A systematic screen for micro-RNAs regulating the canonical Wnt pathway

Abstract

MicroRNAs (miRs) and the canonical Wnt pathway are known to be dysregulated in human cancers and play key roles during cancer initiation and progression. To identify miRs that can modulate the activity of the Wnt pathway we performed a cell-based overexpression screen of 470 miRs in human HEK293 cells. We identified 38 candidate miRs that either activate or repress the canonical Wnt pathway. A literature survey of all verified candidate miRs revealed that the Wnt-repressing miRs tend to be anti-oncomiRs and down-regulated in cancers while Wnt-activating miRs tend to be oncomiRs and upregulated during tumorigenesis. Epistasis-based functional validation of three candidate miRs, miR-1, miR-25 and miR-613, confirmed their inhibitory role in repressing the Wnt pathway and suggest that while miR-25 may function at the level of â-catenin (β-cat), miR-1 and miR-613 act upstream of β-cat. Both miR-25 and miR-1 inhibit cell proliferation and viability during selection of human colon cancer cell lines that exhibit dysregulated Wnt signaling. Finally, transduction of miR-1 expressing lentiviruses into primary mammary organoids derived from Conductin-lacZ mice significantly reduced the expression of the Wnt-sensitive β-gal reporter. In summary, these findings suggest the potential use of Wnt-modulating miRs as diagnostic and therapeutic tools in Wnt-dependent diseases, such as cancer.

Figure 1. Validation of the Pre-miR/Wnt3a/STF-report screen in HEK293 cells.

(A) Schematic overview of the screening and validation procedure. (B) Target listing of all validated synthetic cherry-pick Hsa-Pre-miRs and the average of all measurements with SEM (N = 2, n = 8). Color-code: Hsa-Pre-miRs that significantly repressed (green tab) or activated (red tab) the STF reporter of canonical Wnt-pathway activity. Green disc: anti-oncomiR; red disc: oncomiR. Inset: Representative performance of control siRNAs in one of the assays with LiCl or Wnt3a pathway induction. (C) Simplified sketch of important Wnt/β-catenin pathway nodes/switches and its activation sites by axin1/2 siRNAs or LiCl (D) Correlation between Wnt-pathway repressor (R) and activator (A) miRs being anti-oncogenic (anti-oncomiRs) or oncogenic (oncomiRs). Within each column the fraction of Pre-miRs whose expression is up- or down-regulated in cancer is indicated by color.

Figure 2. Alignment and tree of all validated Wnt-regulating Pre-miRs to represent and visualize sequence similarities.

Numbers indicate Pre-miR designation. I. e. 233 stands for the mature hsa-Pre-miR-223 strand. Green: Pre-miR that was identified and verified as Wnt-inhibitory in the STF reporter assay. Red: Pre-miR that was identified as Wnt-activator/synergizer (A) Alignment of all mature strand selected verified Hsa-Pre-miRs (B) Quartet puzzling tree phylogram determined with tree-puzzle to visualize the degree of sequence similarities between all Pre-miRs based on nucleotide substitutions (C) Representative examples of miRs that share a high degree of sequence similarities. Note the enrichment of sequence similarity with unidirectional Wnt pathway modulation (see also Fig. S3).

Figure 3. Molecular and functional characterization of candidate Wnt-regulatory miRs.

(A) Epistasis experiments with synthetic human Pre-miR-1, Pre-miR-25 and Pre-miR-613. Influence of miRs on different epistatic pathway stimulations with Wnt3a, LiCl, Axin(1+2)-siRNA or S37A stabilized β-catenin that escapes default ubiquitination and degradation. (B) Semi-quantitative Western blotting to measure relative total β-catenin protein level changes (β-catenin normalized with α-Tubulin level) in HEK293 cells transfected with 50 nM Pre-miR-1, Pre-miR-25 or Pre-miR-613 at day 3 using different pathway inductions (for 1 day) as indicated. (C) Semi-quantitative Western blotting measurement of relative total β-catenin protein level changes in SW480 colon cancer cell (APC deficiency). (D) STF19x reporter assay normalized by CMV/Renilla to measure endogenous pathway activity in Pri-miR-25 stable cancer cell lines (SW480, HCT116) compared to empty vector control stable cells. (E) Psi-check2 reporter assay system to measure the influence of β-catenin CDS mRNA fragment (inserted into the 3′UTR of a Renilla gene) on transcript translation in the presence of indicated miRs. Left: Blue columns represent the empty vector control. Red columns indicate the psi-check2-CTNNB1-CDS vector. Right: normalized psi-check2-CTNNB2 vector values are shown. (F) Viability of transfected cancer cells during G418 selection. Relative amount of cells obtained after the selection procedure to establish HCT116, SW480, and HT29 colorectal cancer cell lines expressing Pri-miR-25-pcDNA3.1(-)-Neomycin compared to empty vector control cells. Asterisk: unpaired student's t-test, (P<0.05).

Figure 4. Characterization of miR-1 overexpression in HT29 and HEK293 cells.

(A) HT29 colon cancer cells expressing pLV-Hsa-Pre-miR-1 or control vectors at day 4 of puromycin selection. (B) HT29 cells expressing pLV-Hsa-Pre-miR-1 or control at day 7 of puromycin selection. (C) Quantification of HT29 cells at day 4 and 7. RFP (red) fluorescence indicates the expression of RFP harboring the intronic miR-fragment. (D) HEK293 cells expressing miR-1 or control at day 3 (left) and day 7 (right) post puromycin selection. RED RFP fluorescence indicates expressing HEK293 cells. (E) LiCl induced Wnt pathway activity in stable miR-1expressing HEK293 cells.

Figure 5. Lentiviral expression of miR-1 inhibits expression of the Wnt reporter, axin2/Conductin-LacZ in primary mouse mammary epithelial organoid cultures.

(A-A″″) Representative image of organoid transduced with pLV-dsRed (Control) vector. (B-B″″) Representative image of organoid transduced with pLV-miR-1-dsRed lentiviral vector. The organoids transduced with control lentiviral vector (A″) shows significantly higher expression of Axin2-β-gal compared to organoids expressing miR-1 (B″). The dotted lines in panel A represent non-cellular auto-fluorescence, which was excluded from analysis of fluorescence intensity measured by NIS Elements Software. (C) Quantitative measurement of mean levels of β-gal (fluorescence intensity) in six individual organoids transduced with pLV-dsRed (Control) and seven individual organoids transduced with pLV-miR1-dsRed lentiviral vector. The average intensities were calculated for Region of interest (ROI) selected for each organoid using dsRed expression (that represents lentiviral infection) using the NIS Elements Software. Error bars denote standard deviation among samples.