ERG deregulation induces PIM1 over-expression and aneuploidy in prostate epithelial cells

Abstract

The ERG gene belongs to the ETS family of transcription factors and has been found to be involved in atypical chromosomal rearrangements in several cancers. To gain insight into the oncogenic activity of ERG, we compared the gene expression profile of NIH-3T3 cells stably expressing the coding regions of the three main ERG oncogenic fusions: TMPRSS2/ERG (tERG), EWS/ERG and FUS/ERG. We found that all three ERG fusions significantly up-regulate PIM1 expression in the NIH-3T3 cell line. PIM1 is a serine/threonine kinase frequently over-expressed in cancers of haematological and epithelial origin. We show here that tERG expression induces PIM1 in the non-malignant prostate cell line RWPE-1, strengthening the relation between tERG and PIM1 up-regulation in the initial stages of prostate carcinogenesis. Silencing of tERG reversed PIM1 induction. A significant association between ERG and PIM1 expression in clinical prostate carcinoma specimens was found, suggesting that such a mechanism may be relevant in vivo. Chromatin Immunoprecipitation experiments showed that tERG directly binds to PIM1 promoter in the RWPE-1 prostate cell line, suggesting that tERG could be a direct regulator of PIM1 expression. The up-regulation of PIM1 induced by tERG over-expression significantly modified Cyclin B1 levels and increased the percentage of aneuploid cells in the RWPE-1 cell line after taxane-based treatment. Here we provide the first evidence for an ERG-mediated PIM1 up-regulation in prostate cells in vitro and in vivo, suggesting a direct effect of ERG transcriptional activity in the alteration of genetic stability.

Figure 1. Overexpression of HA-tagged TMPRSS2/ERG (tERG), EWS/ERG (E/E) and FUS/ERG (F/E) in NIH-3T3 cells.

a) Anti-HA immunoblot of NIH-3T3 whole cell lysates. Each transfection was performed in triplicate. Actin was used as a loading control. b) Evaluation of NIH-3T3 proliferative activity through [3H]Thymidine incorporation assay after 48 hrs of growth. Data for each ERG form represent the average of the results of the three independent populations and are normalized against the average of NIH-3T3_empty vector cells (*p<0.05). c) Venn diagrams of the transcription profiles of NIH-3T3 transfectants considering genes with a cut-off ≥1.5 fold differential expression compared to control cells (NIH3T3-empty). d) Graphic representation of the top 10 highly significant molecular and cellular function categories identified by the common 183 deregulated genes. Categories ranking was obtained by Ingenuity Pathway Software. Significance refers to the −log(p-value), which was obtained by the Ingenuity program using a right-tailed Fisher's exact test.

Figure 2. Significant co-expression of ERG and PIM1.

a) PIM1 expression fold change in NIH-3T3 transfectants compared to empty-vector cells, as determined by Quantitative Real-Time PCR (*** = p<0.001). b) Quantitative Real-Time PCR for PIM1 in human prostate cell lines. PIM1 expression, relative to the non-malignant RWPE-1 cell line is visualized in the graph (*** = p<0.001). tERG positivity is shown. c) Correlation between ERG and PIM1 expression in prostate samples from the Kunderfranco et al dataset . The Spearman rank correlation test was used to estimate the correlation significance (p = 0.0016; r = 0.3779). d) cDNA prepared from human prostate samples (tumors #1 and #2, plus a matched tumoral/non-tumoral pair [#3T/#3NT]) and from the tERG-positive VCaP cell line, was amplified as described in materials and methods. The arrows indicate the tERG amplified products. d–e) Quantitative Real-Time PCR for ERG and PIM1 in prostate cancer samples. Relative ERG and PIM1 expression compared to the tERG-negative prostate cancer #1 (d) or to the non-tumoral sample #3NT (e) is reported. ** = p<0.01; *** = p<0.001; T = tumoral; NT = non-tumoral.

Figure 3. ERG dependent PIM1 induction in RWPE-1 cells.

a–b) PIM1 over-expression in RWPE-1 transfected with HA-tagged TMPRSS2/ERG coding region (RWPE-1_tERG) as evidenced by immunoblot (a) and Real-Time quantitative PCR (b). Actin is shown as a loading control. ERG and PIM1 fold changes were normalized against empty-vector cells. c) siRNA knockdown of ERG (siERG) in RWPE-1_tERG induces a decrease in PIM1 levels compared to non-targeting siRNA (siNT). d) Sequence of the 194 bp-long PIM1 promoter region amplified in ChIP analysis. The 7 bp putative ERG binding site identified by Transcription Element Search System (TESS) is shown in italic. The distances from the ATG starting site are presented. e–f) Chromatin Immunoprecipitation (ChIP) showed a significant enrichment of HA-tERG binding to PIM1 promoter compared to IgG control in RWPE-1_tERG. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) promoter was used as a negative control. ** = p<0.01.

Figure 4. Sensitivity of RWPE-1_tERG cells to chemotherapeutic drugs.

a) Proliferative potential of tERG over-expressing and control (empty) cells measured by thymidine uptake assay up to 72 hrs growth in standard conditions. b–c) 72 hours treatment of RWPE-1 transfectants with taxol or idarubicin-hydrochloride. Proliferative potential was measured by thymidine uptake assay. d–e) Re-growth of RWPE-1 transfectants after 10 nM taxol (d) or 0.5 µM idarubcicin-hydrochloride (e): following 24-hrs treatment, drugs were removed and the cells allowed to grow for additional 72 hours. Proliferation of cells collected at the time of wash-out is set as 100%. * = p<0.05.

Figure 5. The tERG/PIM1 axis favours aneuploidy.

a) Immunoblot assay demonstrates Cyclin B1 increase in RWPE-1_tERG cells. siRNA knockdown of PIM1 (siPIM1) reduces this effect compared to non-targeting siRNA (siNT). The same amount of whole cell lysates were loaded in two different polyacrilamide gels. The tERG_siPIM1 sample for the HA immunoblot was loaded in a different position compared to the gel for PIM1, Actin and Cyclin B1 blot b) Flow cytometric analysis of non-synchronized propidium iodide stained cells. Treatment with taxol for 15/24/42 hours induces an important increase in the >4n fraction of RWPE-1_tERG compared to empty cells. c) Percentage of >4n cells after taxol treatment obtained from the average of at least two independent experiments. ** = p<0.01, *** = p<0.001. d) Annexin-V assay was carried out to evaluate apoptosis after 42 hrs of 10 nM taxol treatment for RWPE-1 transfectants. The percentage of cells in each quadrant is shown.