Increased CD8+ T cell response to Epstein-Barr virus lytic antigens in the active phase of multiple sclerosis

Abstract

It has long been known that multiple sclerosis (MS) is associated with an increased Epstein-Barr virus (EBV) seroprevalence and high immune reactivity to EBV and that infectious mononucleosis increases MS risk. This evidence led to postulate that EBV infection plays a role in MS etiopathogenesis, although the mechanisms are debated. This study was designed to assess the prevalence and magnitude of CD8+ T-cell responses to EBV latent (EBNA-3A, LMP-2A) and lytic (BZLF-1, BMLF-1) antigens in relapsing-remitting MS patients (n = 113) and healthy donors (HD) (n = 43) and to investigate whether the EBV-specific CD8+ T cell response correlates with disease activity, as defined by clinical evaluation and gadolinium-enhanced magnetic resonance imaging. Using HLA class I pentamers, lytic antigen-specific CD8+ T cell responses were detected in fewer untreated inactive MS patients than in active MS patients and HD while the frequency of CD8+ T cells specific for EBV lytic and latent antigens was higher in active and inactive MS patients, respectively. In contrast, the CD8+ T cell response to cytomegalovirus did not differ between HD and MS patients, irrespective of the disease phase. Marked differences in the prevalence of EBV-specific CD8+ T cell responses were observed in patients treated with interferon-β and natalizumab, two licensed drugs for relapsing-remitting MS. Longitudinal studies revealed expansion of CD8+ T cells specific for EBV lytic antigens during active disease in untreated MS patients but not in relapse-free, natalizumab-treated patients. Analysis of post-mortem MS brain samples showed expression of the EBV lytic protein BZLF-1 and interactions between cytotoxic CD8+ T cells and EBV lytically infected plasma cells in inflammatory white matter lesions and meninges. We therefore propose that inability to control EBV infection during inactive MS could set the stage for intracerebral viral reactivation and disease relapse.

Conflict of interest statement

Marco Salvetti, MD, received lecture fees from Biogen-Dompé and research support from Bayer-Schering, Biogen-Dompé, Merck-Serono, Sanofi-Aventis. Claudio Gasperini, MD, has served as a consultant for Merck Serono and Biogen Idec, and has received speaker honoraria from Teva, Merck Serono, Bayer Shering and Biogec Idec. Diego Centonze, MD, is an Advisory Board member of Merck-Serono, Teva and Bayer Shering and received funding for travelling and speaker honoraria or consultation fees from Merck Serono, Teva, Novartis, Bayer Shering, Sanofi-Aventis and Biogen Idec. He is also an external expert consultant of the European Medicine Agency (EMA) and the principal investigator in clinical trials for Novartis, Merck Serono, Teva, Bayer Shering, Sanofi Aventis and Biogen Idec. The other authors have no competing interests to report. This does not alter our adherence to all PLoS Pathogens policies on sharing data and materials.

Figures

Figure 1. Prevalence of EBV-specific CD8+ T cell responses in HD and MS patients.

(A) HLA-A*0201 and HLA-B*0801 pentamer+ CD8+ T cells specific for peptides from EBV latent (left panel) and lytic (left) proteins were investigated in HLA matched HD (n = 43), total untreated MS patients (n = 79) and the same patients subdivided into active (n = 31) and inactive (n = 48) MS patients (a-MS and i-MS, respectively) based on clinical and MRI criteria, as shown in Table 2. (B) IFN-β (n = 20) and natalizumab (NTZ) (n = 14) treated patients were analyzed as whole population or subdivided into active and inactive patients, as above. The numbers within or above the columns correspond to the percentages of individuals with detectable EBV-pentamer+ CD8+ T cells (grey columns) among the total donors tested (white columns); p values were calculated with Pearson's chi-squared test.

Figure 2. Magnitude of EBV-specific CD8+ T cell responses in HD and MS patients.

The frequencies of CD8+ T cells specific for EBV latent and lytic antigens were assessed in HD (n = 17) and total untreated MS patients (n = 26) (A), as well as in untreated active MS (a-MS, n = 13) and inactive MS (i-MS, n = 13) patients, IFN-β-treated active MS patients (a-MS IFN-β, n = 4) and natalizumab (NTZ)-treated inactive MS patients (i-MS NTZ, n = 14) (B, C). The percentages of antigen-specific CD8+ T cells within the CD3+ CD8+ T cell population were analyzed by staining with the corresponding HLA-A*0201 and HLA-B*0801 pentamers. Symbols in B and C refer to individual responses to peptide epitopes from the indicated EBV latent and lytic proteins from each donor. The frequencies were calculated after gating on total, live CD3+ CD8+ T cells. Data are shown in logarithmic scale. In A the bars represent the median ± the minimum and maximum value while in B and C the bars represent mean values ± SD; p values are calculated with unpaired t-test with 95% confidence intervals. (D) Representative flow cytometric expression profiles for pentamer+ CD8+ T cells specific for the indicated EBV latent and lytic antigens in HD, untreated inactive MS and active MS patients are shown in D. Numbers represent percentages of pentamer+ cells within the CD3+ CD8+ T-cell population.

Figure 3. Longitudinal monitoring of EBV-specific CD8 T cell responses in relapsing remitting MS.

The frequencies of CD8+ T cells specific for EBV lytic and latent antigens and CMV pp65 antigen were measured at different time points in 2 untreated MS patients (A), 2 natalizumab-treated MS patients (B) and 3 HD (C) using HLA-A*0201 and HLA-B*0801 pentamers. All patients were monitored clinically and with MRI at the times indicated by the arrows. The untreated MS patients had gadolinium-enhancing MRI lesions 20 days (MSB2/B2-2) and 1 day (MSA14) before the peak of the CD8+ T cell response to EBV lytic antigens while the 2 natalizumab-treated patients (MS BTY5 and MS BTY8) did not show any disease activity. Percentages of pentamer+ cells within the CD3+ CD8+ population are given on the y axis; the numbers on the x-axis indicate months since the start of the observation period (A, C) or of drug therapy (B).

Figure 4. Immunohistochemical detection of BZLF-1 in the inflamed meninges of the MS brain.

(A) A prominent immune infiltrate comprising a perivascular B-cell-follicle-like structure is visualized with anti-CD20 antibody in the meninges lining a cerebral sulcus of a MS case. (B) Immunostaining for the EBV immediate early lytic protein BZLF-1 in a serial section reveals the presence of several BZLF-1+ cells at the periphery of the B-cell follicle. Ig+ cells are present in the same area (inset). (C, D) High magnification pictures of Ig+ plasmablasts/plasma cells (green) coexpressing BZLF1 (red); these groups of cells correspond to those shown in the top left and bottom right parts of panel B. (E) Expression of the EBV early lytic protein BRFR1 (red) in a substantial proportion of Ig+ (red) plasmablasts/plasma cells in the same area shown in panel B; double immunofluorescence staining for Ig and BFRF-1 was performed in a section adjacent to those stained in A and B. (F) Two BZLF-1+ cells in another part of the B-cell-rich infiltrate shown in panel A. Nuclei were stained with DAPI (blue) in panels B-F. (G-K) Other examples of BZLF-1+ cells at the border of a meningeal B-cell follicle (G, H) and in a diffuse meningeal infiltrate (I-K) from a different MS case. Arrows in G and I point to the nuclei that are immunoreactive for BZLF-1 in H and J, respectively. The inset in panel K shows an Ig+ cell (green) with a BZLF-1+ nucleus (red) at high magnification. Bars = 100 µm in A and inset in B; 50 µm in B; 20 µm in C, E, F, K and inset in E; 10 µm in D, G, H and inset in K.

Figure 5. Immunohistochemical detection of BZLF-1 in acute white matter lesions of the MS brain.

(A) A large, B-cell enriched perivascular immune infiltrate surrounding 3 blood vessels in an active white matter lesion of a MS case is visualized with anti-CD20 antibody. (B) Immunostaining for myelin-oligodendrocyte glycoprotein (MOG) in a serial section reveals presence of myelin. (C) The same lesion shows massive microglia/macrophage activation in the parenchyma. Sections shown in panels A–C were counterstained with hematoxylin. (D, E) Double immunofluorescence staining for Ig (green) and BZLF-1 (red) reveals the presence of several Ig+ plasmablasts/plasma cells co-expressing BZLF-1 in the portion of the perivascular cuff marked with a frame in panel A. The upper and lower insets in E show 1 and 2 Ig+ cells co-expressing BZLF-1 at high power magnification, respectively. (F) Double immunofluorescence staining for BFRF-1 (red) and Ig (green) shows presence of double labelled BFRF-1+/Ig+ cells in the same area of the perivascular cuff stained for BZLF-1 in E. (G, H) Cells double stained for Ig (green) and BZLF-1 (red) in smaller perivascular cuffs of the same active lesion shown in A–C. Bars: 200 µm in A and C; 100 µm in B; 50 µm in D, E, G; 20 µm in F and H; 10 µm in the insets in D.

Figure 6. BZLF-1 gene expression in MS brain immune infiltrates.

RNA was extracted from the indicated areas microdissected from sections of a control lymph node and from brain sections of 4 MS cases (MS79, MS92, MS342, MS180). Quantitative real-time RT-PCR for BZLF-1, for CD19 and GAPDH was performed following pre-amplification of cDNAs, as described in Materials and methods. Data for BZLF-1 and CD19 are normalized for the GAPDH level of expression. BZLF-1 mRNA was detected in 3 out of 4 meningeal B-cell follicles and in the perivascular cuff of an active white matter lesion (cuff A-WML), but not in lymph node, perivascular cuffs of chronic active WM lesions (cuff CA-WML), demyelinated white matter parenchyma (lesioned WM), demyelinated grey matter parenchyma (lesioned GM), and normal-appearing white matter (NAWM). The presence of CD19 signal indicates that the microdissected region includes a B-cell containing immune infiltrate. Values are means ± SD of triplicate values.

Figure 7. Cytotoxic attack on EBV infected cells in the MS brain.

(A, B) Double immunofluorescence stainings for granzyme B (GrB, green) and CD8 (red) performed on sections of highly inflamed MS brain samples show granzyme B expression in a substantial proportion of CD8+ T cells present in two perivascular inflammatory cell infiltrates of an active white matter lesion; granzyme B-containing granules are localized in the cytoplasm and on the surface of CD8+ cells (inset in A). (C) Presence of granzyme+ CD8+ cells in a diffuse meningeal immune infiltrate. The arrow points to one of the rare granzyme B+/CD8− cells. (D) Double immunofluorescence staining for CD8 (red) and granzyme B (green) shows that granzyme B−/CD8+ cells accumulate outside and inside a B-cell follicle (marked with a dotted line and stained for CD20 in the left inset), whereas double labelled granzyme B+/CD8+ cells are detectable only at the periphery the B-cell rich area (arrows). Two CD8+ cells, one of which is also granzyme B+, are shown at high power magnification in the inset on the right. (E) Group of CD8+ cells (green) close to an EBV lytically infected, BFRF-1+ cell (red) at the border of a B-cell follicle. (F) The CD8+ cell closer to the BRFR-1+ cell extends cytoplasmic processes (arrow) that adhere to the infected cell (high magnification of the area marked with the frame in E). (G) Three CD8+ cells (green) with a lymphobast-like morphology are close to, but do not touch, a BFRF-1+ cell (red) in the inflamed meninges; two of the CD8+ cells display granzyme B-containing granules on their surface (blue, arrows). (H) Two BFRF-1+ cells (red), one of which is enclosed by a CD8+ cell (green) and both are covered by granzyme B+ granules (blue, arrows) at the border of a meningeal B-cell follicle. (I, J) Double immunofluorescence stainings for perforin (red) and Ig (green) show perforin granules polarized toward Ig+ plasma cells (arrows) in active white matter lesions. Nuclei are stained with DAPI (blue) (A, B, C, D, E, F, I, J). Bars : 100 µm in the left inset in D; 50 µm in D; 20 µm in A–C, E, I; 10 µm in G, H and right inset in D; 5 µm in F, J and inset in A.