miRNA profiling of naïve, effector and memory CD8 T cells

Abstract

microRNAs have recently emerged as master regulators of gene expression during development and cell differentiation. Although profound changes in gene expression also occur during antigen-induced T cell differentiation, the role of miRNAs in the process is not known. We compared the miRNA expression profiles between antigen-specific naïve, effector and memory CD8+ T cells using 3 different methods--small RNA cloning, miRNA microarray analysis and real-time PCR. Although many miRNAs were expressed in all the T cell subsets, the frequency of 7 miRNAs (miR-16, miR-21, miR-142-3p, miR-142-5p, miR-150, miR-15b and let-7f) alone accounted for approximately 60% of all miRNAs, and their expression was several fold higher than the other expressed miRNAs. Global downregulation of miRNAs (including 6/7 dominantly expressed miRNAs) was observed in effector T cells compared to naïve cells and the miRNA expression levels tended to come back up in memory T cells. However, a few miRNAs, notably miR-21 were higher in effector and memory T cells compared to naïve T cells. These results suggest that concomitant with profound changes in gene expression, miRNA profile also changes dynamically during T cell differentiation. Sequence analysis of the cloned mature miRNAs revealed an extensive degree of end polymorphism. While 3'end polymorphisms dominated, heterogeneity at both ends, resembling drosha/dicer processing shift was also seen in miR-142, suggesting a possible novel mechanism to generate new miRNA and/or to diversify miRNA target selection. Overall, our results suggest that dynamic changes in the expression of miRNAs may be important for the regulation of gene expression during antigen-induced T cell differentiation. Our study also suggests possible novel mechanisms for miRNA biogenesis and function.

Figure 1. Distribution of the seven highly expressed miRNAs in the T cell subsets.

The cloning frequency of each miRNA is shown as a percentage of total miRNA clones isolated for each T cell subset.

Figure 2. Differential expression of miRNAs in naïve, effector and memory T cells.

Signal intensity for a particular miRNA in each cell type was divided by that of the pooled RNA reference sample to generate the heat map. Same amount of total RNA is used for each slide (so the signal is actually normalized according to total RNA). The miRNA clustering tree is shown on the left and the sample clustering tree is shown on the top. The color scale shown at the bottom illustrates the relative expression level of the indicated miRNA across all samples: red denotes an expression above the mean, blue denotes an expression lower than the mean and grey represents signal lower than the background.

Figure 3. Comparison of miRNA expression profile in the T cell subsets by 3 different methods.

A) Expression profile by direct cloning. The cloning frequency was normalized to the total hybridization signal in the miRNA microarray. B) Expression profile by microarray hybridization. Ratios of sample/reference hybridization signal is shown. C) Expression profile by real time PCR. Expression level was normalized to that of small non-coding RNA U6B. Mean of triplicate experiments±SD is shown. Similar results were also obtained in 2 additional independent experiments.

Figure 4. Correlation amongst three methods of miRNA expression profiling.

Twenty different miRNAs (indicated on the left) that were picked up by direct cloning and/or microarray hybridization were also tested by real time PCR. A shows miRNAs whose relative expression between naïve, effector and memory T cells correlated in all 3 methods. B shows miRNAs with expression profile correlated only in the microarray and real time RT-PCR. C shows the expression profile that correlated only in cloning and RT-PCR. D shows the expression profile that correlated only in cloning and microarray and E represents the miRNAs that did not correlate in any two methods. The correlation was determined arbitrarily according to their movement in the same direction in the different assays.

Figure 5. Novel miRNA–miR-669d in mouse T cells.

The presence of the stem loop sequences for miRNA-669d in the mouse genome is depicted in (A-B). C) Small RNAs from mouse T cells were tested for the presence of miR-669d by real time RT-PCR. Expression level was normalized to that of small non-coding RNA U6B. Mean of triplicate experiments±SD is shown.

Figure 6. miRNA 3′ end heterogeneity affects detection efficiency of stem-loop primer-based real time PCR.

Detection efficiency was compared using exact miR-16 oligonucleotide and miR-16 oligonucleotide with 1 nt deletion or 1 nt addition at the 3′ end as templates. Mean of triplicate experiments±SD is shown.