| To whom correspondence should be addressed to: Takeshi Nishioka, Laboratory of Radiation Therapy, Graduate School of Health Sciences, Hokkaido University, N12W5, Kita-ku, Sapporo 060-0812, Japan. Tel/Fax: +81–11–706–3411 E-mail: trout@hs.hokudai.ac.jp |
A large number of researchers have reported that the intrinsic status of p53 is closely related to cancer phenotypes (Greenblatt et al., 1994) Various genotoxic reagents, including ionizing- radiation and anti-cancer drugs, are activators of p53-driven cell death pathways. However, it has also been revealed that some proportion of human cancers harboring wild-type p53 exhibit a resistance phenotype for ionizing radiation (Safran et al., 1996; King et al., 2000).
Comprehensive DNA array analysis is a powerful tool for exploring gene expression levels among different tumor cell lines, such as radiosensitive or radioresistant tumor cells. From the list of differentially expressed genes, we propose transcription factor ATF5 as a candidate gene affecting p53-dependent cell death mechanisms.
QRsP cells, a transplantable fibrosarcoma, were cultured with 8% FBS containing DMEM. Ninety percent confluent QRsP cells were irradiated at doses of 10, 20 and 30 Gy utilizing a Cobalt system (Toshiba, Japan). One hour after irradiation, 1×104 trypsinized QRsP cells were seeded onto 35 mm culture dishes.
Genomic DNA was extracted from QRsP and QRsPIR-1 cells using a Gentra Puregen kit (Qiagen, Valencia, CA, USA). Mouse p53 exons 2~11 were amplified and cloned into pCR vector (Invitrogen, Carlsbad, CA, USA) followed by sequencing.
QRsP cells irradiated at a dose of 10 Gy were seeded onto a 10 cm dish and cultured for 14 days. Well-demarcated colonies were trypsinized using a cloning cylinder and grown in DMEM. Twenty-four independent cell lines were established and named QRsPIR-1 through QRsPIR-24.
Ten thousand PBS-suspended cells were injected subcutaneously on the dorsal portion of 6-week-old female C57BL/6 mice (5 animals each). On day 28, all animals were sacrificed and the tumor masses were dissected for pathological and microarray analysis.
The procedures followed our institutional animal experiment guideline.
Total RNA was extracted from semi-confluent cultured cells or implanted tumor tissues. Extracted RNA was labeled and hybridized onto a mouse microarray chip followed by signal detection and computer analysis (Agilent Technologies, Santa Clara, CA, USA).
A phase contrast microscope (TE2000; Nikon Instech, Tokyo, Japan) equipped with a 10× objective, and enclosed in an acrylic resin box at 37°C was used for time-lapse observations. Time-lapse images were captured every 5 min using a CCD camera (QICAM-FAST; QImaging, Surrey, BC, Canada) controlled by Image-Pro software (Media Cybernetics Inc., Bethesda, MD, USA).
To establish irradiation-surviving cell lines, we irradiated the parental clone with 10 Gy, and picked 24 colonies and named them QRsPIR-1 through QRsPIR-24. We re-irradiated these clones and found resistant and sensitive clones compared to the parental clone (QRsPIR-1 for the resistant and QRsPIR-5 for the sensitive). Fig. 1A shows the result of colony assay. There is statistical significance in the number of colonies between QRsP and QRsPIR-5 (p<0.01, Student’s t-test). Fig. 1B shows that these three clonal cells demonstrated the same morphology in vitro and in vivo. Fig. 1C shows western blot of p53; similar amounts of p53 were detected among those three cell lines. To see if there is irradiation-induced up-regulation of p53, we did luciferase reporter assay for p21. Fig. 2 showed that the relative amount of luciferase driven by p21-promoter was strongly expressed 18 hours after 10 Gy irradiation for QRsPIR-5. It seems likely that the DNA-damage-induced up-regulation of p53 cascade is intact for QRsP-IR-5 cells.
 View Details | Fig. 1. A: QRsP, QRsPIR-1 and QRsPIR-5 were irradiated at a dose of 10 Gy and 1×104 irradiated-cells were seeded onto 35 mm culture dishes. Simultaneously, 100 non-irradiated cells were seeded to obtain Plating Efficiency (PE) of each cell line. All irradiated and non-irradiated cells (control) were cultured for 10 days followed by fixation and staining with Gimsae. PE is the number of cells that grow into colonies per 100 cells seeded. Survival fraction (SF) is the number of colonies per 1×104 cells planted (divided by PE/100). Each bar indicates a mean SF of triplicated experiments and error bars standard deviations. Lower panel shows Gimsae-stained colonies that appeared after 10 Gy irradiation and control. B: Phase-contrasted images of each cell line in tissue culture and haematoxylin-eosin stained microscopical images of pathologically processed tumor tissues. No obvious morphological difference was observed. Each pathological image demonstrates frequent mitoses. C: Whole cell lysate of each cell line was subjected to Western blot with anti-p53 monoclonal antibody. A comparative amount of p53 was expressed in all cell lines. |
 View Details | Fig. 2. p21 luciferase reporter assay. Cells were irradiated with 10 Gy and 18 hours later the relative amount of luciferase driven by p21-promoter was measured. Although with a weak statistical significance (p=0.08), QRsPIR-5 cells showed greater promoter activity of p21 compared to the parental QRsP. |
By adenovirus-mediated p53 gene transfer, the radiosensitive clone demonstrated a higher apoptotic fraction (Fig. 3, right bottom). The susceptibility to p53-induced cell death observed in QRsPIR-5 is a possible explanation for the radiosensitive phenotype of this clone.
 View Details | Fig. 3. QRsP, QRsPIR-1 and QRsPIR-5 cells were infected with Ad-p53 at a m.o.i of 20 pfu/cell. Twenty-four h after infection, cells were fixed with 80% ice cold ethanol followed by propium iodide staining. The histograms of mock-infected and Ad-p53 infected cells demonstrate different apoptotic (sub G1) populations among these cell lines. An apoptotic population of the QRsP, QRsPIR-1, and QRsPIR-5 was 7.12%, 13.57%, and 26.51%, respectively. The data were representative of two independent experiments. |
This fact encouraged us to search for a candidate gene which is able to enhance or repress the function of p53.
Five independent materials were analyzed (QRsP vs. QRsPIR-1 and QRsPIR-5, in vitro and in vivo) using GeneSpring software. We identified 23 genes that were expressed similarly between the parental QRsP cells and QRsPIR-1 cells and expressed differently between the parental QRsp and QRsPIR-5, with a fold change between 0.95 and 1.05 and a threshold 2.0 and 0.5 (Table I). A transcription factor ATF5 was expressed at a significantly lower level in QRsPIR-5 cells, leading us to further investigate the function of ATF5.
Fig. 4A shows a higher survival fraction of ATF5 transfectant (QRsPIR-5/ATF5) at a dose of 10 Gy (p<0.01 by Student’s t-test). Fig. 4B shows the results of p53-luciferase-reporter assay with H1299 (a p53-null lung cancer cell line). This experiment revealed that co-transfected ATF5 reduced the transactivational activity of ectopically expressed p53 and p63 (both p<0.05 in comparison to p53/p63 alone transfectants by Student’s t-test, respectively) (Fig. 4B).
 View Details | Fig. 4. A: The stable ATF5 transfectant and mock transfectant of QRsPIR-5 (QRsPIR-5/ATF5 and QRsPIR-5/pBR, respectively) were established. The radioresistance of the QRsPIR-5/ATF5 cells was restored (right: colonies, Gimsae staining; left: quantitative representation of their surviving fractions). Error bars represent standard deviations for triplicate dishes in a single experiment. Lower panel demonstrated PCR results, indicating an integration of human ATF5 cDNA into QRsPIR-5/ATF5 genome. B: H1299 cells were transfected with plasmids together with p53-luciferase reporter and renilla luciferase plasmid. Sixteen hours after transfection, cell lysates were collected and luciferase activity was measured using a Dual-Luciferase reporter assay system. Error bars represent standard deviations for triplicate wells in a single experiment. Repressive effect of ATF5 on p53 and p63 is demonstrated. C: Migration tracks of QRsPIR-5/ATF5 or QRsPIR-5/pBR cells seeded sparsely on fibronectin-coated coverslips and followed for 6 h. The intersection of the x and y axes was taken as the starting point of each cell path. Migration rate was calculated from the accumulated displacement of each cell trajectory. |
We performed a random migration assay for ATF5-transfectanted QRsP cells. Twenty cells were tracked by phase-contrast microscopy every 5 min for 6 h (Fig. 4C). Migration rates of QRsPIR-5/ATF5 and QRsPIR-5/pBR cells were 31±10 and 14±10 mm/h, respectively (p<0.05, Student’s t-test).
Whether p53 status can predict treatment results in cancer therapy is being widely discussed. It seems natural to think that treatment results are not simply related to p53 status. p21 is usually expressed through p53 elevation following DNA damage such as irradiation. However, it was not up-regulated for QRsP, the radiotolerant clone, whereas it was up-regulated for the radiosensitive clone (Fig. 2). Although there might be many other mechanisms that suppress p53 function, in our experiment, ATF5 was shown to be one of the factors related to radiosensitivity.
The transcription factors of mammalian ATF/ CREB family consist of the large group of basic region-leucine zipper (bZIP) proteins (Hai and Harman, 2001). ATF5 is a member of the ATF/CREB family and shares homology to ATF4. Among the ATF/CREB family, the ATF4 subgroup has been reported to be transcriptional repressors. ATF4 knockout mice have been generated by two independent groups and both groups have reported that ATF4-deficient mice displayed microphthalmia due to lens malformation, possibly because of lack of negative effect of ATF4 on p53 required for apoptosis in normal development. The abnormal lens formation can be rescued by p53 deletion or lens-specific expression of ATF4 (Tanaka et al., 1998; Hettmann et al., 2000). It seems likely that the expression level of ATF4 and/or ATF5 is involved in p53-dependent apoptosis. The anti-apoptotic function of ATF5 has also been reported by other groups (Persengiev and Green, 2003; Persengiev et al., 2002) For example, stably expressed ATF5 has been shown to inhibit apoptosis resulting from cytokine deprivation in an IL-3-dependent cell line (Persengiev et al., 2002).
It is noteworthy that exogenous expression of ATF5 restored the radiotolerant phenotype of QRsPIR-5 cells. Luciferase reporter assay also revealed a repressive effect of ATF5 on the transactivational activity of p53. We also demonstrated that ATF5 repressed the transcriptional activity of the p63/p51A plasmid. A recent study reported that p63 and p73, both homologs of p53 (Ikawa et al., 1999; Katoh et al., 2000), are required for p53-dependent apoptosis in response to DNA damage (Flores et al., 2002). As for the mechanisms behind the repressive effects of ATF5, there may be recruitment of the CREB-binding protein CBP/p300, a transcriptional co-activator. Binding activity between ATF4 and CBP/p300 has been previously reported (Yukawa et al., 1999). Considering the sequence homology between ATF4 and ATF5, ATF5 may also be able to bind CBP/p300. However, exogenously expressed p300 could not overcome the repressive effect of ATF5 to p53 or p63 (data not shown). The mechanism of p53 repression by ATF5 needs to be elucidated in further studies.
It is interesting that tumorigenicity was similar between QRsPIR-5 and the parental QRsP cells in our transplantation experiment. There were no significant differences in the morphological features of the two cell lines, either in vitro and in vivo (Fig. 1B). Although IR-5 cells easily die by irradiation or p53 overexpression, the transplantation experiment suggests that these cells are not “weak” cells. Another interesting finding was that stable ATF5 transfection induced greater cell motility compared with mock-transfection. Sablina et al. have reported that exogenous expression of p53 or its homolog p73 in mouse fibroblast caused the inhibition of cell migration (Sablina et al., 2003). Barbieri et al. have also reported that the loss of p63 led to increased cell migration and up-regulation of genes involved in tumor cell invasion and metastasis (Barbieri et al., 2006). A high ATF5 expression level may have a greater risk of local spread or distant metastasis.
Glioblastoma is one of the most malignant tumors in humans, with an average survival time from diagnosis of approximately 9 to 11 months. There is currently no effective therapy for this disease. Interestingly, two-thirds of glioblastomas harbor wild-type p53. On the other hand, a high level of ATF5 expression is frequently seen in glioblastoma cells (Angelastro et al., 2006).
In summary, we found that transcription factor ATF5 was able to repress the transcriptional activity of p53 and its homolog, p63. Moreover, a subclone of mouse fibrosarcoma cell line, which expressed a lower amount of ATF5 mRNA, was radiosensitive. Endogenous expression level of ATF5 could be a novel prognostic marker for radiation therapy, and tumors with a lower level of ATF5 might be a candidate for gene therapy in the future.
This study was supported in part by a Grant-in-Aid for Scientific Research (B20390319) provided by the Ministry of Education, Science, and Culture of Japan. A portion of this study was presented at ASTRO (American Society for Therapeutic Radiology and Oncology), scientific session (Biology II) in Boston, 2008.
None.
|
Index