Co-regulation of intragenic microRNA miR-153 and its host gene Ia-2 β: identification of miR-153 target genes with functions related to IA-2β in pancreas and brain

Abstract

Aims/hypothesis: We analysed the genomic organisation of miR-153, a microRNA embedded in genes that encode two of the major type 1 diabetes autoantigens, islet-associated protein (IA)-2 and IA-2β. We also identified miR-153 target genes that correlated with IA-2β localisation and function.

Methods: A bioinformatics approach was used to identify miR-153's genomic organisation. To analyse the co-regulation of miR-153 and IA-2β, quantitative PCR analysis of miR-153 and Ia-2β (also known as Ptprn2) was performed after a glucose stimulation assay in MIN6B cells and isolated murine pancreatic islets, and also in wild-type Ia-2 (also known as Ptprn), Ia-2β single knockout and Ia-2/Ia-2β double knockout mouse brain and pancreatic islets. Bioinformatics identification of miR-153 target genes and validation via luciferase reporter assays, western blotting and quantitative PCR were also carried out.

Results: Two copies of miR-153, miR-153-1 and miR-153-2, are localised in intron 19 of Ia-2 and Ia-2β, respectively. In rodents, only miR-153-2 is conserved. We demonstrated that expression of miR-153-2 and Ia-2β in rodents is partially co-regulated as demonstrated by a strong reduction of miR-153 expression levels in Ia-2β knockout and Ia-2/Ia-2β double knockout mice. miR-153 levels were unaffected in Ia-2 knockout mice. In addition, glucose stimulation, which increases Ia-2 and Ia-2β expression, also significantly increased expression of miR-153. Several predicted targets of miR-153 were reduced after glucose stimulation in vitro, correlating with the increase in miR-153 levels.

Conclusions/interpretation: This study suggests the involvement of miR-153, IA-2β and miR-153 target genes in a regulatory network, which is potentially relevant to insulin and neurotransmitter release.

Fig. 1

Genomic organisation of miR-153 transcription unit. (a) Schematic representation of the IA-2 and IA-2β genomic loci (not to scale). The locations of miR-153-1 and miR-153-2 are indicated (between exon 19 and exon 20). The IA-2 gene in humans and mice extends over 20 kb and consists of 23 exons. The genomic structure of IA-2β in humans and mice consists of 23 exons and spans approximately 1,000 and 800 kb, respectively. (b) Evolutionary conservation of miR-153-1 and miR-153-2 sequences as determined by ClustalW alignment. Seed sequences are underlined. Black or grey shading indicates conserved or non-conserved residues, respectively. X. tropicalis, Xenopus tropicalis

Fig. 2

Co-regulation of miR153 and Ia-2β expression. (a) Quantitative PCR analysis of miR-153 mRNA expression in wild-type (WT), AKO, BKO and DKO mouse brain. Relative expression (fold) was calculated using the relative cycle threshold quantification method () and normalised to Sno135; WT n = 6, AKO n = 4, BKO n = 4, DKO n = 6. (b) Quantitative PCR analysis of pri-miR-153 mRNA expression in mice as above (a). Relative expression (fold) was calculated as above (a) and normalised to Tbp; WT n = 6, AKO n = 4, BKO n = 4, DKO n = 6. (c) Quantitative PCR analysis of pri-miR-153 mRNA expression in wild-type and DKO mouse pancreatic islets. Relative expression (fold) was calculated as above (a) and normalised to Tbp; WT n = 6, DKO n = 6. (d) Quantitative PCR analysis of Ia-2, Ia-2β and miR-153 transcript levels in cultures of MIN6B cells at 5 mmol/l (black bars) and 25 mmol/l (white bars) glucose stimulation. Relative expression (fold) was calculated as above (a) and normalised to β-actin. n = 3. (e) Quantitative PCR analysis of Ia-2, Ia-2β and miR-153 transcript levels in pancreatic islets from adult wild-type (C57BL/6) mice upon stimulation with 3.3 mmol/l (black bars) and 16.7 mmol/l (white bars) glucose. Relative expression (fold) was calculated as above (a) and normalised to β-actin; n = 6. Experiments (a–e) were performed in triplicate and significance determined by Mann–Whitney U test, *p < 0.05 and **p < 0.01

Fig. 3

miR-153 target gene prediction analysis. (a) Predicted protein×protein interactions using IPA 9.0. Various points of interaction between the candidate miR-153 targets are shown. The arrow from protein tyrosine phosphatase, receptor type, N polypeptide 2 (PTPRN2, also known as IA-2β) to miR-153 indicates that PTPRN2 affects the abundance of miR-153. Red lines with a short line at the end indicate an inhibitory effect of miRNA target gene interaction. Lines between two proteins indicate that only binding was detected. Blue arrows indicate the direction of protein × protein interaction. BSN, bassoon; PCLO, piccolo; SYT, synaptotagmin. (b) Expression of predicted targets as indicated was determined by quantitative PCR on adult wild-type mouse total RNA from pancreas (black bars), heart (grey bars) and brain (white bars). (c) Expression of pri-miR-153 was determined by quantitative PCR in adult wild-type (C57BL/6) mouse total RNA from pancreas (black bar), heart (grey bar) and brain (white bar). (b,c) Relative expression (fold) of miR-153 was calculated using the relative quantification method, taking mouse pancreas levels as calibrator, i.e. onefold. All data were normalised to Tbp in three independent experiments (n = 4)

Fig. 4

miR-153 regulation of SNCA expression in vitro and expression of miR-153 targets after glucose stimulation and in Ia-2β knockout mice. (a) Gene expression analysis of predicted miR-153 target genes as determined by quantitative PCR in pancreatic islets of wild-type (black bars) and Ia-2/Ia-2β knockout (white bars) mice or (b) in brains of wild-type (black bars) and Ia-2β knockout (white bars) mice. The relative expression (as percentage of wild-type) of miR-153 targets was calculated using the relative quantification method (taking the level of wild-type for each target gene as calibrator, i.e. set at 100%) and normalised to Tbp. (c) Gene expression analysis as above (a,b) in MIN6B cells in response to 5 mmol/l (black bars) or 25 mmol/l (white bars) glucose stimulation. The relative expression of miR-153 targets (as percentage of 5 mmol/l glucose treatment) was calculated using the relative quantification method (with 5 mmol/l glucose for each target gene as calibrator, i.e. set at one) and normalised to Tbp. (d) Gene expression analysis as above (a,b) in MIN6B cells or (e) SH-SY5Y cells after induction of miR-153 precursor or SCR overexpression. The relative expression of miR-153 targets (as percentage of SCR treatment) was calculated using the relative quantification method and normalised to Tbp. (a–e) Error bars represent SEM; n > 3 independent duplicate experiments. Significance was determined using the Mann–Whitney U test, *p < 0.05 and **p < 0.01. (f) Western blot analysis of endogenous bassoon, SNAP25, alpha-synuclein (SNCA), parkin and β-actin in SH-SY5Y cells treated with 50 nmol/l (final concentration) of miR-153 or SCR

Fig. 5

Luciferase reporter analysis of predicted miR-153 targets. Luciferase and Renilla luciferase constructs were transfected into SH-SY5Y cells with the miR-153 (black bars) or SCR (grey bars) oligonucleotides at a final concentration of 50 nmol/l. Wild-type or seed mutant SNCA, BSN and PCLO 3′ UTR construct signals (see ESM Fig. 4, luciferase activity normalised to Renilla) are shown as a percentage of the signal in control cells transfected with SCR. Error bars represent SEM; n > 3 independent triplicate experiments. Significance was determined using Mann–Whitney U test, *p < 0.05