Original Article

Cancer Gene Therapy (2012) 19, 19–29; doi:10.1038/cgt.2011.62; published online 16 September 2011

Fibroblast growth factor and ornithine decarboxylase 5′UTRs enable preferential expression in human prostate cancer cells and in prostate tumors of PTEN/ transgenic mice

M Moussavi1,2, N Moshgabadi1, L Fazli1, E Leblanc1, K Zhang1, W Jia3 and P S Rennie1,4

  1. 1Vancouver Prostate Centre, University of British Columbia, Vancouver, BC, Canada
  2. 2Department of Medicine, University of British Columbia, Vancouver, BC, Canada
  3. 3Department of Surgery, University of British Columbia, Vancouver, BC, Canada
  4. 4Department of Urologic Sciences, University of British Columbia, Vancouver, BC, Canada

Correspondence: Professor PS Rennie, Vancouver Prostate Centre, 2660 Oak Street, Vancouver, BC V6H 3Z6, Canada. E-mail: prennie@interchange.ubc.ca

Received 9 October 2010; Revised 18 July 2011; Accepted 6 August 2011; Published online 16 September 2011.



In this study, we have taken advantage of over-expression of eukaryotic translation initiation factor 4E (eIF4E) in prostate cancer cells to design a viral-based targeting system of prostate cancer. Three different lengths of 5′-untranslated regions (5′UTRs) derived from either fibroblast growth factor-2 (FU-FGF2-GW) or ornithine decarboxylase (FU-ODC149-GW and FU-ODC274-GW) were inserted upstream of enhanced green fluorescent protein (GFP) gene in a lentiviral backbone. Both nonmalignant control (PNT1B and BPH-1) and neoplastic (LNCaP, C4-2, DU145 and PC-3) prostate cell lines were transfected with each plasmid or virus alone, or in the presence of siRNA against eIF4E, and their expression was monitored via GFP protein levels. Two 5′UTRs (FU-FGF2-GW and FU-ODC-GW) were selected as being most sensitive to eIF4E status. Lentiviruses containing these sequences were injected directly into the prostates of PTEN/ (tumor-bearing) and control mice. Immunofluorescence data and western blot analyses determined that a lentivirus containing a 5′UTR derived from FGF-2 is the best candidate for directing selective gene expression in the prostate tumors of PTEN/ mice in vivo. This study demonstrates that judicious selection of a complex 5′UTR can enhance selective targeting of viral-based gene therapies for prostate cancer.


prostate cancer; viral therapy; lentivirus; 5′UTR; FGF-2; PTEN knockout



Prostate cancer remains the second leading cause of cancer-related mortality in North American men.1 Patients with localized disease are often treated by surgery or radiation therapy, though almost half of these patients are not cured by these treatments.2 The majority of prostate cancer-related mortality is due to advanced and metastatic disease. Although most patients with advanced metastatic disease initially respond to androgen ablation therapies, over time their tumors become unresponsive and progress to castration-resistant prostate cancer, which has a median life expectancy of ~18 months.3, 4, 5, 6 The high incidence of prostate cancer in the North American population and its ability to progress and become resistant to treatment highlight the urgent need to develop alternative therapeutic options.

Viral-based gene therapies, such as oncolytic viral therapy, have the potential to become a viable alternative to common treatments. Previous results from our laboratory demonstrated that the oncolytic virus, vesicular stomatitis virus, was capable of increasing cell death in locally advanced prostate cancer in vivo.7 In addition, there is clinical evidence that viruses can be used to safely control treatment-resistant prostate cancer.8 Viruses such as vesicular stomatitis virus often demonstrate a preference to infect and replicate in tumor cells owing to the presence of a faulty interferon response.7, 9 These studies demonstrate the potential use of viruses as an option for targeted therapy. However, not all cancer cells have a defective interferon response. Accordingly, it would be advantageous to create recombinant viruses, which preferentially restrict their gene expression and replication to malignant cells.

To enhance tumor specificity of viruses, eukaryotic translation initiation factor 4E (eIF4E) was chosen as a potential selection criteria as it is characteristically over-expressed in tumors.4, 10 Previous studies have illustrated an increase of the intrinsic levels eIF4E in a variety of malignancies, including prostate cancer.4 Elevations in eIF4E lead to increases in cell proliferation, suppression of apoptosis and other characteristics associated with neoplastic transformation. eIF4E has been implicated in the control of translation of a few select proteins involved in developmental processes such as growth factors, proto-oncogenes and transcription factors.11 Typically, the 5′-untranslated region (UTR) of these mRNAs contains excessive secondary structures, which are normally discriminated against by the translational machinery; therefore, higher levels of eIF4E/eIF4F complex are required to unwind their 5′UTRs for efficient translation.12 Similarly, a genetically engineered virus whose replication mechanisms are under the control of eIF4E should be expressed more selectively in tumor cells with higher levels of eIF4E protein. To this end, the goal of this study was to determine which 5′UTR should be incorporated into the genetic backbone of a lentivirus, which could then lead to a tumor-specific viral infection. Three different 5′UTRs were selected, including one derived from fibroblast growth factor- 2 (FGF-2)4, 10 and two derived from ornithine decarboxylase (ODC).13 The 5′UTRs from FGF-2 and ODC have previously been shown to contain extensive, stable secondary structures, and it has previously been reported that their translation is tightly controlled.10, 13, 14

We found that two out of the three 5′UTR tested in vitro showed excellent potential for tumor-specific targeting based on eIF4E levels and the extent of enhanced green fluorescent protein (GFP) expression. Infection of prostate tumors in the PTEN/ transgenic mouse model confirmed this impression.


Materials and methods

Cell culture

BPH-1 and PNT1B non-neoplastic human epithelial prostate cells, and LNCaP, C4-2, DU145 and PC-3 human prostate cancer cells were purchased from the American Type Culture Collection. MPPK-1 cells were previously described.7 PNT1B cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen, Burlington, Ontario, Canada). All other cell lines (LNCaP, C4-2, DU145, PC-3 and MPPK-1) were maintained in RPMI (Invitrogen). All cells were supplemented with 10% fetal bovine serum (Invitrogen) and 100 units per ml penicillin/streptomycin.

Plasmid construction

The lentiviral vector FUGW, which contains an ubiquitin promoter upstream of the enhanced GFP expression gene,15 was a gift from David Baltimore (California Institute for Technology). FUGW was used as the backbone for lentiviral vector construction in this study. 5′UTR from FGF-2 (533bp) was obtained as previously described4 and cloned into the BamHI site of FUGW, downstream of the ubiquitin promoter and upstream of GFP. The PCR product from the full-length 5′UTR derived from ODC (ODC274) (274bp) and the loop portion of the 5′UTR of ODC (ODC149) (149bp) were purified and digested from PCDNA3 vector and inserted at the BamHI site of FUGW.

Cell transfections

Approximately 3 × 105 prostate cell lines were grown in 6-well plates until they were ~80% confluent. Lipofectamine 2000 (Invitrogen) was used for gene transfection according to the manufacturer's instructions. Briefly, 5μgμl−1 of lentiviral plasmid and 20nM siRNA (sense: GGA CGA UGG CUAA UUA CAU; anti sense: AUG UAA UUA GCC AUC GUCC) against eIF4E was diluted with OPTIMEN (Invitrogen) medium for 20min at room temperature. Simultaneously, Lipofectamine 2000 was mixed with OPTIMEM medium. After 20min incubation at room temperature, Lipofectamine 2000 was mixed with plasmid and siRNA against eIF4E and added to the cells. After 8h incubation, the medium was replaced with fresh medium. Experiments were terminated 48h post transfection.

Viral propagation and titration

ProFection Mammalian Transfection System (Promega, Madison, WI, USA) was used according to the manufacturer's instructions for lentiviral propagation. Briefly, 1.5 × 106 293 T cells on 10-cm plate were co-transfected with 10μg of transducing vector (FUGW, FU-UTR-GW), 10μg of packaging vector pR8.91 and 5μg of vesicular stomatitis virus envelope glycoprotein vector in unconditioned Dulbecco's modified Eagle's medium for 12–16h. After this time period, the medium on the cells was replaced with 5% fetal bovine serum containing Dulbecco's modified Eagle's medium. The supernatant was collected at 24h, 48h and 72h, and was then centrifuged at 3750g for 5min at 4°C in a bench top centrifuge (Beckman, Fullton, CA). The clear supernatant containing lentiviruses was then passed through a 0.45μm filter to remove cellular debris and was further concentrated by ultracentrifugation (123000g for 90min using Beckman ultracentrifuge with rotor S28). To quantify the viruses, concentrated lentiviruses were serially diluted and used to infect 3 × 105 of 293 T cells. At 72h post infection, cells were trypsinized and analyzed by flow cytometry. A transducing unit per ml of viruses was calculated as percent GFP positive cells (after infection with 1ml of virus) multiplied by the number of cells plated for infection.

Western blot analysis

Cell lysates prepared in a lysis buffer (50mM Tris HCI, pH 7.5, 150mM NaCl, 1% NP40 and 5mM EDTA) in the presence of a protease and phosphatase inhibitor cocktail (Sigma, St Louis, MO), containing equivalent amounts of protein (60μg), were resolved using 10% SDS-PAGE and transferred onto a nitrocellulose membrane using Bio-Rad (Mississauga, Ontario, Canada) transblot apparatus at 350mA for 90min. The nitrocellulose membrane (Gibco, Burlington, Ontario, Canada) was then blocked with Odyssey blocking buffer for 45min, and stained with primary antibodies to eIF4E (BD Transduction, Mississauga, Ontario, Canada) and GFP (Roche, Mississauga, Ontario, Canada) at a dilution of 1:1000. Antigen-bound primary antibodies were detected by IRDye 800CW goat anti-mouse or goat anti-rabbit IgG, and scanned on an Odyssey scanner (LI-COR Biosciences, Lincoln, NE).

Flow cytometry analysis of lentiviral infectivity

To determine viral infection, ~3 × 104 LNCaP and PNT1B cells were seeded in 12-well plates. Cells were treated with viruses in the presence or absence of 20nM of siRNA or scrambled siRNA against eIF4E. Cells were trypsinized and prepared into single-cell suspensions 72h post infection. Expression of GFP was analyzed on BD FACSCanto II (BD Biosciences, Mississauga, Ontario, Canada) in the FITC-A channel. Cells that were not infected with viruses were used as negative control, and cells infected with FUGW virus were used as positive control.

Prostate-specific PTEN/ mouse tumor model

The ARR2probasin-Cre transgenic line, a gift from Dr P Roy-Burman (University of Southern California),16 was crossed in-house with PTENflox/flox mice, from Dr Tak Mak (University of Toronto),17 to create prostate-specific PTEN knockout mice (PTEN/). To confirm PTEN deletion, genomic DNA was removed from tail clips of F2 offspring. The PTEN/ mice display a high-grade carcinoma between 12–15 weeks.18 For this study, all mice used were 15 weeks of age.

For viral delivery, mice were anesthetized with isoflurane and subsequently injected with 1.5mgkg−1 Metacam as analgesia presurgery. A small incision was made in the abdomen of PTEN/ male mice, and 100μl of lentiviruses at 109pfuml−1 was delivered by intra-prostate injection. At least three mice per group were killed three days post viral delivery. Prostate organs were extracted, and either fixed as frozen organs in optimum cutting temperature solution or snap-frozen in liquid nitrogen. Animal procedures were performed according to the Canadian Council on Animal Care guidelines.

Immunofluorescence staining of tissues

Four micrometer cryosections at −20°C were cut and mounted on a slide. Each section was fixed in 4% formaldehyde at room temperature for 10min. Sections were washed with PBS and incubated with 0.1% Triton in PBS for 5min. Slides were washed again and placed in a prewarmed steamer within Citrate Buffer (4.5ml of 0.1M citric acid; 20.5ml of 0.1M sodium citrate; fill up to 250ml with dH2O; pH 5.6) in a Coplan Jar for 30min. After cooling to room temperature, slides were washed again in PBS and then incubated for 10min in 3% H2O2. Excess H2O2 was washed away with PBS, and Mouse on Mouse (Vector) was used according to the manufacturer's instructions. Dilutions of 1:20 of anti-eIF4E (BD Transduction) or 1:250 of anti-GFP (Abcam) were used as the primary antibody. Either 1:200 Texax Red anti-mouse secondary or 1:200 FITC anti-Rabbit secondary antibody was applied to slides for 30min. Slides were covered with a drop (10μl) of mounting media (Vectorshield) containing 4′-6-diamidino-2-phenyl indole before being covered by a coverslip.

Immunohistochemical staining

Five micrometer sections were prepared from paraffin-embedded tissues, and the tissues were extracted from paraffin as described previously.7, 19 Tissues were stained with 1:25 anti-eIF4E, 1:200 anti-4EBP1, 1:50 anti-phospho-4EBP1, 1:50 anti-phosphor-4G and 1:50 anti-PTEN antibodies. All antibodies were purchased from Cell Signaling (Pickering, Ontario, Canada). All sections were reviewed by a pathologist (L Fazli) and scored blinded.

Statistical analysis

GraphPad InStat 3 (www.graphpad.com) was used for all statistical analyses. To measure statistical differences between mice-stained tissues, the nonparametric Wilcoxon rank test was used, as there was no assumption of a normal distribution of scores.



Levels of eIF4E are elevated in prostate cancer cell lines

Various prostate cell lines were checked for their status of eIF4E protein levels. When levels of eIF4E from prostate cancer cells such as LNCaP, C4-2, PC-3 and DU145 were compared with noncancer-derived cell lines (BPH-1 and PNT1B), there was consistently a significant increase of ~1.5- to 2-fold in eIF4E protein expression (P<0.05) in the human prostate cancer lines (Figure 1). Similarly, levels of eIF4E in PTEN/-derived prostate cancer cells, MPPK-1, were ~1.5-fold higher than those of BPH-1 and PNT1B noncancer-derived cells. Our data confirm previous findings that eIF4E is consistently higher in prostate cancer cell lines relative to nonmalignant cells.20

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Relative expression of eIF4E in control and malignant prostate cancer cells. Relative expression of eIF4E protein in noncancerous controls (BPH-1 and PNT1B) and malignant (C4-2, LNCaP, DU145, PC-3 and MPPK-1) cell lines. Whole-cell protein extracts from each cell line were resolved on SDS-PAGE and compared by western blotting for eIF4E. Density ratios of eIF4E levels were normalized to β-actin. Results are from at least three independent experiments and expressed as means±s.d.

Full figure and legend (42K)

Lentiviral containing ODC and FGF-2 5′UTRs were most sensitive to eIF4E protein levels in vitro

Three different 5′UTRs, known for their stable secondary structures, were inserted downstream of an ubiquitin promoter and upstream of a GFP marker gene in the FUGW plasmid. Two of the UTRs, ODC274 (274bp) and ODC149 (149bp), were derived from ODC. The ODC274 is the full-length 5′UTR of ODC, while the ODC149 contained the loop portion of the ODC 5′UTR.13 The other 5′UTR tested was derived from FGF-2.4, 12 A schematic representation of each lentiviral plasmid constructs is presented in Figure 2.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

The four lentiviral transfer vectors used in this study. (a) The control vector FUGW, contains ubiquitin promoter upstream of a GFP expression gene. (b) Lentiviral transfer vector containing the loop portion of ODC's 5′UTR mRNA (FU-ODC149-GW). (c) Lentiviral transfer vector containing the complete 5′UTR from ODC mRNA (FU-ODC274-GW). (d) Lentiviral transfer vector containing the 5′UTR from FGF-2 growth factor (FU-FGF2-GW).

Full figure and legend (92K)

In order to determine which construct is the most dependent on eIF4E levels for expression of GFP protein in vitro, both non-neoplastic and all neoplastic prostate cell lines were transfected with FUGW, FU-ODC149-GW, FU-ODC274-GW or FU-FGF2-GW plasmids. At 48h post transfection, levels of GFP protein expression between the various constructs containing different types of 5′UTRs were compared and normalized to FUGW control plasmid. GFP protein expression in FU-FGF2-GW-transfected BPH-1 cells was determined to be the lowest, with a 91% decrease when compared with the GFP levels from FUGW plasmid transfection in the same cell line (Figure 3a). Similarly, in PNT1B cells, the level of GFP expression was decreased by 83% in FU-FGF2-GW-transfected cell lines, followed by FU-ODC274-GW GFP levels that were reduced by 34% (Figure 3b).

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression of lentiviral plasmids containing 5′UTRs in control prostate cells. Measurement of GFP expression after transfection of nonmalignant prostate cells with different lentiviral plasmids (FUGW, FU-ODC149-GW, FU-ODC274-GW and FU-FGF2-GW). (a) BPH-1 and (b) PNTIB cells were transfected with lentiviral plasmids containing different 5′UTRs. GFP levels were measured and ratios of GFP levels normalized to β-actin, and FUGW control plasmids were graphed (n=3; as means±s.d.).

Full figure and legend (62K)

Next, we determined the effect of each 5′UTR on GFP expression in prostate cancer cell lines LNCaP, C4-2, PC-3 and DU145, given that they displayed higher levels of eIF4E (Figure 1). Each cell line was either transfected with plasmid alone or co-transfected with 20nM of an eIF4E siRNA antisense, which effectively decreased levels of eIF4E in vitro (Figure 4). When LNCaP cells were co-transfected with 20nM of eIF4E siRNA plus FU-ODC274-GW, GFP protein levels were significantly reduced by 94% (P<0.05) when compared with FU-ODC274-GW alone. However, there was no significant difference (P>0.05) in GFP expression for either FU-FGF2-GW or FU-ODC149-GW in the presence or after knockdown of eIF4E (Figure 5a). In C4-2 cell lines, the levels of GFP expression after siRNA silencing of eIF4E were most drastically reduced by 91% in cells transfected with FU-ODC274-GW (Figure 5b). This was followed by a 65% decrease in GFP expression in FU-FGF2-GW-transfected C4-2 cells. In DU145 cells, GFP levels were decreased by 87% when co-transfected with FU-FGF2-GW together with eIF4E siRNA, followed by 49% decrease in GFP protein expression when transfected with FU-ODC274-GW alone compared with when co-transfected with siRNA (Figure 5c).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

siRNA knockdown of eIF4E in prostate cancer cells. Human prostate cancer cell lines (LNCaP, C4-2, DU145 and PC-3) were transfected with 20nM of siRNA against eIF4E. (a) Western blot analysis demonstrated that eIF4E protein level is reduced in each cell line transfected with siRNA compared with non-transfected control and siRNA-scrambled. (b) Density ratios of eIF4E levels were normalized to β-actin. Results are from at least four independent experiments and expressed as means±s.d.

Full figure and legend (83K)

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression of lentiviral plasmids containing 5′UTRs in prostate cancer cells. To determine which 5′UTR is most dependent on eIF4E expression, prostate cancer cell lines were co-transfected with the lentivirus transfer plasmids both alone and in the presence of 20nM of siRNA against eIF4E. GFP protein levels in (a) LNCaP, (b) C4-2, (c) DU145 and (d) PC-3 prostate cancer cells co-transfected with 20nM of si4E and lentiviral vectors were measured. Density ratios of GFP protein levels were normalized to β-actin and FUGW control plasmid in transfected cells. Results are from at least three independent experiments and expressed as means±s.e.m.

Full figure and legend (94K)

Similar results were seen in PC-3 cell lines. There was approximately a 90% decrease in GFP expression in cells co-transfected with both FU-FGF2-GW and eIF4E siRNA, followed by a 45% reduction in GFP levels in PC-3 cells transfected with FU-ODC149-GW and siRNA. However, there was no significant difference in GFP expression in PC-3 cells that were expressing FU-ODC274-GW with eIF4E knockdown (Figure 5d). Overall, expression of GFP from FU-ODC274-GW and FU-FGF2-GW plasmids were the most sensitive to eIF4E protein levels.

Next, to determine whether viral infection would demonstrate similar eIF4E sensitivity, a representative prostate cancer cell line (LNCaP) and a nonmalignant cell line (PNT1B) were infected with lentivirus FUGW, FU-ODC149-GW, FU-ODC274-GW or FU-FGF2-GW in presence or absence of 20nM siRNA against eIF4E or scrambled siRNA. Cells were then prepared for flow cytometry, and percent GFP positive cells were calculated. After infection of PNT1B cells with FU-ODC274-GW and FU-FGF2-GW, close to 63 and 97% reduction of GFP positive cells were observed, respectively, when compared with FUGW-control viral infection (Figure 6a). However, cells that were infected with FU-ODC149-GW had no significant (P>0.05) change in percent GFP positive cells compared with FUGW control, thus indicating a lack of viral selectivity.

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Relative infectibility of nonmalignant and prostate cancer cell lines by lentiviruses containing different length 5′UTRs. (a) Nonmalignant PNT1B cells were infected with lentivirus FUGW, FU-ODC149-GW, FU-ODC274-GW and FU-FGF2-GW at MOI=10. Cells were collected 72h post infection, and analyzed by flow cytometry. GFP positive cells were detected under the FITC channel. There was a significant decrease in % GFP positive cells seen between FUGW-infected cells compared with FU-ODC274-GW- (*) and FU-FGF2-GW-infected cells (). * represents P<0.008 and represents P<0.005. Results are from three independent trials, represented as average % GFP positive cells ±s.d. (b) LNCaP prostate cancer cell lines were infected with lentivirus (FUGW, FU-ODC149-GW, FU-ODC274-GW and FU-FGF2-GW) at MOI=5 in presence or absence of 20nM siRNA against eIF4E. Cells were collected 72h post infection, and analyzed by flow cytometry. Nontreated cells are labeled as control (Ctl). GFP positive cells were detected under the FITC channel. (*) represents significant reduction in % GFP expression between LNCaP cells infected with FU-ODC274-GW and treated or not with 20nM siRNA against eIF4E. (**) represents significant reduction in % GFP expression between LNCaP cells infected with FU-FGF2-GW and treated or not with 20nM siRNA against eIF4E. P<0.01. Data are reported from three independent experiments as average % GFP positive cells ±s.d.

Full figure and legend (60K)

In LNCaP cells, there was an 80% reduction in GFP positive (viral-infected) cells treated with FU-ODC274-GW and siRNA compared with FU-ODC274-GW virus alone. Similarly, a 58% reduction was seen in LNCaP cells treated with both FU-FGF2-GW virus and siRNA against eIF4E compared with FU-FGF2-GW virus alone. However, there was no significant difference seen (P>0.05) in the infectivity of FU-ODC149-GW virus in presence (51.3±1.3% GFP positive cells) or absence (59.3±0.6% GFP positive cells) of eIF4E (Figure 6b). Thus, indicating that FU-ODC274-GW and FU-FGF2-GW viruses were more sensitive to the expression level of eIF4E. The level of GFP expression of these viruses, in response to the level of eIF4E protein expression, was further corroborated with western blot analysis (data not shown).

Levels of eIF4E are elevated in prostate tissues derived from tumor-bearing transgenic PTEN/ mouse prostates

As eIF4E is a downstream protein in the PTEN/Akt/mTOR signaling pathway, it is expected that there would be elevated eIF4E protein levels in the absence of PTEN. Immunohistochemical analysis of prostates from 15-week-old PTEN/ transgenic mice and control mice were compared in a blinded study and scored by a pathologist (Figure 7). The data confirmed that deletion of PTEN in prostates led to an increase in eIF4E levels when compared with control prostates. Additionally, we studied the level of other components of the eIF4E translational machinery. We observed an increase in the phosphorylated levels of eIF4E binding protein (4E-BP1) in PTEN/ prostates compared with control prostates (Figure 7b). Typically, when 4E-BP1 is bound to eIF4E it renders eIF4E inactive in the cytoplasm. On phosphorylation of 4E-BP1, it dissociates from eIF4E, allowing eIF4E to be phosphorylated and leading to an overall increase in translation. We found that the level of expression of the phosphorylated form of eIF4G (4G) is also higher in the prostates of PTEN/ transgenic mice compared with controls (Figure 7b). 4G, which is also bound to eIF4E, is activated upon phosphorylation of eIF4E in such a way that it is now able to interact with the 40S ribosomal subunit and initiate translation.21, 22 Together, these findings confirm that there is an increase in the activity of eIF4E and a concomitant increase in the translational machinery in PTEN/ prostates.

Figure 7.
Figure 7 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Levels of eIF4E family of translational initiation factors were analyzed in prostate tissues derived from PTEN/ and control mice. (a) Paraffin-embedded tissues were stained with PTEN antibody to demonstrate PTEN knockout in tumor-bearing PTEN/ mice. (b) Paraffin-embedded tissues were stained with anti-eIF4E, anti-eIF4E binding protein 1 (4E-BP1) and anti-phospho-4E-BP1 and anti-phospho-4 G antibodies. Representative slides were prepared and visualized at 20 × magnification. Results were scored by a pathologist (n=3).

Full figure and legend (258K)

Protein expression driven by lentiviruses containing FU-FGF2-GW is much greater in PTEN/ prostates than in controls

To determine which 5′UTR is best expressed in vivo, lentiviruses were packaged with FUGW, FU-ODC274-GW and FU-FGF2-GW plasmids. After titration, viruses were injected (109pfu per 100μl) into PTEN/ and control prostates. Three days after viral injection, the prostates were removed and stained with anti-eIF4E (red), anti-GFP (green) and 4′-6-diamidino-2-phenyl indole (blue). FUGW lentivirus was used as a control virus, and its expression was monitored through immunofluorescence staining (Figure 8). Protein expression of FUGW lentivirus did not stringently depend on cellular eIF4E protein levels. The level of GFP protein expression was further confirmed through western blot analyses (Figure 9). Although there was approximately a 2-fold increase in eIF4E protein level in PTEN/ prostates compared with control mouse prostates (data not shown), there was no significant difference in viral protein expression as determined by GFP levels (P<0.05) in FUGW-infected mice.

Figure 8.
Figure 8 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Representative immunofluorescence staining of prostate tissues from PTEN/. (a) Control and (b) mouse prostate tissue. Tissues were stained for eIF4E and GFP proteins 72h post intra-prostatic lentiviral injection with FUGW, FU-ODC274-GW or FU-FGF2-GW virus. The nuclei of the prostate cells have been stained with 4′-6-diamidino-2-phenyl indole. Data are representative of three independent experiments at × 40 magnification.

Full figure and legend (429K)

Figure 9.
Figure 9 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

GFP protein levels in prostates of PTEN/ and control mice after intra-prostatic injection of lentivirus containing 5′UTRs. GFP protein levels in PTEN/ and control prostates infected with 100μl of 109pfuml−1 of lentivirus containing FUGW, FU-ODC274-GW or FU-FGF2-GW. In each group, control mice were intra-prostatically injected with 100μl PBS, labeled as control. (a) GFP and β-actin expression was determined by western blot analysis. (b) Density ratio of GFP levels were normalized to both β-actin and control virus containing FUGW27 transfer plasmid. Results are from three independent experiments and expressed as means±s.e.m.

Full figure and legend (60K)

FU-ODC274-GW-lentivirus was also injected into prostates of both PTEN/ and control mice (Figure 9). Western blot analysis from three independent replicates demonstrated that there is no significant (P>0.05) increase in GFP expression in FU-ODC274-GW-treated PTEN/ mice compared with control. Conversely, PTEN/ mice treated with FU-FGF2-GW showed an almost 2-fold increase in viral protein expression, (P<0.05), as determined by GFP protein level, compared with control (Figure 9), indicating that, after infection with FU-FGF2-GW, lentivirus viral protein expression is more sensitive to eIF4E levels when compared with FU-ODC274-GW.



Currently, there are few treatment options available for prostate cancer patients with locally advanced or metastatic disease who have failed androgen ablation therapies.23 While docetaxel may extend survival for a few months, the patients are subjected to toxicity of normal tissue and may suffer from several adverse side effects.6 By comparison, targeted virotherapy has the potential to selectively kill cancer cells while sparing normal cells.8, 24

In this regard, our laboratory has previously demonstrated that eIF4E is a good discriminatory factor between normal and cancer cells as eIF4E levels are substantially elevated in the latter.4, 15, 20 eIF4E is the rate-limiting subunit of the eIF4F complex, which is comprised of eIF4E, the cap binding protein; eIF4A, an ATPase and RNA helicase; and eIF4G, a scaffolding protein. eIF4E is critical for initiating the process of translation by binding to the 5′-cap 7-methylguanosine (m7G) present on all nuclear transcribed mRNAs,25 and together with the eIF4F complex, unwinds the excess secondary structures within the 5′UTR to facilitate the loading of the 40S ribosomal subunit to the mRNA.22, 26 In normal cells with low levels of eIF4E, mRNAs with long G/C rich 5′UTRs, such as growth factors and other oncogenic proteins, are translated less efficiently.27 However, in cancer cells, which generally have intrinsically higher eIF4E levels, these complex secondary 5′UTR structures are more readily translated.4

In the present study, we constructed three different lentiviruses containing different types and lengths of 5′UTRs derived from FGF-2 or ODC. These 5′UTR regions were chosen because of their high G/C rich secondary structures, which require elevated eIF4E for translation. Comparison of eIF4E levels in prostate cancer cell lines to those in nonmalignant prostate cells (Figure 1) revealed that eIF4E levels were higher in cancer cell lines, in agreement with what has been previously shown.20 To establish that our lentiviral plasmids containing complex 5′UTRs are translated at a lower rate in nonmalignant cells, lentiviral transfer plasmids (FUGW, FU-ODC149-GW, FU-ODC274-GW and FU-FGF2-GW) were transfected into nonmalignant prostate cells. Expression of each plasmid was estimated through GFP protein expression and normalized to control FUGW plasmid. We found (Figure 3) that in lentiviral expression plasmids containing complex secondary 5′UTR structures, upstream of protein coding sequence (GFP), the transcripts were expressed at lower levels as compared with control (FUGW). To ensure that this lower expression of complex 5′UTR-containing expression plasmids correlated with eIF4E levels, prostate cancer cell lines were co-transfected with the plasmids together with siRNA against eIF4E. As expected, in the presence of siRNA, eIF4E levels were markedly reduced in all cell lines (Figure 4). To determine which 5′UTR was more sensitive to eIF4E levels, the expression of GFP protein in prostate cancer cells transfected with different lengths of 5′UTR lentiviral plasmids expressing transcripts with different 5′UTRs was measured in the presence or absence of the eIF4E siRNA. Our results indicated that FU-ODC149-GW expression was not strictly dependent on eIF4E levels in the majority of the cell lines. However, FU-FGF2-GW expression was reduced substantially in conjunction with decreased eIF4E in the majority of cell lines tested (Figure 5). FU-ODC274-GW expression was also significantly decreased in cells that were co-transfected with siRNA against eIF4E. Expression levels of FU-ODC274-GW plasmid in LNCaP cells upon eIF4E knockdown with siRNA showed the greatest reduction.

These data were further corroborated through viral infection of LNCaP prostate cancer cells and nonmalignant PNT1B cells with lentivirus containing the three aforementioned transfer plasmids (Figure 6). As expected, FU-ODC274-GW and FU-FGF2-GW had greater tumor cell (LNCaP) selection compared with control (PNT1B). Hence two of the three lentivirus (FU-ODC274-GW and FU-FGF2-GW), containing complex secondary 5′UTR structures, were selected for further testing in vivo. This finding is consistent with other work demonstrating the ability to target cancer cells using FGF-2 deived 5′UTR in adenovirus system.28

Prostate-specific PTEN/ mice are transgenic models that develop prostate cancer with disease progression, which mimic that seen in patients. Several studies have demonstrated that deletion and/or mutation of the PTEN gene leads to upregulation of the AKT/mTOR pathway, which can lead to increased eIF4E expression.29, 30, 31 To confirm this increase in translation machinery activity in our transgenic prostate-specific PTEN/ mouse model, protein levels of eIF4E, along with other components of translational machinery (such as phosphorylated 4G (p4G) and phosphorylated 4E-BP (p-4EBP), were evaluated and compared between PTEN/ and control mice (Figure 7). Our results confirmed that eIF4E along with other p-4G and p-4EBP were upregulated in prostates of PTEN/ mice.

To test the efficacy and specificity of expression of each of the modified 5′UTR-containing plasmids, FU-ODC274-GW, FU-FGF2-GWand FUGW (control) lentiviruses were engineered and injected directly into PTEN/ and control prostates. Three days later, GFP levels were evaluated by immunofluorescence and western blot analyses. Although both FU-ODC274-GW and FU-FGF2-GW were expressed predominantly in the prostates of tumor-bearing mice (Figure 8), the FU-FGF2-GW lentivirus showed a more selective preference for PTEN/ prostates compared with the FU-ODC274-GW virus (Figure 9). Thus, GFP reporter gene expression from a virus containing a 5′UTR derived from FGF-2 proved to be more discriminatory for tumors with high levels of eIF4E.

In conclusion, we have determined that by using a 5′UTR derived from FGF-2 in genetically engineered viruses, it is possible to select tumors with high levels of eIF4E. This finding is in keeping with recent studies, which have suggested that direct targeting of the translational machinery, specifically eIF4E, is required for efficient targeted therapies of androgen-independent prostate cancers.3, 32, 33 Hence, the careful testing and selection of 5′UTRs such as that derived from FGF-2 may enable the construction of a virus suitable for gene therapy, in that its therapeutic proteins are selectively expressed only in tumors and not in normal tissues.


Conflict of interest

The authors declare no conflict of interest.



  1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T et al. Cancer statistics, 2008. CA Cancer J Clin 2008; 58: 71–96. | Article | PubMed | ISI |
  2. Mazhar D, Waxman J. Early chemotherapy in prostate cancer. Nat Clin Pract Urol 2008; 5: 486–493. | Article | PubMed |
  3. Andrieu C, Taieb D, Baylot V, Ettinger S, Soubeyran P, De-Thonel A et al. Heat shock protein 27 confers resistance to androgen ablation and chemotherapy in prostate cancer cells through eIF4E. Oncogene 2010; 29: 1883–1896. | Article | PubMed | ISI |
  4. Yu D, Scott C, Jia WW, De Benedetti A, Williams BJ, Fazli L et al. Targeting and killing of prostate cancer cells using lentiviral constructs containing a sequence recognized by translation factor eIF4E and a prostate-specific promoter. Cancer Gene Ther 2006; 13: 32–43. | Article | PubMed | ISI | ChemPort |
  5. Fusi A, Procopio G, Della Torre S, Ricotta R, Bianchini G, Salvioni R et al. Treatment options in hormone-refractory metastatic prostate carcinoma. Tumori 2004; 90: 535–546. | PubMed | ISI | ChemPort |
  6. Petrylak DP. Future directions in the treatment of androgen-independent prostate cancer. Urology 2005; 65(6 Suppl): 8–12. | Article | PubMed | ISI |
  7. Moussavi M, Fazli L, Tearle H, Guo Y, Cox M, Bell J et al. Oncolysis of prostate cancers induced by vesicular stomatitis virus in PTEN knockout mice. Cancer Res 2010; 70: 1367–1376. | Article | PubMed | ISI |
  8. Thirukkumaran CM, Nodwell MJ, Hirasawa K, Shi ZQ, Diaz R, Luider J et al. Oncolytic viral therapy for prostate cancer: efficacy of reovirus as a biological therapeutic. Cancer Res 2010; 70: 2435–2444. | Article | PubMed | ISI |
  9. Lichty BD, Power AT, Stojdl DF, Bell JC. Vesicular stomatitis virus: re-inventing the bullet. Trends Mol Med 2004; 10: 210–216. | Article | PubMed | ISI | ChemPort |
  10. DeFatta RJ, Li Y, De Benedetti A. Selective killing of cancer cells based on translational control of a suicide gene. Cancer Gene Ther 2002; 9: 573–578. | Article | PubMed | ISI | ChemPort |
  11. van der Velden AW, Thomas AA. The role of the 5′ untranslated region of an mRNA in translation regulation during development. Int J Biochem Cell Biol 1999; 31: 87–106. | Article | PubMed | ISI | ChemPort |
  12. DeFatta RJ, Chervenak RP, De Benedetti A. A cancer gene therapy approach through translational control of a suicide gene. Cancer Gene Ther 2002; 9: 505–512. | Article | PubMed | ISI | ChemPort |
  13. Manzella JM, Blackshear PJ. Regulation of rat ornithine decarboxylase mRNA translation by its 5′-untranslated region. J Biol Chem 1990; 265: 11817–11822. | PubMed | ISI | ChemPort |
  14. Grens A, Scheffler IE. The 5′- and 3′-untranslated regions of ornithine decarboxylase mRNA affect the translational efficiency. J Biol Chem 1990; 265: 11810–11816. | PubMed | ISI | ChemPort |
  15. Yu D, Jia WW, Gleave ME, Nelson CC, Rennie PS. Prostate-tumor targeting of gene expression by lentiviral vectors containing elements of the probasin promoter. Prostate 2004; 59: 370–382. | Article | PubMed | ISI | ChemPort |
  16. Wu X, Wu J, Huang J, Powell WC, Zhang J, Matusik RJ et al. Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation. Mech Dev 2001; 101: 61–69. | Article | PubMed | ISI | ChemPort |
  17. Suzuki A, Yamaguchi MT, Ohteki T, Sasaki T, Kaisho T, Kimura Y et al. T cell-specific loss of Pten leads to defects in central and peripheral tolerance. Immunity 2001; 14: 523–534. | Article | PubMed | ISI | ChemPort |
  18. Wang S, Gao J, Lei Q, Rozengurt N, Pritchard C, Jiao J et al. Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer Cell 2003; 4: 209–221. | Article | PubMed | ISI | ChemPort |
  19. Margiotti K, Wafa LA, Cheng H, Novelli G, Nelson CC, Rennie PS. Androgen-regulated genes differentially modulated by the androgen receptor coactivator L-dopa decarboxylase in human prostate cancer cells. Mol Cancer 2007; 6: 38. | Article | PubMed | ChemPort |
  20. Lee CY, Bu LX, DeBenedetti A, Williams BJ, Rennie PS, Jia WW. Transcriptional and translational dual-regulated oncolytic herpes simplex virus type 1 for targeting prostate tumors. Mol Ther 2010; 18: 929–935. | Article | PubMed | ISI |
  21. Gingras AC, Raught B, Sonenberg N. Regulation of translation initiation by FRAP/mTOR. Genes Dev 2001; 15: 807–826. | Article | PubMed | ISI | ChemPort |
  22. Pestova TV, Kolupaeva VG, Lomakin IB, Pilipenko EV, Shatsky IN, Agol VI et al. Molecular mechanisms of translation initiation in eukaryotes. Proc Natl Acad Sci USA 2001; 98: 7029–7036. | Article | PubMed | ChemPort |
  23. Vishnu P, Tan WW. Update on options for treatment of metastatic castration-resistant prostate cancer. Onco Targets Ther 2010; 3: 39–51. | PubMed |
  24. Cozzi PJ, Burke PB, Bhargav A, Heston WD, Huryk B, Scardino PT et al. Oncolytic viral gene therapy for prostate cancer using two attenuated, replication-competent, genetically engineered herpes simplex viruses. Prostate 2002; 53: 95–100. | Article | PubMed | ISI | ChemPort |
  25. Shatkin AJ. mRNA cap binding proteins: essential factors for initiating translation. Cell 1985; 40: 223–224. | Article | PubMed | ISI | ChemPort |
  26. Gingras AC, Raught B, Sonenberg N. eIF4 initiation factors: effectors of mRNA recruitment to ribosomes and regulators of translation. Annu Rev Biochem 1999; 68: 913–963. | Article | PubMed | ISI | ChemPort |
  27. Zimmer SG, DeBenedetti A, Graff JR. Translational control of malignancy: the mRNA cap-binding protein, eIF-4E, as a central regulator of tumor formation, growth, invasion and metastasis. Anticancer Res 2000; 20: 1343–1351. | PubMed | ISI | ChemPort |
  28. Stoff-Khalili MA, Rivera AA, Nedeljkovic-Kurepa A, DeBenedetti A, Li XL, Odaka Y et al. Cancer-specific targeting of a conditionally replicative adenovirus using mRNA translational control. Breast Cancer Res Treat 2008; 108: 43–55. | Article | PubMed | ISI | ChemPort |
  29. Puli S, Jain A, Lai JC, Bhushan A. Effect of combination treatment of rapamycin and isoflavones on mTOR pathway in human glioblastoma (U87) cells. Neurochem Res 2010; 35: 986–993. | Article | PubMed | ISI |
  30. Ghayad SE, Cohen PA. Inhibitors of the PI3K/Akt/mTOR pathway: new hope for breast cancer patients. Recent Pat Anticancer Drug Discov 2010; 5: 29–57. | Article | PubMed |
  31. Bianchini A, Loiarro M, Bielli P, Busa R, Paronetto MP, Loreni F et al. Phosphorylation of eIF4E by MNKs supports protein synthesis, cell cycle progression and proliferation in prostate cancer cells. Carcinogenesis 2008; 29: 2279–2288. | Article | PubMed | ISI |
  32. Tasseff R, Nayak S, Salim S, Kaushik P, Rizvi N, Varner JD. Analysis of the molecular networks in androgen dependent and independent prostate cancer revealed fragile and robust subsystems. PLoS One 2010; 5: e8864. | Article | PubMed |
  33. Balakumaran BS, Porrello A, Hsu DS, Glover W, Foye A, Leung JY et al. MYC activity mitigates response to rapamycin in prostate cancer through eukaryotic initiation factor 4E-binding protein 1-mediated inhibition of autophagy. Cancer Res 2009; 69: 7803–7810. | Article | PubMed | ISI |


We thank Dr Latif Wafa for helping in manuscript preparation. We thank Mr Howard Tearle for help in the animal studies. This work was supported by a grant from the Terry Fox Foundation of Canada. MM was supported by a scholarship from the Department of Defense USA, PC073406.