If it becomes evident that a molecule is important for immune regulation, its mode of action immediately becomes a focus of interest. Sometimes, however, the issue of how it is produced to begin with may be overlooked. Aire, a gene responsible for a hereditary type of autoimmune disease, known as autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), may represent one such example, because of its well-known role in the establishment of self-tolerance. Since the cloning of Aire in 1997, researchers have found that it functions in two essential events within the thymus, through the representation of a wide array of tissue-restricted Ags (TRAs) from medullary thymic epithelial cells (mTECs) [1]: clonal deletion [2,3] and the production of functional Treg cells [3][4][5][6]. Furthermore, the means by which the nonorthodox transcription regulator Aire produces so many TRAs have been thoroughly investigated at the molecular Correspondence: Prof. Mitsuru Matsumoto e-mail: mitsuru@tokushima-u.ac.jp level [7, 8]. In contrast to these efforts to clarify the cellular immunology and molecular biology of Aire, our understanding of the mechanisms controlling Aire expression itself is still far from complete. In particular, we have only limited information on the promoter/enhancer element(s) for Aire within the mammalian genome.In this issue, Haljasorg et al. [9] investigate the molecular regulation of Aire expression by focusing on the region ß3 kb upstream of Aire, which was highlighted by in silico analysis: the corresponding sequence was highly conserved among many mammalian species, containing two putative NF-κB binding sites spaced less than 10 bp apart. Using a reporter gene assay together with an electrophoretic mobility shift assay (EMSA), the authors confirm that these NF-κB-binding motifs indeed bind to classical (RelA and p50) and nonclassical (RelB and p52) NF-κB components. The authors named this region "conserved noncoding sequences" (CNS1), with possible enhancer activity for Aire. In order to prove their hypothesis, they adopted the tough approach of generating mice lacking CNS1 (CNS1-KO) through a targeted deletion. Sure enough, CNS1-KO mice showed no Aire expression in the thymus or secondary organs [9]. Furthermore, CNS1-KO mice showed many characteristic features of Aire-deficient mice, including reduced expression of many so-called Aire-dependent TRAs, such as salivary protein 1 and insulin 2, and altered development of mTECs (i.e. changes in the composition of immature and mature mTECs), although no obvious autoimmune phenotypes were observed on a C57BL/6 background [9].Given that it was demonstrated that mice deficient in various TNF receptor superfamily members activating the NF-κB signal pathway have defective Aire expression, Haljasorg et al.[9] next examined the possible cytokine signals responsible for NF-κB-mediated Aire expression. The authors found that Aire was induced from wild-type (WT) mouse thymi but not from CNS1-KO thymi upon stimulation with RANKL (receptor activator o...