Supplementary Figure 5: Analysis of NKG2C+ NK cell proliferation.
From: Peptide-specific recognition of human cytomegalovirus strains controls adaptive natural killer cells

(a) Purified CD56dim NK cells from HCMV– donors were cultured with peptide-pulsed RMA-S–HLA-E–LFA-3 cells in the presence of IL-15 combined with treatment with IL-12 plus IL-18 during the initial 20 h of culture. Precursor frequency of NKG2C+ NK cells over time. Symbols indicate individual donors (n=6 in 2 independent experiments) and lines median. Two-way repeated-measures ANOVA with Bonferroni correction. (b-f) Mathematical analysis of NKG2C+ NK cell proliferation dynamics. (b-c) Symbols and error bars indicate mean±SEM of experimentally obtained precursor frequencies of NKG2C+ NK cells (b) with (data from Supplementary Fig. 5a) or (c) without (data from Supplementary Fig. 4e) treatment with IL-12 plus IL-18 during the initial 20 h of culture. Lines indicate best-fit gamma distributions, which are used as input for Fig. 5f and Supplementary Fig. 5d. (d) Modified Gett–Hodgkin model describing NKG2C+ NK cell proliferation and accumulation dynamics in the absence of treatment with IL-12 plus IL-18. Symbols and error bars indicate mean±SEM of experimentally obtained NKG2C+ NK cell numbers after normalization to day 1 values (set as 1); lines indicate best-fit curves of the model. Precursor frequencies were experimentally obtained (Supplementary Fig. 4e, Supplementary Fig. 5c), while division times and death rates (both mean±SEM) were inferred as best-fit parameters by non-linear optimization. (e-f) Modified Gett–Hodgkin models with fixed input parameters in the (e) presence or (f) absence of treatment with IL-12 plus IL-18. Symbols and error bars indicate mean±SEM of experimentally obtained NKG2C+ NK cell numbers after normalization to day 1 values (set as 1); lines indicate curves of the model. Precursor frequencies were experimentally obtained; division time and death rate values were inferred by non-linear optimization for the VMAPQSLLL peptide (as in Fig. 5f and Supplementary Fig. 5d, respectively) and set as fixed parameters for both peptides. NS not significant, ***P < 0.005, ****P < 0.0001.