To better evaluate a role for miR-155 expression with respect to mTOR signaling and the ERα gene signature, we next analyzed expression of the miR-155 host gene (miR-155HG) across TCGA tumor data. In opposition to Rictor expression, the miR-155HG, which encodes the mature miR-155 sequence, correlated with an ERα
− status in TCGA breast tumor samples and mature miR-155 expression correlated with an ERα
− status in breast cancer cell lines (Figure
1B and Additional file
4: Figure S2B). As miR-155 expression correlated an ERα
− phenotype and Rictor expression correlated with ERα
+ tumors, we next investigated whether the observed high levels of miR-155 expression in ERα
− breast cancers was a driving force for the repression of Rictor. The MDA-MB-157 breast cancer cell line demonstrated the highest levels of miR-155 expression (Additional file
4: Figure S2B), so we chose this cell line and transfected a doxorubicin inducible red fluorescent protein (RFP)-miR-155sponge designed to inhibit miR-155 expression. Following transfection and RFP induction, we performed qPCR to determine Rictor expression levels. qPCR was performed and results demonstrated an increase (p < 0.06) in Rictor expression levels following miR-155 inhibition (Additional file
4: Figure S2C). In order to investigate the relationship between miR-155, mTOR, and ERα signaling; we used the ER
+ MCF-7 cell line transfected with miR-155 as this cell line demonstrated repressed Rictor expression levels and expressed levels of miR-155 equivalent to that of ER
− cell lines (Additional file
4: Figure S2A and S2D respectively). To better understand the relationship between miR-155 expression and the mTOR signaling cascade, we uploaded all miR-155 predicted targets using Pathway Interaction Database (PID) [
24] and obtained network maps for predicted miR-155 target genes and pathways (Table
1). Strikingly, many of these pathways were mediated by PI3K signaling or growth factors, which have been shown to crosstalk with mTOR signaling and indeed many components of both mTOR signaling complexes (mTORC1 and 2) were, predicted targets of miR-155 (Table
1) [
25].
Table 1
Pathways Predicted to be altered by miR-155 Target Regulation
PDGFR-beta signaling pathway | 2.77E-11 |
TGF-beta receptor signaling | 1.74E-09 |
Signaling events mediated by hepatocyte growth factor receptor (c-Met) | 6.00E-09 |
IL4-mediated signaling events | 9.61E-08 |
CXCR4-mediated signaling events | 2.22E-08 |
mTOR signaling pathway | 4.37E-08 |
IGF1 pathway | 7.71E-08 |
Regulation of retinoblastoma protein | 9.30E-08 |
Signaling events mediated by stem cell factor receptor (c-Kit) | 2.75E-07 |
AP-1 transcription factor network | 2.76E-07 |
ErbB1 downstream signaling | 3.70E-07 |
Neurotrophic factor-mediated Trk receptor signaling | 5.56E-07 |
Direct p53 effectors | 5.80E-07 |
FGF signaling pathway | 6.75E-07 |
GMCSF-mediated signaling events | 7.06E-07 |
Given that the TCGA tumor data demonstrated an inverse relationship between the loss of Rictor expression and miR-155HG expression in relation to ERα status and that Rictor expression was repressed in our MCF7-miR155 cell line, we next sought to determine the effects of miR-155 on mTOR signaling. By combining our in-house Seedfinder program (identifies isoform specific seedsites across the genome for miR-155) with previously published deep sequencing data for MCF-7 cells and the UCSC Genome Browser [
26,
27]. Appropriate miR-155 targets were chosen for further investigation based on evaluation of isoforms with 3’UTR being expressed in MCF-7 cell line (Table
2). We determined that the p70s6K 3’UTR possessed an 8-mer site its 3’UTR and the 3’UTRs of Deptor, Rheb, and TSC1 each possessed 7-mer sites (Table
2). Based on this, qPCR was performed for Deptor, Rheb, TSC1, Raptor, and p70s6K. Results demonstrate that in MCF-7-miR-155 cells, significantly increased p70s6 kinase expression was observed (Figure
1C), and significantly decreased expression of the mTOR repressor Deptor was seen (Figure
1C). Western blot analysis further confirmed increased mTOR activity demonstrated through the increased total and phospho-mTOR (S2448) in MCF-7-miR-155 cells (Figure
1D). In addition, decreased Rictor and TSC1 protein levels were observed in MCF-7-miR-155 cells (Figure
1D). The combined loss of Rictor (a critical mTORC2 component) and TSC1an mTORC2 activator and mTORC1 suppressor) suggests that miR-155 induced mTOR signaling through the mTORC1 complex. Evaluation of downstream targets of mTORC1 and mTORC2 were next evaluated. Western blot demonstrated enhanced phosphorylation of p-eEF2K and p-eIF4B downstream targets of mTORC1 (Figure
1E), there was however no noticeable change in p-p70s6K or p-S6 ribosomal protein. mTORC2 is known to enhance PKCα expression, so we next evaluated PKCα gene expression and saw repressed expression of PKCα in the MCF-7-miR-155 cell line (Figure
1F).
Table 2
mTOR Associated miR-155 Target Sequences Expressed in MCF-7 Brest Cancer Cells
Rheb | 1 | 1 | 0 | 1 | 0 |
TSC1 | 8 | 3 | 0 | 1 | 0 |
p70s6K | 6 | 5 | 1 | 0 | 1 |
Rictor | 5 | 3 | 1 | 0 | 1 |
PRKAA2 | 1 | 1 | 1 | 0 | 1 |
PML | 17 | 3 | 0 | 0 | 1 |
EEF2 | 1 | 1 | 0 | 1 | 0 |
Deptor | 2 | 2 | 0 | 1 | 0 |
EIF4E | 4 | 3 | 0 | 1 | 0 |
ULK2 | 2 | 1 | 0 | 1 | 0 |
YWHAE | 4 | 4 | 0 | 0 | 1 |