Fig 1.
CD25+Foxp3+ Treg cells are associated with poor outcome.
Identification of CD25+Foxp3+ Treg cells by flow cytometry (A). At the time of diagnosis, frequencies and (B) absolute cell counts (C) of CD25+Foxp3+ cells are higher in eBL patients who died than in survivors and healthy controls.
Fig 2.
Higher numbers of both CD45RA+Foxp3lo naïve Treg (nTreg cells) and CD45RA-Foxp3hi effector Treg (eTreg) cells are associated with poor outcome.
(A) Gating strategy differentiating nTreg and eTreg cells from CD45RA-Foxp3lo non-Treg cells. At the time of diagnosis, eBL non-survivors have higher absolute cell counts of both nTreg (B) and eTreg (C) cell subtypes compared to eBL survivors and healthy controls.
Fig 3.
Elevated Treg frequencies in patients prior to relapse.
Frequencies of CD25+Foxp3+ Treg cells (A) and EBV viral load (B) at monthly follow-up time points in 3 event-free survivors and 3 patients who relapsed. Proposed Treg frequency threshold indicative of future relapse was established by measuring Treg frequencies in 8 other event-free survivors 4–9 months post discharge, and frequencies never rose above 0.5% (dotted line). Designated as a red ♦ = BL001 (relapse), designated as a red ● = BL002 (relapse), designated as a red ■ = BL003 (relapse), designated as a blue ▲ = BL004 (event-free survivor), designated as a blue ▼ = BL005 (event-free survivor) designated as a blue ■ = BL006 (event-free survivor).
Fig 4.
T cells in eBL patients have higher PD-1 expression than in healthy controls.
Gating strategy for the identification of PD-1+ cells (A). Frequency of CD4+ and CD8+PD-1+ cells at the time of diagnosis in eBL non-survivors, eBL survivors and healthy controls (B and C, respectively). In general PD1 expression on CD4 and CD8 T cells was higher in eBL non-survivors compared to age-matched malaria exposed yet health controls. eBL survivors had intermediate levels of PD1 expressed on their CD4 T cells yet levels as high as non-survivors for their CD8 T cells. Patients who relapsed tended to have peaks of elevated CD8+PD-1+ frequencies over time (D). Designated with a red ♦ = BL001 (relapse), designated with a red ● = BL002 (relapse), designated with a red ■ = BL003 (relapse), designated with a blue ▲ = BL004 (event-free survivor), designated with a blue ▼ = BL005 (event-free survivor) designated with a blue ■ = BL006 (event-free survivor). FMO; Fluorescence Minus One.
Fig 5.
Non-survivors produce less EBV-specific IFN-γ, and from difference T cell subsets.
Among IFN-γ producers, there was no difference in the magnitude of CD4+ T cell IFN-γ production in response to EBNA-1 or PfSEA-1 stimulation between eBL survivor and non survivors (A), non-survivors produced less CD8+ T cell IFN-γ from EBNA-1 stimulated cells (B). Among EBNA-1-specific CD8+ T cell IFN-γ producers, survivors produced more CD8+ T cell IFN-γ from the CD45RA-CCR7- effector memory (TEM) cell subset than non-survivors (C). Among PfSEA-1-specific CD8+ T cell IFN-γ producers, there was no difference between survivors and non-survivors in the cell type that produced IFN-γ (D). Among IFN-γ T cell responders in both survivors and non-survivors, CD25+Foxp3+ Treg frequency negatively correlated with CD8+ T cell EBNA-1 IFN-γ production (E).
Fig 6.
Non-survivors produce more IL-10 than survivors.
A) Comparing only patients who produced IL-10 above background upon stimulation, the magnitude of CD4+ T cell IL-10 response was greater in eBL non-survivors when stimulated with either EBNA-1 peptide pool or PfSEA-1-1 peptides. B) Among non-survivors, some IL-10 was from different Treg sources, including CD45RA+Foxp3lo naïve Treg (nTreg) and CD45RA-Foxp3hi effector Treg (eTreg), but was also produced from Foxp3- CD4+ T cells. Pie charts represent numbers of patients who produced IL-10 from one of the three sources of CD4+ T cell comparing EBNA1 and PfSEA1.