An evolutionarily conserved mutual interdependence between Aire and microRNAs in promiscuous gene expression

Abstract

The establishment and maintenance of central tolerance depends to a large extent on the ability of medullary thymic epithelial cells to express a variety of tissue-restricted antigens, the so-called promiscuous gene expression (pGE). Autoimmune regulator (Aire) is to date the best characterised transcriptional regulator known to at least partially coordinate pGE. There is accruing evidence that the expression of Aire-dependent and -independent genes is modulated by higher order chromatin configuration, epigenetic modifications and post-transcriptional control. Given the involvement of microRNAs (miRNAs) as potent post-transcriptional modulators of gene expression, we investigated their role in the regulation of pGE in purified mouse and human thymic epithelial cells (TECs). Microarray profiling of TEC subpopulations revealed evolutionarily conserved cell type and differentiation-specific miRNA signatures with a subset of miRNAs being significantly upregulated during terminal medullary thymic epithelial cell differentiation. The differential regulation of this subset of miRNAs was correlated with Aire expression and some of these miRNAs were misexpressed in the Aire knockout thymus. In turn, the specific absence of miRNAs in TECs resulted in a progressive reduction of Aire expression and pGE, affecting both Aire-dependent and -independent genes. In contrast, the absence of miR-29a only affected the Aire-dependent gene pool. These findings reveal a mutual interdependence of miRNA and Aire.

Keywords: Aire; Thymic epithelium; gene expression; microRNA (miRNA).

Figure 1

miRNAs are differentially expressed in thymic cell subsets. (A) Heat maps representing relative expression of selected miRNAs, ordered by their expression level in mTEC^high. The top 25 shows the highest, the bottom 10, the lowest, ratio of expression value between mTEC^high and thymocytes in mouse (left) and human (right). For full lists of miRNA expression in mouse and human samples, see Supporting Information Tables 1 and 2. Data shown are generated from three (mouse) and two (human) independent experiments. (B) Venn diagrams showing overlap in miRNA expression between different thymic epithelial subsets in mouse and human. (C) Comparison of microarray and qPCR results for miRNAs expressed at higher levels in mTEC^high versus mTEC^low in both mouse and human (mouse data representative of three independent experiments are shown).

Figure 2

miRNA expression dynamics in mTEC subsets. (A) Candidate miRNA expression as measured by quantitative RT-PCR in sorted CD80^low, CD80^highAire^neg and CD80^highAire^pos cells. Data are presented as the gene expression relative to the level in CD80^highAire^neg cells set to 100% and shown as mean ± SD of three technical replicates, representative of two independent experiments. (B) miR-129, miR-302b and miR-499 expression was downregulated (left), whereas miR-202 expression was upregulated (right) in Aire-deficient mutants as measured by quantitative RT-PCR in sorted CD80^low and CD80^high TECs of Aire-deficient mice and control littermates. Data are presented as the gene expression relative to the level in WT CD80^high and shown as mean ± SD of three technical replicates, representative of two independent experiments.

Figure 3

Lack of Dicer in the thymic epithelium leads to discordant maturation marker expression and a progressive loss of pGE. (A) Dicer mutants showed a decrease in the percentage of EpCAM-positive stromal cells, with mTECs (CD45^neg EpCAM^pos Ly51^neg) being more affected than cTECs (CD45^neg EpCAM^pos Ly51^pos). Dicer mutant mTECs showed a shift to lower CD80 and higher MHCII expression as compared with control littermates. Upper two rows show CD45^neg cells. Lower row shows CD45^neg EpCAM^pos Ly51^neg cells. Representative profiles of single animals of three samples are shown. (B) Dicer null mutants exhibited a progressive reduction of the levels of promiscuously expressed antigens, as demonstrated by quantitative RT-PCR on sorted TECs. Data are presented as percentage of the gene expression level obtained for the same subset of control littermates and shown as mean ± SD of three technical replicates, representative of two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, Student's t-test.

Figure 4

Reduction in Aire-dependent pGE in miR-29a-deficient mice. (A) miR-29a null mTECs showed normal CD80 maturation profile at 4 weeks of age and a shift to lower CD80 and higher MHCII expression as compared with control littermates at 12 weeks. Flow cytometry plots shown are representative of n > 3 mice analysed in two independent experiments. (B) Expression of Aire and Aire-dependent genes was reduced in miR-29a null mutant mice at 12 weeks (but not at 4 weeks) of age, whereas Aire-independent genes expression is comparable with that of control littermates. Data are presented as percentage of the gene expression level obtained for the same subset of control littermates and are shown as mean ± SD of three technical replicates, representative of two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, Student's t-test.