Fig. 3.

Effect of single-dose IL-2 treatment on the expression of IL-6R on the CD4⁺ T cell compartment in vivo. (A,B) Data shown depict the variation (Mean +/− SEM) of the frequency of IL-6R^hi cells on: (i) CD127⁺CD25⁻ CD45RA⁻ Teffs (A); and (ii) IL-6R^hi CD127⁺CD25⁻ CD45RA⁺ naïve T cells (B) compared to the pre-treatment baseline, following IL-2 treatment in 22 T1D patients enrolled in the “Adaptive study of IL-2 dose on regulatory T cells in type 1 diabetes” (DILT1D) [2], [3]. Median pre-treatment baseline frequencies of IL-6R^hi cells were 39.3% (range: 21.5–59.9%) and 5.39% (range: 1.49–30.4%) within CD4⁺ CD45RA⁻ Teffs and CD4⁺ CD45RA⁺ naïve T cells, respectively. (C) Data shown depict the variation (Mean +/− SEM) of the IL-6R mean fluorescence intensity (MFI) levels on the surface of total CD127^lowCD25⁺ Tregs compared to the pre-treatment baseline MFI levels (median = 600; range: 392–836). Patients were stratified based on whether they received: (i) the lower IL-2 doses of 0.04–0.045 x 10⁶ U/ml (N = 8; depicted in black); or (ii) the higher IL-2 doses of 0.16–0.737×10⁶ U/ml (N = 14; depicted in red). The MFI cuttoff to define IL-6R^hi cells was the same as the one used for the Treg subset (detailed in Fig. 1). The maximum increases over the baseline pre-treatment frequencies achieved during the course of the study are indicated for each IL-2 dosing group. P values for the maximum increase in the frequency of the assessed parameter in response to a single dose of IL-2 was calculated using a two-tailed paired non-parametric Wilcoxon signed rank test comparing the frequencies observed at the timepoint where the maximal increase was achieved with the respective baseline pre-treatment frequencies. P values for the IL-2 dose-dependent effects were calculated using a two-tailed non-parametric a two-tailed non-parametric Mann-Whitney test comparing the frequency of IL-6R^hi cells between the low and high dose groups at each timepoint. The DILT1D data from individuals prior to normalization as a group are available, however they cannot be anonymized sufficiently to be able to put into the public domain without risk of participant identification. Data are available on request, through the Cambridge University institutional repository (DOI link: https://doi.org/10.17863/CAM.832). *P < 0.05; **P < 0.01; ns = not significant.