MMP-9 from sublethally irradiated tumor promotes Lewis lung ... - Nature

Oncogene (2012) 31, 458–468

& 2012 Macmillan Publishers Limited All rights reserved 0950-9232/12 www.nature.com/onc

ORIGINAL ARTICLE

MMP-9 from sublethally irradiated tumor promotes Lewis lung carcinoma cell invasiveness and pulmonary metastasis CH Chou1,2, C-M Teng3, K-Y Tzen4,5, Y-C Chang6, J-H Chen7 and JC-H Cheng1,2,7,8 1 Graduate Institute of Oncology, National Taiwan University College of Medicine, Taipei, Taiwan; 2Department of Oncology, Division of Radiation Oncology, National Taiwan University Hospital, Taipei, Taiwan; 3Pharmacological Institute, National Taiwan University College of Medicine, Taipei, Taiwan; 4Department of Nuclear Medicine, National Taiwan University Hospital, Taipei, Taiwan; 5Molecular Imaging Center, National Taiwan University, Taipei, Taiwan; 6Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan; 7 Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University College of Electrical Engineering and Computer Science, Taipei, Taiwan and 8Graduate Institute of Clinical Medicine, National Taiwan University College of Medicine, Taipei, Taiwan

Matrix metalloproteinases (MMPs) associate with tumor progression and metastasis. We sought to investigate the role of MMP-9 from sublethally irradiated tumor in accelerated pulmonary metastasis of Lewis lung carcinoma (LLC-LM) and the corresponding anti-metastasis strategies in C57BL/6 mice. We used Matrigel-coated Boyden chamber assays and chicken chorioallantoic membrane assays to evaluate the invasion capability of irradiated LLC-LM cells (7.5 Gy), reverse transcription– polymerase chain reaction and the western blot assay to investigate the expression of MMPs by irradiated cells, and small interfering RNA duplexes to inhibit MMP-9 expression. LLC-LM cells differing in MMP-2 or -9 expression were subcutaneously injected into right thighs and the resulting tumors were irradiated (10 Gy 5) to induce pulmonary metastasis. Radiation significantly enhanced MMP-9 at both the transcriptional and translational levels. MMP-9 siRNA significantly inhibited in vitro radiation-enhanced invasiveness. The number of radiation-accelerated pulmonary metastases was significantly reduced by MMP-9 knockdown and MMP-2/9 knockdown. Reverse transcription–polymerase chain reaction of LLC-LM cells in the blood and lung tissue revealed MMP-9 involvement in radiation-enhanced intravasation. Either higher-dose irradiation (30 Gy 2) or pretreatment with prototypical MMP-9 inhibitor, zoledronic acid, significantly reduced the number of pulmonary metastases. The viability of irradiated tumor was seen on both positron emission tomography and magnetic resonance imaging, and tumor/serum MMP-9 levels suggested the association of local control of primary tumor and inhibition of timedependent MMP-9 activities. Our results demonstrate that MMP-9 is crucially involved in radiation-enhanced LLCLM cell invasiveness in vitro and in pulmonary metastasis from inadequately irradiated primary tumor in vivo. Oncogene (2012) 31, 458–468; doi:10.1038/onc.2011.240; published online 27 June 2011

Correspondence: Dr JC-H Cheng, Department of Oncology, Division of Radiation Oncology, National Taiwan University Hospital, No. 7 Chung-Shan South Road, Taipei 100, Taiwan. E-mail: [email protected] Received 5 February 2011; revised 9 May 2011; accepted 11 May 2011; published online 27 June 2011

Keywords: MMP-9; radiation; Lewis lung carcinoma; invasiveness; metastasis

Introduction Although local disease control has improved with advances in surgical and radiotherapeutic technologies, distant metastasis remains the greatest barrier to the cure of cancer. Several factors affect metastasis, including tumor factors, tumor–stromal interaction, inflammatory and immune reactions, and so on. The role of one of these, the matrix metalloproteinases (MMPs), in tumor progression and metastasis has been extensively investigated (Vihinen et al., 2005). MMPs are a family of zincdependent enzymes that degrade the major components of extracellular matrix (Martin et al., 1999) and affect tumor growth, invasion and metastasis (Bode and Maskos, 2003; Chakraborti et al., 2003; Rundhaug, 2005). MMPs have been correlated with cancer prognosis clinically and histopathologically (Leeman et al., 2003). One, in particular, MMP-9 (a 92-kDa type IV collagenase or gelatinase B) has a role in tumor invasion, angiogenesis and metastasis (Bernhard et al., 1994), and its level is related to angiostatin synthesis and systemic metastasis in animal models (Basile et al., 2004; Chen et al., 2005). A pulmonary metastasis model has been established in which primary Lewis lung carcinoma (LLC-LM) eradication by irradiation (five 10-Gy fractions) is followed by the development of pulmonary metastasis and decreased urinary levels of MMP-2, the enzyme responsible for angiostatin processing in this tumor model (Camphausen et al., 2001). On the other hand, MMP-9 is specifically induced in premetastatic lung endothelial cells and macrophages by distant primary tumors of LLC-LM, and significantly promotes lung metastasis (Hiratsuka et al., 2002). In vitro data suggested the role of radiationactivated MMP-9 expression in the enhanced invasiveness of sublethally irradiated liver cancer cells (Cheng et al., 2006). Besides, recent clinical data (Gnant et al., 2009) surprisingly demonstrate improvement in the disease-free survival of breast cancer patients receiving adjuvant

MMP-9 in LLC-LM invasiveness and metastasis CH Chou et al

459

zoledronic acid (ZA), a prototypical MMP-9 inhibitor (Giraudo et al., 2004). Moreover, the interaction of bone marrow-derived cells with MMP-9 in the tumor microenvironment is now considered a seminal event for tumor recurrence after irradiation (Ahn and Brown, 2008). In this study, the role of MMP-9 was investigated in both the in vitro LLC-LM cell invasiveness model using real-time– polymerase chain reaction (PCR), western blotting and gelatin zymography, and the in vivo pulmonary metastasis model using positron emission tomography (PET) and magnetic resonance imaging (MRI). Furthermore, high-dose irradiation (30 Gy 2) and ZA were used for MMP-9 blocking to study the role of MMP-9 in the in vivo model. Results Sublethal radiation enhances in vitro and in vivo LLC-LM cell invasiveness LLC-LM cells were seeded on the Matrigel-coated inserts of Boyden chambers and irradiated with different

doses (0, 2.5, 5, 7.5 and 10 Gy). After 24 h, the invading cells were fixed, stained and counted. Significantly more invading cells were found after irradiation, especially at 7.5 Gy (P ¼ 0.005; Figures 1a and b). This effect on LLC-LM cell invasiveness was significantly enhanced after irradiation with 2.5 Gy at 40 h (P ¼ 0.007), as well as with 7.5 Gy at 8 (P ¼ 0.00003), 24 (P ¼ 0.0006) and 40 h (P ¼ 0.0002) (Figure 1c), respectively, implying that activation of one or more invasiveness-related proteins is needed. In contrast, the enhancement of migration ability of the irradiated LLC-LM cells was not as significant as that of the invasion ability (data not shown). In subsequent experiments, 7.5 Gy was chosen because it caused the most significant enhancement of invasiveness without significantly affecting LLC-LM cell viability in Trypan blue exclusion assay (Figure 1d), but not in clonogenic assay (Figure 1e). Besides, 2.5 Gy moderately enhanced LLC-LM cell invasiveness with no viability inhibition in both assays. In contrast, 10 Gy decreased the number of viable cells, suggesting it may cause cell growth arrest. The chorioallantoic membrane (CAM) assay was used in

Figure 1 Radiation enhances LLC-LM cell invasiveness. (a) LLC-LM cells were seeded in Matrigel-coated inserts of Boyden chambers and irradiated with the indicated doses. After 24 h, the invading cells were viewed microscopically (high-power field, 200). (b) Invading cells were counted. *Po0.01, irradiation group vs sham group. (c) Invading cells were counted at 8, 24 and 40 h with sham irradiation, 2.5- or 7.5-Gy irradiation treatment. *Po0.01, irradiation group vs sham group. (d) LLC-LM cells (105 cells per dish) were seeded and irradiated with the indicated doses. The number of viable cells was determined 24 and 48 h later by Trypan blue exclusion assay. (e) LLC-LM cells (103 cells per dish) were seeded and irradiated with the indicated doses. The number of colonies with more than 50 cells was counted at 7 days for the clonogenic assay. (f) GFP-LLC-LM cells were irradiated with the indicated doses before being seeded on CAM. Total genomic DNA was isolated from the lower CAM. Data are the quantitative results of five independent experiments and the ratio of GFP/chGAPDH. *Po0.01, irradiation group vs sham group. Oncogene


460

conjunction with a semi-quantitative PCR assay to measure protease-dependent cancer cell intravasation. Green fluorescent protein (GFP)-LLC-LM cells were irradiated with different doses and then seeded on CAMs. After 48 h of incubation, total DNA was isolated from the lower CAM, and quantitative analyses of invading GFP-LLC-LM cells were conducted using GFP specific primers. The ratio of GFP/chGAPDH (chicken glyceraldehyde 3-phosphate dehydrogenase) was significantly higher in CAMs seeded with irradiated (7.5 Gy) cells than in CAMs seeded with shamirradiated cells (P ¼ 0.0001) (Figure 1f). Thus, a dose of 7.5 Gy or less enhanced both in vitro and in vivo tumor cell invasiveness. Radiation enhances MMP-9-dependent LLC-LM cell invasiveness Expression of mmp-9 genes in LLC-LM cells were tested by reverse transcription–PCR 6 h after irradiation (Figure 2a). Radiation, especially 7.5 Gy (P ¼ 0.000006), prompted increased expression of mRNA for mmp-9, but not mmp-2 (Figure 2b), increased MMP-9 protein expression (western blotting; Figure 2c) in cell lysates and increased MMP-9 activity (gelatin zymography; Figure 2d) in the supernatants. Furthermore, the increase in total MMP-9 concentration was significant in the supernatant of irradiated (7.5 Gy) cells after 12 h (P ¼ 0.000002) (Figure 2e). These results demonstrated that radiation enhances transcriptional and translational MMP-9 expression in LLC-LM cells. To confirm the effect of radiation-induced MMP-2 and MMP-9 on LLC-LM cell invasiveness, small interfering RNA duplexes were used to knock down

(KO) mmp-2 and mmp-9 mRNA. Four cell lines (wildtype LLC-LM (parental), MMP-2KO, MMP-9KO and MMP-2/9KO) were compared and found to have different expression patterns of mmp-2 and/or mmp-9 mRNA and protein in culture supernatant (Figures 3a and b). Furthermore, KO of MMP-9 (MMP-9KO) reduced the number of invading cells (Figure 3c), whereas MMP-2 and/or MMP-2 depletion had no effect on percentage viability or radiosensitivity over a 3-day post-treatment period (Figure 3d). Radiation enhances MMP-9 dependent LLC-LM cell pulmonary metastasis in vivo C57BL/6 mice were injected with the four cell lines, with all the lines showing similar tumorigenicity and sensitivity to radiotherapy at the primary site (data not shown). Interestingly, radiation-accelerated lung metastasis was suppressed in mice injected with the MMP9KO and MMP-2/9KO cell lines (Figure 4a), with these metastases confirmed pathologically (Figure 4b). Both the lung weights (Figure 4c) and the number of lung metastases (Figure 4d) were significantly higher in mice injected with wild-type (P ¼ 0.008 and P ¼ 0.0002) and MMP-2KO (P ¼ 0.006 and P ¼ 0.00004) cells. Taken together, the results demonstrate that MMP-9 expression has a critical role in the in vivo metastasis of irradiated LLC-LM cells. MMP-9 has a critical role in tumor cell invasiveness We stably transfected the four LLC-LM cell lines to express GFP, and used them to verify the invasiveness of primary LLC-LM tumors in C57BL/6 mice. Invasiveness was assessed by measuring GFP mRNA in the

Figure 2 Radiation enhances LLC-LM cell invasiveness through increased MMP-9 protein expression. (a) LLC-LM cells were irradiated with the indicated doses or not (sham). After 6 h, total RNA was isolated and reverse-transcribed to cDNA for mmp-2 and mmp-9 mRNA detection by PCR, with actin as a loading control. (b) mmp-9 mRNA expression at the indicated doses as compared with the sham group. (c) Expression of MMP-9 in total cell lysate protein detected by western blot with specific antibodies, with a-tubulin as a loading control. (d) After 12 h, MMP-9 as well as MMP-2 activities were measured in culture supernatant using gelatin zymography. (e) Total MMP-9 protein in culture supernatant. *Po0.01, irradiation group vs the sham group. Oncogene


461

Figure 3 Radiation enhances MMP-9-dependent LLC-LM cell invasiveness. (a) The expression patterns of mmp-2 and mmp-9 mRNA in wild-type, MMP-2KO, MMP-9KO and MMP-2/9KO LLC-LM cell lines were evaluated by reverse transcription–PCR, with actin as a loading control. (b) The expression patterns of MMP-2 and MMP-9 protein level in cell culture supernatant of each line. *Po0.01. (c) The cells of each line were irradiated with 7.5 Gy or sham-irradiated. After 24 h, the invaded cells were fixed, stained and counted. The average number from five different high-power fields on each filter is presented. *Po0.01. (d) The cells (2 104 cells per dish) of each line were irradiated with 7.5 Gy. After 72 h, the number of viable cells was determined by Trypan blue exclusion assay. The percent viability is expressed relative to the wild-type (sham-irradiated) control.

blood (Figures 5a and c) and lung (Figures 5b and d) using real-time quantitative reverse transcription–PCR. With radiotherapy to the thigh tumor (10 Gy per day for 5 consecutive days), significant radiation-enhanced invasiveness of wild-type and MMP-2KO cells (but not MMP-9KO and MMP-2/9KO cells) was demonstrated in whole blood (P ¼ 0.00001 and 0.00002 on day 9, and P ¼ 0.003 and 0.001 on day 11) and lung tissue (P ¼ 0.0002 and 0.000004 on day 11), indicating that MMP-9 but not MMP-2 impacts the process of intravasation. Pretreatment but not post-treatment with ZA inhibits radiation-enhanced MMP-9 activity in serum and inhibits radiation-enhanced LLC-LM lung metastasis ZA, an anti-MMP-9 agent (Giraudo et al., 2004), was used to clarify the potential of MMP-9 as treatment. The mice were randomized into six groups as described in the legend to Figure 6. Without radiation, ZA itself had no apparent effect on the primary tumor. The limb tumors were smaller in the irradiated groups (RT, preZA þ RT or RT þ post-ZA) than in the non-irradiated groups (volume around 2000 mm3 at the time of being killed). RT (30.3±3.2) in contrast to sham RT (5.0±2.1) induced a greater number (P ¼ 0.0003) of pulmonary metastases. Pretreatment with ZA 25 mg/kg per day (8.0±4.0, P ¼ 0.002) or 100 mg/kg per day (2.3±2.1, P ¼ 0.0001), but not post-treatment with ZA 25 mg/kg per day (24.6±3.2, P ¼ 0.10) or 100 mg/kg per day (21.0±7.0, P ¼ 0.10) prevented the development of

radiation-induced pulmonary metastasis, especially in the early phase of radiation-accelerated metastasis from the primary tumor. The number of pulmonary metastases was similar in all three non-irradiated groups (Figure 6a). Metastases on the lung surface are shown in Figure 6b. In vivo, MMP-9 activity (Figure 6c) and levels of total MMP-9 (Figure 6d) in serum samples increased within a week after completion of radiotherapy in the RT and RT þ post-ZA groups as compared with other groups, which all had similar serum MMP-9 activity and level. High-dose irradiation reduces radiation-enhanced MMP-9 activity in the serum and inhibits acceleration of LLC-LM lung metastasis by sublethal radiation Primary tumor response, serum MMP-9 level and pulmonary metastasis were evaluated in mice treated with two 30-Gy fractions and mice treated with five 10Gy fractions each to the thigh tumor. The number of pulmonary metastases was greater after the 10-Gy regimen (23±5.9 vs 0±0, P ¼ 0.0002). On the second (day 9) and fourth days (day 11) of irradiation, thigh tumors were imaged by PET/computed tomography with [18F]-2-fluoro-2-deoxy-D-glucose, the uptake of which is an indicator of tumor metabolism. As shown in Figure 7a, the tumor viability was more vulnerable to 30-Gy radiotherapy. Thigh tumors were also imaged by weekly dynamic contrast-enhanced-MRI (Figure 7b). Tumor growth suppression was maintained (tumor continued to shrink) after 30 Gy for 2 fractions, but Oncogene


462

Figure 4 Radiotherapy (RT)-induced acceleration of pulmonary metastasis was dependent on MMP-9. (a) Cells (1 106) from each line were injected subcutaneously into the right hind limb of C57BL/6 mice. On day 8, RT (10 Gy per day for 5 consecutive days) was given (RT) or not (sham) to the thigh tumors. On day 28, all mice were killed and their lungs were removed. Two representative sets of lungs from each group are shown. (b) Histological examination of pulmonary metastases from LLC-LM cells. (c) The lung weights and (d) the number of pulmonary surface metastases were compared as indicated. *Po0.01.

Figure 5 MMP-9 but not MMP-2 has a critical role in the process of intravasation in vivo. Cells of the four GFP-LLC-LM cell lines (1 106) were injected subcutaneously into the right hind limb of C57BL/6 mice. Mice were killed on the day before radiation (RT) (day 7), on the second day of radiation (day 9) and on the fourth day of radiation (day 11). (a, c) Whole blood was collected from the heart, and (b, d) the lungs were dissected. The intravasation of LLC-LM cells in the blood or lung tissue was traced by determining GFP mRNA expression using fluorescence real-time PCR. N ¼ 3 at each time point in each group. *Po0.01.

was transient (began to re-grow at 2 weeks) after 10 Gy for five fractions. Moreover, two 30-Gy fractions were associated with less radiation-induced increase in serum Oncogene

MMP-9 level (Figure 8a). In contrast to five 10-Gy fractions, two 30-Gy fractions significantly reduced the mmp-9 mRNA level in the primary tumor (P ¼ 0.002


463

Figure 6 MMP-9 inhibitor, ZA, inhibits radiotherapy (RT)-induced acceleration of pulmonary metastasis. Injection of LLC-LM cells (1 106) subcutaneously in the right hind limb of C57BL/6 mice resulted in primary tumors on day 8. Mice were randomized into sham, RT, pre-ZA, pre-ZA þ RT, post-ZA and RT þ post-ZA groups. In the RT, pre-ZA þ RT and RT þ post-ZA groups, the tumors were irradiated (10 Gy daily on 5 consecutive days). ZA (25 or 100 mg/kg per day) was given intraperitoneally from day 5 (pre-ZA þ RT and pre-ZA groups) for 24 days or from day 15 (post-ZA þ RT and post-ZA groups) for 14 days. (a) The number of pulmonary surface metastases was compared between the indicated groups. *Po0.01. (b) A photograph showing two representative sets of lungs with surface metastasis from each treatment group. (c) Gelatin zymography was used to determine total MMP-9 activity in the serum. The percentage of MMP-9 gelatinase activity at each time point is presented relative to day 1. (d) Serum level of total MMP-9 was determined by enzyme immunoassay. The dosage of ZA was 100 mg/kg per day in (b–d).

on day 11) (Figure 8b) and significantly inhibited entry of GFP-LLC-LM cells into the blood (P ¼ 0.000004 on day 9 and P ¼ 0.000003 on day 11) (Figure 8c) and lung tissue (P ¼ 0.000002 on day 11) (Figure 8d).

Discussion A model for radiation-induced pulmonary metastasis using LLC-LM cells was established by Camphausen et al. (2001). This model resembles the model of pulmonary metastasis after surgical removal of the primary tumor (O’Reilly et al., 1994). They proposed that irradiation of the primary tumor leads to a net imbalance of proangiogenic over antiangiogenic factors (including angiostatin). In their experiments, urinary MMP activity, a surrogate marker for angiostatin production, increased with primary tumor progression, but decreased with radiation-induced local control. However, our pretreatment with ZA, an MMP-9 inhibitor, resulted in a similar pattern of decreased serum MMP-9 level and activity after radiotherapy-induced primary tumor regression, but surprisingly it suppressed the development of pulmonary metastasis. Instead, radiation followed by ZA treatment failed to prevent pulmonary metastasis with only the moderate decrease in MMP-9 activity. Notably, serum MMP-9 activity declined in the

RT and RT þ post-ZA groups, but remained much higher than that in the pre-ZA þ RT group after completion of radiotherapy. The time dependence of the ZA effect implies that both the metastatic process and the primary tumor’s response to sublethal radiation occur simultaneously. This metastatic process begins soon after irradiation starts and may become irreversible later. Besides, the imaging findings demonstrated the poor local control of the primary tumor by radiation. With high-dose radiotherapy and better local control of the primary tumor, the tumor/serum MMP-9 levels and pulmonary metastasis were ameliorated. All these data imply that pulmonary metastasis in this model might be at least partly attributed to inadequately irradiated primary tumor and the associated MMP-9 surge following radiotherapy. Our findings raise the possibility that radiationinduced metastasis involves more than changes in the systemic balance of angiostatin. This possibility has been suggested by other researchers. Pozzi et al. (2002) found that low plasma levels of MMP-9 do not prevent increased tumor angiogenesis. They emphasized that excessive MMP-9 activity generates inhibitors of endothelial cell proliferation, including but not necessarily limited to angiostatin, resulting ultimately in auto-inhibition of angiogenesis (Pozzi et al., 2002). In a model of LLC-LM growth in animals depleted of macrophages, Dong et al. (1997) found that despite lipopolysacchaOncogene


464

Figure 7 Higher-dose radiotherapy (RT) better controls the primary tumor. LLC-LM cells (1 106) injected subcutaneously in the right hind limb of C57BL/6 mice resulted in primary tumors on day 8. Mice were randomized into sham, sublethal dose (5 10 Gy from days 8 to 12) RT and high dose (2 30 Gy on days 8 and 10) RT groups. (a) On days 9 and 11, mice were scanned by PET/CT using [18F]-2-fluoro-2-deoxy-D-glucose to determine primary tumor viability. (b) On days 7, 10, 17 and 24, primary tumor size and other parameters were assessed by dynamic contrast-enhanced magnetic resonance imaging. Representative images show tumor viability expressed as maximum standard uptake value (max SUV) and tumor volume expressed in cm3.

ride-induced MMP-9 activity, the level of angiostatin was decreased. They concluded that metalloelastase but not MMP-9 was responsible for the generation of angiostatin by macrophages (Dong et al., 1997). Surgical resection of legs bearing PC-3 tumor was found Oncogene

to accelerate distal angiogenesis; however, radiotherapy to the primary tumor retarded the angiogenesis (Hartford et al., 2000). Variation in tumor types, different latency periods for tumors occurring after different radiotherapy fractionation regimens and the complex rela-


465

Figure 8 Higher-dose radiotherapy (RT) reduces primary tumor and serum MMP-9 levels as well as eliminates disseminated GFP-LLC-LM cells in blood and lungs. Cells of GFP-LLC-LM lines (1 106) were injected subcutaneously into the right hind limb of C57BL/6 mice. Mice were killed on the second (day 9) and fourth days (day 11) of RT. (a) Serum MMP-9 level was determined by enzyme immunoassay. (b) The irradiated thigh tumors were dissected on day 11 and assayed for mmp-9 mRNA by real-time-reverse transcription–PCR. Whole blood was collected from the heart, and the lungs were dissected on days 9 and 11. The GFP-LLC-LM cells in (c) the blood and (d) lung tissue were traced by determining the GFP mRNA expression using fluorescence real-time PCR. The data are mean ratio of GFP to mouse GAPDH. N ¼ 5 at each time point in each group. *Po0.01.

tionship between MMP-9 and angiogenesis may all confound interpretation of the relationship between the radiation response of primary tumor and distant cancer metastasis. In the current model, processes other than angiostatin regulation may be activated by increased MMP-9 activity owing to sublethal radiation, and lead to pulmonary metastasis. In this study, 7.5 Gy increased the in vitro invasiveness and MMP-9 expression of LLC-LM. mmp-9 gene KO by small interfering RNA duplexes suppressed LLC-LM invasiveness in vitro and reduced radiation-accelerated pulmonary metastasis in vivo. These data support the role of radiation-activated MMP-9 in the invasion and metastasis of LLC-LM cells. Zhang et al. (2006) similarly found that secretion of active and pro-MMP9 was higher from the high-metastatic sublines of LLC that they established than from the parent lines (Zhang et al., 2006). MMP-9 is recruited to the glycolipidenriched microdomain/rafts, facilitating its secretion and activation, and resulting in increased invasion and metastatic potential. Osinsky et al. (2005) found the correlation between the numbers of LLC metastasis and the activities of MMP-2/MMP-9 in primary tumor. MMP-9 has been proposed to have an important role in the induction of angiogenesis and in the integration of bone marrow-derived endothelial cells in the tumor vasculature. Upregulation of MMP-9 and vascular endothelial growth factor has been shown to increase LLC cell invasiveness in vitro. Hiratsuka et al. (2002) demonstrated that for lung-specific metastasis, the distant primary tumor must induce vascular endothelial growth factor receptor-1-dependent MMP-9 in pulmo-

nary endothelial cells (Hiratsuka et al., 2002). In their study, MMP-9 expression was evident in various primary tumors and lung tissues, but the link between MMP-9 in the primary tumor and pulmonary metastasis remained undefined. Similarly in the study by Huang et al. (2002), host-derived stromal MMP-9 expression mainly in tumor-infiltrating macrophages contributed to the angiogenesis and progression of mouse ovarian tumors. Both their ovarian cancer cell lines expressed MMP-9, but no difference in the tumor MMP-9 level was seen between MMP-9-competent and -deficient nude mice. Given the absence of comparison to MMP9-deficient cancer cells, the role of tumor-derived MMP9 in the metastatic process has not been determined. Other studies using the models of cervical cancer (Giraudo et al., 2004), breast cancer (Owen et al., 2003), skin cancer (Coussens et al., 2000) or neuroblastoma (Jodele et al., 2005) have emphasized the crucial role of MMP-9 from bone marrow-derived cells, rather than MMP-9 from cancer cells, in initiating the angiogenic switch. In this study, we focused on the crucial role of MMP-9 derived from LLC-LM cells, especially after sublethal irradiation, in initiating the metastatic cascade in lung cancer. The metastases were suppressed either by genetic KO of MMP-9, the MMP-9 inhibitor (ZA) or higher-dose irradiation. All these interventions depleted or reduced the MMP-9 in primary LLC-LM tumor, and ultimately prevented pulmonary metastasis. The quantitative detection of systemic LLC-LM cells in mouse blood and lungs shortly following irradiation of the primary tumors also coincided with the serum MMP-9 increase, likely supporting the systemic spread Oncogene


466

of LLC-LM cells through the MMP-9 activation of the unsatisfactorily controlled primary tumor. Notably, sublethal (5 10-Gy) irradiation was deemed effective in control of the primary tumor control by gross inspection (Camphausen et al., 2001). Using modern imaging tools, including PET and MRI, we found that this scheme of primary tumor irradiation and the associated MMP-9 surge were unsatisfactory, whereas high-dose (2 30-Gy) irradiation better suppressed tumor viability, reduced the MMP-9 increase and inhibited metastasis. MMP-9 has been suggested as a requirement for tumor vasculogenesis and re-growth after irradiation, through the recruitment of CD11b þ myelomonocytic cells (Ahn and Brown, 2008). The investigators also noticed that the tumor recurs after irradiation even in the presence of low levels of MMP-9 provided by other cells, and suggested the monitoring of sources and levels of MMP-9 before the inhibitory strategy (Ahn and Brown, 2009). Therefore, high-dose lethal irradiation should be an effective way to prevent the MMP-9 secretion from the persistently viable tumor and prevent the subsequent metastatic cascade. There are two limitations of this study. First, only one cell line (LLC-LM) was tested in this study. LLC-LM cells and the generated tumor were relatively resistant to irradiation and might not represent all the post-irradiation processes for other cancer cells. Second, the in vitro single dose of 7.5 Gy is not totally reflecting most of the clinical fractionated radiotherapy using 2–3 Gy per fraction. However, the in vivo difference between the doses of 10 Gy 5 and 30 Gy 2 may support the clinical use of high-dose/intensity, shortcourse radiosurgery to ablate the resistant tumor and prevent unwanted secondary effect from sublethally irradiated tumor refractory to long-course fractionated radiotherapy. Such an intensified radiotherapeutic mode has been increasingly used with promising results. Materials and methods Cell culture LLC-LM cells were grown at 37 1C in a humidified atmosphere of 5% CO2/95% air in Dulbecco’s modied Eagle’s medium containing 10% heat-inactivated fetal bovine serum plus penicillin–streptomycin under sterile tissue culture conditions. Irradiation and treatment The flask received different doses of g-radiation (0–10 Gy) from a cobalt-60 unit. The distance from the source to the bottom of the flask was 80 cm. The dose rate was 1 Gy/min. ZA was provided by Novartis (Basel, Switzerland). Boyden chamber invasion assay Invasion chambers were prepared by coating the membranes (8-mm pore) of 24-well inserts with 50 ml (10 mg/ml) of Matrigel (Becton-Dickinson, Bedford, MA, USA). A total of 105 cells were added to the upper chamber. After cell attachment, the medium was changed to serum-free medium and the cells were Oncogene

irradiated. After the indicated incubation time, the cells on the upper surface were removed using a cotton bud. The remaining invading cells were fixed, stained with 0.1% crystal violet for 1 h at room temperature and counted at 200 magnification in five different high-power fields per filter. Experiments were repeated three times. Intravasation and invasion experiments For intravasation and invasion experiments, LLC-LM cells were stably transfected with the pEGFP-1 vector (Clontech, Palo Alto, CA, USA) by Lipofectamine (Invitrogen, Carlsbad, CA, USA) to generate a GFP-LLC-LM cell line. GFP-LLCLM cells (106 cells) were irradiated (7.5 Gy) 1 h before detachment from the culture dish using 0.25% Tris-ethylenediaminetetraacetic acid, inoculated onto the CAM of a 9-dayold chicken embryo, in which an artificial air sac was created (upper CAM). Following incubation for 48 h, the lower half of the CAM (lower CAM) was removed for genomic DNA isolation (DNA Isolation Kit; Qiagen Systems, Hilden, Germany). The quantitative analysis method was as described previously (Cheng et al., 2006). Cell growth curve determination Cells (105 cells per dish) were seeded in 6-cm plates, irradiated with different doses and maintained at 37 1C in a humidified 10% CO2 atmosphere. After 24, 48 and 72 h, the number of viable cells was determined by exclusion of Trypan blue (0.12% (wt/vol)). MMP-9 and MMP-2 enzyme-linked immunoabsorbent assays The pro-MMP-9 and total MMP-2 were measured using the mouse MMP-9 (Total) ELISpot kit and MMP-2 (Total) Quantikine ELISA kit (R&D Systems, Minneapolis, MN, USA). Reverse transcription–polymerase chain reaction Reverse transcription of RNA isolated from cells was performed in a final reaction volume of 20 ml containing 5 mg of total RNA in Moloney murine leukemia virus reverse transcriptase buffer (Promega, Madison, WI, USA). Specific gene cDNA was cloned and amplified by PCR with the following primers: b-actin (sense 50 -CTCCTATGTGGGTGACGAGG-30 and antisense 50 -CTTTTCACGGTTGGCCTT-30 amplified a 202-bp fragment), mouse mmp-2 (sense 50 -ACAAGTGGTCCGCGT AAAGT-30 and antisense 50 -CGGTCATCATCGTAGTTG GTT-30 amplified a 501-bp fragment) and mouse mmp-9 (sense 50 -AACCCTGTGTGTTCCCGTT-30 and antisense 50 -GGAT GCCGTCTATGTCGTCT-30 amplified a 486-bp fragment). The reaction products were analyzed as described above. Western blot analysis Total protein was extracted using Mammalian Protein Extraction Reagent (Pierce, Rockford, IL, USA). The isolated protein was separated by electrophoresis on a 10% SDS–Tris glycine polyacrylamide gel, transferred to a polyvinylidene difluoride membrane and immunoblotted with antibodies (2 mg/ml) against mouse MMP-9 (sc6841) or a-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Bound antibodies were detected using appropriate peroxidase-coupled secondary antibodies and an enhanced chemiluminescence detection system (Boehringer Mannheim, Mannheim, Germany). MMP-2 and MMP-9 gene KO by small interfering RNA duplexes To generated MMP-2KO, MMP-9KO and MMP-2/9KO lines, the short hairpin RNA plasmids, TRC-2 mouse


467 short hairpin RNA (clone TRCN0000031224) and CSHL Mm short hairpin RNA mir 7.10 (clone V2MM-50831) against mouse mmp-2 mRNA and mmp-9 mRNA (GenDiscovery Biotechnology, Taipei, Taiwan), were transfected into LLCLM cells. The stable transfection clones were selected using puromycin (1 mg/ml) and evaluated for MMP-2 and MMP-9 expression in each cell line. Furthermore, an MMP-2/9 double KO line was established by transfecting the MMP-9 KO line with TRC-2 mouse short hairpin RNA plasmid. Gelatin zymography The supernatant of LLC-LM cells (5 ml) or mouse serum (5 ml) was analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 9% polyacrylamide gels containing 1 mg/ml gelatin. The gels were washed for 30 min at room temperature in 2.5% Triton X-100, washed several times in ddH2O, incubated in 50 mM Tris (pH 7.6), 1 mM ZnCl2, 0.15 M NaCl and 10 mM CaCl2 for 18 h at 37 1C, and then stained with 0.2% Coomassie blue R250. Bands of lysis representing gelatinase activity were then visualized against a dark background. Gelatinase activity level at each point was calculated relative to that of the basal level. Animal model and tumor irradiation Male, 5- to 6-week-old, C57BL/6 mice (National Taiwan University Animal Center, Taipei, Taiwan) were used. For each experiment (which was repeated three times), 1 106 cells from one of several different LLC-LM cell lines were injected subcutaneously into the right hind limb. At 8 days after implantation, mice were immobilized in a customized harness that left the right hind leg exposed. The remainder of the body was shielded by five times the half-value thickness of lead. A cobalt-60 unit was used to irradiate the primary tumor with 50 Gy (five 10-Gy daily fractions, at the dose rate of 1 Gy/min on days 8–12) or 60 Gy (two 30-Gy fractions on days 8 and 10). On day 28, we killed the mice, removed and weighed both lungs and counted the number of surface metastases. All the animal care and handling procedures and experimental protocols were approved by the Committee of Experimental Animal Management at College of Medicine, National Taiwan University. Determination of mmp-9 mRNA in primary tumor by quantitative real-time RT–PCR Total RNA was isolated from the primary tumor. The RT of RNA, amplification, detection of DNA, data acquisition, primer design and quantitative real-time PCR analysis were all conducted as described previously (Thwin et al., 2009). PCR primers (forward/reverse) for mmp-9 and GAPDH were as follows: 50 -CTCGAACTTTGACAGCGACA-30 /50 -CCCT CAGTGAAGCGGTACAT-30 ; and 50 -CAAGGTCATCCAC GACCACT-30 /50 -CCAGTGAGTTTCCCGTTCAG-30 . GAPDH expression was used as an internal calibrator for equal RNA loading and to normalize relative expression data for the mmp-9 gene. The real-time PCR data were quantified using relative quantification (the 2DDCT method). Detection of GFP-LLC-LM cells in whole blood and lung tissue To trace the intravasation of LLC-LM cells, the stable clones of MMP-2KO, MMP-9KO and MMP-2/9KO cells were further transfected with pEGFP-1 vector, selected for stable GFP expression and cloned as described above. Different LLC-LM-GFP cell lines were used in animal experiments. Mice were killed on the indicated days. Whole blood and lungs were collected for

GFP mRNA detection. Total RNA was isolated for reverse transcription. The GFP and GAPDH cDNA were analyzed using a fluorescein quantitative real-time PCR detection system (LightCycler DNA Master SYBR Green I, Roche Molecular Biochemicals, Indianapolis, IN, USA). The primer pairs were as follows: for GFP, 50 -CTGCTCTCTTGGGT CCACTGG-30 and 50 -CACCGCCTTGGCTTGTCACAT-30 ; and for mouse GAPDH, 50 -TGTCAAGCTCATTTCCTGGT ATGA-30 and 50 -CTTACTCCTTGGAGGCCATGTAG-30 . Amplification was followed by melting curve analysis to verify the correctness of the amplicon. A negative control without cDNA was run with every PCR to assess the specificity of the reaction. Analysis of data was performed using LightCycler software (Roche Diagnostics Ltd, Burgess Hill, UK). PCR efficiency was determined by analyzing a dilution series of cDNA (external standard curve). The amount of GFP mRNA was normalized to the amount of GAPDH mRNA and expressed in arbitrary units. Small animal PET/computed tomography After the mice fasted overnight with free access to water, PET/ computed tomography scans with [18F]-2-fluoro-2-deoxyD-glucose (in saline; 14 MBq; 378 mCi; injected intravenously via tail vein for 60 min in one-bed position) were performed on days 9 and 11 of the experiment to measure viability and metabolic activity of the tumors. The mice were anesthetized with isofluorane gas (about 2% (v/v)) (Baxter Healthcare, Deerfield, IL, USA) and placed in a prone position during the scan. Body-mode tumor scans were taken with a dedicated, eXplore Vista DR (GE Healthcare, Waukesha, WI, USA), small animal PET/computed tomography scanner 40 min after the injection of [18F]-2-fluoro-2-deoxy-D-glucose. The images were processed and reconstructed using a vendor-supplied two-dimensional filtered back projection. Small animal MRI A superconductive 7 T MRI scanner (Biospec, Bruker Corp., Ettlingen, Germany) dedicated to animal imaging was used. The magnetic resonance parameters included modified PVM FLASH, TR 100 ms, TE 3.661 ms, flip angle 301, NEX 1, 9 slices, FOV 4 2 cm2, slice thickness 1 mm, matrix size 256 128, resolution 156.25 mm and scan time 12.8 s per acquisition. There were a total of 60 acquisitions and contrast medium was injected beginning on the fifth scan (that is, 64 s after the first acquisition). Dynamic contrast-enhanced-MRI images were analyzed using Tofts model (Tofts and Kermode, 1991; Tofts et al., 1999). After the cine perfusion images from dynamic contrast-enhanced-MRI were registered, contours were manually drawn around the tumor. The magnetic resonance constants included Ktrans (the volume transfer constant), kep (rate constant of backflux from extravascular extracellular space to plasma), ve (total volume of extravascular extracellular space per unit of tissue) and vp (total blood plasma volume). The tumor volume was obtained by summation of the data from all slices containing the tumor. Statistics Data are presented as the mean±s.d. for the indicated number of separate experiments. Differences between pairs of treatment groups were tested using the Student’s t-test. P-value of less than 0.05 was considered significant. Conflict of interest The authors declare no conflict of interest. Oncogene


468 Acknowledgements This work was supported by the National Science Council, Execute Yuan, Taiwan, ROC (Grant numbers NSC97-2314B-002-113-MY3, NSC98-2627-B-002-017 and NSC99-2627B-002-008), and the National Taiwan University Hospital

(Grant number NTUH 98S1128). We appreciate the contribution of the Molecular Imaging Center, National Taiwan University, in providing the technical support of the dedicated small animal PET/CT scanner for imaging. We also thank Mr Samuel C Leu for his help with the PET/CT scans and image analysis.

References Ahn GO, Brown JM. (2008). Matrix metalloproteinase-9 is required for tumor vasculogenesis but not for angiogenesis: role of bone marrow-derived myelomonocytic cells. Cancer Cell 13: 193–205. Ahn GO, Brown JM. (2009). Influence of bone marrow-derived hematopoietic cells on the tumor response to radiotherapy. Cell Cycle 8: 970–976. Basile DP, Fredrich K, Weihrauch D, Hattan N, Chilian WM. (2004). Angiostatin and matrix metalloprotease expression following ischemic acute renal failure. Am J Physiol Renal Physiol 286: F893–F902. Bernhard EJ, Gruber SB, Muschel RJ. (1994). Direct evidence linking expression of matrix metalloproteinase 9 (92-kDa gelatinase/ collagenase) to the metastatic phenotype in transformed rat embryo cells. Proc Natl Acad Sci USA 91: 4293–4297. Bode W, Maskos K. (2003). Structural basis of the matrix metalloproteinases and their physiological inhibitors, the tissue inhibitors of metalloproteinases. Biol Chem 384: 863–872. Camphausen K, Moses MA, Beecken WD, Khan MK, Folkman J, O’Reilly MS. (2001). Radiation therapy to a primary tumor accelerates metastatic growth in mice. Cancer Res 61: 2207–2211. Chakraborti S, Mandal M, Das S, Mandal A, Chakraborti T. (2003). Regulation of matrix metalloproteinases: an overview. Mol Cell Biochem 253: 269–285. Chen X, Su Y, Fingleton B, Acuff H, Matrisian LM, Zent R et al. (2005). Increased plasma MMP9 in integrin a1-null mice enhances lung metastasis of colon carcinoma cells. Int J Cancer 116: 52–61. Cheng JC, Chou CH, Kuo ML, Hsieh CY. (2006). Radiationenhanced hepatocellular carcinoma cell invasion with MMP-9 expression through PI3K/Akt/NF-kappaB signal transduction pathway. Oncogene 25: 7009–7018. Coussens LM, Tinkle CL, Hanahan D, Werb Z. (2000). MMP-9 supplied by bone marrow-derived cells contributes to skin carcinogenesis. Cell 103: 481–490. Dong Z, Kumar R, Yang X, Fidler IJ. (1997). Macrophage-derived metalloelastase is responsible for the generation of angiostatin in Lewis lung carcinoma. Cell 88: 801–810. Giraudo E, Inoue M, Hanahan D. (2004). An amino-bisphosphonate targets MMP-9-expressing macrophages and angiogenesis to impair cervical carcinogenesis. J Clin Invest 114: 623–633. Gnant M, Mlineritsch B, Schippinger W, Luschin-Ebengreuth G, Po¨stlberger S, Menzel C et al. (2009). Endocrine therapy plus zoledronic acid in premenopausal breast cancer. N Engl J Med 360: 679–691. Hartford AC, Gohongi T, Fukumura D, Jain RK. (2000). Irradiation of a primary tumor, unlike surgical removal, enhances angiogenesis suppression at a distal site: potential role of host–tumor interaction. Cancer Res 60: 2128–2131. Hiratsuka S, Nakamura K, Iwai S, Murakami M, Itoh T, Kijima H et al. (2002). MMP9 induction by vascular endothelial growth factor receptor-1 is involved in lung-specific metastasis. Cancer Cell 2: 289–300.

Oncogene

Huang S, Van Arsdall M, Tedjarati S, McCarty M, Wu W, Langley R et al. (2002). Contributions of stromal metalloproteinase-9 to angiogenesis and growth of human ovarian carcinoma in mice. J Natl Cancer Inst 94: 1134–1142. Jodele S, Chantrain CF, Blavier L, Lutzko C, Crooks GM, Shimada H et al. (2005). The contribution of bone marrow-derived cells to the tumor vasculature in neuroblastoma is matrix metalloproteinase-9 dependent. Cancer Res 65: 3200–3208. Leeman MF, Curran S, Murray GI. (2003). New insights into the roles of matrix metalloproteinases in colorectal cancer development and progression. J Pathol 201: 528–534. Martin FM, Beckett RP, Bellamy CL, Courtney PF, Davies SJ, Drummond AH et al. (1999). The synthesis and biological evaluation of non-peptidic matrix metalloproteinase inhibitors. Bioorg Med Chem Lett 9: 2887–2992. O’Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M et al. (1994). Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by Lewis lung carcinoma. Cell 79: 315–328. Osinsky SP, Ganusevich II, Bubnorskaya LN, Valkovskaya NV, Kovelskaya AV, Sergienko TK et al. (2005). Hypoxia level and matrix metalloproteinases-2 and -9 activity in Lewis lung carcinoma: correlation with metastasis. Exp Oncol 27: 202–205. Owen JL, Iragavarapu-Charyulu V, Gunja-Smith Z, Herbert LM, Grosso JF, Lopez DM. (2003). Up-regulation of matrix metalloproteinase-9 in T lymphocytes of mammary tumor bearers: role of vascular endothelial growth factor. J Immunol 171: 4340–4351. Pozzi A, LeVine WF, Gardner HA. (2002). Low plasma levels of matrix metalloproteinase 9 permit increased tumor angiogenesis. Oncogene 21: 272–281. Rundhaug JE. (2005). Matrix metalloproteinases and angiogenesis. J Cell Mol Med 9: 267–285. Thwin MM, Douni E, Arjunan P, Kollias G, Kumar PV, Gopalakrishnakone P. (2009). Suppressive effect of secretory phospholipase A2 inhibitory peptide on interleukin-1beta-induced matrix metalloproteinase production in rheumatoid synovial fibroblasts, and its antiarthritic activity in hTNFtg mice. Arthritis Res Ther 11: R138. Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp MV et al. (1999). Estimating kinetic parameters from dynamic contrastenhanced T1-weighted MRI of a diffusible tracer: standardized quantities and symbols. J Magn Reson Imag 10: 223–232. Tofts PS, Kermode A. (1991). Measurement of the blood–brain barrier and leakage space using dynamic MR imaging 1. Fundamental concepts. Magn Reson Med 17: 357–367. Vihinen P, Ala-aho R, Kahari V-M. (2005). Matrix metalloproteinases as therapeutic targets in cancer. Curr Cancer Drug Targets 5: 203–220. Zhang Q, Furukawa K, Chen HH, Sakakibara T, Urano T, Furukawa K. (2006). Metastatic potential of mouse Lewis lung cancer cells is regulated via ganglioside GM1 by modulating the matrix metalloprotease-9 localization in lipid rafts. J Biol Chem 281: 18145–18155.