Lymphocryptovirus Infection of Nonhuman Primate B Cells Converts Destructive into Productive Processing of the Pathogenic CD8 T Cell Epitope in Myelin Oligodendrocyte Glycoprotein

Abstract

EBV is the major infectious environmental risk factor for multiple sclerosis (MS), but the underlying mechanisms remain obscure. Patient studies do not allow manipulation in vivo. We used the experimental autoimmune encephalomyelitis (EAE) models in the common marmoset and rhesus monkey to model the association of EBV and MS. We report that B cells infected with EBV-related lymphocryptovirus (LCV) are requisite APCs for MHC-E-restricted autoaggressive effector memory CTLs specific for the immunodominant epitope 40-48 of myelin oligodendrocyte glycoprotein (MOG). These T cells drive the EAE pathogenesis to irreversible neurologic deficit. The aim of this study was to determine why LCV infection is important for this pathogenic role of B cells. Transcriptome comparison of LCV-infected B cells and CD20(+) spleen cells from rhesus monkeys shows increased expression of genes encoding elements of the Ag cross-presentation machinery (i.e., of proteasome maturation protein and immunoproteasome subunits) and enhanced expression of MHC-E and of costimulatory molecules (CD70 and CD80, but not CD86). It was also shown that altered expression of endolysosomal proteases (cathepsins) mitigates the fast endolysosomal degradation of the MOG40-48 core epitope. Finally, LCV infection also induced expression of LC3-II(+) cytosolic structures resembling autophagosomes, which seem to form an intracellular compartment where the MOG40-48 epitope is protected against proteolytic degradation by the endolysosomal serine protease cathepsin G. In conclusion, LCV infection induces a variety of changes in B cells that underlies the conversion of destructive processing of the immunodominant MOG40-48 epitope into productive processing and cross-presentation to strongly autoaggressive CTLs.

FIGURE 1.

RNA sequence data. High-throughput Illumina single-end sequencing was performed on three replicates of BLC, CD20⁺, CD20⁻, and PBMC samples. (A) PCA was done on the samples, and the first and second components are depicted as a scatter plot. (B) Differential gene-expression analysis was performed on BLCs versus CD20⁺ and BLCs versus CD20⁻, and the number of differentially expressed genes (false discovery rate < 0.01, logFC > 1) is depicted as a Venn diagram. (C) In these differential gene-expression lists, selected genes associated with Ag presentation, lysosome, and proteasome are depicted as a sample clustered heat map with Z-scores across samples. A list of all genes with their fold change (BLC versus CD20⁺, BLC versus CD20⁻, and CD20⁺ versus CD20⁻) is shown in Supplemental Table I.

FIGURE 2.

qPCR of cathepsin transcripts in MNC subsets from rhesus monkey and marmosets. Total RNA was extracted from rhesus monkey (A) and marmoset (B) PBMCs and spleen MNCs. Also, CD20⁺ and CD20⁻ fractions were isolated from rhesus monkey samples. cDNA was synthesized for qPCR using primer–probe combinations (Table II). Transcript levels were normalized against ABL. Data are presented as mean ± SEM.

FIGURE 3.

Proteolytic degradation of rhMOG. A fixed concentration (2 μg/ml) of rhMOG was incubated for 3, 24, or 48 h with different concentrations (5, 2.5, and 0.5 μg) of lysates from Mm-SCs and Cj-SCs (A), Mm-SC CD20⁺ and Mm-SC CD20⁻ fractions (D), and Mm-BLC and Cj-BLC lysates (G). After the indicated time intervals aliquots of the lysates were run on a 4–12% Bis-Tris gradient gel, which was stained with Coomassie Blue to visualize degradation. For each gel, the percentage degradation of rhMOG was calculated as described in Materials and Methods. The percentage degradation of rhMOG by the different lysates is shown for Mm-SCs (B), Cj-SCs (C), Mm-SC CD20⁺ (E), Mm-SC CD20⁻ (F), Mm-BLCs (H), and Cj-BLCs (I). Second to last lane contained lysate with OVA as an irrelevant Ag. OVA was not degraded and is not shown in this figure. Data shown are a representative of at least three replicates.

FIGURE 4.

Involvement of CatG and CatH in rhMOG degradation. (A) Activity measurements of CatG (left panel) and CatH (right panel) in the indicated cell lysates. Measurements were performed in the absence or presence of the inhibitors CMK and E64. Data are presented as mean ± SEM. Lysates were prepared from total Mm-SCs, as well as from CD20⁻ and CD20⁺ cell fractions. (B) Lysates were incubated with rhMOG for 24 h in the presence or absence of CMK or E64. (C) Percentage degradation of rhMOG. (D and E) The same analysis was performed for nonfractionated Cj-SCs. Data shown are a representative of at least three replicates. *p < 0.05.

FIGURE 5.

Alignment of recombinant MOG sequences. Recombinant MOG sequences encoding from 1 to 125 were aligned for the following species: human, rhesus monkey, marmoset, mouse, rat, and naked mole rat. The pathogenically relevant T cell epitope in the rhMOG sequence MOG24–36 (MHC class II/Caja-DRB1*W1201–binding Th1 epitope) is single underlined, whereas the one in MOG40–48 (MHC class I/Caja-E–binding NK-CTL epitope) is double underlined. The two epitopes are located in close proximity within the highly conserved sequence 21–58. The only difference between the sequence in mouse/marmoset/rat versus rhesus monkey/human MOG is a substitution of proline (P) by a serine (S) at position 42 (indicated in green). The Arg (R) residues, which can be citrullinated by posttranslational modification, are indicated in red. The pound sign (#) marks the only N-glycosylation site at the asparagine (N) at position 31.

FIGURE 6.

Degradation of bead-coated Mm-MOG35–51 and Cj-MOG35–51. Peptides that were labeled at the N terminus with biotin and at the C-terminal end with FITC were coated on magnetic beads via streptavidin. The peptide-coupled beads were incubated for 24 h at 37°C in lysates of CD20⁺ (upper panel) and CD20⁻ (lower panel) SC fractions from rhesus monkeys (A) or Mm-BLCs (upper panel) and Cj-BLCs (lower panel) (B). The indicated inhibitors were tested at the optimal concentration. Depicted is the release of FITC in arbitrary units (AU) from the peptide into the supernatant, as a measure of proteolytic activity.

FIGURE 7.

Mm-MOG35–51 and Cj-MOG35–51 degradation by different cell lysates. (A) Cj-MOG35–51 and Mm-MOG35–51 peptides were incubated for 24 h at a fixed concentration (10 μg/ml) with total SC lysate and with CD20⁺ and CD20⁻ spleen fractions from rhesus monkeys. (B) Percentage degradation of the MOG peptides. Total Cj-SCs (C and D) and Cj-BLCs and Mm-BLCs (E and F) were tested under the same conditions. All lysates were incubated in the presence or absence of the inhibitors CMK and E64. Data shown are a representative of at least three replicates.

FIGURE 8.

Citrullination prevents degradation of MOG peptides. MOG35–51 peptides derived from the marmoset and rhesus monkey sequence were synthesized; Arg at positions 41 and/or 46 were replaced by Cit. The peptides were incubated for 24 h at a fixed concentration (10 μg/ml) with total SC lysate, CD20⁺ and CD20⁻ spleen fractions from rhesus monkeys (A), marmoset total SCs (B), or Mm-BLC lysates and Cj-BLC lysates (C). All lysates were incubated in the presence or absence of the inhibitors CMK and E64. The graphs represent the percentage degradation of the indicated peptide. Data shown are a representative of at least three replicates. SDS-PAGE gels are shown in Supplemental Fig. 1.

FIGURE 9.

Expression of PAD and LC3 in LCV-infected B cells. Total RNA was extracted from different cell lysates to determine expression of mRNA encoding PAD2 (A) and PAD4 (B) by qPCR. Transcript levels were normalized against the reference gene ABL. Median of each cell population is shown. (C) Cell lysates were immune-blotted for LC3 and GAPDH as loading control. As positive controls, Abs against LC3-I and LC3-II were run along with the other samples. (D) The density of the LC3-I and LC3-II band in (C) is related to the GAPDH band. (E) Immunofluorescence staining was performed with rhesus monkey cells (left panel) and marmoset cells (right panel) with anti-PAD2, anti-PAD4, and anti-LC3. Original magnification ×400. *p < 0.05, **p < 0.001, Cj-BLC/Mm-BLC versus other cell subsets, Mann–Whitney U test.