. Wang, University of South Dakota), and rabbit anti-LMP7 (1:5000; E. Krüger, Biochemistry, Greifswald). HRP-conjugated secondary Abs (1:15,000–30,000; Jackson ImmunoResearch Laboratories) detected binding, which was visualized with ECL SuperSignal (Pierce) using Amersham Imager 600 (GE Healthcare).
DUB activity assay
DUB activity was assessed by incubation of 20 μg lysate from bmDCs with 0.8 μg of the HA-tagged ubiquitin/vinyl methyl ester probe (HA-Ub-VME) in labeling buffer (50 mM Tris [pH 7.4], 5 mM MgCl2, 250 mM sucrose, 1 mM DTT, and 1 mM ATP) for 1 h at 37°C. Proteins were then resolved on SDS-PAGE 4–20% gradient gels, and blots were subsequently probed with rat anti-HA Ab. Labeled proteins were identified based on their migration on SDS-PAGE and by comparison with previous published data in which the specific bands were analyzed by mass spectroscopy (15, 23).
Chymotrypsin-like activity assay
For measurement of proteasomal activity, 10 μg total protein was diluted in incubation buffer (20 mmol/l HEPES, 0.5 mmol/l EDTA, 5 mmol/l DTT, 0.1 mg/ml OVA, 6 mmol/l ATP, 50 mmol/l phosphocreatine, and 0.2 U phosphocreatine kinase in water [pH 7.8]) to a final volume of 50 μl. Samples were preincubated in incubation buffer for 2 h at 4°C. After preincubation, the substrate for chymotrypsin-like activity Suc-LLVY-AMC (Calbiochem; EMD Chemicals, division of Merck, Darmstadt, Germany) was added to the samples at a final concentration of 60 μmol/l. Proteasomal activity was measured at 355 and 460 nm using a fluorescent spectrophotometer (Mithras LB 940; Berthold Technologies, Wildbad, Germany) after incubation at 37°C for 1 h in the dark.
Whole-cell or tissue RNA was extracted using the RLT Plus Kit (Qiagen) according to the manufacturer’s instructions. RNA was reverse transcribed into cDNA with random hexamer primer (Invitrogen) and Moloney murine leukemia virus reverse transcriptase (New England Biolabs). mRNA was amplified by 40 cycles using SYBR Green (Life Technologies) on a StepOnePlus Detector (Applied Biosystems). The following primer sequences were used: mouse UCH-L1–forward (fw) 5′-AGC TGG AAT TTG AGG ATG GA-3′; mouse UCH-L1–reverse 5′-GGC CTC GTT CTT CTC GAA A-3′; human UCH-L1–fw 5′-GGT GGC ACA GAC AAC CAA AAA-3′; human UCH-L1–reverse 5′-GTT CTC TCC TCG AAA GCT CCT-3′; mouse LMP7-fw 5′-TCG CTC GGA CCC AGG ACA CTA C-3′; mouse LMP7-reverse 5′-TTC TCC GTC CCC ACC CAG GGA-3′; mouse β5-fw 5′-TGG GGT GTC CCA GAA GAG CCA-3′; mouse β5-reverse 5′-CCC GCT GTA GCC CTG GAG TCA-3′; mouse/human 18s-fw 5′-TTC GAA CGT CTG CCC TAT CAA-3′; and mouse/human 18s-reverse 5′-CTG CCT TCC TTG GAT GTG GTA-3′. All samples were run in duplicate and normalized to 18s rRNA. Relative quantification of gene expression was calculated using the ΔΔ cycle threshold method.
Data are expressed as mean ± SEM. All statistical analyses were performed with GraphPad Prism 5 software. A p value <0.05 was considered to be statistically significant. The two-tailed Student t test was used for comparison between two groups unless otherwise stated.
UCH-L1 is expressed in DCs, and its expression and enzymatic activity is regulated by LPS and IFN-γ stimulation
UCH-L1 is one of the most abundant proteins found in the mammalian brain and is also highly expressed in the peripheral nervous system (24, 25). We have previously shown that UCH-L1 is expressed in glomerular and tubulointerstitial cells of the kidney closely resembling DCs (16). This study focuses on the novel expression of UCH-L1 in DCs and its impact on the immune response. Western blot analysis demonstrated the constitutive expression of UCH-L1 in bmDCs, as well as in CD11c+ DCs isolated from mouse kidney and spleen, whereas UCH-L1 protein was not detected in UCH-L1–deficient DCs from our constitutive UCH-L1–deficient mouse line littermates (Fig. 1A). To assess expression levels of UCH-L1 transcript in other immune cells, we analyzed FACS-sorted splenic CD8 T, CD4 T, CD8+ DCs, CD8− DCs, B, NK, NKT, and γδ T cells by quantitative PCR (qPCR) (Fig. 1B, Supplemental Fig. 1A–C). The results show that B cells have the highest mRNA expression of UCH-L1, followed by CD8− DCs and γδ T cells. All other populations have lower expression levels in comparison with bmDCs as reference (Fig. 1B). We also assessed the expression levels of UCH-L1 by qPCR in immune cells sorted from human PBMCs. For this analysis, T cells, monocytes, plasmacytoid DCs, and subsets of DCs were sorted according to the gating scheme (Supplemental Fig. 1D). DCs cultured from monocytes in GM-CSF and IL-4 for 6 d were also included in the analysis. The results show that cultured DCs have the highest expression of UCH-L1. DC subsets including the cDC1 cross-presenting subset have similarly low levels of UCH-L1 relative to PBMCs as reference, whereas T and B lymphocytes have even lower expression of UCH-L1 relative to PBMCs as reference (Fig. 1C). At the protein level, UCH-L1 staining appeared diffuse within the cytoplasm in FACS-sorted murine splenic DCs and bmDCs (Fig. 1D, 1E), whereas no specific staining was seen in FACS-sorted non–B/DC cells or in FACS-sorted splenic DCs of UCH-L1–deficient mice (Fig. 1D). Stimulation with LPS, resulted in a time dependent upregulation of UCH-L1 expression, whereas stimulation with IFN-γ decreased UCH-L1 expression at protein and mRNA levels (Fig. 1F, 1H, 1I), demonstrating a specific time-dependent regulation of UCH-L1 in response to DC stimulation. Indeed, DUB activity measured using the ubiquitin-derived, activity-based HA-UB-VME showed that LPS not only enhanced UCH-L1 transcript and protein levels, but also the amount of active UCH-L1 enzyme (Fig. 1G).
CD8 T cell response is impaired in UCH-L1–deficient mice following listeria infection
To assess whether UCH-L1 had an impact on the innate and/or the adaptive immune response, we infected UCH-L1–deficient mice and littermate controls with listeria and analyzed the early innate and adaptive responses. Infection with the LmEGD resulted in a strong recruitment of neutrophils, DCs, and inflammatory monocytes into the spleen as measured by flow cytometry in naive animals and on day 3 of infection, representing the early innate response (Fig. 2A, 2B, Supplemental Fig. 2A). Significantly fewer neutrophils were recruited in UCH-L1–deficient animals compared with controls (Fig. 2B); however, this did not result in a significant alteration in liver and spleen bacterial load between the groups (Fig. 2C) on day 3 postinfection. To examine whether UCH-L1 deficiency affected the adaptive response to listeria, we performed infections with the listeria strain LmOVA, which expresses OVA and may be used to analyze the Ag-specific T effector response (17). On day 9 postinfection, cells from spleen and liver were incubated with the peptides OVA257–264 and LLO189–201 to stimulate listeria-induced CD8 and CD4 T cells, respectively (Fig. 3A). The frequencies of IFN-γ– and TNF-α–secreting, OVA257–264-specific CD8 T cells generated in response to listeria infection were measured by flow cytometry and were significantly lower in UCH-L1–deficient spleens compared with controls (Fig. 3B, Supplemental Fig. 2B) with a similar trend for liver (Fig. 3C). In contrast, the frequencies of IFN-γ– and TNF-α–secreting, LLO189–201-specific CD4 T cells were not significantly different between the groups in spleen or liver (Fig. 3D, 3E). We further assessed CD8 T cell proliferation in vivo following LmOVA infections by measuring T cell BrdU incorporation, as outlined in the experimental scheme (Fig. 3F). The proliferation of CD8 T cells was significantly reduced on day 5 in spleens of UCH-L1–deficient animals compared with littermate controls (Fig. 3G, 3H). Taken together, the data show that loss of UCH-L1 affects the ability to mount an Ag-specific CD8 T cell response to listeria infection.
UCH-L1 is required for DC-mediated cross-priming of the CD8 T cell response
To investigate the impairment of the CD8 T cell response in UCH-L1–deficient animals and considering the expression of UCH-L1 in DCs, an in vivo approach was undertaken to examine cytotoxicity of Ag-specific activated CD8 T cells (Fig. 4A). In this setting, OVA was injected with the adjuvants α-galactosylceramide (α-Galcer) or CpG oligonucleotides. This approach results in uptake of OVA by professional APCs and cross-presentation of OVA peptides on MHC I to CD8 T cells (26). On day 4, CFSE-labeled OVA257–264 loaded and unloaded control targets were injected and OVA257–264-specific lysis was determined. Specific killing was significantly lower in UCH-L1–deficient animals compared with littermate controls for both α-Galcer and CpG (Fig. 4B). To distinguish whether the reduction in cytotoxic killing occurred as a result of a defect in the APCs or a defect in the T cells, we set up a second in vivo protocol to measure the CD8 T cell response following phagocytic uptake of cell-associated Ags, which occurs through cross-priming (19). According to the scheme (Fig. 4D), animals were injected with CFSE-labeled, TCR-transgenic, OVA-specific OT-I CD8 T cells. One day later, MHC-mismatched, apoptotic cell–associated OVA was injected. In this scenario, the Ag must first be taken up by host APCs and be processed and presented on MHC I to stimulate a specific CD8 T cell response, a process that is primarily performed by DCs. We found robust proliferation of OT-I cells in littermate controls with significant reduction of proliferation in UCH-L1–deficient animals, as measured by dilution of incorporated CFSE label in OT-I (Fig. 4E, 4F). Animals injected with cell-associated Ag lacking OVA served as negative controls. Taken together, the results show that UCH-L1 is affecting the priming of the CD8 T cell response. Because CD8 T cells exhibit low to undetectable levels of UCH-L1 protein (Figs. 1D, 4C), this suggests that the effects of UCH-L1 on CD8 priming are mediated through UCH-L1’s function in DCs.
To demonstrate, that the observed in vivo effects on CD8 T cell priming were a result of an altered DC function in UCH-L1–deficient mice, we explored the role of UCH-L1 in DC-mediated cross-priming in vitro with control or UCH-L1–deficient littermate bmDCs. For this, day 8 bmDCs were incubated with cell-associated OVA according to the scheme (Fig. 5A). Naive bmDCs from littermate control and UCH-L1–deficient DCs used for these experiments were shown to have equivalent frequencies of CD11c+ MHC II+ DCs as well as equivalent surface levels of MHC and costimulatory molecules (Supplemental Fig. 3A–D). CFSE-labeled OT-I CD8 or OT-II CD4 T cells were added, and processing, presentation, and priming were allowed to develop over 3 d. The resulting proliferation of OVA-specific OT-I T cells was significantly reduced in cultures with UCH-L1–deficient DCs (Fig. 5B, 5C), showing that, indeed, UCH-L1 deficiency in DCs affects their ability to mount a strong CD8 T cell response. To measure the CD4 T cell response, upregulation of CD69 was used as a readout of CD4 T cell activation because OT-II cells have restricted proliferation. In this setting, littermate or UCH-L1–deficient bmDCs incubated with cell-associated, OVA-induced equivalent activation of OVA-specific OT-II CD4 T cells (Fig. 5D, 5E). We conclude that UCH-L1 expression in DCs is necessary to mount an efficient CD8, but not CD4 T cell response.
Formation of H-2Kb–OVA complexes following exogenous uptake is perturbed in UCH-L1–deficient DCs
Cross-presentation describes the process whereby exogenous Ags are taken up by endocytosis or phagocytosis and follow the phagosome-to-cytosol or vacuolar pathways for loading Ag on MHC I molecules. To examine the loading of MHC I molecules through cross-presentation, we performed phagocytosis assays using OVA-coated, 3-μM beads as exogenous Ag plus 0.1 μg/ml LPS capable of inducing cell maturation and thus promoting cross-presentation (27, 28). Littermate control and UCH-L1–deficient bmDCs were pulsed with OVA-coated beads for 20 min, and the development of MHC I-OVA complexes was assessed after 5 h. Confocal imaging of the formation of H-2Kb–OVA complexes showed a specific vesicular staining pattern in control bmDCs with no nonspecific staining in bmDCs exposed to beads without OVA as negative controls (Fig. 6A). In contrast, the presence of H-2Kb–OVA complexes was significantly reduced in UCH-L1–deficient bmDCs (Fig. 6). This difference was not due to a defect in OVA-coated bead uptake, as we did not find significant differences in phagocytosis or receptor-mediated endocytosis (Supplemental Fig. 3E–H). Additionally, this difference was most likely not due to a defect in Ag processing through the ubiquitin proteasomal system, as we did not find significant differences in the overall ubiquitination status or proteasome levels/activity in bmDCs (Supplemental Fig. 4).
Phagosome maturation is equivalent in control and UCH-L1–deficient bmDCs
Following phagocytosis, the phagosome undergoes intracellular trafficking and sequential fusion with endosomes and lysosomes to allow the acquisition of NADPH oxidases, vacuolar ATPases, and hydrolases for proteolysis and degradation of the phagocytic content, a process termed phagosome maturation (29). In DCs, phagosome maturation results in the generation of immunogenic peptides for Ag presentation. In this study, we have used single-phagosome flow cytometry to assess LAMP-1 acquisition and phagocytic content degradation simultaneously. Control and UCH-L1–deficient bmDCs were incubated with OVA-coated beads and LPS at 16°C for 30 min to allow internalization and further incubated for different chase times at 37°C to allow phagosome maturation. Phagosomes were then isolated and labeled with Abs against LAMP-1 and OVA for flow cytometry analysis. Phagosomes were gated on single-bead-containing phagosomes, noninternalized beads were excluded, and phagosomes were analyzed for acquisition of LAMP-1 and degradation of OVA protein (Fig. 7A). Phagosomal LAMP-1 showed a similar percentage increase on control and UCH-L1–deficient DC phagosomes over the time course of phagosome maturation (Fig. 7B). Degradation of OVA within control and UCH-L1–deficient phagosomes also showed a similar increasing percentage over the time course of phagosome maturation (Fig. 7C). We conclude that DC loss of UCH-L1 does not interfere with phagosome maturation and must therefore rather affect trafficking of MHC I molecules.
MHC I and TfR recycling is significantly reduced in UCH-L1–deficient bmDCs
During cross-presentation, MHC I molecules are recruited either from the plasma membrane or from the ER, for both phagosome-to-cytosol and vacuolar pathways of cross-presentation. Subsequently, peptide is loaded onto MHC I in the phagosome, before MHC I/Ag complexes are then shuttled to the plasma membrane for presentation. To assess the trafficking dynamics of MHC I molecules, we performed recycling and internalization assays of MHC I (internalized by clathrin‐independent endocytosis) and, as comparison, TfR (internalized by clathrin‐dependent endocytosis). Recycling was analyzed by primary Ab labeling, internalization, surface acid stripping, and time course incubation to allow receptor recycling back to the surface where secondary Ab labeling could occur. The recycling of both MHC I and TfR was significantly reduced on UCH-L1–deficient bmDCs over the time course examined (Fig. 8A, 8B). In contrast, the internalization of MHC I and TfR from the plasma surface were not altered between control and UCH-L1–deficient bmDCs (Fig. 8C, 8D). These results show that UCH-L1 is perturbing the recycling of MHC I molecules necessary to reach the cross-presenting compartment for Ag loading to take place.
Colocalization of MHC I with late endosomal/lysosomal vesicles is perturbed in UCH-L1–deficient DCs
Targeting of MHC I to a compartment within the endocytic pathway competent for cross-presentation is required for loading of cross-presented Ag. Therefore, we examined the surface and intracellular distribution of MHC I in littermate control and UCH-L1–deficient bmDCs treated with OVA-coated beads plus 0.1 μg/ml LPS for 20 min, followed by a 5 h chase. High-resolution confocal imaging of intracellular MHC I (green) colocalizing to endosomal and lysosomal compartments revealed that colocalization to Rab7+ late endosomal and LAMP-1+ endolysosomal compartments was strongly reduced in UCH-L1–deficient (UCH-L1−/−) bmDCs, whereas colocalization to EEA1+ early endosomal and Rab11a+ recycling/storage endosomal compartments was not affected (Fig. 9A–D). Interestingly, overall surface levels of MHC I were not significantly different between control and UCH-L1–deficient bmDCs for both naive and treated conditions (Fig. 9E). These results show that loss of UCH-L1 affects trafficking of a proportion of recycling MHC I, which localizes to the cross-presenting compartment for loading. This is visualized with a simplified scheme showing MHC I trafficking within the endocytic pathway and indicating where UCH-L1 is proposed to exert its effects (Fig. 9F).
In this study, we describe the expression of the DUB UCH-L1 in DCs and reveal a key role for UCH-L1 in cross-presentation and activation of the CD8 T cell response. Following stimulation with LPS, we could show a strong upregulation of UCH-L1 expression and enzymatic activity in bmDCs, whereas stimulation with IFN-γ–decreased UCH-L1 expression showing that UCH-L1 expression is strongly regulated by immune stimuli. LPS stimulation results in a phase of massive protein synthesis (<4 h), followed by strong downregulation of synthesis during DC maturation (30). We show that LPS-mediated upregulation of UCH-L1 begins after 4 h when protein synthesis is downregulated. We and others have shown that UCH-L1 affects protein synthesis pathways via mTORC1 activity in neurons and the brain (31, 32). This has consequences for Ag processing, in particular cross-presentation, whereby UCH-L1–mediated suppression of protein synthesis activity would favor loading of exogenous Ags on recycling MHC I molecules, an important source of MHC I for cross-presentation (33–35).
To assess the relevance of UCH-L1 in the immune response, we adopted a well-established listeria infection model with the possibility to examine Ag-specific responses. A significant reduction in the generation and proliferation of Ag-specific CD8, but not CD4, T cells was observed in UCH-L1–deficient animals compared with controls, showing that loss of UCH-L1 affects the ability to mount an Ag-specific CD8 T cell response. The immune response to listeria involves both innate and adaptive arms of immunity with the CD8 T cell response necessary for clearance and long-term protection from the bacterium (36, 37). Bacterially infected macrophages locate to the splenic T cell zone, where they encounter TNF-α and NO-producing DCs, cDCs, and T cells, producing Ags that undergo cross-presentation and CD8 T cell activation. We have previously shown the importance of cross-presentation for the CD8 T cell response to listeria (38). Assessment of the CD8 T cell response following phagocytosis of cell-associated Ag or receptor-mediated endocytosis of Ag showed that indeed UCH-L1 expression affected the ability of animals to cross-prime CD8 T cells in vivo. It is possible that the in vivo effects of UCH-L1–deficient mice may involve other cell types, including B cells, that also express UCH-L1. Indeed, B cells have been shown to play a supportive role in the T cell response to listeria infection, however, not directly involving their Ag presentation function (39). However, B cells or γδ T cells are not thought to have the ability to cross-prime T cell responses in vivo (4). Further, our in vitro cross-priming experiments show that it is UCH-L1’s expression specifically in DCs that affects the ability to mount a strong CD8 T cell response, whereas the CD4 T cell response is not significantly altered. Intriguingly, another DUB, YOD1, has previously been implicated in cross-presentation, whereby expression of a dominant-negative form (YOD1-C160S) resulted in enhanced cross-presentation by DCs (40). The authors postulated that YOD1 might be involved in broader membrane-associated functions as part of an endosome/phagosome-associated machinery. Cellular functions of DCs required for cross-presentation such as phagocytosis and receptor-mediated endocytosis were not affected by loss of UCH-L1. Surprisingly, proteasome activity and composition, as well as changes to overall ubiquitin levels, were also unaffected in UCH-L1–deficient bmDCs. Of note, UCH-L1 has been shown to have hydrolase-independent activity interacting with several cellular proteins to affect autophagy and cell proliferation (41, 42), suggesting that UCH-L1 function in DCs could indeed be different from in neuronal tissue where it is thought to function primarily in maintaining monoubiquitin. In humans, UCH-L1 mutations lead to early onset neurodegeneration with progressive neurologic dysfunction with no immunological phenotype outside of the nervous system being described to date. However, our findings of an immunological function of UCH-L1 in DCs and the expression of UCH-L1 in human immune cells including DCs, warrants future investigation of the immune response in humans with UCH-L1 mutations.
The formation of cross-presented MHC I/OVA complexes depends on two main events, namely phagosomal maturation to generate antigenic peptides for direct loading or egress to the cytosol and, second, trafficking of MHC I molecules to meet Ags within the cross-presenting compartment. In vitro cross-presentation assays with OVA-coated beads showed significant reduction in the formation of MHC I/OVA complexes in UCH-L1–deficient bmDCs. However, when phagosome maturation was examined, both phagosomal LAMP-1 acquisition and OVA degradation were unaffected by UCH-L1 loss over the time course of phagosome maturation, suggesting that it is the trafficking of MHC I that is specifically affected by UCH-L1. Trafficking of MHC I within DCs for cross-presentation takes several routes, including recycling from the surface to endolysosomes (33–35); Ags from the phagosome-to-cytosol pathway may be loaded on newly synthesized MHC I molecules in the ER (4), and alternatively, a proportion of newly synthesized MHC I molecules may be targeted via binding to the CD74 chaperone invariant chain into endolysosomal compartments for cross-presentation (43). Our studies identify UCH-L1 playing a specific role in the recycling route of the plasma membrane proteins MHC I and TfR. We show that recycling is necessary for the localization of MHC I to the late endosomal/lysosomal compartment, where cross-presentation takes place. Thus, according to our simplified scheme UCH-L1 is proposed to act on MHC I by removing/editing its ubiquitin-based signal to promote recycling at the junction between recycling and endocytic maturation. It is well appreciated that ubiquitination provides an important signal in promoting distinct trafficking routes of many plasma membrane proteins whereby ubiquitinated proteins within the endocytic system are delivered to the multivesicular body pathway, recognized by endosomal sorting complex required for transport (ESCRT) machinery, and sequestered to intraluminal vesicles of multivesicular bodies for degradation (44). What our study indicates and what is just beginning to be understood is how the diversity of ubiquitin-based recognition in endosomal trafficking is controlled by DUBs (45). Of note, membrane-associated RING-CH (MARCH) family of E3 ligases induce ubiquitination and degradation of plasma membrane proteins involved in the immune response, including MHC II and CD86. Importantly, the opposition of ubiquitination promotes recycling of these receptors (46, 47). We conclude that UCH-L1 alters MHC I recycling necessary for the localization and loading of MHC I molecules into the cross-presenting compartment for subsequent activation of the CD8 T cell response.
This work was supported by the Else Kröner Fresenius Stiftung (to A.T.R., J.R., and C.M.-S.) and by Deutsche Forschungsgemeinschaft SFB1192 (to C.M.-S.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
bone marrow–derived DC
conventional DC 1
HA-tagged ubiquitin/vinyl methyl ester probe
L. monocytogenes strain expressing OVA
- MHC I
MHC class I
ubiquitin C-terminal hydrolase.
The authors have no financial conflicts of interest.