FOXO1 inhibits Runx2 transcriptional activity and prostate cancer cell migration and invasion

Abstract

Prostate cancer patients with regional lymph node involvement at radical prostatectomy often experience disease progression to other organs, with the bone as the predominant site. The transcription factor Runx2 plays an important role in bone formation and prostate cancer cell migration, invasion, and metastasis. Here we showed that the forkhead box O (FOXO1) protein, a key downstream effector of the tumor suppressor PTEN, inhibits the transcriptional activity of Runx2 in prostate cancer cells. This inhibition was enhanced by PTEN but diminished by active Akt. FOXO1 bound to Runx2 in vitro and in vivo and suppressed Runx2's activity independent of its transcriptional function. FOXO1 inhibited Runx2-promoted migration of prostate cancer cells, whereas silencing of endogenous FOXO1 enhanced prostate cancer cell migration in a Runx2-dependent manner. Forced expression of FOXO1 also inhibited Runx2-promoted prostate cancer cell invasion. Finally, we found that expression of PTEN and the level of FOXO1 in the nucleus is inversely correlated with expression of Runx2 in a cohort of prostate cancer specimens from patients with lymph node and bone metastasis. These data reveal FOXO1 as a critical negative regulator of Runx2 in prostate cancer cells. Inactivation of FOXO1 due to frequent loss of PTEN in prostate cancer cells may leave the oncogenic activities of Runx2 unchecked, thereby driving promiscuous expression of Runx2 target genes involved in cell migration and invasion and favoring prostate cancer progression.

Figure 1

FOXO1 inhibits the transcriptional activity of Runx2. A, Effect of FOXO1-WT and FOXO1-A3 on Runx2. PC-3 cells were transfected with a luciferase-based OC promoter reporter (mOG2-Luc), a Renilla luciferase reporter and plasmids as indicated. At 24 h after transfection, luciferase activities were measured. Error bars, SD among three individual experiments. * P<0.05. Expression of FOXO1 and Runx2 proteins was analyzed by immunoblotting (IB). ERK2 was used as a loading control. B, Left, effect of wild-type and mutated PTEN on Runx2. LNCaP cells were transfected with plasmids as indicated and luciferase activities were measured and analyzed as described in (A). * P<0.05. # P>0.05. Right, a diagram summarizing the lipid and protein phosphatase activity of wild-type and mutated PTEN. C, Effect of FOXO1 on 6×OSE2-Luc. PC-3 cells were transfected with 6×OSE2-Luc, a Renilla luciferase reporter and plasmids as indicated. Luciferase activities were measured as described in (A). * P<0.05. D, Effect of Akt on FOXO1 inhibition of Runx2. DU145 Runx2-stable cells (clone #10) were transfected with the plasmids as indicated and luciferase activities were measured as described in (A). * P<0.05.

Figure 2

FOXO1 binds to and inhibits Runx2 in a manner independent of its transcriptional activity. A, DU145 cells were transfected with plasmids as indicated and luciferase activities were measured as described in Materials and Methods. * P<0.05. B, C, In vivo interaction between FOXO1 and Runx2 proteins. Cell lysates from 293T (B) and DU145 (C, left) cells transfected with plasmids as indicated or from untransfected DU145 Runx2-stable cells (clone #10) (C, right) were subjected to immunoprecipitation and western blot analysis with antibodies as indicated. D, In vitro interaction between FOXO1 and Runx2 proteins. Top, a diagram showing five GST-FOXO1 recombinant proteins. FKH, forkhead domain; TA, transactivation domain. Middle and Bottom, DU145 cells were transfected with HA-Runx2 and cell lysates were subjected to GST pull-down assay. GST and GST-FOXO1 recombinant proteins were analyzed by Coomassie blue staining. GST and GST-FOXO1 recombinant proteins at correct molecular mass are indicated by arrowheads.

Figure 3

FOXO1 inhibits expression of endogenous Runx2 target genes. A, LNCaP cells were transfected with plasmids as indicated for 48, and cells were harvested for expression of OP, IL8, VEGF and MMP13 mRNA using quantitative real-time RT-PCR. Expression levels of each gene in cells transfected with Runx2 and/or FOXO1 are shown relative to that in cells transfected with pcDNA3.1 empty vector, which is arbitrarily set to 1. * P<0.05, lane 2 versus 1. ** P<0.05, lane 3 versus 2. B, PC-3 cells were transfected with plasmids and/or siRNAs as indicated for 48 h, and cells were harvested for analysis of OC mRNA expression. * P<0.05, lane 2 versus 1. ** P<0.05, lane 3 versus 1. # P>0.05, lane 4 versus 3. C, DU145 Runx2-stable cells (clone #10) were transfected with siRNAs for FOXO1 and/or Runx2. Expression levels of Runx2 target genes VEGF and IL8 were analyzed as described in (A). * P<0.05, lane 2 versus 1. ** P<0.05, lane 3 versus 2. D, Effect of FOXO1 on the binding of Runx2 to the promoter of OP gene. Top, a diagram of primers (a/b and c/d sets) in the human OP promoter used for ChIP assay. Bottom, PC-3 cells were transfected with plasmids as indicated for 24 h and harvested for ChIP assay. The density of PCR bands from ChIP samples in lane 10, 11, and 12 was normalized by that of the bands from input samples in lane 7, 8, and 9, respectively. The relative ratio was determined by dividing the normalized values in lane 10, 11, and 12 by the value in lane 10.

Figure 4

FOXO1 inhibits Runx2-induced migration of PCa cells. PC-3 cells (A) and DU145 Runx2-stable cells (Clone #10) (C) were transfected with indicated plasmids or siRNAs and cultured to confluence on 6-well plates. Cell layer was scratched with a 200-μl pipette tip. For each sample, at least three scratched fields were photographed immediately (0 h) or at 16 h (for PC-3) or 12 h (for DU145) after scratching. Photographs of representative images were taken at ×100 magnification. B and D, Quantification of cell migration after scratching of PC-3 (B) and DU145 Runx2-stable cells (D). Error bars, SD among the average width of three independent scratches. An additional set of cells were harvested for western blot analysis of expression of proteins as indicated. In (B), * P<0.05, lane 2, 3 versus 1; ** P<0.05, lane 4 versus 1; *** P<0.05, lane 5, 6 versus 4. In (D), * P<0.05, lane 2 versus 1; ** P<0.05, lane 3 versus 2; *** P<0.05, lane 4 versus 1.

Figure 5

FOXO1 inhibits Runx2-induced invasion of PCa cells. PC-3 cells were transfected with indicated plasmids and cultured in the insert of BioCoat Matrigel invasion chamber for 22 h and stained with Diff-Quik stain. At least three fields for each sample were photographed after stain. A, Photographs of representative images were taken at ×200 magnification. B, Cell invasion was evaluated by counting the invaded cells in three independent fields. Error bars represent SD among three independent experiments. * P<0.05, lane 2, 3 versus 1. ** P<0.05, lane 5, 6 versus 4.

Figure 6

Immunohistochemistry of Runx2, PTEN, and FOXO1 proteins in PCa specimens. A, Scoring data of the density of Runx2, PTEN and FOXO1 proteins. * The density of FOXO1 protein in the nucleus. B, Tumor cells in one case express high levels of Runx2 in the nucleus (upper left) but not PTEN (lower left). Tumor cells in another case express high levels of PTEN in the cytoplasm (lower right) but little or no Runx2 (upper right). C, FOXO1 protein was not expressed in one case (left) but expressed in the nucleus of tumor cells in another case (right). D, A diagram depicts a hypothetical model wherein activation of Akt due to loss of PTEN results in FOXO1 phosphorylation and nuclear exportation, thereby promoting hyperactivation of Runx2 and PCa metastasis.