The noncoding RNA, miR-126, suppresses the growth of neoplastic cells by targeting phosphatidylinositol 3-kinase signaling and is frequently lost in colon cancers

Abstract

MicroRNAs (miRNA/miR) are a class of small noncoding RNAs implicated in the pathogenesis of various malignancies. In the current study, using micro(RNA) arrays, we found a ubiquitous loss of miR-126 expression in colon cancer lines when compared to normal human colon epithelia. Reconstitution of miR-126 in colon cancer cells resulted in a significant growth reduction as evidenced in clonogenic assays. A search for miR-126 gene targets revealed p85beta, a regulatory subunit involved in stabilizing and propagating the phosphatidylinositol 3-kinase (PI3K) signal, as one of the potential substrates. Restoration of miR-126 in cancer cells induced a > or =3-fold reduction in p85beta protein levels, with no concomitant change in p85alpha, a gene that is functionally related to p85beta but not a supposed target of miR-126. Additionally, using reporter constructs, we show that the p85beta-3' untranslated region is directly targeted by miR-126. Furthermore, this miR-126 mediated reduction of p85beta was accompanied by a substantial reduction in phosphorylated AKT levels in the cancer cells, suggesting an impairment in PI3K signaling. Finally, in a panel of matched normal colon and primary colon tumors, each of the tumors demonstrated miR-126 down-regulation together with an increase in the p85beta protein level. Taken together, we propose that miR-126 regulates PI3K signaling partly by targeting p85beta, and that the loss of miR-126 may provide a selective growth advantage during colon carcinogenesis.

Figure 1. Loss of miR-126 in colon cancer cells

miR-126 expression levels in colon cancer cell lines were shown as fold changes relative to the normal colon epithelial crypts (NCC) pooled from 5 individuals. Note the marked reduction of miR-126 expression in 12 of 12 cancer lines. Amplification products of the qRT-PCR reaction from NCCs were cloned, sequenced and confirmed as miR-126 (data not shown). 5S-rRNA was used as a normalization control.

Figure 2. miR-126 suppresses the growth of colon cancer cells

A) Quantification of cell growth in V429 and V703 cell lines transfected with a control or miR-126 precursor molecule at 3 different concentrations (Ambion). The bar graphs represent the percentage of live cells, normalized to the control group, at 72 hrs after control or miR-126 transfection. Data was obtained from 3 independent wells and each experiment repeated 3 times. Note the dose-dependent suppression of cell growth by miR-126 in both V429 and V703 cells. B) Quantification of colony numbers in cell lines. The bar graphs represent colony numbers expressed as a percentage of the colony counts observed in the control group for each concentration. The colony counts for control and miR-126 groups at different doses were obtained from 3 independent wells and each experiment was repeated 3 times. Note the dose-dependent reduction in the colony number following miR-126 reconstitution in both cell lines. Statistical analysis was performed using an unpaired t test and the error bars indicate standard error of the means. (*) represents significant differences (*p<0.05, **p<0.005, ***p<0.0005) between means of the control verses miR-126 treated groups for each concentration.

Figure 3. miR-126 represses PI3K signaling by targeting p85β

A) Sequence alignment of miR-126 with 3′ UTR of p85β (Griffiths-Jones et al., 2006). Note the complimentarity of miR-126 seed sequence to p85β 3′ UTR displayed as vertical lines. B) Rows 1–5 show the RNA levels analyzed by qRT-PCR and rows 6–10 display the protein levels determined by Western blots after miR-126 induction. Row 1 demonstrates the successful restoration of miR-126 in cell lines transfected with its precursor molecule (Ambion). Note the reduction in p85β (row 6) and phospho-AKT (row 8) proteins in miR-126 treated cells. β–actin, β₂microglobulin (β₂m) and 5S-rRNA were used as loading controls for protein, mRNA and miRNA respectively. All analyses were performed 48 hrs after miR-126 transfection. C) Dual luciferase assays of luciferase reporter constructs carrying a wild-type p85β 3′ UTR verses a 3′ UTR mutation in the potential miR-126 binding site. Note the dose-dependent reduction in the luciferase activity of wild-type p85β when compared to the mutant p85β following miR-126 induction. (**) indicates significant differences between the means of wild-type and mutant p85β 3′ UTR luciferase activities within each group (t-test, p<0.005). All analyses were performed in triplicates after 48 hr incubation and values expressed as the ratio of luciferase/renilla activities and normalized to the control miRNA precursor within the respective groups.

Figure 4. Inverse correlation between miR-126 and p85β in primary colon tumors

The upper panel shows the qRT-PCR quantification of miR-126 and lower panel shows the western blot analysis of p85β protein levels in matched colon normal (N) and primary colon tumor (T) pairs. miR-126 in tumors is expressed as fold change relative to their matched normals. Note the significant downregulation of miR-126 (≥2-fold, t-test p<0.05) and overexpression of p85β protein in each of the colon tumors. 5S-rRNA and β–actin protein were used as normalization controls for miR-126 and p85β protein, respectively.