Longitudinal study of T and natural killer cell responses in Epstein-barr virus-associated lymphoproliferative diseases in children

Abstract

T and natural killer (NK) cells are recognized as the major immune effector cells in the long-term control of Epstein-Barr virus (EBV) which persists through the lifetime of the human host. We hypothesize that distinct patterns of deficiencies of T and NK cell responses are associated with EBV-associated lymphoproliferative diseases (EBV-LPDs) in immunocompetent and immunocompromised hosts. The following cohorts of children with EBV-LPDs, namely, infectious mononucleosis (IM; n=31), EBV-associated hemophagocytic lymphohistiocytosis (EBV-HLH; n=9), systemic EBV-positive T-cell lymphoma of childhood (SEBVTL; n=8), chronic active EBV infection (CAEBV; n=4) and post-transplant lymphoproliferative disorder (PTLD; n=15), were followed longitudinally. Plasma and PBMC viral loads, EBV-infected cell types and the function and development of distinct NK cell subsets and polyfunctional T cells (PFCs) were determined by quantitative PCR, flow cytometric RNA and specific flow cytometric assays, respectively.

The IM patients had EBV-infected CD19+ B cells, elevated PBMC and plasma viral loads which declined over time. The potent degranulating CD56dimNKG2A+KIR- NK cell subset was detected at diagnosis and persisted into 12 months post-diagnosis. Both lytic and latent antigen-specific CD4+ and CD8+ PFCs of diverse function emerged at 6-12 months. In contrast, EBV-HLH patients harbored EBV-infected CD8+ (n=4) and CD4+ T (n=1) cells as well as CD19+ B (n=3) cells. The SEBVTL patients had predominantly EBV-infected CD8+ T cells whereas CAEBV patients had variable EBV-infected cell types (CD4+, CD8+ or gd T cell or NK cell). The proliferative capacity of NK cells was retained in EBV-HLH whereas that of NK cells was transiently impaired in SEBVTL and severely impaired in CAEBV. A low frequency of BZLF1-specific CD8+ PFCs was found in EBV-HLH whilst the time of emergence of latent antigen-specific PFCs was comparable to that of IM. SEBVTL patients could not generate lytic and latent antigen-specific CD8+ PFCs at diagnosis. However, PFC responses to EBV could be restored in both EBV-HLH and SEBVTL patients with gradual decrease in viral loads over time. In contrast, CAEBV patients showed markedly defective NK cell proliferation and inability to generate EBNA1 and LMP2-specific CD4+ and CD8+ PFCs, correlating with the persistently elevated viral loads. PTLD patients had diminished frequency of CD56dimNKG2A+KIR- NK cells from the time of diagnosis through remission and almost undetectable level of CD8+ PFCs at 24 months. PTLD patients co-infected with CMV and EBV had significantly higher levels of a CMV-associated CD56dimNKG2ChiCD57+NKG2A-KIR+ than the NKG2A+KIR- NK cell subset.

In conclusion, IM have functional CD56dimNKG2A+KIR- NK cell responses from the time of diagnosis and can develop diverse lytic and latent antigen-specific CD4+ and CD8+ PFC responses at 6-12 months. EBV-HLH can retained NK cell function but show defective lytic CD8+ PFC responses whilst SEBVTL have multiple defects in NK cell as well as lytic and latent CD8+ PFC responses. Nevertheless, both EBV-HLH and SEBVTL can restore long-term T and NK cell responses towards EBV. In contrast, CAEBV have persistently defective T and NK cell responses leading to a chronic immunodeficient state towards EBV and PTLD have diminished frequency of CD56dimNKG2A+KIR- NK cells and delayed development of EBV-specific PFCs.