REP1 inhibits FOXO3-mediated apoptosis to promote cancer cell survival

Abstract

Rab escort protein 1 (REP1) is a component of Rab geranyl-geranyl transferase 2 complex. Mutations in REP1 cause a disease called choroideremia (CHM), which is an X-linked eye disease. Although it is postulated that REP1 has functions in cell survival or death of various tissues in addition to the eye, how REP1 functions in normal and cancer cells remains to be elucidated. Here, we demonstrated that REP1 is required for the survival of intestinal cells in addition to eyes or a variety of cells in zebrafish, and also has important roles in tumorigenesis. Notably, REP1 is highly expressed in colon cancer tissues and cell lines, and silencing of REP1 sensitizes colon cancer cells to serum starvation- and 5-FU-induced apoptosis. In an effort to elucidate the molecular mechanisms underlying REP1-mediated cell survival under those stress conditions, we identified FOXO3 as a binding partner of REP1 using a yeast two-hybrid (Y2H) assay system, and we demonstrated that REP1 blocked the nuclear trans-localization of FOXO3 through physically interacting with FOXO3, thereby suppressing FOXO3-mediated apoptosis. Importantly, the inhibition of REP1 combined with 5-FU treatment could lead to significant retarded tumor growth in a xenograft tumor model of human cancer cells. Thus, our results suggest that REP1 could be a new therapeutic target in combination treatment for colon cancer patients.

Figure 1

Cell survival was impaired in the intestine of rep1^−/− mutant zebrafish embryos. (a and c) Dorsal view images of wild-type zebrafish (a) and rep1^−/− mutant zebrafish embryos (c). Lateral view images of wild-type (b) and rep1^−/− mutant zebrafish embryos (d) at 5 days post fertilization (dpf). (d) Defective eye morphology was detected in rep1^−/− mutant embryos. The length of the intestine of rep1^−/− mutant embryo was shortened and it was malformed (white arrow in d). (e and f) TUNEL assays with wild-type (e) and Rep1^−/− mutant embryos (f). Extensive TUNEL labeling was observed in the intestine of rep1^−/− mutant embryos. (g and h) Immunostaining with anti-caspase-3 in wild-type Tg(flk:gfp) (g) and rep1^−/− mutant Tg(flk:gfp) embryos (h). Apoptotic cell death was detected in the intestine of rep1^−/− mutant Tg(flk::egfp) embryos (red fluorescence indicated with white arrow in h) as well as in the stomach of these embryos (yellow arrow in h). Blood vessels were visualized with green fluorescence (g and h). All images were lateral views except for (a) and (c) at 5 dpf. Scale bar, 200 μm

Figure 2

Silencing of REP1 sensitizes colon cancer cells to serum starvation- and 5-FU-induced apoptosis. (a) Immunohistochemical staining of REP1 was performed on tissue microarray of human colon cancer specimens. Representative images of immunohistochemical staining of REP1 in colon tissues from normal and carcinoma patients. High-magnification images are shown in inset. Scale bar, 100 μm. (b) Protein expression of REP1 in normal colon cells (FHC) and colon cancer cells (LoVo, SNUC4, HT29, HCT116) was determined by immunoblotting. β-ACTIN was included as an internal loading control. Numbers below blots indicate the expression as measured by fold change. (c–e) HCT116 and LoVo cells were transfected with siRNAs targeting GFP or REP1. (c) Growth rate of siGFP- versus siREP1-transfected HCT116 and LoVo cells in serum starvation culture (0.1% FBS). Cells were harvested at the indicated times and counted after trypan blue staining to exclude dead cells. (d) Flow cytometry analysis of the frequency of apoptotic (active caspase-3+) cells among siGFP- or siREP1-transfected HCT116 and LoVo cells. Isotype control staining is indicated by solid gray regions and anti-active caspase-3 staining is indicated by black dotted-lines. (e) Expressions of REP1, BIM, BAK, BID, BAX, BAD, and p27 were analyzed by immunoblotting. β-ACTIN was included as an internal loading control. Numbers below blots indicate the expression as measured by fold change. (f and g) HCT116 and LoVo cells were transfected with siRNAs targeting GFP or REP1. Cells were treated with the indicated concentrations of 5-FU for 24 h. (f) Viability of siGFP- or siREP1-transfected HCT116 and LoVo was measured by an MTT assay, and then the concentrations resulting in 50% inhibition of cell viability (IC50 values) were determined. Each experiment was performed in triplicate, and error bars represent S.D. from the mean. (g) Flow cytometry analysis of the frequency of apoptotic (active caspase-3+) cells among siGFP- or siREP1-transfected HCT116 and LoVo cells. Isotype control staining is indicated by solid gray regions and anti-active caspase-3 staining is indicated by black dotted-lines. All Graphs represent two independent experiments performed in triplicate. Error bars represent S.D. from the mean. *P<0.01, **P<0.001

Figure 3

REP1 confers resistance to serum starvation- and 5-FU-induced apoptosis. HEK293 cells were transfected with empty vector (no), REP1 wild-type (WT) or REP1 mutant (Mut). (a) Expressions of FLAG (REP1), BIM, BAK, and p27 were analyzed by immunoblotting. β-ACTIN was included as an internal loading control. Numbers below blots indicate the expression as measured by fold change. (b) Growth rate of transfected cells in serum-starvation culture (0.1% FBS). Cells were harvested at the indicated times and counted after trypan blue staining to exclude dead cells. (c) The frequency of apoptotic (active caspase-3+) cells among the transfected cells cultured in serum-starvation condition (0.1% FBS) was estimated by flow cytometry analysis. (d) Transfected cells were treated with 50 uM of 5-FU for 24 h and then the frequency of apoptotic (active caspase-3+) cells was determined by flow cytometry analysis. All graphs represent two independent experiments performed in triplicate. Error bars represent S.D. from the mean. *P<0.01, **P<0.001

Figure 4

REP1 interacts with FOXO3. (a and b) Positive protein–protein interactions were determined by monitoring cell growth over 3 days on a medium lacking leucine (a), and by the formation of blue colonies on X-gal plates containing galactose at 30 °C (b). The values of β-galactosidase activity (unit) estimated by adding o-nitrophenyl β-_D-galactopyranoside (ONPG) are indicated below their corresponding lanes. Graph represents two independent experiments performed in triplicate. Error bars represent S.D. from the mean. **P<0.001. (c) Co-immunoprecipitation of REP1 and FOXO3. HEK293 cells were co-transfected with the indicated constructs. Lysates of the transfected cells immunoprecipitated with anti-HA antibody (left) or anti-FLAG antibody (middle), followed by western blotting using anti-FOXO3 and anti-REP1 antibodies. The input represents 5% of whole-cell lysates (WCL) used in immunoprecipitation (right). (d–e) Endogenous REP1 binds to endogenous FOXO3 in HCT116 cells. Immunoprecipitation was performed using rabbit IgG, anti-FOXO3 (d) or anti-REP1 (e), followed by western blotting using the indicated antibodies

Figure 5

siREP1-mediated cell death was alleviated with inhibition of FOXO3 in colon cancer cells. (a–e) HCT116 cells were transfected with siRNAs targeting REP1 and FOXO3 as indicated. (a) Expressions of FOXO3, REP1, BIM, BAK, and p27 were analyzed by immunoblotting. β-ACTIN was included as an internal loading control. Numbers below blots indicate the expression as measured by fold change. (b and c) Transfected cells were cultured in serum-starvation condition (0.1% FBS). (b) Number of live cells was counted by trypan blue staining. (c) The frequency of apoptotic cells (active caspase-3+) was determined by flow cytometry analysis. (d–e) Transfected cells were treated with 20 uM of 5-FU for 24 h. (d) Number of live cells was counted by trypan blue staining. (e) The frequency of apoptotic (active caspase-3+) cells was determined by flow cytometry analysis. All graphs represent two independent experiments performed in triplicate. Error bars represent S.D. from the mean. *P<0.01, **P<0.001

Figure 6

REP1 negatively regulates the transcriptional activity of FOXO3 and the expression of FOXO3 target genes. (a and b) HCT116 cells were transfected with siRNAs targeting GFP or REP1. (a) FOXO3 transcriptional activity was determined using the FHRE-Luc reporter system. The cells were transfected with the FHRE-Luc reporter. A vector expressing β-galactosidase was co-transfected to ensure transfection efficiency and to normalize luciferase activity values. For analysis of the FOXO3 reporter activity in the cells, it was normalized to the control cells transfected with the empty vector. (b) mRNA expression of BAK, BIM, and p27 was analyzed by real-time quantitative RT-PCR. (c) FOXO3 transcriptional activity was determined using the FHRE-Luc reporter system. HEK293 cells were transfected with the FHRE-Luc reporter, together with the indicated plasmids. (d) HEK293 cells were transfected with empty vector (no), REP1 wild-type (WT), or REP1 mutant (Mut). mRNA expression of BAK, BIM, and p27 was analyzed by real-time quantitative RT-PCR. Fold change was calculated relative to the expression level of mRNA in control cells. All graphs represent two independent experiments performed in triplicate. Error bars represent standard deviations from the mean. *P<0.01, **P<0.001

Figure 7

REP1 inhibits nuclear trans-localization of FOXO3. HEK293 cells were transfected with RFP-FOXO3 together with empty vector (no), GFP-REP1 WT, or GFP-REP1 Mut. Confocal fluorescent microscopy was used to evaluate the subcellular distribution of FOXO3 and REP1. Representative pictures are shown in (a), and the experimental quantitation of subcellular localization of FOXO3 is shown in (b). Graph represents three independent experiments and error bars represent S.D. from the mean. **P<0.001

Figure 8

Inhibition of REP1 enhances the anti-tumor effect of 5-FU. (a) Schematic representation of the therapy regimen in mice implanted with HCT116 colon cancer cells. Nude mice were inoculated subcutaneously with 1 × 10⁶ HCT116 cells per mouse. Ten days following tumor challenge, siRNA CNP targeting GFP or REP1 (7 ug/mouse) was injected intravenously, twice for 2 consecutive days. Chitosan hydrogels loaded with 5-FU (0.1 mg/kg) were administered intratumorally at day 14. (b) Tumor growth in mice inoculated with HCT116 treated with the indicated reagents (five mice/group). (c) Tumor mass in mice at 23 days after challenge. Error bars represent S.D. from the mean. *P<0.01, **P<0.001