The impact of a long non-coding RNA at the Pax6 locus on beta cell identity and function

Abstract

Paired box 6 (Pax6) is an essential transcription factor in pancreatic development of α- and β-cells along with maintaining the functionality of mature β-cells. The expression of Pax6 is controlled by several regulatory regions within, and distal to the locus. More recently a long non-coding RNA (lncRNA) upstream of Pax6 has been showed to regulate Pax6 expression in α-cells, in an isoform specific manner. Paired box 6 opposite strand 1 (Pax6os1) is another lncRNA expressed in pancreatic β- and δ-cells, that overlaps with Pax6 in the opposite orientation. Therefore, in this study, I sought to investigate the role of Pax6os1 within the context of β-cell function and glucose homeostasis, dependently and independently of Pax6. In vitro analyses showed that the expression of Pax6os1 was highly matched to that of Pax6 but with a greater degree of tissue specificity. The silencing of Pax6os1 suggested it was a functionally repressive antisense lncRNA, as an increase in Pax6 expression was observed within MIN6 β-cells, which had a positive effect of the β-cell genetic signature. The subcellular localisation of the lncRNA transcript, expressed both within the cytoplasm and nucleus, was seen to be highly dynamic, though remained unresponsive to glucose, despite the expression of Pax6os1 being upregulated by increasing glucose in MIN6 cells and CD1 primary mouse islets. Interestingly, Pax6 was also upregulated by glucose, suggesting a co-regulation of the two transcripts. It is possible that this co-regulation may form the basis of a feedback loop in which Pax6 regulates its own expression, with a Pax6 binding motif being identified within the upstream UTR of Pax6os1. Despite the repressive effect in vitro, the negative relationship between Pax6os1 and Pax6 was inconclusive in vivo following Pax6os1 deletion, as Pax6 mRNA remained unaltered, while a tendency of PAX6 protein lowering was seen in a gender specific manner. Minimal transcriptional changes existed between control and Pax6os1 null mice, with only genes Tnc and Slc28a2 being differentially expressed. However, further small changes in the key β-cell genes Glut2 (Slc2a2), ZnT8 (Slc30a8), G6pc2 and Pcsk1 hint towards an improvement in β-cell identity. Notably, the negligible differential gene expression occurred alongside a lack of Pax6 upregulation, despite a greater reduction in Pax6os1 expression in vivo, which may suggest that Pax6 mediates the transcriptional effects of Pax6os1. It is hypothesised that the discrepancies in the in vitro and in vivo expression changes may be due to the Pax6os1 deletion, disrupting a cis-regulatory region required for Pax6 regulation. However, Pax6os1 likely exerts effects independent of Pax6, as phenotypic traits existed in Pax6os1 null mice. All changes were modest, while showing gender, age, and dietary intervention dependent effects. The most prominent change was a reduction in body mass of Pax6os1 null females, with an improved glucose tolerance under diabetic challenge at 12 weeks of age. Glucose tolerance in males was preserved irrespective of age and diet, despite a reduction in β-cell mass and insulin sensitivity under HFD, due to an increased insulin output. The mechanistic action of Pax6os1 remains undetermined, while protein binding partner identification point towards a chromatin associated role, potentially through the tethering of intermediate molecules and chromatin modifying enzymes. Taken together, the data demonstrates a novel role for Pax6os1 in the control of body mass and glucose homeostasis. Further studies investigating the specific mechanistic function of Pax6os1 and its effect on local chromatin architecture would be a valuable addition to the research and aid its translatability to diabetes pathogenesis.