NADPH oxidase modulates MHC class II antigen presentation by B cells

Summary

Phagocyte NADPH oxidase plays a key role in pathogen clearance via reactive oxygen species (ROS) production. Defects in oxidase function result in chronic granulomatous disease (CGD) with hallmark recurrent microbial infections and inflammation. The oxidase′s role in the adaptive immune response is not well-understood. Class II presentation of cytoplasmic and exogenous Ag to CD4⁺ T cells was impaired in human B cells with reduced oxidase p40^phox subunit expression. Naturally arising mutations which compromise p40^phox function in a CGD patient also perturbed class II Ag presentation and intracellular ROS production. Reconstitution of patient B cells with wild-type, but not a mutant, p40^phox allele restored exogenous Ag presentation and intracellular ROS generation. Remarkably, class II presentation of epitopes from membrane Ag was robust in p40^phox-deficient B cells. These studies reveal a role for NADPH oxidase and p40^phox in skewing epitope selection and T cell recognition of self Ag.

Introduction

The phagocyte NADPH oxidase plays a critical role in microbial killing by catalyzing electron transfer from NADPH to molecular oxygen giving rise to superoxide and other forms of ROS (1). This oxidase contains several phox (phagocyte oxidase) subunits including gp91^phox and p22^phox which comprise flavocytochrome b₅₅₈ as well as cytoplasmic p40^phox, p47^phox, and p67^phox. The binding of ligands to phagocyte receptors stimulates cytoplasmic subunit translocation to membrane-bound flavocytochrome b₅₅₈ facilitating superoxide production (1). Phox subunit mutations are associated with CGD, characterized by the absence of or marked reduction in ROS production, recurrent pathogenic infections, and granulomatous inflammation (2, 3).

While the microbicidal role of phagocyte NADPH oxidase during the innate immune response to pathogenic infections is established (2), its function in APC-T cell interactions is less well-understood. In gp91^phox-deficient dendritic cells, but not macrophages, NADPH oxidase tempers phagosome acidification, preserving internalized Ag for efficient MHC class I-mediated cross-presentation to CD8⁺ T cells (4, 5). By contrast, CD4⁺ T cell activation was enhanced in response to murine macrophages with p47^phox mutations (6). Thus, mutation of distinct oxidase subunits may differentially affect cellular immune responses.

Evidence is conflicting regarding a direct role for NADPH oxidase in regulating MHCII Ag presentation. Presentation of exogenous ovalbumin but not tetanus toxoid Ag was altered in APC from CGD patients (7, 8). Neither study identified the defective oxidase subunits in the CGD patient-derived APC. ROS produced by NADPH oxidase can regulate autophagy in phagocytes (9), and oxidase subunits are detected in endosomes and phagosomes (10). Presentation of exogenous Ag via MHCII requires Ag transit and proteolysis in endosomes and lysosomes to yield peptide ligands (11). Cytoplasmic or nuclear Ag can also access MHCII via several autophagy pathways (12). MHCII αβ complexes are directed via the invariant chain (Ii) to endosomes where proteases fragment Ii (13). HLA-DM, whose function is regulated by HLA-DO, facilitates the removal of these fragments and antigenic peptide capture by MHCII. The resulting peptide-MHCII complexes then traffic to the cell surface for immune surveillance by CD4⁺ T cells.

The p40^phox subunit has been associated with Crohn′s disease and rheumatoid arthritis in genome wide association studies (14–16), yet its role in Ag presentation has not been investigated. Studies here examine how loss of p40^phox in human B cells affects MHCII Ag presentation and oxidase function. In the absence of functional p40^phox, human B cells displayed a reduced capacity for cytoplasmic Ag presentation. Perturbations in p40^phox also disrupted MHCII exogenous Ag presentation, yet presentation of membrane autoantigens was efficient. Disruption in p40^phox compromised intracellular but not extracellular ROS production by B cells. These results suggest roles for NADPH oxidase and its regulatory subunit p40^phox in skewing epitope selection by MHCII.

Materials and Methods

Cell lines

Human B lymphoblastoid cell lines (B-LCL) and T cells have been described (17, 18). Lentiviral shRNA targeting human p40^phox (19) or β2-microglobulin (hβ2M) transcripts (Sigma-Aldrich) were used to transduce B-LCL to generate p40^phox- or β2M-deficient cells. Institutional approval was obtained for human blood collection. A B-LCL, AR40 from a p40^phox-deficient patient (3) (DRβ1*0101, DRβ1*0701) was generated and transduced to express DRβ1*0401 (17). Retroviruses encoding p40^phox WT or p40^phoxR105Q were used to transduce AR40.DR4 (3). Frev and AR40 B-LCL synthesize Igλ but not Igκ chains.

Western blotting

Cell lysates were analyzed by immunoblotting (10, 17) using Ab for p22^phox (10), p40^phox(Upstate), p47^phox, p67^phox (BD Biosciences, San Jose, CA), gp91^phox (10), and GAD (Sigma-Aldrich). The mAb DA6.147 detects HLA-DRα and αβ dimers (20). The mAb PIN1.1 detects Ii (21). Membranes were stripped and reprobed for β-actin (Sigma-Aldrich) or GAPDH (Chemicon) as controls for sample loading. Quantity One®1-D Analysis Software (BioRad) was used to quantify protein expression.

Antigen presentation

APC were incubated +/− 10–20 μM human serum albumin (HSA), human IgG Ag (Sigma-Aldrich) or peptides HSA_64–76, κI_188–203, and κII_145–159 (Quality Controlled Biochemicals) for 6 hr at 37°C. APC were washed and incubated with epitope-specific T cells for 24 hr at 37°C prior to analysis for T cell activation (17). Data shown is the average of triplicate samples for each assay, and the error bars indicate the mean T cell activation +/− the standard deviation. Statistical comparisons between two groups were performed using an unpaired t test while comparison among three groups was performed using a one-way ANOVA. In each case, p≤0.01 was considered to be significant. Adjustment for multiple comparisons was made using the Bonferroni correction.

Flow cytometry and ROS production

Cells were fixed, permeabilized, and incubated with the mAb MaP.DM1 or HLA-DO (BD Biosciences) (17). To detect intracellular ROS, viable cells were incubated with 5 μM CM-H₂DCFDA (Invitrogen) and stimulated +/− 10 μg/mL PMA (Sigma) for 30 min at 37°C (22). Cellular ROS production was sensitive to the oxidase inhibitor diphenyleneiodonium.

Results and Discussion

Diminished cytoplasmic Ag presentation in B-LCL with reduced p40^phox expression

Autophagy promotes cytoplasmic Ag presentation by MHCII (12, 23). To test if p40^phox plays a role in cytoplasmic GAD Ag presentation, p40^phox expression was disrupted in PriessGAD B-LCL using shRNA. Expression of p40^phox was reduced ~80% in these cells compared to parental PriessGAD while GAD Ag expression was unperturbed (Supplemental Fig. 1A–B). The ability of p40^phox-deficient B-LCL to present GAD epitopes was substantially reduced (Fig. 1A). PriessGAD cells transduced with control hβ2M shRNA (Supplemental Fig. 1C) stimulated GAD-specific T cells comparably with parental PriessGAD (Fig. 1A). Levels of LC3-II, a marker of autophagosome formation (24), were similar in each B cell line tested, suggesting little change in autophagy with disruption of p40^phox (Supplemental Fig. 1D). Surface expression of HLA-DR4 was equivalent for PriessGAD and shRNA-treated cells (Supplemental Fig. 1E–F). Notably, PriessGAD cells with reduced p40^phox levels stimulated T cells specific for endogenous Ag Ig κ more efficiently than the parental PriessGAD cells (Fig. 1B). These results suggest reduced p40^phox expression in B-LCL may compromise cytoplasmic Ag presentation while favoring epitope presentation from endogenous Ig κ.

FIGURE 1.

Altered Ag presentation in B-LCL with reduced p40^phox expression. A, PriessGAD, PriessGAD + p40^phox shRNA, or PriessGAD + hβ2M shRNA cells were cultured with GAD-specific T cells to measure cytoplasmic GAD presentation. B, PriessGAD and PriessGAD + p40^phox shRNA cells were cultured with either κI- or κII-specific T cells to measure endogenous Ig κ presentation. In data not shown, APC were titrated from 1000–125 cells/well in order to determine the effect of reduced p40^phox expression on endogenous Ig κ presentation. Data in A–B representative of 3 independent experiments. (**p≤0.01, ***p≤0.001, ****p≤0.0001)

Reducing B-LCL p40^phox expression disrupted MHCII exogenous Ag presentation

To evaluate the role of p40^phox in exogenous Ag presentation, p40^phox expression was diminished ~80% by treating Frev B-LCL with a p40^phox-specific shRNA (Supplemental Fig. 1G). T cell responses to the exogenous Ag HSA and Frev cells with diminished p40^phox levels were significantly reduced compared to Frev cells transduced with hβ2M shRNA (Fig. 2A, Supplemental Fig. 1H) or the parental line Frev (data not shown). The ability of Frev cells with reduced p40^phox expression to present exogenous IgG Ag was also evaluated. Frev cells with reduced p40^phox levels were less effective in presenting exogenous IgG to T cells compared with Frev transduced with control hβ2M shRNA (Fig. 2B–C) or the parental line Frev (data not shown). Surface expression of HLA-DR4 was equivalent in each cell line (Supplemental Fig. 1I–J). Only minor differences were observed in exogenous peptide presentation by Frev cells with diminished p40^phox expression compared to cells treated with hβ2M shRNA (Fig. 2A–C). These results suggest that a reduction in p40^phox expression in B-LCL may perturb multiple routes for Ag presentation by MHCII.

FIGURE 2.

Reduced exogenous Ag presentation in B-LCL with diminished p40^phox expression. A, Frev + p40^phox shRNA or Frev + hβ2M shRNA cells were incubated with HSA Ag or HSA_64-76 peptide and cultured with HSA-specific T cells to measure MHCII presentation. Data representative of 2 independent experiments. B and C, Frev + p40^phox shRNA or Frev + hβ2M shRNA cells were incubated with human IgG Ag, κI_188-203 peptide, or κII_145-159 peptide and cultured with either κI-specific T cells (B) or κII-specific T cells (C) to measure MHCII presentation. Minor differences in peptide presentation were observed using B-LCL treated with p40^phox shRNA based on statistical analyses. Data in B–C representative of 3 independent experiments. (**p≤0.01, ***p≤0.001, ****p≤0.0001)

Analysis of the MHCII pathway in B-LCL with mutations in p40^phox

A new genetic subgroup of CGD with mutations in the gene NCF4 encoding p40^phox was described in a patient and linked to functional defects in the neutrophil NADPH oxidase (3). In this patient, one NCF4 allele harbors a frame-shift mutation with a premature stop codon while the other allele encodes a point mutation (R105Q) resulting in a nonfunctional form of p40^phox. A B-LCL from this patient (AR40) expressing HLA-DR4 was further transduced with either wild-type p40^phox (AR40.DR4.p40^phox WT) or the R105Q mutant allele (AR40.DR4.p40^phox R105Q) to evaluate the effects of p40^phox mutation and reconstitution. Immunoblots demonstrated reduced p40^phox expression in AR40.DR4 cells, consistent with the frame-shifted NCF4 allele as seen in patient neutrophils (3), and higher levels of wild-type or mutant p40^phox in the reconstituted cells (Fig. 3A). Levels of gp91^phox and p67^phox were comparably reduced in the reconstituted cells relative to the patient line, likely due to clonal variation. Extracellular ROS production upon PMA stimulation was similar in the patient and wild-type p40^phox reconstituted B cells (Supplemental Fig. 2A) as observed in p40^phox-deficient neutrophils (3). However, basal and PMA-inducible intracellular ROS production were reduced in the AR40.DR4 cells compared to cells expressing wild-type p40^phox (Fig. 3B), thus suggesting a defect in the ability of the oxidase in the patient B cells to produce intracellular ROS similar to p40^phox-deficient neutrophils (3).

FIGURE 3.

Expression of oxidase subunits and MHCII in p40^phox-deficient and reconstituted B-LCL. A, Cell lysates were immunoblotted to detect oxidase subunits and GAPDH. Values indicate the ratio of oxidase subunit levels to a loading control. B, B cells were incubated with CM-H₂DCFDA to detect intracellular ROS production and left untreated (UT) or PMA stimulated. C, Cell lysates were immunoblotted for HLA-DRα, HLA-DRαβ, Ii, and GAPDH. Values indicate the ratio of MHCII or Ii to GAPDH. D, B cells were stained to detect HLA-DM or HLA-DO. Data in panels A–D representative of at least 3 independent experiments.

Whether the absence of functional p40^phox in AR40.DR4 B-LCL influenced the expression of molecules in the MHCII Ag presentation pathway was tested. A slight reduction in the levels of total HLA-DRα, HLA-DRαβ, and Ii were observed in AR40.DR4.p40^phox WT and AR40.DR4.p40^phox R105Q compared to AR40.DR4 (Fig. 3C). Maturation of Ii was not impaired in AR40.DR4 as detected by the presence of mature glycosylated forms of Ii (* in Fig. 3C) in the p40^phox-deficient and reconstituted B-LCL. Changes in HLA-DM and HLA-DO can alter Ag presentation without perturbing T cell responses to synthetic peptides (25–27). AR40.DR4, AR40.DR4.p40^phox WT, and AR40.DR4.p40^phox R105Q expressed equivalent levels of HLA-DM and HLA-DO (Fig. 3D). Taken together, these results suggest that the absence of functional p40^phox in AR40.DR4 did not substantially alter the levels of HLA-DR, -DM, -DO, and Ii.

Reconstitution of p40^phox-deficient B-LCL partially restored MHCII exogenous Ag presentation

Naturally occurring mutations in p40^phox impacted the ability of B-LCL to efficiently present Ag to MHCII-restricted T cells (Fig. 4). Reconstitution of the patient-derived AR40.DR4 cells with wild-type p40^phox, but not the mutant p40^phox R105Q allele, enhanced exogenous HSA Ag presentation (Fig. 4A). AR40.DR4 cells were unable to present either Ig κI or κII epitopes to T cells (Fig. 4B–C). Reconstitution of AR40.DR4 cells with p40^phox WT restored κI epitope presentation to a greater extent than with p40^phox R105Q expression (Fig. 4B). Only reconstitution of AR40.DR4 with p40^phox WT facilitated presentation of the κII epitope from exogenous hIgG (Fig. 4C). Although AR40.DR4 cells were able to present exogenously added synthetic peptides to T cells (Fig. 4), the level of κI peptide presentation was reduced compared to either AR40.DR4.p40^phox WT or AR40.DR4.p40^phox R105Q (Fig. 4B). Surface expression of HLA-DR4 was equivalent in each cell line (Supplemental Fig. 2B). Exogenous tetanus toxoid Ag presentation was reduced not only in AR40.DR4 but in B cells deficient in another oxidase subunit gp91^phox (Supplemental Fig. 2C), consistent with perturbations in oxidase function modulating MHCII Ag presentation. Reconstitution of AR40.DR4 with wild-type p40^phox restored tetanus presentation (Supplemental Fig. 2C). Addition of an extracellular source of ROS failed to reconstitute tetanus presentation by B cells deficient in p40^phox or gp91^phox (Supplemental Fig. 2D). The ability of AR40.DR4 to endocytose a model exogenous Ag, FITC-albumin, was equivalent for p40^phox-deficient or reconstituted B-LCL (Supplemental Fig. 2E). In data not shown, we observed a similar persistence of the FITC-albumin at longer time points (6–18 hr) in each cell line. These results suggest that p40^phox-deficiency does not substantially affect the internalization or initial degradation of a model exogenous protein.

FIGURE 4.

Exogenous but not endogenous Ag presentation was defective in p40^phox-deficient B-LCL. A–C, B cells were incubated with HSA Ag or HSA_64-76 peptide (A), human IgG Ag (B–C), κI_188-203 peptide (B), or κII_145-159 peptide (C) as in Fig. 2 to measure MHCII presentation. In data not shown, cells without Ag or peptide failed to activate epitope-specific T cells. (D) B cells were cultured with T cells to measure endogenous HLA-A epitope presentation. Data in A–D representative of 3 independent experiments. (**p≤0.01, ***p≤0.001, ****p≤0.0001)

The presence of assembled HLA-DRαβ dimers on the surface of AR40.DR4 suggested these MHCII may acquire peptides from a source other than exogenous Ag. The ability of these cells to present antigenic peptides derived from an endogenous transmembrane protein was evaluated using an HLA-DR4-restricted T cell recognizing an epitope from MHC class I HLA-A. AR40.DR4 cells activated the HLA-A-specific T cells while reconstitution with either the p40^phox WT or R105Q mutant allele did not enhance this Ag presentation (Fig. 4D). Total MHC class I expression in each of the AR40.DR4-derived cells was equivalent (data not shown). These results suggest that while MHCII-restricted exogenous Ag presentation was impaired in the p40^phox-deficient cells, the presentation of an endogenous transmembrane protein in the context of MHCII could be readily detected.

In conclusion, the microbicidal role of phagocyte NADPH oxidase in neutrophils and macrophages during the innate immune response to pathogenic infections is well-established, but the oxidase's role in APC during the adaptive immune response is less clear. Here, functional p40^phox was shown to be required for the efficient presentation of cytoplasmic GAD and multiple exogenous Ag by MHCII in B cells. Additionally, studies suggest that functional gp91^phox is important for MHCII presentation of exogenous Ag. In phagocytes, cytoplasmic p40^phox binds to membrane phosphatidylinositol 3-phosphate (PI3P) and promotes assembly of the NADPH oxidase complex on phagosomes (19). Association of p40^phox and the oxidase with PI3P found in endosomal/lysosomal membranes (3) could influence MHCII Ag processing and presentation within these organelles. The cytosolic localization of the mutant p40^phox R105Q and lack of PI3P binding (3) may explain its inability to completely restore exogenous Ag presentation in p40^phox-deficient B cells. In B cells, the oxidase may regulate BCR signaling (28) and as revealed here, MHCII Ag presentation. Constitutive and inducible intracellular ROS production was higher in B cells expressing functional p40^phox, supporting a direct role for p40^phox in regulating B cell intracellular ROS generation. Interestingly, p40^phox-deficient B cells were capable of presenting epitopes derived from endogenous membrane-resident proteins suggesting that p40^phox may modulate the peptide repertoire displayed by MHCII on B cells and subsequently, CD4⁺ T cell activation. Alterations in the function of the oxidase in B cells may therefore contribute to genetic predisposition to autoimmunity in some CGD patients. Increased incidences of rheumatoid arthritis, inflammatory bowel disease, as well as discoid lupus have been associated with oxidase subunit mutations and CGD (29).