Novel targets of miR-30, a microRNA required for biliary development

Background

MicroRNAs are short non-coding RNAs that regulate cell function, differentiation, organ development, and disease states (reviewed in^1–4). MiRNAs regulate gene expression via the RNA-induced silencing complex (RISC), leading to translational repression and/or transcript degradation⁵. The identification of gene targets for individual miRNAs is key to understanding their function, thus necessitating accurate methods to identify significant mRNA targets. The prevailing methods of miRNA target identification begin with computational algorithms based on free energy change calculations of complementary binding, other sequence features, and evolutionary conservation^6,7. However, false positive rates are high, as interactions may be predicted in physiological settings in which targeting does not naturally occur⁸. The data gathered from these computational methods are typically validated by luciferase reporter assays in cultured cells, in which the activity of the relevant miRNA is altered (i.e., by miRNA over-expression or repression).

MiR-30a, expressed as part of an intronic cluster with miR-30c-2, is required for hepatobiliary development in zebrafish⁹. In our previous study⁹, we found that within the fetal mouse liver and infant human liver, members of the miR-30 family are predominantly expressed within the biliary primordium, post-natal bile ducts, and hepatocytes. Using a zebrafish model of miR-30a deficiency, we confirmed that miR-30a plays a critical role in vertebrate bile duct development⁹. The next step in understanding the function of the miR-30 family in biliary development is to identify the genes that it targets.

Using gene expression profiling in cultured hepatoblasts⁹, we previously identified a set of mRNAs whose levels are increased following antisense oligonucleotide (ASO)-mediated inhibition of miR30a⁹. From this dataset, we have selected 7 novel candidate targets (Ccne2, Celsr3, Mdm2, Mtdh, Smad1, Twf, and Timp3) on the basis of biological relevance and two bioinformatic prediction methods: Targetscan and PITA^6,7. In addition, we evaluated four candidate targets previously tested via ASO-mediated miR-30a inhibition in cultured cells (Ak1, Tnrc6a, Egfr, and Inhba)⁹ (Table 1). All 11 candidates were tested using luciferase reporter assays coupled with over-expression of miR-30a. Here we show that several of these candidate target genes - selected on the basis of their known roles in growth and morphogenesis - are direct miR-30a targets.

Methods

Results

Over-expression of miR-30a and miR-30c2 in cell culture

To determine the regulatory effect of the miR-30a and miR-30c cluster (“mir-30ac2”) on the candidate target genes, we induced the over-expression of mir-30ac2 in cultured cells. We transfected a human embryonic kidney cell line transformed with the SV40 large T antigen (HEK293FT) with a doxycycline-inducible green fluorescent protein (GFP) and miR-30ac2 expression plasmid. We observed GFP-positive cells 24 hours post-induction. Mir-30a transcript levels were quantified using qRT-PCR, and revealed a 28-fold increase of miR-30a in cells transfected with the miR-30ac2 expression plasmid versus a control encoding a scrambled sequence (Figure 1). The unrelated control miRNA let-7c was also assayed and expression levels were not significantly altered between the two treatment groups, confirming that miR-30ac2 over-expression did not have an effect on miRNA production in general.

Figure 1. Over-expression of miR-30ac2 in HEK293FT cells.

A plasmid encoding miR-30a cluster miRNAs (miR-30a and miR-30c2) or a scrambled miRNA sequence was transfected into HEK293T cells. The resulting levels of miR-30a and let-7a were measured by qRT-PCR. Error bars indicate the standard error of the mean. * p < 0.05. The data represent three separate transfections, each with three technical replicate wells.

Confirmation of candidate miR-30 target genes

We selected nine potential miR-30 target genes for further investigation and included two confirmed miR-30 target genes in our study to serve as positive controls. We constructed luciferase reporter plasmids for each of the genes of interest by subcloning each 3′-UTR immediately downstream of the Renilla (hRluc) luciferase cDNA (Data File 1). Following co-transfection of HEK293FT cells with the miR-30ac2 over-expression plasmid and each reporter plasmid, we observed that reporters containing the 3′-UTRs of all the candidate genes were significantly down-regulated relative to the control reporter, with the exception of Inhba (Figure 2). The degree of down-regulation ranged from 28–88% of basal activity.

Figure 2. Confirmation of miR-30ac2 target genes.

HEK293T cells we transfected with the miR-30ac2 expression plasmid and the indicated reporter plasmids. The ratio of firefly to Renilla luciferase activity for each reporter was normalized to that of an empty reporter. Error bars indicate the standard error of the mean. * p < 0.05. The data represent three separate transfections, each with three technical replicate wells.

MiRNA target prediction

To determine if the degree of down-regulation correlated to predicted miRNA target affinity, we correlated their PITA and TargetScan prediction scores with the reporter assay results. A slight positive correlation (R² = 0.68) was observed between the PITA target score and the reporter assay (Figure 3). No correlation was observed between the Targetscan score and the degree of repression.

Figure 3. Correlation of PITA and Targetscan prediction scores with repression by miR-30ac2.

The PITA (left) and Targetscan (right) scores are plotted on the x-axis, with the relative level of luciferase activity following miR-30ac2 over-expression on the y-axis.

Discussion

The regulation of gene expression plays a critical role in establishing the complex protein patterning that is a hallmark of development and growth. While transcriptional control determines which genes have the potential to be expressed, translational control presents an additional mechanism to fine tune protein level output spatially and temporally. In the fetal liver, miR-30a and miR-30c2 are expressed in ductal plate cholangiocytes and decrease dramatically in the postnatal bile duct⁹.

To investigate what role differential miR-30ac2 expression may be having on biliary development, we wanted to determine if miR-30ac2 directly targeted the 3′-UTR of genes that have an important function in cell growth and morphogenesis. The panel of candidate miR-30 target genes selected for validation was based on the results of gene expression profiling by microarray following miR-30a knockdown in miR-30a ASO-treated mouse embryonic liver (BMEL) cells⁹. From this data, nine genes were selected for further validation due to the potential role they play in hepato-biliary development. Of the nine experimental targets, eight were identified as direct miR-30ac2 targets, with three of these eight targets (Mtdh¹², Smad1¹³, and Twf¹⁴) validated as direct miR-30ac2 targets in separate studies during the preparation of this manuscript.

The novel genes identified as directly regulated by miR-30ac2 are Ccne2, Celsr3, Egfr, Mdm2, and Timp3. Additionally, in our study, Mtdh, Smad1, and Twf, were identified as miR-30ac2 targets, corroborating the results of the previous studies and validating our method. Inhba was the only gene included in our study that was not identified as a direct target. The targeting of these genes by miR-30ac2 support the hypothesis that the miR-30 family is required for biliary development, as the targeted genes all also have roles in hepatocyte proliferation. CcnE2 (Cyclin E2) plays a role in promoting S-phase entry from quiescence, and has also been implicated in regulatory pathways promoting liver regeneration and fibrogenesis^15,16. Celsr3 (cadherin, EGF LAG seven-pass G-type receptor 3) is a cadherin important for the maintenance of planar cell polarity¹⁷, an essential feature of hepatic epithelial cells¹⁸. Egfr (Epidermal growth factor receptor) is an essential activator of liver regeneration¹⁹, and its over-expression has also been associated with cholangiocarcoma²⁰. Mdm2 (E3 ubiquitin protein ligase) regulates tumor suppressor proteins, such as p53, by marking them for proteasomal degradation, and its aberrant expression is linked to hepatocellular carcinoma, indicating a role in liver cell proliferation²¹. Timp3 (tissue inhibitor of metallopeptidase 3) inhibits matrix metalloproteinases which function to degrade extracellular matrix, and has also been implicated in hepatocellular carcinoma²² .

During the preparation of this manuscript, three of our confirmed miR-30ac2 targets were independently verified as miR-30 targets and described in other publications. Mtdh¹², Smad1¹³, and Twf¹⁴. Tnrc6a and Ak1 were included in this study as verified positive controls and have been previously described⁹. Inhba (Inhibin, beta A) was not targeted by miR-30ac2 in our assay suggesting that it is not a direct target, but does not rule out the possibility of indirect regulation.

The degree of target repression as a result of miR-30ac2 over-expression is loosely correlated to the predicted PITA score which calculates aggregate ∆∆G of all 3′-UTR binding sites. Those targets with the greatest PITA score demonstrated a high level of repression, in the order of 28–42% expression relative to our controls. The Targetscan scores did not correlate with the reporter assay results. This difference between the two methods most likely reflects their distinct algorithms; PITA relies on thermodynamic prediction of miRNA:mRNA binding affinity, whereas Targetscan places more emphasis on evolutionary conservation^6,7. Overall, our results highlight the necessity of confirming predicted miRNA target genes by experimental approaches.

In conclusion, this study presents five new novel targets of miR-30a and miR-30c, all of which play roles in biliary and liver development, and additionally have implications in cancers of the liver. Given the crucial role of miR-30 in biliary development, we speculate that aberrant expression of miR-30 may be a factor in the development of human biliary diseases, and can therefore be considered as a potential target for therapeutic manipulation.