Radiation enhances long-term metastasis potential of residual hepatocellular carcinoma in nude mice through TMPRSS4-induced epithelial–mesenchymal transition

Article metrics


Recurrence and metastasis are frequently observed after radiotherapy for hepatocellular carcinoma (HCC), although upregulation of matrix metalloproteinases (MMPs) and vascular endothelial growth factor (VEGF) induced by radiation has been claimed to be involved, the mechanism is not clarified yet. In the present study, by using MHCC97L, a human HCC cell line with metastatic potential, and its xenograft in nude mice, we found that radiation induced a 48- to 72-h temporary increase in the expression of MMP-2 and VEGF both in vitro and in vivo, but only the in vitro invasiveness of MHCC97L cells was enhanced, while the in vivo metastatic potential of tumors was suppressed. Whereas, 30 days after radiation, when the expression of MMP-2 and VEGF decreased to unirradiated control levels, the in vivo dissemination and metastatic potential of residual tumors have just begun to increase with overexpression of TMPRSS4, which induced loss of E-cadherin through induction of Smad-Interacting Protein 1 (SIP1), an E-cadherin transcriptional repressor, and led to epithelial–mesenchymal transition (EMT). This process was blocked by treatment of siRNA-TMPRSS4. In conclusion, our study revealed novel findings regarding the biphasic effect of radiation on the metastatic potential of residual HCC. Overexpression of TMPRSS4 has a critical role in radiation-induced long-term dissemination and metastasis of residual HCC by facilitating EMT. These findings may provide new clues to suppress the radiation-induced dissemination and metastasis, thereby improve the prognosis of HCC patients.


Recent advances have made radiotherapy to be one of the therapeutic options of unresectable hepatocellular carcinoma (HCC).1, 2, 3 However, even with high dose of radiation and improving delivery method, a significant number of HCC patients still have local failure and exhibit a substantially higher rate of distant metastasis than those who achieve permanent local control. Similar observations have also been found in the clinical study of other tumors.4 In Lai et al.5 study, distant metastasis was developed in 33% of the patients after radiation for prostate carcinoma. And the study by Bonner et al.6 showed that the 2-year cumulative rate of incidence of distant metastases, mostly in the lungs and bones, was as high as 17% after high-dose radiotherapy for squamous-cell carcinoma of the head and neck. Therefore, to improve the effect of radiotherapy, addition of chemotherapy to radiotherapy is now the standard for locally advanced tumors and improves the survival of patients. However, as the mechanism of radiation-induced metastasis has not been identified and drugs are not directly targeted at related molecules, in many studies including Bonner’s,6 even adding chemotherapy to radiotherapy has no effect on distant metastases.

Since the first major work examining the effect of local radiation on distant metastasis in 1949 by Kaplan and Murphy,7 a variety of studies have been performed, and several possible mechanisms have been suggested, including DNA changes, abscopal effects, microvascular damage and tumor necrosis induced by the radiation therapy;8, 9, 10, 11 however, convincing data are not yet available. Recent studies have shown that radiation can enhance the invasiveness of a variety of cancer cells in vitro by activating several genes, such as matrix metalloproteinases (MMPs) and vascular endothelial growth factor (VEGF),12 both of which have a vital role in tumor progression. And inhibition of radiation-enhanced MMPs protein expression significantly suppressed radiation-induced invasion.13 But whether it will cause increased frequency of metastasis in vivo is still not clarified. Furthermore, in some studies the invasiveness of cancer cells appears to be inhibited or unaffected by radiation.14, 15 Therefore, it is necessary to elucidate the effect of radiation on the invasion and metastatic potential of cancer cells, as well as the relevant molecular mechanisms involved. The better understanding of radiotherapy on cancer biology will certainly provide clues for intervention and improve the outcome of radiotherapy.

In this work, we investigated the effects of radiation on the invasion and metastatic potential of HCC and analyzed the underlying mechanism. To rule out the influence of radiation on tumor load and immune system, which may affect the metastasis, reimplantation of residual cancer after radiation and their unirradiated control counterpart were specially designed for the study of metastatic potential of residual HCC after radiotherapy.

Materials and methods

Reagents and cell culture

MHCC97L cell line was used in this study. MHCC97L cell line is a human HCC cell line with metastatic potential, derived from MHCC97 HCC cell line.16 Cells were maintained in Deulbecco's Modified Eagle Medium, supplemented with 10% fetal bovine serum and 1% streptomycin/penicillin at 37 °C with humidified 95% air and 5% CO2. All MMP-2-related experiments were done under serum-free conditions.

Animals and tumor model

Male BALB/c nu/nu mice, 4–6 weeks old, were obtained from Shanghai Institute of Materia Medica, Chinese Academy of Science. All studies on mice were conducted in accordance with the National Institutes of Health ‘Guide for the Care and Use of Laboratory Animals’ and were approved by Shanghai Medical Experimental Animal Care Committee. A patient-like human HCC nude mouse model that established (by orthotopic inoculation of histologically intact MHCC97L tumor tissue into the liver) in authors’ institution was employed. Each group contained 12–18 mice.


For in vitro study, cells were grown to 70% confluence and were irradiated with different doses (0, 4 or 8 Gy) of 6 MV X-ray at room temperature using a linear accelerator at a dose rate of 200 MV per minute. For in vivo study, when the intrahepatic tumor grew to the maximal diameter of 1.0–1.5 cm and was palpable, radiation port was designed on the basis of tumor location. Mice were anesthetized and immobilized to make the tumor area exposed, whereas the remainder of the body including a partial normal liver was shielded by lead. The tumors were irradiated with 9 MV electrons at 2 Gy per fraction, 5 days per week for 2 weeks, using a radiation field of 20 × 20 mm, which covered the entire tumor.

In vitro matrigel invasion assay

Invasiveness of MHCC97L cells was measured by the invasion of cells through Matrigel-coated transwell inserts (Corning, Corning, NY) as previously described.17 Briefly, transwell inserts with 8 μm pores were coated with Matrigel (40 μg per well; Becton Dickinson, Bedford, MA). A total of 1 × 105 cells were added to the upper chamber, suspended in 100 μl of Deulbecco's Modified Eagle Medium. The lower chamber was filled with 600 μl of Deulbecco's Modified Eagle Medium supplemented with 0.1% bovine serum albumin. After attaching the cells to the insert, the medium in the upper chamber was changed to serum-free medium and was exposed to radiation. After 48 h of incubation, the cells on the upper surface were removed using a cotton bud. The remaining invaded cells were fixed and stained with 0.1% crystal violet for 1 h at room temperature. Finally, invaded cells were counted at × 200 magnification in five different fields of each filter. Experiments were repeated for three times.

Gelatin zymography

MHCC97L cells grown to 70% confluence were washed and refreshed with serum-free Deulbecco's Modified Eagle Medium, and then were irradiated and incubated for 24, 48, 72 and 96 h. The supernatant of the unirradiated and irradiated MHCC97L cells was collected and concentrated to 2 mg ml−1. Samples were added to each lane and subjected to 10% SDS-PAGE using 10% polyacrylamide gel containing 1 mg ml−1 gelatin. After electrophoresis, the gel was washed in 2.5% Triton X-100, and incubated in 50 mM Tris-HCl buffer (pH 8.0) containing 0.5 mM CaCl2 and 1 mM ZnCl2 at 37 °C for 20 h. The gel was stained with 0.5% Coomassie Brilliant Blue R-250 and destained with destaining buffer (10% acetic acid and 45% methanol).

RNA isolation and quantitative RT-PCR

Total RNA was extracted from tumors using Trizol reagent (Invitrogen, Carlsbad, CA). In all, 2 μg of total RNA was reverse transcribed using RevertAid first strand cDNA synthesis kit (Fermentas, Glen Burnie, MD). TMPRSS4 mRNA expression was quantified using the SYBR Green PCR MASTER MIX kit (Takara, Otsu, Japan) and ABI PRISM 7300 Sequence Detection System, Foster City, CA. PCR amplification cycles were programmed for 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. The efficiency of cDNA synthesis was estimated by PCR with β-actin-specific primers. The following sense (S) and antisense (AS) primer were used in the RT-PCR reactions: TMPRSS4 (S: 5′-IndexTermCCGTTACCTATGTTGTCATTG-3′, AS: 5′-IndexTermCTCACTCTTGCTGCCACTCA-3′); β-actin (S: 5′-IndexTermCCTCTATGCCAACACAGTGC-3′, AS: 5′-IndexTermGTACTCCTGCTTGCTGATCC-3′). Relative mRNA levels were calculated based on the Ct values, according to the equation:

Western blot analysis

The conditioned medium was collected for in vitro MMP-2 analysis. Tumor specimens and cells were lysed in lysis buffer for VEGF, MMP-2, TMPRSS4, E-cadherin, N-cadherin, vimentin and Smad-Interacting Protein 1 (SIP1) analysis. After a brief sonication, the lysates were clarified by centrifugation at 12 000 g for 15 min at 4 °C, and protein content was measured by the Bradford method. An aliquot (20 μg protein per lane) of the total protein was separated by 12% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA). The membrane was incubated with monoclonal mouse anti-human VEGF and MMP-2 (1:200; Chemicon, Billerica, MA), rabbit anti-human E-cadherin, N-cadherin (1:1000; Cell Signaling, Danvers, MA), TMPRSS4 (1:200; ProteinTech Group, Chicago, IL) and vimentin (1:200; Santa Cruz Biotechnology, Santa Cruz, CA), and goat anti-human SIP1 (1:200; Santa Cruz Biotechnology). Then probed with anti-mouse/rabbit/goat IgG (Chemicon) at 1:10 000 dilution for 1 h. The blots were developed using an enhanced chemiluminescence (ECL) kit (Pierce, Rockford, IL). Each experiment was repeated at least for three times.


Immunohistochemistry was carried out using a two-step protocol. Briefly, after microwave antigen retrieval, tissues were incubated with primary antibodies for 60 min at room temperature, and were incubated with secondary antibody for 30 min. The sections were developed in diaminobenzidin solution under microscopic observation and counterstained with hematoxylin. Negative control slides with the primary antibodies omitted were included in all assays.

siRNA design and assay

MHCC97L cells were incubated without antibiotics for 24 h before transfection. Control siRNA labeled with FAM and specific TMPRSS4 siRNA were mixed with Lipofectamine2000 (Invitrogen) according to the manufacturer’s recommendation and added to the cells. After 6 h at 37 °C, the medium was changed and the cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum. Transfection efficiency was observed under a fluorescence microscope. The silencing level of TMPRSS4 was determined by western blot 48 h after transfection. To minimize the off-target effect of RNAi, three TMPRSS4 siRNAs that have been demonstrated capable of suppressing TMPRSS4 expression were mixed together as a pool that was named siRNA-TMPRSS4. The sequences for the three siRNAs were siRNA-1, sense 5′-IndexTermGGAUCCUGACAGUGAUCAATT-3′ and antisense 5′-IndexTermUUGAUCACUGUCAGGAUCCTG-3′; siRNA-2, sense 5′-IndexTermCCGAUGUGUUCAACUGGAATT-3′ and antisense 5′-IndexTermUUCCAGUUGAACACAUCGGTA-3′; siRNA-3, sense 5′-IndexTermAGGUGAUUCUGGAUAAAUATT-3′ and antisense 5′-IndexTermUAUUUAUCCAGAAUCACCUTG-3′; FAM-labeled siRNA was used as a control (sense 5′-IndexTermUUCUCCGAACGUGUCACG-3′ and antisense 5′-IndexTermACGUGACACGUUCGGAGA-3′). The siRNAs were chemically synthesized at the Laboratory of RNA Chemistry (Shanghai GenePharma, Shanghai, China).

Immunofluorescence assay

TMPRSS4 expression in MHCC97L cells transfected with siRNA-TMPRSS4 or siRNA-NC was detected by immunofluorescence assay. Briefly, 48 h after transfection, cells cultured on glass slides were fixed by acetone for 15 min. After treated with 0.2% Triton X-100 for 2 min, the fixed cells were blocked with bovine serum albumin and stained with rabbit anti-human TMPRSS4 fluorescein isothiocyanate conjugated monoclonal antibody (1:200; Chemicon) for 1 h at 37 °C. After rinsing in phosphate-buffered saline, the slices were counterstained with DAPI (Vector Laboratories, Burlingame, CA) and examined under fluorescent microscope (Olympus BX-40, Tokyo, Japan). Each experiment was repeated at least for three times.

Tumor regrowth and metastasis assay

In vivo study, the irradiated mice and their unirradiated control group were killed 2 and 30 days after radiation and the residual tumors were resected and prepared to the same size of 1 × 2 mm, then were reimplanted into the liver of other normal nude mice. The mice were treated with siRNA-TMPRSS4 twice weekly (0.5 μg kg−1 body weight) by tail vein injection after tumor reimplantation, whereas the mice in control group were treated with siRNA-NC. Six weeks after the implantation, animals were killed and autopsied. Tumor volume (V ) was measured with calipers and was calculated as V=ab2/2, where a is the longer and b is the shorter of two orthogonal diameters.18 The lungs were removed for standard histopathological examinations. Serial sections were made for every tissue block of the lung, and frequency of lung metastases was counted under the microscope. Other organs suspected of tumor involvement were also sampled for histopathological studies.

Statistical analyses

Statistical analyses were performed by SPSS 13.0 for windows (SPSS, Chicago, IL). The χ2 test, Fisher's exact probability and Student's t-test were used for comparison between groups. If variances in groups were not homogeneous, the non-parametric Mann–Whitney U-test was used. All tests were two-tailed, and data were considered to be statistically significant at P<0.05.


Radiation-enhanced invasiveness of MHCC97L cells in vitro

Before investigating the effect of radiation on invasiveness of MHCC97L cells, we first examined the effect of radiation on proliferation of cells using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as described in previous study.19 Radiation suppressed the proliferation of HCC cells in a dose-dependent manner and almost completely inhibited the proliferation at a dose of 8 Gy.

Changes in the invasiveness of HCC cells after radiation were examined by using the Matrigel invasion assay. The invasiveness of MHCC97L cells was significantly increased after radiation in a dose- and time-dependent manner, the highest rate (fourfold greater compared with unirradiated controls) was seen 48 h after 8 Gy radiation (P<0.05; Figure 1). But 96 h after radiation, there was no significant difference in the invasiveness between the radiated and unirradiated HCC cells.

Figure 1

Radiation enhances the invasiveness of HCC cells. MHCC97L cells were seeded in Matrigel-coated transwell inserts. Cells were irradiated with different doses (0, 4 and 8 Gy). After 48 h, the cells invading the membranes were photographed under a light microscope and counted. Student's t-test was used to evaluate the difference between groups. *P<0.05

Radiation increased MMP-2 activity

To determine the role of gelatinases in the radiation-induced invasiveness, the conditioned medium was collected 24, 48, 72 and 96 h after radiation and was subjected to gelatin zymography. Zymographic analysis revealed a time- and dose-dependent increase in enzyme activity of MMP-2 in conditioned medium of irradiated cells, attaining a maximum at 48 h after 8 Gy radiation (Figure 2a). Consistent with the changes of cell invasiveness, MMP-2 activity also began to decrease after that and dropped below the control level after 72 h. These results suggest that the increased MMP-2 activity may have a role in the radiation-enhanced temporary invasiveness in vitro.

Figure 2

Effect of radiation on the expression or activity of MMP-2 and VEGF in MHCC97L cell lines. (a, b) The supernatant of the unirradiated and irradiated MHCC97L cells were collected 24, 48, 72 and 96 h after radiation and were subjected to gelatin zymography and western blot analysis, to evaluate the effect of radiation on the activity and expression of MMP-2; (c) MHCC97L cells were irradiated by 8 Gy radiation and were lysed 24, 48, 72 and 96 h later to detect the protein level of VEGF. Tubulin was used as a loading control. Quantitation of VEGF protein abundance normalized to Tubulin is shown as mean and s.e. The protein levels were significantly (*P<0.05) enhanced 48 h after radiation. (d) Western blot analysis of the tumor expression of MMP-2 and VEGF 2 and 30 days after radiation. The protein levels of MMP-2 and VEGF were significantly enhanced 2 days after radiation (*P<0.05).

Radiation increased MMP-2 and VEGF secretion in MHCC97L cells

The expression of MMP-2 and VEGF in irradiated cells was determined by western blot analysis. Compared with unirradiated control group, the expression of MMP-2 and VEGF of the irradiated group was significantly induced after radiation, in a time- and dose-dependent manner, reached the maximum at 48 h after radiation and began to decline after 72–96 h (Figures 2b and c).

Effects of radiation on MMP-2 and VEGF expression, tumor regrowth, dissemination and metastasis

Similarly to in vitro study, the expression of MMP-2 and VEGF in HCC tissues was significantly increased 2 days after radiation but dropped to unirradiated control levels within 30 days (Figure 2d). To clarify whether the temporary overexpression of MMP-2 and VEGF will promote metastasis in vivo and exclude other factors that may influence metastasis, we specially designed that the irradiated tumors were resected 2 and 30 days after radiation and reimplanted into the liver of other nude mice. Six weeks after reimplantation, the mice were killed to evaluate tumor regrowth, intrahepatic dissemination and metastasis. Compared with control group, both tumor regrowth (1.61±0.51 vs 2.25±0.52 cm3, P<0.05; Figure 3a) and incidence of lung metastasis (Figure 3b; 12.5 vs 66.7%, P<0.05) were significantly inhibited 2 days after radiation. However, 30 days after radiation, tumor growth recovered (2.60±0.61 vs 2.15±0.71 cm3, P>0.05; Figure 3a) and lung metastasis began to increase (63.6 vs 100%, P<0.05; Figure 3b). Intrahepatic disseminations from tumors 30 days after radiation were also significantly increased than control group, which did not develop any intrahepatic disseminations (18±8.05 vs 0, P<0.05; Figure 3c).

Figure 3

Effect of radiation on tumor regrowth, lung metastasis, intrahepatic dissemination and TMPRSS4 mRNA transcription. (a) The tumor regrowth was significantly inhibited 2 days after radiation (P<0.05), but recovered 30 days after radiation; (b) the incidence of lung metastasis from tumors 2 days after radiation was significantly inhibited (P<0.05), while that from tumors 30 days after radiation was significantly increased (P<0.05); (c) pathological examination showed that the intrahepatic dissemination was significantly increased after reimplantation of tumors 30 days after radiation (P<0.05); (d) quantitative RT-PCR analysis of TMPRSS4 days after radiation. The transcription of TMPRSS4 mRNA was significantly increased 30 days after radiation (P<0.05).

Morphologic changes of tumors after radiation

Tumors of control group were solitary and developed no disseminations. The control tumors pushed aside the surrounding liver tissue and formed a distinct border between tumor and liver tissue (Figures 4a and b, arrows). Compared with control group, after reimplantation of the tumors 30 days after radiation, there were many disseminations in the liver, the border between tumor and liver tissue was unclear and some cells turned from round into a spindle-like mesenchymal phenotype (Figures 4c–e, arrows). Lung metastases were easily observed (Figure 4f).

Figure 4

Macroscopic and microscopic changes of MHCC97L tumor after radiation. (a, b) The tumor of control group displayed an exophytic growth pattern and the border between the tumor and liver tissue was clear (arrow); (c, d) the tumor 30 days after radiation showed an infiltrative growth pattern. Some cells turned from round into a spindle-like mesenchymal phenotype, and intrahepatic dissemination was significantly increased (arrow); (e, f) HE staining of intrahepatic disseminations and lung metastasis from tumors 30 days after radiation.

Radiation-induced epithelial–mesenchymal transition

As the above-mentioned increases in MMP-2 and VEGF expression did not enhance the metastasis and dissemination, we further investigated the expression of epithelial–mesenchymal transition (EMT)-related genes. Western blot analysis showed that 30 days after radiation, tumors displayed significantly reduced E-cadherin expression, whereas E-cadherin expression maintained in the tumors of other two groups (Figure 5a). Furthermore, SIP1, an E-cadherin transcriptional repressor,20 and TMPRSS4, which has been proved to induce EMT through ERK1/2 and MAPK pathway,21 were significantly induced in tumors 30 days after radiation (Figure 5a). Other mesenchymal markers, such as N-cadherin and vimentin, were also significantly enhanced in this group (Figure 5a). Immunohistochemistry and quantitative RT-PCR also confirmed that the expression of TMPRSS4 was highly upregulated in tumors 30 days after radiation (Figures 3d and 5b).

Figure 5

(a) Western blot analysis of expression of TMPRSS4, SIP1, E-cadherin, N-cadherin and vimentin in tumors 2 and 30 days after radiation. GAPDH was used as a loading control. Quantitation of protein abundance normalized to GAPDH is shown as mean and s.e. The protein levels were significantly (P<0.05) changed 30 days after radiation. (b) Immunohistochemistry showed that the expression of TMPRSS4 in tumors 30 days after radiation (B) was significantly higher than that of control group (A).

TMPRSS4 siRNA inhibited radiation-enhanced dissemination and metastasis

To clarify the role of TMPRSS4 in radiation-induced EMT, a pool of siRNA-TMPRSS4, which was effective in suppressing TMPRSS4 expression (Figure 6a), was given twice weekly by tail vein injection for 6 weeks. The results showed that, compared with control group, siRNA-TMPRSS4 significantly inhibited the radiation-induced expression of TMPRSS4 and SIP1 (Figure 6b), and abolished the intrahepatic dissemination lesions (0.88±1.13 vs 18±8.05, P<0.05). Furthermore, western blot and immunohistochemistry demonstrated that the expression of E-cadherin was enhanced in siRNA-TMPRSS4-treated group than in control group (Figures 6b and c). These results suggest that radiation-induced EMT may be mediated by overexpression of TMPRSS4.

Figure 6

(a) Immunofluorescence and western blot analysis showed that compared with negative control siRNA (siRNA-NC) treatment group, siRNA-TMPRSS4 significantly inhibited the expression of TMPRSS4 in MHCC97L cells; (b) western blot analysis showed that siRNA-TMPRSS4 significantly inhibited radiation-induced upregulation of TMPRSS4 and SIP1, and enhanced the expression of E-cadherin (P<0.05); (c) immunohistochemistry showed that the tumor expression of E-cadherin was significantly higher in siRNA-TMPRSS4 group (B) than in siRNA-NC group (A).


Converse data have been reported concerning whether radiation enhances metastatic potential of residual cancer. In a review by Von Essen,22 the data from 41 different experiments were gathered to examine the effect of local irradiation on distant metastases. Though most of these experiments demonstrated a greater metastasis load after irradiation; however, some still showed a decreased metastasis load. In addition, in many of these studies, the mice were examined for metastases when moribund, and because irradiated mice had lower tumor load and longer survival times than unirradiated control mice, the increased incidence of metastases after irradiation may primarily be a result of the prolonged survival.

In our study, in order to investigate the direct effect of radiation on the metastatic potential of residual HCC, and exclude the influence of radiation on tumor load and immune system, the reimplantation of the same size of short-term (2 days after radiation) and long-term (30 days after radiation) residual cancer after radiation was specially designed. Different to the previous studies, we found a biphasic effect of radiation on metastasis. In short, 2 days after radiation, the regrowth and metastatic potentials of residual tumors were significantly inhibited; while 30 days after radiation, not only tumor growth recovered, but the potential of dissemination and metastasis also begun to increase. These findings were similar to the clinical observation that dissemination and metastasis seldom happened shortly after radiotherapy but usually developed after a period of relief or incubation.

Recently, several studies have demonstrated that radiation can promote the in vitro invasion of cancer cells by activating several genes involved in tumor invasion and metastasis. Jadhav23 has reported that radiation enhances the invasiveness of human neuroblastoma cells in association with enhanced expression or activity of MMP-9 and VEGF. Cheng's24 study also indicates that radiation-enhanced invasion capability is evident in HCC cells through upregulation of MMP-9, which he supposed to be correlated with metastasis due to radiotherapy in clinical practice. Similarly to the previous studies that only investigate the short-term effect of radiation on cancer cells in vitro, our in vitro study also found a 48- to 72-h increase in the expression of MMP-2 and VEGF as well as the invasiveness of HCC cells. However, whether these temporary increases induced by radiation will contribute to the in vivo metastasis of HCC needs to be clarified.

Therefore, in our in vivo study, the effects of radiation on MMP-2 and VEGF were also studied. The expression of MMP-2 and VEGF in HCC tissues was increased 2 days after radiation, but dropped to unirradiated control levels within 30 days. These findings are compatible with the clinical observation of Speake,25 who found in clinical practice that radiation increased MMP expression was short lived and limited to a matter of days as opposed to weeks. To our surprise, the change in the expression of MMP-2 and VEGF was not in accordance with the metastatic potential, which was significantly inhibited 2 days after radiation and begun to enhance 30 days after radiation. This indicates that there may be some factors other than MMP-2 and VEGF that are involved in the radiation-induced late-stage metastasis and dissemination.

In our ensuing study, we detected a reduced expression of E-cadherin and upregulation of mesenchymal markers, including N-cadherin and vimentin, in tumors 30 days after radiation, together with some morphologic changes of tumors, such as no distinct border between tumor and liver tissue, a spindle-like mesenchymal phenotype change, and increased intrahepatic disseminations and lung metastasis. All these findings were highly suggestive of EMT, and were not observed in tumors 2 days after radiation. This suggests that EMT is induced by radiation after a period of time.

EMT is a process implicated in the conversion of early-stage tumors to invasive malignancies whereby epithelial cell layers lose polarity and cell–cell contact and undergo a dramatic remodeling of the cytoskeleton.26, 27 Induction of EMT in tumors will result in weakened intercellular adhesion and enhanced cell motility, thereby allow tumor cells to metastasize and establish secondary tumors at distant site.28 For radio-oncologist, radiation-enhanced cell motility is an important fact should not be ignored, because increase in motility will lead to a local evasion of tumor cells from the target field without receiving an appropriate radiation dose, and may thus be a source of local recurrence and distant metastasis. The mechanism underlying radiation-enhanced motility varies in different cells; it can not only be directly enhanced through Rac or FAK signaling pathways,29, 30 but can be induced by hypoxia-mediated upregulation of urokinase-type plasminogen activator receptor.31 In our study, radiation-induced EMT is the main cause of enhanced cell motility and blocking the process of EMT can drastically suppress cell motility and dissemination.

Until recently, the association between radiation and EMT has not been intensively investigated, and only a few studies have examined the underlying mechanism. Andarawewa32 has proved that radiation can predispose non-malignant human mammary epithelial cells to undergo transforming growth factor-β mediated EMT through MAPK signaling pathways, thereby elicits heritable phenotypes that could contribute to neoplastic progression. Studies of Tsukamoto et al.33 also indicate that radiation can induce EMT through promoting the expression of Twist, an organizer of EMT, thus enhance the invasive potential of endometrial carcinoma cells.

In our study, 30 days after radiation, we observed strong SIP1 induction in TMPRSS4-overexpression tumors. TMPRSS4 is a novel type II transmembrane serine protease found at the cell surface and is highly expressed in a variety of cancer tissues.21, 34 SIP1 is one of the prominent EMT-inducing regulators that repress E-cadherin transcription via interaction with specific E-boxes of the proximal E-cadherin promoter.35 Since siRNA-TMPRSS4 has been proved to inhibit radiation-induced expression of TMPRSS4 and SIP1, thereby enhance E-cadherin expression and suppress radiation-enhanced dissemination and metastasis in our study, we speculate that radiation-induced EMT may be mediated by TMPRSS4-induced SIP1. Though it is well established that TMPRSS4 can activate SIP1 and stimulate EMT via MAPK signaling pathway,21, 36 it is the first time by our study to demonstrate that TMPRSS4 is involved in radiation-induced EMT. However, since we could not rule out the possibility of other E-cadherin regulators such as Snail, Slug and Twist that may be involved in radiation-induced EMT, further study is essential to elucidate the mechanism involved in radiation-induced EMT.

Importantly, radiation-induced EMT and enhancement in dissemination and metastasis in our study are observed at doses used clinically as a fractionated radiation for tumors. Therefore, possibility of metastasis due to radiotherapy deserves serious attention in clinical practice, and additional work is needed to eliminate the adverse effect of radiation. According to our results, one of the most important strategies is targeting TMPRSS4. Furthermore, since loss or reduction of E-cadherin expression is a key hallmark of EMT,26 it is not surprising that treatment aimed at enhancing the expression of E-cadherin will also suppress irradiation-induced metastasis. And the observation that induction of EMT begins 30 days after radiotherapy implies that chemotherapy aimed at inhibiting radiation-induced metastasis should be administered as early as possible since clinical standard treatment protocol of fractionated therapy usually takes 4–6 week to finish. In addition, though temporary increases in MMP-2 and VEGF expression do not promote the metastasis of radiated tumors, they may be a paracrine proliferative stimulus to accelerate the growth of microtumors not included in the radiotherapy field, so temporary MMP-2 or VEGF inhibition around radiotherapy may also have additional benefits.

To sum up, our study reveals novel findings regarding the biphasic effect of radiation on the metastatic potential of residual HCC and the related molecular mechanism involved, and may have significant implications for the radiotherapy of HCC as well as for cancer treatment in general. The observation that radiation can enhance long-term metastatic potential of residual HCC through TMPRSS4-induced EMT will provide new clues to suppress the post-radiation disseminations and metastasis, thereby improve the prognosis of HCC patients.


  1. 1

    Krishnan S, Dawson LA, Seong J, Akine Y, Beddar S, Briere TM et al. Radiotherapy for hepatocellular carcinoma: an overview. Ann Surg Oncol 2008; 15: 1015–1024.

  2. 2

    Kulik LM, Carr BI, Mulcahy MF, Lewandowski RJ, Atassi B, Ryu RK et al. Safety and efficacy of 90Y radiotherapy for hepatocellular carcinoma with and without portal vein thrombosis. Hepatology 2008; 47: 71–81.

  3. 3

    Hassoun Z, Gores GJ . Treatment of hepatocellular carcinoma. Clin Gastroenterol Hepatol 2003; 1: 10–18.

  4. 4

    Maeta M, Koga S, Kanayama H, Murakami A, Ikeda Y, Inoue Y . Does preoperative radiation for thoracic esophageal cancer promote intramural lymphatic invasion. Jpn J Surg 1986; 16: 84–89.

  5. 5

    Lai PP, Perez CA, Lockett MA . Prognostic significance of pelvic recurrence and distant metastasis in prostate carcinoma following definitive radiotherapy. Int J Radiat Oncol Biol Phys 1992; 24: 423–430.

  6. 6

    Bonner JA, Harari PM, Giralt J, Azarnia N, Shin DM, Cohen RB et al. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck. N Engl J Med 2006; 354: 567–578.

  7. 7

    Kaplan HS, Murphy ED . The effect of local roentgen irradiation on the biological behavior of a transplantable mouse carcinoma. Increased frequency of pulmonary metastasis. J Natl Cancer Inst 1949; 9: 407–413.

  8. 8

    Baumann M, Becker S, Kruger HJ, Vogler H, Maurer T, Beck-Bornholdt HP . Flow cytometric determination of the time of metastasis during fractionated radiation therapy of the rat rhabdomyosarcoma R1H. Int J Radiat Biol 1990; 58: 361–369.

  9. 9

    de Ruiter J, Cramer SJ, Lelieveld P, van PLM . Comparison of metastatic disease after local tumour treatment with radiotherapy or surgery in various tumour models. Eur J Cancer Clin Oncol 1982; 18: 281–289.

  10. 10

    Heisel MA, Laug WE, Stowe SM, Jones PA . Effects of X-irradiation on artificial blood vessel wall degradation by invasive tumor cells. Cancer Res 1984; 44: 2441–2445.

  11. 11

    Bonfil RD, Bustuoabad OD, Ruggiero RA, Meiss RP, Pasqualini CD . Tumor necrosis can facilitate the appearance of metastases. Clin Exp Metastasis 1988; 6: 121–129.

  12. 12

    Kaliski A, Maggiorella L, Cengel KA, Mathe D, Rouffiac V, Opolon P et al. Angiogenesis and tumor growth inhibition by a matrix metalloproteinase inhibitor targeting radiation-induced invasion. Mol Cancer Ther 2005; 4: 1717–1728.

  13. 13

    London CA, Sekhon HS, Arora V, Stein DA, Iversen PL, Devi GR . A novel antisense inhibitor of MMP-9 attenuates angiogenesis, human prostate cancer cell invasion and tumorigenicity. Cancer Gene Ther 2003; 10: 823–832.

  14. 14

    Bauman GS, Fisher BJ, McDonald W, Amberger VR, Moore E, Del Maestro RF . Effects of radiation on a three-dimensional model of malignant glioma invasion. Int J Dev Neurosci 1999; 17: 643–651.

  15. 15

    Gliemroth J, Feyerabend T, Gerlach C, Arnold H, Terzis AJ . Proliferation, migration, and invasion of human glioma cells exposed to fractionated radiotherapy in vitro. Neurosurg Rev 2003; 26: 198–205.

  16. 16

    Tian J, Tang ZY, Ye SL, Liu YK, Lin ZY, Chen J et al. New human hepatocellular carcinoma (HCC) cell line with highly metastatic potential (MHCC97) and its expressions of the factors associated with metastasis. Br J Cancer 1999; 81: 814–821.

  17. 17

    Tavian D, Salvi A, De Petro G, Barlati S . Stable expression of antisense urokinase mRNA inhibits the proliferation and invasion of human hepatocellular carcinoma cells. Cancer Gene Ther 2003; 10: 112–120.

  18. 18

    Labonte P, Kadhim S, Bowlin T, Mounir S . Inhibition of tumor growth with doxorubicin in a new orthotopically implanted human hepatocellular carcinoma model. Hepatol Res 2000; 18: 72–85.

  19. 19

    Song Y, Dong MM, Yang HF . Effects of RNA interference targeting four different genes on the growth and proliferation of nasopharyngeal carcinoma CNE-2Z cells. Cancer Gene Ther 2011; 18: 297–304.

  20. 20

    Vandewalle C, Comijn J, De Craene B, Vermassen P, Bruyneel E, Andersen H et al. SIP1/ZEB2 induces EMT by repressing genes of different epithelial cell-cell junctions. Nucleic Acids Res 2005; 33: 6566–6578.

  21. 21

    Jung H, Lee KP, Park SJ, Park JH, Jang YS, Choi SY et al. TMPRSS4 promotes invasion, migration and metastasis of human tumor cells by facilitating an epithelial-mesenchymal transition. Oncogene 2008; 27: 2635–2647.

  22. 22

    von Essen CF . Radiation enhancement of metastasis: a review. Clin Exp Metastasis 1991; 9: 77–104.

  23. 23

    Jadhav U, Mohanam S . Response of neuroblastoma cells to ionizing radiation: modulation of in vitro invasiveness and angiogenesis of human microvascular endothelial cells. Int J Oncol 2006; 29: 1525–1531.

  24. 24

    Cheng JC, Chou CH, Kuo ML, Hsieh CY . Radiation-enhanced hepatocellular carcinoma cell invasion with MMP-9 expression through PI3K/Akt/NF-kappaB signal transduction pathway. Oncogene 2006; 25: 7009–7018.

  25. 25

    Speake WJ, Dean RA, Kumar A, Morris TM, Scholefield JH, Watson SA . Radiation induced MMP expression from rectal cancer is short lived but contributes to in vitro invasion. Eur J Surg Oncol 2005; 31: 869–874.

  26. 26

    Javle MM, Gibbs JF, Iwata KK, Pak Y, Rutledge P, Yu J et al. Epithelial-mesenchymal transition (EMT) and activated extracellular signal-regulated kinase (p-Erk) in surgically resected pancreatic cancer. Ann Surg Oncol 2007; 14: 3527–3533.

  27. 27

    Ding W, You H, Dang H, LeBlanc F, Galicia V, Lu SC et al. Epithelial-to-mesenchymal transition of murine liver tumor cells promotes invasion. Hepatology 2010; 52: 945–953.

  28. 28

    Thiery JP, Sleeman JP . Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol 2006; 7: 131–142.

  29. 29

    Hwang SY, Jung JW, Jeong JS, Kim YJ, Oh ES, Kim TH et al. Dominant-negative Rac increases both inherent and ionizing radiation-induced cell migration in C6 rat glioma cells. Int J Cancer 2006; 118: 2056–2063.

  30. 30

    Nalla AK, Asuthkar S, Bhoopathi P, Gujrati M, Dinh DH, Rao JS . Suppression of uPAR retards radiation-induced invasion and migration mediated by integrin beta1/FAK signaling in medulloblastoma. PLoS One 2010; 5: e13006.

  31. 31

    Rofstad EK, Mathiesen B, Galappathi K . Increased metastatic dissemination in human melanoma xenografts after subcurative radiation treatment: radiation-induced increase in fraction of hypoxic cells and hypoxia-induced up-regulation of urokinase-type plasminogen activator receptor. Cancer Res 2004; 64: 13–18.

  32. 32

    Andarawewa KL, Erickson AC, Chou WS, Costes SV, Gascard P, Mott JD et al. Ionizing radiation predisposes nonmalignant human mammary epithelial cells to undergo transforming growth factor beta induced epithelial to mesenchymal transition. Cancer Res 2007; 67: 8662–8670.

  33. 33

    Tsukamoto H, Shibata K, Kajiyama H, Terauchi M, Nawa A, Kikkawa F . Irradiation-induced epithelial-mesenchymal transition (EMT) related to invasive potential in endometrial carcinoma cells. Gynecol Oncol 2007; 107: 500–504.

  34. 34

    Wallrapp C, Hähnel S, Müller-Pillasch F, Burghardt B, Iwamura T, Ruthenbürger M et al. A novel transmembrane serine protease (TMPRSS3) overexpressed in pancreatic cancer. Cancer Res 2000; 60: 2602–2606.

  35. 35

    Horsman MR, Murata R . Combination of vascular targeting agents with thermal or radiation therapy. Oncogene 2002; 54: 1518–1523.

  36. 36

    Choi SY, Shin HC, Kim SY, Park YW . Role of TMPRSS4 during cancer progression. Drug News Perspect 2008; 21: 417–423.

Download references


This work was supported by grants from the National Natural Science Foundation of China (Grant No. 30901444).

Author information

Correspondence to Z-Y Tang.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article


  • radiation
  • hepatocellular carcinoma
  • metastasis
  • epithelial–mesenchymal transition

Further reading