MicroRNA expression profiles and target-gene predictionTwo-hundred Treg and Tconv cells were purified by fluorescence-activated cell sorting (Supplementary Figure 1) and microRNAs were analyzed with the human Megaplex RT-stem-loop microRNA Pool A v2.1 (Life Technologies, CA-USA). Expression of hsa-miR-125a-5p (here named miR-125a-5p) was also tested by TaqMan microRNA single assay qPCR on preamplified products using the following assays: hsa-miR-125a-5p (ID002198), snRNAU6 (ID001973), snRNAU44 (ID001094), and snRNAU48 (ID1006) (Life Technologies). Expression levels of each microRNA are reported as Cycles to Threshold (Ct) of PCR and Ct are normalized (dCt) using small RNAs endogenous controls (RNU6A, RNU48, RNU44). MicroRNAs were considered differentially expressed when the cutoff fold change was <0.5 or >2.0 and when the cutoff p-value was <0.05 using the 2-tailed “Student t-test” on normally distributed dCt values. Undetermined values were set to a maximum Ct of 40. Each amplification plot for every microRNA was manually checked to avoid false positive Ct and only microRNAs with a Ct <35 and with adequate efficiency amplification plots in all sample replicates were taken into consideration for 2−dCT calculations and subsequent statistical analyses. Targetscan6.2 and PICTAR algorithms were used to identify potential microRNA target genes.