 Index | ||||||||||
|
| To whom correspondence should be addressed: Takeshi Nishioka, Department of Biomedical Sciences and Engineering, Faculty of Health Sciences, Hokkaido University, N12-W5, Kita-ku, Sapporo 060-0812, Japan. Tel/Fax: +81–11–706–3411 E-mail: trout@hs.hokudai.ac.jp |
|
Radiotherapy is one of the most effective treatments for various malignancies. It can cure tumors arising in the head and neck, uterine cervix, and recently even the lung (Inoue et al., 2009). It also plays an important role (combined with surgery or chemotherapy) in the management of breast cancer and lung cancer; both are now leading causes of death worldwide. Several attempts have been made to increase the tumor-killing effects of radiotherapy. Among them, shortening the overall treatment time of radiotherapy to counterattack tumor repopulation has been the focus of attention for the past decade. Recently, molecular-targeted medicines have been shown to be effective. Repopulation or regeneration is a widely recognized phenomenon in which treatment with any cytotoxic agent can trigger surviving cells in a tumor to divide faster than before (Kim and Tannock, 2005).
Recent reports on the use of accelerated irradiation regimens are encouraging, but inevitably accompanied by increased toxicities (Bernier, 2005). Achieving a high degree of tumor control with fewer complications is a difficult proposition, which makes it necessary to understand tumor behavior during radiotherapy at the genetic and molecular levels. Despite widespread awareness of tumor regeneration among oncologists and biologists, little is known about its basic mechanism. Why and how do tumor cells survive irradiation and start to regenerate? Does a certain percentage of cells survive just by chance or because those cells had a tolerant genetic profile in the first place (i.e., tolerant cell selection)? If selection is the case, are surviving cells like cancer stem cells (Gilbert and Ross, 2009)? Addressing these issues will help us better understand tumor kinetics during radiotherapy and could provide us with new insights towards the development of novel therapies.
QRsP transplantable fibrosarcoma cells, p53 wild type, (Okada et al., 1992), were cultured in 8% FBS containing DMEM. Ninety percent confluent QRsP cells were irradiated at a dose of 10 Gy using a Cobalt system (Toshiba, Japan). The dose rate was kept at 1.8 Gy/min, and the distance from the source to the mice was 80 cm. One hour post irradiation, 1×104 of trypsinized QRsP cells were seeded on 35 mm culture dishes. At the same time, 100 non-irradiated QRsP cells were also seeded on 35 mm culture dishes to confirm plating efficiency. These cells were cultured for 10 days with DMEM followed by methanol fixation and Giemsa staining.
QRsP cells irradiated at a dose of 10 Gy were seeded onto 10 cm dishes and cultured for 14 days. Well-demarcated colonies were trypsinized using a cloning cylinder and grown in DMEM. Six independent cell lines were established and named QRsPIR-1 through QRsPIR-6. A subset of these cells was irradiated at a dose of 10 Gy and 1×104 of trypsinized cells were cultured for 10 days, and the number of methanol-fixed and Giemsa-stained colonies were counted. The rest of the cells were stored at –80°C for future experiments.
For in vitro comparison, total RNA was extracted from semi-confluent QRsP parental cells and QRsPIR-1 cells. A comparison also made for the parental cells and QRsP-5 cells. We also prepared RNA from both QRsP- and QRsPIR-1-derived tumors (one animal each). Extracted RNA was labeled and hybridized onto a mouse microarray chip, followed by signal detection and computer analysis according to the manufacturer’s protocol (Agilent Technologies, Santa Clara, CA). Genes up- or down-regulated more than 2 fold are summarized in tables (hereafter referred to as “array results”). NCBI Map Viewer was used to locate the genes on chromosomes.
Semi-confluent QRsP parent cells and QRsPIR-1 cells were trypsinized and resuspended in PBS. One thousand PBS-suspended cells were injected subcutaneously into the dorsal side of 6-week-old female C57BL/6 mice (5 animals each). On the 28th day, all animals were sacrificed and the tumor masses were dissected. The experiments strictly followed the animal care guidelines of Hokkaido University. Tumor tissues were fixed with 10% formaldehyde and embedded in paraffin according to routine pathological procedures. A 5 micrometer-thick section of each specimen was stained with hematoxylin and eosin. Two pathologists independently examined the samples. Mitotic cells were counted in five randomly selected on high-power fields, and the average numbers of mitoses were obtained.
Total RNA was extracted from QRsP parental cells, QRsPIR-1 and QRsPIR-5 cells using the Isogen RNA extraction kit (Nippon Gene) according to the manufacturer’s instructions. cDNA was prepared from the total RNA with a reverse transcriptase (Invitrogen), oligo dT and dNTP mixture (Promega). The indicated cDNA was specifically amplified by thermal cycler (LightCycler; Roche Applied Science, Indianapolis, IN) using the corresponding primer pairs and probes for mouse MMP3, MMP13, LOX, p57 and β-actin. The following primers were used: MMP-3: sense, 5'-TTGTTCTTTGATGCAGTCAGC-3'; anti-sense, 5'-GATTTGCGCCAAAAGTGC-3'; MMP13: sense, 5'-GCCAGAACTTCCCAACCAT-3'; anti-sense, 5'-TCAGAGCCCAGAATTTTCTCC-3'; LOX: sense, 5'-CAGGCTGCACAATTTCACC-3'; anti-sense, 5'-CAAACACCAGGTACGGCTTT-3'; p57: sense, 5'-CAGGACGAGAATCAAGAGCA-3'; anti-sense, 5'-GCTTGGCGAAGAAGTCGT-3'; and β-actin: sense, 5'-AAGGCCAACCGTGAAAAGAT-3'; GTGGTACGACCAGAGGCATAC-3'.
QRsP is a transplantable fibrosarcoma-cell-line derived from C57BL/6 mice (Kobayashi et al., 2002). This gave us an advantage over the usual nude mouse experiments in the sense that “host vs. tumor” reactions are occurring. On culture condition, 28 out of 1×104 QRsP cells survived 10 Gy irradiation and made well-demarcated colonies. This figure is comparable to other sarcoma cell lines (Kranjc et al., 2005). The results of cDNA analyses are shown in Table 1a and b, in which the top 30 genes are listed in the order of the fold magnitude for QRsPIR-1. Highlighted genes by red or green suggest genes up- or down-regulated for QRsPIR-5 as well. Seven (23%) or 15 (50%) genes are commonly up- or down-regulated between QRsPIR-1 and QRsPIR-5. If we set 2.0-fold as a threshold, 132 or 238 genes were up- or down-regulated for QRsPIR-1. Among these genes, 11 (8.3%) or 63 (26.5%) genes were commonly up- or down-regulated respectively, between QRsPIR-1 and QRsPIR-5. The gene expression levels of MMP3, MMP13, LOX, and p57 were further evaluated by quantitative real-time RT-PCR. As shown in Fig. 1, MMP3, MMP13, and LOX were up-regulated (643.6-, 22.9-, and 20.8-fold for QRsPIR-1, 40.2-, 10.6-, and 6.9-fold for QRsPIR-5, respectively). p57 was down-regulated 0.8-fold for QRsPIR-1, 0.3-fold for QRsPIR-5,respectively. These data were quite consistent with our array results except for p16, for which real-time PCR gave us neither positive nor negative data for some unknown reasons. The fact that our commonly down-regulated genes outnumbered up-regulated ones is probably explained by DNA damage by irradiation, which, in nature, means the “destruction” of genes. There was an interesting finding that came out from the Monte Carlo simulation (Date and Shimozuma, 2001) suggesting that genes prone to irradiation are not uniformly separated on chromosomes but presumably are located within a particular “target volume” as described below. Such a target has a volume of 3.2 μm3 (a cube of 1.47 μm on each side), in which radicals (i.e. -OH, ·OH, etc.) make a cluster (Date, in preparation). An example of the assumption is shown in Fig. 2a and b, in which 7 out of 7 down-regulated genes on chromosome 1 are located in two separate “target volumes”, and 8 out of 9 down-regulated genes on chromosome 11 in two such volumes. It is interesting that cyclin-dependent kinase inhibitors (CDKIs), p16/INK4A and p57/Kip2, were particularly down-regulated in QRsPIR-1 cells (14.8- and 12.0-fold, respectively). CDKIs are keys to regulate cell cycle (Le et al., 2010). Indeed, cell doubling time measured under phase contrast microscopy at five min intervals was shorter for QRsPIR-1 (Fig. 3). p16 is also known to cause radiation-induced premature senescence (Rodier et al., 2009; Muthna et al., 2010). On the side of up-regulated genes, the following genes caught our attention: matrix proteinase (MMP) 13 and 3 (25.8- and 22.5-fold), lysl oxidase (LOX, 4.8-fold), and integrin beta 7 (3.4-fold). MMPs and LOX are highly associated with the invasive nature of malignant tumors (Zhang et al., 2010; Brekhman and Neufeld, 2009). Integrin beta 7 is also a key molecule of cancer cell adhesion and invasion (Kielosto et al., 2009). To verify the invasive activity of MMPs, a collagen zymography was performed and a band probably corresponding to MMP13 was detected (Fig. 4). These data encouraged us to proceed to in vivo study. Upon macroscopic examination on day 28 after tumor implantation, 2 mice out of 5 that were implanted with QRsP parental cells demonstrated a palpable tumor mass whereas all 5 mice implanted with QRsPIR-1 cells developed obvious tumors. The average size of the tumor mass derived from QRsPIR-1 cells was larger than that of those derived from QRsP parental cells (Fig. 5). Two out of five (40%) QRsPIR-1 tumor showed muscular invasion while no QRsP-derived tumors showed any invasion to the surrounding tissues. Under histopathological examination, QRsP parental and QRsSPIR-1-derived tumor tissue showed the same morphological features (Fig. 6). However, QRsPIR-1-derived tumor tissue showed frequent cell mitoses. The mean mitotic cell number was 4.0+/–3.9(SD) for QRsP, and 12.8+/–3.4 for QRsPIR-1 (Fig. 5; p<0.01, Student’s t-test). These in vivo results clearly demonstrated that QRsPIR-1 cells have an aggressive nature compared with the parental QRsP cells, and reflected the results of cDNA analyses. One of the limitations of the present study may be too much emphasis on cDNA analyses. A living creature, a collection of cells, is always in a dynamic state, therefore its mRNA expression can change as a function of a given environment. This implies the data presented here might have been different if sampled even a few minutes earlier or later. A typical example is cDNA data obtained from QRsPIR-1-derived tumor tissue. The array results was significantly different from the in vitro one (Table 2a and b). Several cell-cycle related genes were up-regulated: cyclin A2 (2.5-fold), cyclin B2 (2.2-fold), and PCNA (2.3-fold). Angiogenesis-related genes were also up-regulated: T-box 1 transcription factor (7.3-fold) (Kobayashi et al., 2002) and PDGF (2.5-fold) (Kumar et al., 2010). Interestingly, one of the cancer stem cell markers, CD34 (Pierce et al., 2008), was up-regulated (4.0-fold). One might think of normal tissue involvement for these up-regulated genes. The pathologists are fully experienced in handling tumor sampling, and we believe the array results came mostly from the tumor tissues.
 View Details | Fig. 1. Results of real-time RTPCR for p57, MMP3, MMP13, and LOX are shown. Note that the genes that are associated with invasiveness (MMP3, MMP13, and LOX) are strongly expressed. |
 View Details | Fig. 2. The loci of genes that are down-regulated for GRsPIR-1 on chromosome 1 and 11. There are two regions that could be within a “target volume” of ionizing irradiation. The target volume can be translated into a cube of 1.5 μm on each side (arrow). Numbers highlighted in green indicate fold changes. |
 View Details | Fig. 3. Average and standard deviations of cell cycle time are as follows: 831.8+/–180.7 min for QRsP and 738.3+/–72.7 min for QRsPIR-1 (p<0.05, Student’s t-test). |
 View Details | Fig. 4. Collagen zymography. A band corresponding to MMP13 is observed. |
 View Details | Fig. 5. QRsP or QRsPIR-1 cells (1×104 cells) were implanted. All animals implanted with QRsPIR-1 cells showed a palpable tumor mass, whereas 3 out of 5 animals implanted with QRsP parent cells had no signs of tumor. The average weights of the dissected tumor masses were plotted. The volumes of the QRsPIR-1-derived tumor masses were larger than the QRsP tumor masses, but the difference was not statistically significant. |
 View Details | Fig. 6. Histopathological features of the tumors are indicated. Both QRsP and QRsPIR-1 tumors showed the same histological features despite different mitotic cell number. A significant difference in the number of mitosis per high-power field was observed. |
In conclusion, irradiation is an effective tool for cancer treatment; however, if tumor cells survive ionizing irradiation, those cells may be more aggressive (i.e., rapid proliferation and higher invasiveness). The higher transplantation ratio might be indicative of the presence of cancer stem cell in the QRsPIR-1-derived tumor, though the up-regulation of CD34 alone is not enough to support the notion.
This study was supported in part by a Grant-in-Aid for Scienific Research (B20390319) provided by the Ministry of Education, Culture, Sports, Science and Technology of Japan. We wish to thank Dr. Futoshi Okada (Department of Biochemistry and Molecular Biology, Yamagata University Graduate School of Medical Science, Yamagata, Japan) for providing us with the QRsP cells used in this research and Mr. Neil Colley for his kind assistance in manuscript preparation.