Tonsillar homing of Epstein-Barr virus-specific CD8+ T cells and the virus-host balance

Abstract

Patients with infectious mononucleosis (IM) undergoing primary EBV infection show large expansions of EBV-specific CD8+ T cells in the blood. While latent infection of the B cell pool is quickly controlled, virus shedding from lytically infected cells in the oropharynx remains high for several months. We therefore studied how responses localize to the tonsil, a major target site for EBV, during primary infection and persistence. In acute IM, EBV-specific effectors were poorly represented among CD8+ T cells in tonsil compared with blood, coincident with absence of the CCR7 lymphoid homing marker on these highly activated cells. In patients who had recently recovered from IM, latent epitope reactivities were quicker than lytic reactivities both to acquire CCR7 and to accumulate in the tonsil, with some of these cells now expressing the CD103 integrin, which mediates retention at mucosal sites. By contrast, in long-term virus carriers in whom both lytic and latent infections had been controlled, there was 2- to 5-fold enrichment of lytic epitope reactivities and 10- to 20-fold enrichment of latent epitope reactivities in tonsil compared with blood; up to 20% of tonsillar CD8+ T cells were EBV specific, and many now expressed CD103. We suggest that efficient control of EBV infection requires appropriate CD8+ T cell homing to oropharyngeal sites.

Figure 1

EBV genome load quantitation in IM patients, post-IM patients, and long-term carriers. (A) EBV genome copies in 1 ml of concentrated throat washing (upper panel) and 10⁶ PBMCs (lower panel) from an IM patient (IM-171) from whom samples were taken during acute disease (days 0–14) and at intervals up to 4 months later. Upper limits of genome levels in throat washing and blood samples from healthy carriers are indicated by the dashed lines. (B) EBV genome copies were estimated in 10⁶ tonsillar cells (upper panels) or 10⁶ PBMCs (lower panels) from tonsillectomy patients who were either undergoing acute IM or recently recovered from IM (post-IM) or were long-term carriers.

Figure 2

Frequency of EBV-specific CD8⁺ T cells in PBMCs and matched tonsil preparations from IM, post-IM, and long-term carrier tonsillectomy patients. (A) PBMCs (left panels) and tonsillar cells (right panels) from acute IM patient IM-5 were stained with either HLA-B*0801 tetramers containing the lytic cycle epitope RAK or latent cycle epitope QAK peptides or the HLA-A*0201 tetramer containing the lytic cycle epitope YVL peptide and subsequently stained with anti-CD8 mAbs. Values shown refer to the percentage of CD8⁺ T cells that stained with the tetramer. (B) PBMCs and tonsillar cells from post-IM-1 were stained as described above with HLA-A*0201 tetramers containing either the lytic cycle epitope YVL or GLC or the latent cycle epitope CLG peptides. (C) PBMCs and tonsillar cells from carrier-32 were stained as described above with the HLA-B*0801 tetramers containing the lytic cycle epitope RAK or latent cycle epitope FLR peptides or the HLA-A*0201 tetramer containing the lytic cycle epitope GLC peptide.

Figure 3

Summary of the frequency of EBV-specific responses in the PBMCs and tonsils of 6 IM, 2 post-IM, and 11 long-term carrier tonsillectomy patients; symbols connected by a line refer to a particular epitope-specific response in an individual donor. (A) Percentage of CD8⁺ T cells staining with EBV lytic epitope–specific tetramers in IM (left panel) and post-IM (middle panel) patients and long-term carriers (right panel). (B) Percentage of CD8⁺ T cells staining with EBV latent epitope–specific tetramers in IM (left panel) and post-IM (middle panel) patients and long-term carriers (right panel). P values shown were obtained using linear mixed model analysis as described in Methods. Significant differences are indicated by an asterisk. In addition, in acute IM, the reduction in epitope-specific T cell representation in tonsil relative to blood was significantly greater for lytic than for latent reactivities (P = 0.004); also, in carriers, the increase in epitope-specific T cell representation in tonsil relative to blood was significantly greater for latent than for lytic epitopes (P = 0.003).

Figure 4

Representation of EBV- and CMV-specific CD8⁺ T cells in PBMCs and tonsil of a long-term EBV/CMV carrier. PBMCs (left panels) and tonsillar cells (right panels) were stained as before. (A) Cells were stained with HLA-B*0801 tetramers containing the EBV lytic epitope RAK and latent epitope FLR peptides. (B) Cells were stained with HLA-A*0101 tetramers containing the CMV peptides VTE and YSE. Results are presented as in Figure 2.

Figure 5

Analysis of EBV-specific CD8⁺ T cells from an acute IM patient for CD38, CD45RO, CCR7, and CD103 status. PBMCs (left panels) or tonsillar cells (right panels) from IM-5 were stained with the HLA-B*0801 tetramer containing the lytic epitope RAK peptide, followed by mAbs specific for CD8 and the relevant marker. Flow cytometric analysis was performed after gating on CD8⁺ cells, and profiles show tetramer staining versus the third marker (CD38, CD45RO, CCR7, and CD103). Values shown refer to the percentage of tetramer-positive cells that express the relevant marker.

Figure 6

Analysis of EBV-specific CD8⁺ T cells from a patient recently recovered from acute IM for CD38, CD45RO, CCR7, and CD103 status. PBMCs (left panels) or tonsillar cells (right panels) from post-IM-1 were stained with HLA-A*0201 tetramers containing either the lytic cycle epitope YVL peptide (A) or the latent cycle epitope CLG peptide (B), as well as with antibodies specific for CD38, CD45RO, CCR7, and CD103. Flow cytometric analysis was performed after gating on CD8⁺ cells, and profiles are presented as in Figure 5.

Figure 7

Analysis of EBV-specific CD8⁺ T cells from a long-term carrier for CD38, CD45RO, CCR7, and CD103 status. PBMCs (left panels) or tonsillar cells (right panels) from carrier-32 were stained with HLA-B*0801 tetramers containing either the lytic cycle epitope RAK peptide (A) or the latent cycle epitope FLR (B) as well as with antibodies specific for CD38, CD45RO, CCR7, and CD103. Flow cytometric analysis was performed after gating on CD8⁺ cells, and profiles are presented as in Figure 5.

Figure 8

Compilation of results of phenotypic analysis of EBV-specific CD8⁺ T cells in tonsil and PBMCs from acute IM, post-IM, and long-term carrier tonsillectomy patients. (A) Percentage of tonsillar EBV-specific CD8⁺ cells expressing the relevant marker. (B) Percentage of PBMC EBV-specific CD8⁺ cells expressing the relevant marker. Left panels show results from acute IM patients; middle panels show results from post-IM patients; and right panels show results from long-term carriers. Filled symbols represent results from EBV lytic cycle epitopes, while open symbols represent results from EBV latent cycle epitopes. Statistical analysis of the data was carried out using a model similar to that described in the legend to Figure 3. CD38 expression was significantly related to infection state both in tonsil (P = 0.0003) and in blood (P < 0.0001); furthermore, CD38 was significantly higher in tonsil than in blood of post-IM patients (P = 0.04). CD103 expression was significantly related to infection state in tonsil only (P = 0.05) and was elevated in tonsil compared with blood in both post-IM patients and long-term carriers (P = 0.008). Significant differences in CCR7 expression were only seen in long-term carriers in whom CCR7 was significantly higher on latent compared with lytic epitope–specific cells in blood (P = 0.003).