Treatment of Radioresistant Stem-Like Esophageal Cancer Cells by an Apoptotic Gene-Armed, Telomerase-Specific Oncolytic Adenovirus

Materials and Methods

Cell lines and cell culture. Cells of the human esophageal adenocarcinoma cancer cell line Seg-1 were obtained from David Beer (University of Michigan, Ann Arbor, MI). The human esophageal squamous cancer cell line TE-2 was obtained from the Cell Resource Center for Biomedical Research Institute of Development, Aging and Cancer (Tohoku University, Sendai, Japan; ref. 32). The two cell lines were maintained in DMEM supplemented with 10% heat-inactivated fetal bovine serum, 1% glutamine, and 1% antibiotics and cultured at 37°C in a humidified incubator containing 5% CO₂. Normal human fibroblast cells (NHFB; Clonetics) were cultured in medium recommended by the manufacturer.

Establishment of radioresistant cell lines. The cells were first grown to ∼50% confluence in 10-cm dishes. Cells were treated with 2 Gy of γ-radiation in a Cs-137 unit (model E-0103; U.S. Nuclear Corp.) at room temperature and then returned to the incubator. When they reached ∼80% confluence, the cells were trypsinized and subcultured into new dishes. When they reached ∼50% confluence, the cells were again irradiated (second fraction). The fractionated irradiations were continued until the total dose reached 60 Gy. The parental cells were subjected to identical trypsinization, replating, and culture conditions but were not irradiated. For all assays on irradiated cells, there was at least a 2-wk interval between the last 2-Gy FIR and the experiment.

Adenoviral vectors. The hTC promoter was created as described previously (33). Ad/CMV-GFP, Ad/TRAIL-RGD, Ad/GFP-E1, and Ad/TRAIL-E1 have been described previously (30, 31, 34–36). The recombinant oncolytic vector Ad/TRAIL-E1 was constructed by placing hTC upstream of both the TRAIL expression cassette and the E1A gene in a shuttle plasmid designated pAd/hTC-TRAIL-E1. The plasmid was then transfected with a ClaI-digested 35-kb fragment of adenovirus type 5 into 293 cells, and Ad/TRAIL-E1 was identified by PCR analysis. The vector was plaque-purified twice, expanded in 293K cells, and purified by two cycles of cesium chloride banding. The sequences in the hTC-TRAIL cassette and the hTC-E1 region were verified by DNA sequencing. Vector amplification, purification, titration, and quality tests were done as previously reported (37). Unless otherwise specified, Ad/CMV-GFP was used as the vector control, and PBS was used as the mock control. Ad/TRAIL-RGD was used as a control of non-oncolytic adenovirus expressing the TRAIL gene, and Ad/GFP-E1 was used as a control of oncolytic adenovirus without TRAIL.

Cell viability and clonogenic assay. The viability of the cell lines was determined by using the sulforhodamine B colorimetric assay, as previously described (38). For clonogenic assays, cells derived from parental cells or established radioresistant cell were trypsinized, both passed through a 40-μm sieve, and immediately irradiated at room temperature with the Cs-137 laboratory irradiator at a dose rate of 3.07 Gy/min for the time required to generate a dose curve of 0, 2, 4, 6, and 8 Gy. Corresponding controls were sham irradiated. Colony-forming assays were done immediately after irradiation by replating cells into triplicate 100-mm culture dishes using same method described previously (31, 32). Each experiment was done in triplicate and repeated at least twice.

To measure long-term cell killing effects of various adenoviral vectors in vitro, cells were seeded into six-well plates at a density of 300 per well in normal culture medium overnight and were then treated with Ad/CMV-GFP, Ad/TRAIL-F/RGD, or Ad/TRAIL-E1 at a multiplicity of infection (MOI) of 100. Cells treated with PBS alone were used as the control. The cells were then cultured in an incubator containing 5% CO₂ at 37°C for 14 d.

Flow cytometric assay. We used the flow cytometric assay to determine TRAIL, β-catenin, Oct3/4, β₁ integrin, and Coxsackie-adenovirus receptor (CAR) expression, and apoptosis. Flow cytometric evaluation of apoptosis was done as previously described (31, 35). Both floating and attached cells were collected and washed twice with cold PBS. The cells were then fixed with cold 70% ethanol and kept overnight at 4°C. Thirty minutes before the flow cytometric assay, propidium iodide staining (1 mL propidium iodide, 10 μL RNase, and 9 mL PBS, for a final concentration of 50 μg/mL) was done. The percentages of cells in the sub-G₁, G₁, S, G₂, and M phases were calculated using Cell Quest software (Becton Dickinson). Fluorescence-activated cell sorting (FACS) analysis for TRAIL or CAR expression on the cell surface was done as described previously (34, 39). In brief, 1 × 10⁶ cells suspended in 100 μL PBS were incubated at 4°C for 1 h with a rabbit polyclonal anti-TRAIL antibody (H257, Santa Cruz Biotechnology) or a mouse monoclonal anti-CAR (RmcB, Upstate) at a concentration of 10 μL/mL. The cells were washed twice with 1% fetal bovine serum in PBS and then incubated in the dark at 4°C for 30 min with a FITC-labeled goat anti-rabbit or anti-mouse immunoglobulin antibody (PharMingen) at a concentration of 10 μL/mL. After two more washes with 1% fetal bovine serum in PBS, the cells were suspended in 1% formaldehyde in PBS and subjected to FACS analysis. Normal rabbit or mouse IgG (Santa Cruz Biotechnology) was used as a control for primary antibodies; the levels detected by this control antibody were used as a basal background.

Western blot analysis. Expression of E1A, β-catenin, β₁ integrin, caspase-3, and caspase-8 was analyzed by Western blotting using the same method described previously (30, 31, 35). Briefly, cells were cultured overnight in 10-cm dishes in normal culture medium and then treated with Ad/CMV-GFP, Ad/TRAIL-F/RGD, or Ad/TRAIL-E1 at a MOI of 100. Two days after the treatment, cells were washed with cold PBS and lysed in Laemmli lysis buffer. Equal amounts of protein and lysate were separated by 10% SDS-PAGE and then transferred to enhanced chemiluminescence membranes (Hybond; Amersham Corp.). These membranes were blocked, washed, and then incubated with primary antibodies. Rabbit anti-human caspase-3 and anti-E1A were purchased from Santa Cruz Biotechnology. Mouse anti-human β-catenin and caspase-8 were obtained from BD PharMingen. After a second washing, membranes were incubated with peroxidase-conjugated secondary antibodies and developed with an enhanced chemiluminescence kit (Amersham Bioscience). β-Actin was used as a loading control.

Side population analysis using flow cytometry. The cells were detached from the dishes with trypsin-EDTA (Invitrogen) and suspended at 2 × 10⁶/mL in HBSS supplemented with 2% FCS and 10 mmol/L HEPES. These cells were then incubated at 37°C for 90 min with 10 g/mL Hoechst 33342 (Sigma Chemical), either alone or in the presence of 50 g/mL verapamil (Sigma Chemical). After incubation, 1 g/mL propidium iodide (BD PharMingen) was added. Cell analysis and purification were done using MoFlo carrying a triple-laser (DakoCytomation). Hoechst 33342 was excited with the UV laser at 350 nm and fluorescence emission was measured with 405/BP30 (Hoechst blue) and 570/BP20 (Hoechst red) optical filters. Propidium iodide labeling was measured through the 630/BP30 filter for the discrimination of dead cells.

Telomerase activity measurement. Telomerase activity was assessed using the TeloTAGGG Telomerase PCR ELISA kit (Roche Applied Science) according to the manufacturer's instructions. In brief, 2 × 10⁵ cells per sample were harvested and transferred; the pelleted cells were resuspended in 200 μL Lysis reagent and incubated on ice for 30 min. After centrifugation at 16,000 × g for 20 min at 4°C, the supernatants of lysates were collected. Three microliters of cell extract (corresponding to 3 × 10³ cell equivalents) were then incubated with biotinylated telomerase substrate oligonucleotide (P1-TS primer) at 25°C for 20 min. Telomerase inactivation was done at 94°C for 5 min. The extended products were amplified by PCR using P1-TS and P2 primers for 30 cycles. Biotinylated telomeric repeat amplification products were incubated for 2 h at 37°C with a digoxigenin-labeled detection probe complementary to the telomeric repeat sequence and immobilized onto streptavidin-coated microtiter plates. Next, the wells were incubated for 30 min at room temperature with peroxidase-labeled antidigoxigenin polyclonal antibody. Finally, the amount of telomeric repeat amplification products was determined after the addition of the peroxidase substrate (3,3′,5,5′-tetramethylbenzidine). The absorbance of each batch was measured at a wavelength of 450 nm (with a reference wavelength of ∼690 nm; V_max kinetic microplate reader, Molecular Devices) within 30 min after addition of the Stop reagent. The 293 cell line was used as the positive control; heat-treated (85°C for 10 min) 293 cells served as negative controls. Additionally, samples and controls were tested in the sample batch throughout the study. Each experiment was done in quadruplicate during the same day and repeated thrice.

Animal study. All animals were maintained and animal experiments were done under NIH and institutional guidelines established for the Animal Core Facility at The University of Texas M. D. Anderson Cancer Center.

Before tumor cell inoculation, female nu/nu mice (4-6 wk old; Charles River Laboratories) were subjected to 3.5 Gy of total body irradiation from a Cs-137 radiation source (model E-0103; U.S. Nuclear Corp.). Seg-1R cells were used to establish s.c. tumors in the mice. Briefly, 3 × 10⁶ cells were injected into the right hind leg of each mouse. When the average size of tumors reached 4 to 5 mm in diameter, mice were randomly divided into five groups: PBS, Ad/CMV-GFP, Ad/GFP-E1, Ad/TRAIL-F/RGD, and Ad/TRAIL-E1 (10 mice per group). For each treatment, 100 μL of PBS with or without 5 × 10¹⁰ vp adenovirus were injected intratumorally with a 27-gauge needle. The treatments were repeated every 3 d for a total of six treatments. Mice were ear-tagged so that data obtained from individual animals could be traced. Tumor dimensions were measured twice per week using digital calipers. Tumor volume was calculated using the equation V (mm³) = a × b²/2, where a is the largest diameter and b is the smallest diameter. The mice were killed when their tumors reached 15 mm in diameter, in which time mice were expected to die within 2 wk. The survival times of mice were estimated based on the date of killing. In another set of experiment, 25 mice bearing Seg-1R were randomly assigned into same five treatment groups (n = 5). Each tumor was treated only once, and after 3 d the whole tumor was removed for the adenovirus replication assay and terminal deoxyribonucleotide transferase–mediated nick-end labeling (TUNEL) assay.

Adenovirus replication assay in vivo. To assess the efficacy of adenovirus replication in vivo, we removed whole tumors from mice 3 d after they were treated only once with various adenoviruses; the tumors were homogenized in 0.5 mL of cold PBS and centrifuged. The viral vector replication in the supernatant of tumor lysate was determined using the tissue culture infectious dose 50 (TCID₅₀) assay in fresh 293 cells as described previously (40).

TUNEL assay. To detect apoptotic cells in tumors, we used the In situ Cell Death Detection kit, POD (Roche Applied Science). The cells were stained according to the manufacturer's instructions, counterstained with hematoxylin, and viewed under a light microscope. A brown color indicates apoptotic nuclei as visualized using 3,3′-diaminobenzidine substrate. Apoptotic cells were counted under a light microscope (×400 magnification) in randomly chosen fields, and the apoptosis index was calculated as a percentage of at least 1,000 scored cells.

Statistical analysis. The data were expressed as means and 95% confidence intervals. The statistical significance of the differences in the in vitro was determined by using the Student's t test (two-tailed). A difference was considered statistically significant when the P value was 0.05 or less. Differences in tumor growth in vivo among the treatment groups were assessed by ANOVA with a repeated measurement module. ANOVA was done to determine statistical significance between treatment groups by using the SAS procedure mixed with SAS version 6.12 software (Statistica, StatSoft). Survival was assessed by using the Kaplan-Meier method.

Results

Identification of CSC-like cell characteristics and expression of higher telomerase activity in established radioresistant esophageal cancer cell lines. To explore the molecular mechanisms involved in radiation resistance, we established radioresistant esophageal cancer cell lines by applying repeated 2-Gy FIR. The clonogenic survival curves after different total doses for the two pairs of cell lines are shown in Fig. 1A and B . There was a significant increase of radioresistance in Seg-1R compared with Seg-1 cells (P = 0.005) and in TE-2R compared with TE-2 cells (up to 25 times; P = 0.006). To determine whether radioresistant esophageal cancer cell lines contain candidate CSCs, we measured the stem cell survival factor markers (Fig. 1C). The percentages of cells positive for β-catenin, Oct3/4, and β₁ integrin in the Seg-1R cell line were 29.38%, 27.5%, and 97.34%, respectively (Fig. 1C), compared with 4.82%, 14.9%, and 45.26% in the Seg-1 cell line (P < 0.05). Figure 1D showed increased expression of β-catenin and β₁ integrin in the Seg-1R and TE-2R cell lines compared with parental cell lines by Western blot.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Establishment of radioresistant esophageal cancer cell lines by continuous FIR and identification of radioresistant cells with cancer stem-like cell characteristics. A, clonogenic survival assay of Seg-1 cells and of Seg-1R cells that were derived from Seg-1 cells by applying continuous 2-Gy fractionated irradiation (total dose of 60 Gy). B, clonogenic survival assay of TE-2 cells and of TE-2R cells that were derived from TE-2 cells. To determine surviving fractions, counts were normalized using the plating efficiency of the unirradiated corresponding control. Points, mean for three experiments, each done in triplicate (n = 9); bars, 95% confidence interval. C, the stem cell survival factor markers β-catenin, Oct3/4, and β₁ integrin were measured in Seg-1 cells (white columns) and Seg-1R cells (black columns) using flow cytometry and specific FITC-conjugated anti-catenin, anti-Oct3/4, and anti–β₁ integrin antibodies. Columns, mean of three independent experiments (n = 3); bars, 95% confidence interval. P values comparing data from Seg-1 and Seg-1R cells were determined using a paired two-sided Student's t test; n.s., not statistically significant. D, the expression of β-catenin and β₁ integrin in Seg-1, Seg-1R, TE-2, and TE-2R cells was analyzed by Western blot. β-Actin was used as a loading control. These experiments were repeated at least thrice with similar results.

To further determine whether any of the two pairs of established radioresistant cancer cell lines contained SP cells and whether such SP cells could self-renew, we stained the cells with the fluorescent dye Hoechst 33342, which has been shown to be extruded actively by the SP cells in various tissues by means of verapamil-sensitive ABC transporters; we then analyzed the cells by flow cytometry. The percentages of Hoechst^Low SP cells were greater in the Seg-1R (14.61%) and TE-2R (2.73%) cell lines than in the Seg-1 (3.47%) and TE-2 (0.30%) lines (Fig. 2A ). In each case, the SP population was greatly decreased by treatment with verapamil (Fig. 2A), indicating that the populations were bona fide SPs. Thus, radioresistant cancer cell lines (Seg-1R and TE-2R) possess SP with Hoechst efflux characteristics similar to those defined in hematopoietic stem cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Identification of SP cells with stem cell–like characteristics in established radioresistant esophageal cancer cell lines. A, Seg-1, Seg-1R, TE-2, and TE-2R cell lines were labeled with Hoechst 33342 dye and analyzed by flow cytometry before and after treatment with verapamil. Data from one of three independent experiments. B, Seg-1R cell line sorted SP and non-SP cells recovered in culture and photographed with an inverted ×10 phase-contrast microscope. SP cells from the Seg-1R cell line formed tight colonies after 4 to 7 d in culture, whereas non-SP cells were scattered and did not proliferate. C, Seg-1R sorted SP cells were cultured for 7 to 10 d, resorted by flow cytometry, recovered for an additional 7 to 10 d, and then re-analyzed by flow cytometry. Each successive sort showed the enrichment of SP cells and the presence of non-SP cells. D, telomerase activity was assessed using the TeloTAGGG Telomerase PCR ELISA kit in Seg-1, Seg-1R; SP/1, NSP/1 (non-SP from sort 1); SP/2, NSP/2; SP/3, NSP/3; TE-2, TE-2R; and NHFB cells. 293 cells were used as positive controls; heat-treated (85°C for 10 min) 293 cells were used as negative controls. Each experiment was done in quadruplicate during the same day and repeated thrice.

To compare the ability of Seg-1R SP cells and non-SP cells in producing SP cells, we stained Seg-1R cells with Hoechst 33342 and sorted them into SP and non-SP fractions by flow cytometry; we then cultured equal numbers of SP and non-SP cells separately. As they proliferated, the cells in the two populations had different morphologies. SP cells formed characteristic compact circular colonies with a cobblestone appearance and survived numerous passages (Fig. 2B). Non-SP cells were sparse and failed to proliferate beyond 2 weeks (Fig. 2B). These differences were not a consequence of prolonged Hoechst retention in the non-SP cells because propidium iodide was used to gate out all nonviable cells. When we restained the cells with Hoechst 33342 and re-analyzed them by flow cytometry, we found that the cultures initiated with SP cells contained both SP and non-SP cells. Serial sorting and reanalysis (total passages, 3) of SP cells showed enrichment of the SP cells (from 11.57% to 19.80% to 35.03%) and the presence of non-SP cells (Fig. 2C), suggesting asymmetrical division occurred during culture. Thus, SP cells are able to self-renew, be enriched, and produce non-SP cells when recovered and serially sorted in culture.

To show the difference of telomerase between parental and radioresistant, SP and non-SP cells, we measured the telomerase activity using the TeloTAGGG telomerase PCR-ELISA kit. As shown in Fig. 2D, the telomerase activity of Seg-1R, Seg-1R SP, and TE-2R cells was significantly higher than Seg-1, Seg-1R non-SP, and TE-2 cells, respectively (P < 0.05). We further tested the telomerase activity of the serial sorted cells. Figure 2D showed that the telomerase activity of SP/1 (SP from sort 1), SP/2 (SP from sort 2), and SP/3 (SP from sort 3) was also significantly higher than NSP/1, NSP/2, and NSP/3, respectively (P < 0.05). However, the telomerase activity of SP/3 was only slightly higher than that of SP/1 and SP/2 with no significant difference (P > 0.05). Certainly, the telomerase activity of these cancer cell lines was very significantly higher than NHFB cells (P < 0.001).

Preferentially targeted radioresistant esophageal CSC-like cells by telomerase-specific oncolytic adenovirus. To detect the cell killing effects of Ad/TRAIL-E1, we treated Seg-1 and Seg-1R, TE-2 and TE-2R, and NHFB cells with Ad/CMV-GFP, Ad/TRAIL-F/RGD, and Ad/TRAIL-E1 at MOIs ranging from 30 to 3,000 vp. Treatment with Ad/TRAIL-F/RGD or Ad/TRAIL-E1 resulted in a dose-dependent cell killing effect at MOIs of 30 to 3,000 vp in all cancer cell lines tested but not in NHFB cells (Fig. 3A ). The result also showed that the cell killing effects of Ad/TRAIL-E1 on esophageal carcinoma cell lines were dramatically better than those of Ad/TRAIL-RGD. Increased cell killing effects on radioresistant cells by Ad/TRAIL-E1 compared with the parental cell line were observed. These findings support the notion that Ad/TRAIL-E1 selectively targets radioresistant CSC-like cells. In NHFB cells, however, minimal toxicity was observed in treatment with Ad/TRAIL-E1 at MOIs of 1,000 to 3,000 vp. This suggests that Ad/TERT-TRAIL-E1 is highly oncolytic but minimally toxic to normal cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

High cell killing and long-term effects of Ad/TRAIL-E1 on radioresistant esophageal cancer cells. A, cell viability determined by sulforhodamine B colorimetric assay in Seg-1, Seg-1R; TE-2, TE-2R; and NHFB cells on the 4th day after treatment by Ad/CMV-GFP (GFP), Ad/TRAIL-F/RGD (TRAIL), or Ad/TRAIL-E1 (TRAIL-E1) at MOIs ranging from 0 to 3,000 vp. Cells treated with PBS was used as a control, with their viability set at 1. Each experiment was done in quadruplicate and repeated at least twice. Points, mean of quadruplicate assay results; bars, SD. B, the relative clonogenic survival of the four cancer cell lines after exposure to PBS, Ad/CMV-GFP, Ad/TRAIL-F/RGD, or Ad/TRAIL-E1 at 300 MOI. Each experiment was done in triplicate and repeated at least twice. Columns, mean results of triplicate assays; bars, SD.

To further determine the long-term effects of various vectors in vitro, we did a clonogenic formation assay. Ad/TRAIL-E1 inhibited colony formation of esophageal carcinoma cells and radioresistant CSC-like cells significantly more than Ad/TRAIL-RGD or Ad/CMV-GFP (Fig. 3B). We also found that long-term effects of Ad/TRAIL-E1 on radioresistant CSC-like cells (Seg-1R and TE-2R) were more effective than on parental esophageal carcinoma cells (Seg-1 and TE-2).

To test whether the enhanced cell killing produced by Ad/TRAIL-E1 was associated with apoptosis, we quantified the sub-G₁ population of cells after treatment with various vectors in FACS. Seg-1 and Seg-1R, TE-2 and TE-2R, and NHFB cells were treated with our tested viruses at MOIs of 100 and 1,000 vp/cell for 48 hours. The result showed that treatment with Ad/TRAIL-E1 or Ad/TRAIL-RGD induced apoptosis in all four cancer cell lines tested, as evidenced by a marked increase of sub-G₁ cells compared with cells treated with Ad/CMV-GFP or PBS (Fig. 4A and B ), whereas the cells treated with PBS or Ad/CMV-GFP had only background levels of apoptosis. No obvious apoptosis was found in NHFB cells treated with the vectors compared with PBS. In all four cancer cell lines tested, Ad/TRAIL-E1 elicited apoptosis at a MOI of 100 vp/cell as effectively as Ad/TRAIL-RGD–induced apoptosis at a MOI of 1,000 vp/cell, suggesting that the cell killing potency of Ad/TRAIL-E1 is at least 10-fold higher than that of Ad/TRAIL-RGD. Notably, both Ad/TRAIL-E1 and Ad/TRAIL-RGD induced higher apoptosis in radioresistant cancer stem-like cells (Seg-1R and TE-2R) than in the parental esophageal carcinoma cells (Seg-1 and TE-2) at the same MOIs.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Apoptosis induction and transgene expression of Ad/TRAIL-E1 in radioresistant esophageal cancer cells. A and B, the sub-G₁ population of Seg-1, Seg-1R; TE-2, TE-2R; and NHFB cells after treatment with various vectors at MOIs of 100 and 1,000 vp/cell for 48 h in a FACS analysis. C, Western blot analysis of expression of E1A and cleavage of caspase-8 and caspase-3 in Seg-1 and Seg-1R cells infected with 100 MOI of Ad/CMV-GFP, Ad/TRAIL-RGD, or Ad/TRAIL-E1 for 48 h. D, the expression of TRAIL was measured in Seg-1 and Seg-1R cells using flow cytometry and specific FITC-conjugated anti-TRAIL antibody. Columns, mean results of triplicate assays; bars, SD.

To verify that tumor killing is associated with expression of the apoptotic genes TRAIL and E1A, Seg-1 and Seg-1R cells infected with 100 MOI of Ad/CMV-GFP, Ad/TRAIL-RGD, or Ad/TRAIL-E1 were incubated for 48 hours, harvested, and then analyzed by Western blotting. Ad/CMV-GFP and Ad/TRAIL-RGD do not contain E1A and are not capable of replicating on their own, whereas Ad/TRAIL-E1–treated cells highly expressed E1A (Fig. 4C). The expression of the TRAIL protein on the cell surface was determined by FACS. Increased TRAIL expression was observed in Seg-1 and Seg-1R cells treated with 100 MOI of Ad/TRAIL-RGD and Ad/TRAIL-E1 vectors but not in cells treated with PBS or Ad/CMV-LacZ (Fig. 4D). Moreover, both Seg-1 and Seg-1R cells treated with Ad/TRAIL-E1 had higher expression levels of TRAIL than cells treated with Ad/TRAIL-RGD, and expression levels of both E1A and TRAIL were higher in Seg-1R cells compared with Seg-1 cells. To confirm the induction of apoptosis by Ad/TRAIL-E1, we tested caspase-3 and caspase-8 by Western blot analysis as well. After the cells were treated with various vectors at 100 MOI for 48 hours, cleaved caspase-3 and cleaved caspase-8 were observed in cells treated with Ad/TRAIL-RGD or Ad/TRAIL-E1 (Fig. 4C). However, in comparison with cells treated with Ad/TRAIL-RGD, both Seg-1and Seg-1R cells treated with Ad/TRAIL-E1 had notably increased cleavage of caspase-3 and caspase-8. Moreover, Seg-1R cells treated with either Ad/TRAIL-E1 or Ad/TRAIL-RGD were detected to have more expression of cleaved caspase-3 and cleaved caspase-8 than that of Seg-1 cells.

To explore the other mechanism that radioresistant cancer cells were more sensitive to adenovirus than parental cells, we analyzed the expression of CAR, α_Vβ₃ and α_Vβ₅ integrins, and DR4 in these cell lines using FACS. As shown in Fig. 5A and B , the expression of CAR, but not expressions of α_Vβ₃ and α_Vβ₅ integrins, or DR4, was significantly higher in the two radioresistant cancer cell lines than in the parental cells (P < 0.05). Thus, we first elucidated that the variation of adenovirus efficacy among different cell lines is related to the fact that radioresistance in cancer cells is associated with a higher expression of CAR. We further compared the CAR levels in the SP with the non-SP population of cells. Interestingly, the expression of CAR in the SP fraction was significantly higher than that of the non-SP population of cells (Fig. 5C; P < 0.05).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Higher expression of CAR in radioresistant esophageal cancer cells. A, the expressions of CAR, α_Vβ₃, α_Vβ₅ integrins, and DR4 were measured in Seg-1 cells (white columns) and Seg-1R cells (black columns) using flow cytometry and specific FITC-conjugated anti-CAR, anti-α_Vβ₃, anti-α_Vβ₅, and anti-DR4 antibodies. B, the expressions of CAR, α_Vβ₃ and α_Vβ₅ integrins, and DR4 were measured in TE-2 cells (white columns) and TE-2R cells (black columns). C, the expression of CAR in the SP fraction (Seg-1RSP) and the non-SP (Seg-1RNSP) population of Seg-1R cells. Columns, mean of three independent experiments (n = 3); bars, 95% confidence intervals. P values comparing data from Seg-1 and Seg-1R cells were determined using a paired two-sided Student's t test; n.s., not statistically significant.

Ad/TRAIL-E1 suppressed tumor growth and prolonged survival in mice bearing radioresistant esophageal CSC-like cells. To further evaluate the antitumor activity of Ad/TRAIL-E1 in vivo, we established human radioresistant esophageal adenocarcinoma xenografts in 6-week-old female nude mice by inoculating them with Seg-1R cells. As shown in Fig. 6A and B , Ad/TRAIL-RGD led to significant tumor suppression and prolonged survival time in tumor-bearing mice compared with mock treatment or control vector (Ad/CMV-GFP) treatment (P = 0.0008 for tumor volume and P = 0.0002 for survival). Ad/GFP-E1 also led to significant tumor suppression and prolonged survival time in tumor-bearing mice compared with mock treatment or control vector (Ad/CMV-GFP) treatment (P = 0.0025 for tumor volume and P = 0.0334 for survival). Remarkably, treatment with Ad/TRAIL-E1 suppressed tumor growth and enhanced survival in tumor-bearing mice compared with mock treatment or control vector (Ad/CMV-GFP) treatment (P = 0.00004 for tumor volume and P = 0.00002 for survival). Furthermore, treatment with Ad/TRAIL-E1 significantly retarded tumor growth and prolonged survival time compared with treatment with Ad/TRAIL-RGD (P = 0.00196 for tumor volume and P = 0.02620 for survival) and Ad/GFP-E1 (P = 0.00103 for tumor volume and P = 0.0008 for survival). However, treatment with Ad/TRAIL-RGD was not significantly different compared with Ad/GFP-E1 (P = 0.33 for tumor volume and P = 0.057 for survival). The mean survival times for mice treated with PBS or Ad/CMV-GFP, Ad/TRAIL-RGD, Ad/GFP-E1, and Ad/TRAIL-E1 were 55, 64, 86, 74, and 125 days, respectively. Treatment with Ad/TRAIL-E1 led to tumor-free survival in 40% mice for 180 days, the longest time point observed in this study. We also evaluated possible treatment-related toxicity by measuring serum liver enzymes and blood cell counts (RBC, WBC, and platelets) 5 days after the final treatment injection. All of the results were within the reference ranges, and no substantial differences were found between the groups (data not shown). In addition, no weight loss was noted in the mice.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Ad/TRAIL-E1 suppressed tumor growth and prolonged survival in mice bearing radioresistant esophageal adenocarcinoma cells. Seg-1R cells were grown as xenograft tumors in nude mice, and mice bearing tumors that reached 4 to 5 mm in diameter were intratumorally injected with PBS, Ad/CMV-GFP, Ad/TRAIL-F/RGD, Ad/GFP-E1, or Ad/TRAIL-E1 for a total of six injections, with vectors of 5 × 10¹⁰ vp each time. Arrows, times when treatments were given. A, tumor volume was monitored over time after the inoculation of tumor cells. Points, means; bars, SE. ANOVA was done to determine statistically significant differences between treatment groups. B, Kaplan-Meier survival curves for the treatment groups. C, representative fields showing H&E-stained sections (top) and tumor cell apoptosis by TUNEL assay (bottom) in Seg-1R tumors from mice. Mice received the various treatments only once, and tumors were removed 3 d after the treatment. Brown, apoptotic nuclei as visualized using 3,3′-diaminobenzidine substrate. Apoptotic cells were counted under a light microscope in randomly chosen fields. D, the viral vector replication in supernatants of tumor lysates was determined using the tissue culture infectious dose 50 (TCID₅₀) assay in fresh 293 cells.

Adenovirus replication and apoptosis induction by Ad/TRAIL-E1 in vivo. As shown in Fig. 6C, H&E staining revealed necrosis within tumors from mice treated with Ad/TRAIL-RGD and Ad/GFP-E1. Furthermore, Ad/TRAIL-E1 induced more tumor necrosis. To assess apoptosis induction in vivo, we did TUNEL staining on tumor sections. As shown in Fig. 6C, Ad/TRAIL-E1 resulted in a significantly higher apoptotic index (43.7%) compared with Ad/TRAIL-F/RGD (19.6%) or Ad/GFP-E1 (17.5%), after subtracting the background level of 3.5% (P = 0.00269). To assess the efficacy of adenovirus replication in vivo, we determined the viral vector replication in supernatants of tumor lysate using the tissue culture infectious dose 50 (TCID₅₀) assay in fresh 293 cells. We found 10⁷ to 10⁸ infectious viral units in the lysates of tumor from mice treated with Ad/TRAIL-E1 and 10⁶ to 10⁷ infectious viral units in the lysates of tumor from mice treated with Ad/GFP-E1 (Fig. 6D). No detectable infectious units (i.e., <10³, the lowest dilution the assay can measure) were found in the lysates of tumors from mice treated with Ad/CMV-GFP or Ad/TRAIL-RGD, demonstrating that the oncolytic vectors but not the E1-defective vectors were replicating in those tumor cells.

Discussion

The existence of CSCs has profound implications for cancer biology and therapy (41). It has been proposed that CSCs may be particularly resistant to chemotherapy and radiation therapy, although good evidence supporting this notion has been lacking (7). Recently, one published report suggested that glioblastoma stem cells are radioresistant and may therefore contribute to treatment failures (42); another article showed that breast cancer–initiating cells are relatively radioresistant compared with the remainder of breast cancer cells (43). We examined whether radioresistant cell lines could reflect biological characteristics of stem-like cells. We found that the putative stem cell markers β-catenin, Oct3/4, and β₁ integrin expression were significantly increased in radioresistant cells. In addition, we noticed that several so-called stemness genes were up-regulated and apoptosis genes were down-regulated in Seg-1R cells compared with Seg-1 cells in microarray analysis (data not shown). Moreover, serial sorting and reanalysis of SP cells from the Seg-1R cell line showed enrichment of the SP and the presence of non-SP cells, suggesting that asymmetrical division occurred during culture. These results indicate that a tumor hierarchy exists in which SP cells can generate both SP and non-SP cells. SP cells seem to be able to self-renew and proliferate but not in non-SP cells. This is in accordance with previous observations that the SP fraction can divide asymmetrically and display a capacity for self-renewal (23, 44, 45). Finally, higher telomerase activity was detected in radioresistant CSCs, particularly SP cells compared with parental cells, indicating that CSCs may present a reservoir with unlimited proliferation potential for generating cancer cells. Our results show that radioresistant stem-like solid cancer cell lines can be established in radioresistant esophageal cancer cells by continuous FIR. It would be interesting to see whether the accumulation of CSCs after FIR can also be documented in esophageal cancer patients treated with standard (50.4 Gy) radiotherapy and chemotherapy and whether the percentage of CSCs within cancer tissues predicts radiosensitivity.

The radioresistance may be caused by radiation-induced gene mutation, altered gene expression, or enrichment of CSCs because CSCs may be resistant to radiotherapy. FIR mimics the clinical situation of development of radioresistance after standard chemoradiotherapy. Elimination of radioresistance is crucial to achieving a pathologic complete response and a cure. There are a variety of markers and techniques to isolate and eliminate these rare CSCs (21, 44–47). In primary breast cancer, it has been shown that the CSCs are telomerase positive (47). The telomerase inhibitors being developed might therefore target CSCs as well as the more mature cancer cells (27). In the current study, we found that oncolytic adenoviral vector Ad/TRAIL-E1 seems to preferentially target esophageal cancer cells, particularly CSCs.

Factors determining the susceptibility of the targeting radioresistant cancer cells to the adenovirus are not entirely known. In this study, we observed that radioresistant cancer cells were more sensitive to adenovirus than parental cells. Overexpression of telomerase activity in esophageal radioresistant CSC-like cells may partially explain the increased sensitivity of esophageal radioresistant cancer cell lines to Ad/TRAIL-E1. Further study revealed increased CAR expression and elevated transgene expression in the radioresistant cancer cell lines. Increased CAR expression on the cell surface of radioresistant cancer cells may contribute to the efficiency of adenovirus-mediated gene transfer. Generally, one major weakness of adenovirus treatment is that adenoviral infection efficiency depends on the expression of CAR, which is not highly expressed on the cell surface of many types of cancer cells (48, 49). However, our study showed that the CAR was highly expressed on the cell surface of radioresistant CSC-like cells. In addition, we found that the telomerase activity of SP/1, SP/2, and SP/3 was significantly higher than NSP/1, NSP/2, and NSP/3, respectively (Fig. 2D). We also found that the expression of CAR in the SP fraction was significantly higher than in the non-SP population of cells (Fig. 5C). It seems that both increased telomerase activity and expression of CAR play role in the efficacy of Ad/TRAIL-E1. Certainly, the relative contribution of these two processes and the detailed mechanism of the tumoricidal effect of Ad/TRAIL-E1 remain to be further elucidated.

In terms of toxicity, our previous report showed that hTERT promoter activity was low in mesenchymal stem cells and Ad/TRAIL-F/RGD was associated with very low toxicity to these mesenchymal stem cells in vitro even at high doses (36). In the current study, we evaluated potential treatment-related toxicity of adenovirus vectors in nude mice model using local injection. We found no substantial differences among the treatment groups, indicating low toxicity of Ad/TRAIL-E1. However, the potential toxicity of Ad/TRAIL-E1 using i.v. injection could be higher and further study is needed.

CAR has primarily been studied in its role as the initial cell surface attachment receptor for Coxsackie and group C adenoviruses. However, CAR is a multifaceted molecule expressed by many cell types and fulfilling many but yet poorly defined cellular functions (50). It behaves typically as an adhesion molecule promoting cell aggregation and may play role in embryogenesis and stem cell lineage (50). Overexpression of CAR in radiation-resistant cancer stem-like cell may be caused by preferential killing of cancer cells with low CAR expression by radiation or radiation-induced CAR expression. The role of tumorigenesis of CAR remains controversial. Recent data indicated that silencing surface CAR expression abrogated xenograft tumorigenesis in vivo and colony formation in vitro (51). These data indicated that CAR expression may be needed for the efficient formation of tumors by a subset of cancer cells. It would be interesting to further explore the relationship between CAR-mediated signaling pathways and CSC properties.

Taken together, our results indicated that radioresistant cells could be established by repeated FIR and that they display CSC-like properties. Ad/TRAIL-E1 specifically targeted esophageal radioresistant CSC-like cells and prolonged survival. Identification of the esophageal CSCs would provide a critical step in advancing the development of future anti-CSCs targeted therapies.