Irradiated fibroblast‐induced bystander effects on invasive growth of squamous cell carcinoma under cancer–stromal cell interaction

Abstract

The irradiated fibroblast‐induced response of non‐irradiated neighboring cells is called ‘radiation‐induced bystander effect’, but it is unclear in non‐irradiated human squamous cell carcinoma (SCC) cells. The present study shows that irradiated fibroblasts promoted the invasive growth of T3M‐1 SCC cells, but not their apoptosis, more greatly than non‐irradiated fibroblasts, using collagen gel invasion assay, immunohistochemistry and Western blot. The number of irradiated fibroblasts decreased to about 30% of that of non‐irradiated fibroblasts, but irradiated fibroblasts increased the growth marker ki‐67 display of SCC cells more greatly than non‐irradiated fibroblasts. Irradiated fibroblasts did not affect the apoptosis marker ss‐DNA expression of SCC cells. Irradiated fibroblasts enhanced the display of the following growth‐, invasion‐ and motility‐related molecules in SCC cells more greatly than non‐irradiated fibroblasts: c‐Met, Ras, mitogen‐activated protein kinase (MAPK) cascade (Raf‐1, MEK‐1 and ERK‐1/2), matrix metalloproteinase‐1 and ‐9, laminin 5 and filamin A. Irradiated fibroblasts, but not non‐irradiated ones, formed irradiation‐induced foci (IRIF) of the genomic instability marker p53‐binding protein 1 (53BP1) and expressed transforming growth factor‐β1 (TGF‐ β1). Irradiated fibroblasts in turn enabled SCC cells to enhance 53BP1 IRIF formation more extensively than non‐irradiated fibroblasts. Finally, effects of irradiated fibroblasts on growth and apoptosis of another HEp‐2 SCC cell type were similar to those of T3M‐1. These results suggest that irradiated fibroblasts promotes invasion and growth of SCC cells by enhancement of invasive growth‐related molecules above through TGF‐ β1‐mediated bystander mechanism, in which irradiated fibroblast‐induced genomic instability of SCC cells may be involved. (Cancer Sci 2008; 99: 2417–2427)

The major malignant tumor of the oral cavity is squamous cell carcinoma (SCC). Radiotherapy has been frequently applied for patients with SCC.^{( 1} , ^{2 )} In the cancer tissue, cancer cell–fibroblast interaction is critical for the behavior of SCC cells.^{( 3} , ^{4 )} Despite the fact that both SCC cells and fibroblasts undergo irradiation by radiotherapy, little attention has been paid to effects of irradiated fibroblasts on the behavior of SCC cells. Previous reports have shown that irradiated fibroblasts are involved in carcinogenesis and invasive growth of both normal and abnormal epithelial cell types that are not exposed to irradiation.^{( 5} , ⁶ , ^{7 )} This irradiated fibroblast‐induced response of non‐irradiated neighboring cells is called ‘radiation‐induced bystander effect’.^{( 8} , ^{9 )} However, it is unclear whether irradiated fibroblasts may affect the behavior of non‐irradiated SCC cells through bystander mechanism under cancer–stromal cell interaction.

p53‐binding protein‐1 (53BP1), a DNA damage checkpoint protein,^{( 10} , ^{11 )} forms irradiation‐induced foci (IRIF) in nuclei in response to irradiation. 53BP1 functions in activation of ATM (mutated in ataxia–telangiectasia), which activates signaling pathways of DNA double‐stranded break repair.^{( 10 )} This response, which leads to cell‐cycle delay, apoptosis and senescence, is critical for maintaining genomic stability.^{( 11} , ¹² , ^{13 )} Thus, IRIF formation of 53BP1 indicates a genomic instability and DNA damage.^{( 10} , ¹² , ¹⁴ , ^{15 )} However, it is unclear whether irradiated fibroblasts themselves enable non‐irradiated SCC cells to form 53BP1 IRIF through the bystander mechanism.

To address these critical issues, we examined the effects of irradiated fibroblasts on the apoptosis, growth and invasion of SCC cells using collagen gel invasion assay system.^{( 3} , ⁴ , ¹⁶ , ¹⁷ , ¹⁸ , ^{19 )} Cellular growth‐, invasion‐ and motility‐related molecules such as c‐Met, Ras, mitogen‐activated protein kinase (MAPK) cascade proteins (Raf‐1, MEK‐1, and ERK‐1/2),^{( 20 )} matrix metalloproteinase‐1, ‐9 (MMP‐1, ‐9),^{( 21 )} laminin 5^{( 22 )} and filamin A^{( 23 )} were analyzed by immunohistochemistry and Western blot. Also, IRIF formation of the genomic instability marker 53BP1 was studied by immunofluorescence.

Materials and Methods

Cell lines.  All procedures involving animal and human materials were performed in accordance with the regulations laid down by the ethical guidelines of Saga University (Saga, Japan). T3M‐1 cells derived from SCC of human oral cavity were used as a SCC cell type.^{( 24 )} This cell line was gifted by Prof. K. Satoh (Institute of Clinical Endocrinology, Tokyo Women's Medical College, Japan). As another SCC cell type, HEp‐2 cells (CCL 23, ATCC, Rockville, MD, USA) that originated from laryngeal SCC were partly used in this study. As a stromal cell type, mouse NIH 3T3 fibroblasts (CCL‐92, ATTC), which have been used in many studies regarding epithelial–mesenchymal and cancer–stromal cell interactions,^{( 25} , ^{26 )} were mainly utilized in this study for the reason described below. Human WI‐26 VA4 fibroblasts (JCRB9042, Health Science Research Resources Bank, Osaka, Japan)^{( 27 )} were in part used to confirm whether another fibroblast type other than NIH 3T3 fibroblasts was able to replicate NIH 3T3 fibroblast‐induced phenomena of SCC cells. These cell types were cultured and maintained in a complete medium: Ham F‐12 medium supplemented with 10% fetal calf serum (FCS) and 50 µg/mL gentamicin.

Culture system.  To examine effects of fibroblasts with or without irradiation on the invasive growth of SCC cells, we carried out collagen gel invasion assay system as described previously.^{( 3} , ⁴ , ¹⁶ , ¹⁷ , ¹⁸ , ^{19 )} Briefly, 1.5 mL of type I collagen gel containing 10 × 10⁵ fibroblasts were poured into a 30‐mm‐diameter culture dish (inner dish) of which the bottom was made with nitrocellulose membrane (Millicell‐CM, Millipore, Bedford, MA, USA). This fibroblast layer was exposed to a single irradiation of 12 Gy produced using γ‐rays generated by Gammacell 40 Atomic Energy of Canada (Ontario, Canada), with 1.043 Gy/min of dose rate, as previously described,^{( 18} , ^{28 )} because 12 Gy irradiation among dosages of 1, 6, 12 and 24 Gy induced the most clear effects on the biological behavior of SCC cells (Fig. 1). Non‐irradiated fibroblast layer, and fibroblast‐non‐containing collagen gel layers with and without irradiation were used as references. After the irradiated and non‐irradiated fibroblast layers and irradiated and non‐irradiated collagen gel layers without fibroblasts were cultured in the complete medium for 24 h, 10 × 10⁵ SCC cells were seeded on these layer types above. Thereafter, this inner dish was placed into a 90‐mm‐diameter culture dish (outer dish), and then the complete medium was added to both dishes. In this way, collagen gel invasion assay system was prepared, and SCC cells were cultured on the four collagen gel layer types. As described below, the number of non‐irradiated NIH 3T3 fibroblasts was about three times that of the irradiated fibroblasts during culture term because irradiation induced cell death in some fibroblasts. Thus, three times the amount of the complete medium of irradiated fibroblast‐containing cultures was used in non‐irradiated fibroblast‐containing cultures to prepare almost an equal nutrient condition between non‐irradiated and irradiated fibroblast‐containing culture assemblies.

Figure 1.

Effects of various dose‐irradiated NIH 3T3 fibroblasts on invasion, and matrix metalloproteinase‐1 (MMP‐1), ERK‐1/2 and p53‐binding protein‐1 (53BP1) irradiation‐induced foci (IRIF) expression of T3M‐1 squamous cell carcimona (SCC) cells at 10 days in culture. Twelve Gy‐irradiated fibroblasts (d) induce the greatest invasion in SCC cells among 0 (a), 1 (b), 6 (c) and 24 Gy (e)‐irradiated ones. Also, 12 Gy‐irradiated fibroblasts promote expression of MMP‐1, ERK‐1/2 and 53BP1 IRIF more greatly than other dose‐irradiated ones (f and g), although there is no significant difference of 53BP1 IRIF expression among 6, 12 and 24 Gy‐irradiated fibroblasts. Arrow, invasion. *P < 0.05.

Morphology and fibroblast number.  We examined the cells with hematoxylin–eosin (HE) staining, using deparaffinized sections of the cellular layer gel fixed with 4% formalin and routinely processed and embedded vertically in paraffin. Cellular stratification of SCC cells on the gel and their downgrowth into the gel were analyzed on the HE staining sections by light microscopy as described elsewhere.^{( 3} , ⁴ , ¹⁷ , ^{19 )} To observe a basement membrane formed by SCC cells, deparaffinized sections were subjected to reticulin silver impregnation staining. In general, irradiation induced cell death in various cell types. Thus, we examined the number of fibroblasts with or without irradiation at 5, 10 and 15 days in culture as follows. The number of fibroblasts was counted in at least 10 randomly chosen non‐contiguous and non‐overlapping fields (at low power view, ×10 objective) of the staining sections with both histochemistry and immunohistochemistry, as described below.

Immunohistochemistry.  SCC cells and fibroblasts were identified by monoclonal pancytokeratin AE1/AE3 and vimentin, respectively, antibodies (Dako Japan, Kyoto, Japan). To characterize SCC cell growth, we examined expression of MAPK cascade molecules (Raf‐1, MEK‐1, and ERK‐1/2), using mouse monoclonal Raf‐1, MEK‐1 and p‐ERK‐1/2 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA). To estimate invasion and motility of cancer cells, we examined expression of MMP‐1, MMP‐9,^{( 21 )} laminin 5^{( 22 )} and filamin A,^{( 23 )} using rabbit polyclonal MMP‐1 and MMP‐9, and mouse monoclonal filamin A (Lab Vision, Fremont, CA, USA) and laminin 5 (D4B5, Chemicon, Temecula, CA, USA) antibodies. We also estimated the expression of tumor necrosis factor‐α (TNF‐α) and transforming growth factor‐β1 (TGF‐β1), which are thought to mediate some radiation‐induced bystander effects,^{( 9 )} using their mouse monoclonal antibodies (Santa Cruz Biotechnology). Immunohistochemistry was carried out on deparaffinized sections by an avidin–biotin complex immunoperoxidase method, as described previously.^{( 29 )}

Proliferation.  SCC cell growth was examined by immunohistochemistry with Ki‐67 (clone MIB‐1, mouse monoclonal antibody, Dako Japan), as described above. To obtain the rate of Ki‐67‐positive SCC cells, 1000 cells were counted, and the percentage of Ki‐67‐positive nuclei was calculated.^{( 4 )}

Apoptosis.  To detect the apoptosis of SCC cells, we carried out immunohistochemistry with anti‐single stranded DNA (ssDNA) antibody (Dako Japan), as described previously.^{( 18 )} To determine the rate of apoptotic SCC cells, 1000 cells were counted, and the percentage of ssDNA‐positive nuclei was calculated.^{( 18 )} As described below, apoptosis of SCC cells had no significant difference between irradiated and non‐irradiated fibroblasts. The reason for this may be that the number of irradiated fibroblasts decreased to about 30% of that of non‐irradiated fibroblasts, as described below. Thus, apoptosis of SCC cells with 30 × 10⁵ irradiated fibroblasts was also compared with that of SCC cells with 10 × 10⁵ non‐irradiated fibroblasts.

Cancer cell invasion.  The invasion of SCC cells into collagen gel was examined by the following method, as described previously.^{( 3} , ⁴ , ^{19 )} Normal laryngeal squamous cells cultured on the gel formed a linear borderline at the contact points between the cells and gel, and the cells had no downgrowth.^{( 30 )} This borderline corresponded to a basement membrane structure, which was visualized with a reticulum silver impregnation stain. Thus, we judged both the smoothly linear borderline and non‐downgrowth of the cells into the gel as a non‐invasion. In contrast, both the irregular borderline and downgrowth were judged as an invasion. The degree of local invasiveness of SCC cells into gel was determined by the following two parameters: (1) invasion frequency; and (2) invasion depth. The invasion frequency was examined by the following formula: the number of invasion parts per 1 mm length of basement membrane. On 10 sections, these invasive parts were examined in 10 randomly chosen non‐contiguous and non‐overlapping fields (at low power view, ×10 objective) by light microscopy. The invasion depth was measured from the surface line to the deepest invasive front of SCC cells by the same method above.

Western blot.  SCC cells were cultured by a double‐dish culture system. In this system, irradiated and non‐irradiated fibroblasts were cultured as a feeder layer on the bottom of the outer dish, and SCC cells were grown on inserts with 0.4‐µm pore size (Falcon cell culture insert, Becton Dickinson, Franklin Lakes, NJ, USA). These inserts were placed on the underlying fibroblasts. After they were cultured for 10 days, SCC cells were scraped from two inner dishes. SCC cells were homogenized in 10‐mM sodium orthovanadate (Na₃VO₄) supplemented with protease inhibitors (Protease Inhibitor Cocktail Set, Boehringer–Mannheim, Tokyo, Japan). In total, 15 µg proteins was loaded to sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE) on 4–12% Bis–Tris Gel (Invitrogen, Carlsbad, CA) and transferred to nitrocellulose membrane. The sheet was incubated overnight at 4 C with anti‐Raf‐1, MEK‐1, p‐ERK‐1/2, MMP‐1, MMP‐9, laminin 5 and filamin A antibodies (described above in immunohistochemistry). The antigen on the membrane was visualized by the standard method. The density of the bands was determined by densitometry. The results were presented as a ratio of the control values. The signal intensity of p‐ERK‐1/2 blotting was normalized to the signal of the corresponding total protein.

Detection of 53BP1 focus formation by immunofluorescence.  To estimate whether fibroblasts with or without irradiation affect a genomic instability of SCC cells, we examined the nuclear irradiation‐induced foci (IRIF) of 53BP1 in SCC cells by immunofluolecence, as described previously.^{( 11} , ¹² , ^{15 )} After antigen retrieval with microwave treatment in citrate buffer, deparaffinized sections were preincubated with 10% normal goat serum. Then, tissues were reacted with anti‐53BP1 rabbit polyclonal antibody (Bethyl Labs, Montgomery, TX, USA) at a 1:200 dilution. The slides were subsequently incubated with Alexa Fluor 488‐conjugated goat anti‐rabbit antibody (Invitrogen). Specimens were counterstained with 4′,6‐diamidino‐2‐phenylindole dihydrochloride (DAPI‐I, Vysis Inc., IL, USA), and were visualized and photographed with a fluorescence microscope (Zeiss Axioplan2, Carl Zeiss Japan, Tokyo, Japan), equipped with a charge‐coupled device (CCD) camera, and then analyzed with IPLab/MAC image software (Scanalytics, Fairfax, VA, USA). To obtain the rate of SCC cells with intranuclear IRIF of 53BP1, 1000 cells were counted in 10 randomly chosen non‐contiguous and non‐overlapping fields (at high power view, ×40 objective), and the percentage of 53BP1 IRIF‐positive SCC cells with or without irradiated fibroblasts was calculated. Also, 53BP1 IRIF formation of fibroblasts with or without irradiation was investigated by the same method above.

Reversibility or irreversibility of radiation‐induced bystander effects.  To estimate whether radiation‐induced bystander effects are reversible or irreversible, we carried out the following study. SCC cells were cultured on irradiated or non‐irradiated fibroblast‐containing gel for 5 days, and then the irradiated or non‐irradiated fibroblast‐affected SCC cells were isolated by trypsin. They were seeded and cultured for 10 days on collagen gel alone. The invasion and 53BP1 IRIF formation of SCC cells were analyzed as described above.

Statistical analysis.  The data obtained from five independent experiments were analyzed by analysis of variance (ANOVA) with post hoc analysis for multiple comparisons. Values represented mean ± SD. P < 0.05 was considered significant.

Results

Effects of various dose‐irradiated fibroblasts on T3M‐1 SCC cell behavior.  To select an appropriate dose of irradiation to fibroblasts, NIH 3T3 fibroblasts were irradiated at 1, 6, 12 or 24 Gy and co‐cultured with T3M‐1 SCC cells. At 10 days, 12 Gy‐irradiated fibroblasts induced the greatest invasion and expression of MMP‐1, ERK‐1/2 and 53BP1 IRIF in SCC cells among various dose‐irradiated ones (Fig. 1). In this study, 12 Gy irradiation was thus selected.

Histology of T3M‐1 SCC cells in collagen gel invasion assay system.  Firstly, to elucidate effects of collagen gel with or without irradiation on T3M‐1 SCC cell behavior, SCC cells were cultured on the non‐irradiated or irradiated collagen gel layer without fibroblasts. SCC cells on the non‐irradiated collagen gel layer as well as the irradiated counterpart formed a thin stratified layer, and they did not infiltrate into the gel at 5 and 10 days in culture. Even at 15 days, SCC cells infiltrated only minimally into the gel. This suggests that both non‐irradiated and irradiated collagen gels themselves do not induce the active downgrowth of SCC cells into gel. Next, to evaluate effects of NIH 3T3 fibroblasts with or without irradiation on SCC cell behavior, SCC cells were cultured on non‐irradiated or irradiated fibroblast‐containing collagen gel layer. The stratification of SCC cells with irradiated fibroblasts tended to be slightly higher than that with non‐irradiated fibroblasts, but the downgrowth of SCC cells with irradiated fibroblasts into gel (Fig. 2b,d,f) was apparently greater than that of SCC cells with non‐irradiated fibroblasts (Fig. 2a,c,e). Also, WI‐26 VA4 fibroblasts allowed SCC cells to almost replicate NIH 3T3 fibroblast‐induced phenomena above (Fig. 3). In addition, irradiation induced cell death in NIH 3T3 fibroblasts more effectively than WI‐26 VA4 fibroblasts, because the number of irradiated NIH 3T3 fibroblasts decreased to about 30% of that of non‐irradiated fibroblasts (Fig. 4), wheresa the number of irradiated WI‐26 VA4 fibroblasts decreased to about 50% of that of non‐irradiated WI‐26 VA4 fibroblasts (data not shown). Thus, we investigated effects of fibroblasts with or without irradiation on the invasive growth of SCC cells, using more irradiation‐affective NIH 3T3 fibroblasts, as described below.

Figure 2.

Histology of T3M‐1 squamous cell carcimona (SCC) cells with non‐irradiated (a, c and e) or irradiated NIH 3T3 fibroblasts (b, d and f). At 5 (a and b), 10 (c and d) and 15 days (e and f) in culture, stratification of SCC cells with non‐irradiated fibroblasts tends to be slightly thicker than that with irradiated fibroblasts. SCC cells with irradiated fibroblasts undergo greater downgrowth into gel than those with non‐irradiated fibroblasts. Note that number of irradiated fibroblasts is apparently lower than that of non‐irradiated fibroblasts. Arrow, downgrowth. HE staining.

Figure 3.

Histology of T3M‐1 squamous cell carcimona (SCC) cells with non‐irradiated (a, c and e) or irradiated WI‐26 VA4 fibroblasts (b, d and f). At 5 (a and b) 10 (c and d) and 15 days (e and f) in culture, WI‐26 VA4 fibroblasts replicate almost NIH 3T3 fibroblast‐induced stratification and downgrowth of SCC cells (Fig. 2). Arrow, downgrowth. HE staining.

Figure 4.

Number of NIH 3T3 fibroblasts with or without irradiation in collagen gel invasion assay. Irradiation drastically decreases number of fibroblasts, and does not allow them to proliferate. In contrast, non‐irradiated fibroblasts gradually grow. There is a significant difference of number of fibroblasts between with and without irradiation at 5, 10 and 15 days in culture (P < 0.001).

Proliferation and apoptosis of T3M‐1 SCC cells.  To elucidate effects of NIH 3T3 fibroblasts with or without irradiation on the growth and apoptosis of T3M‐1 SCC cells, we carried out immunhistochemistry with Ki‐67 and ss‐DNA antibodies, respectively. The rate of Ki‐67‐positive SCC cells with irradiated fibroblasts was greater than that of SCC cells with non‐irradiated fibroblasts at both 10 and 15 days in culture (Fig. 5a). Fibroblasts with or without irradiation did not affect apoptosis of SCC cells. There was no significant difference between the apoptotic rates of SCC cells with irradiated and non‐irradiated fibroblasts at 5, 10 and 15 days in culture (Fig. 5b). As the number of irradiated fibroblasts decreased to about 30% of that of non‐irradiated fibroblasts (Fig. 4), the apoptotic rate of SCC cells with 30 × 10⁵ irradiated fibroblasts was compared with that with 10 × 10⁵ non‐irradiated fibroblasts. However, the rate of SCC cells had no significant difference between both conditions above (Fig. 5c).

Figure 5.

Effect of NIH 3T3 fibroblasts with or without irradiation on growth (a) and apoptosis (b–c) of T3M‐1 squamous cell carcimona (SCC) cells. The growth marker Ki‐67‐positive nuclei (arrowheads in upper panel of a) of SCC cells are presented in brown. At 10 and 15 days in culture, there is significant difference of the rate of Ki‐67‐positive nuclei of SCC cells between with irradiated and non‐irradiated fibroblasts (P < 0.01), indicating that irradiated fibroblasts accelerate proliferation of SCC cells more greatly than non‐irradiated fibroblasts at 10 and 15 days in culture. However, SCC cells with irradiated and non‐irradiated fibroblasts have no significant difference of values at 5 days in culture. This suggests that irradiated fibroblasts do not yet enable SCC cells to induce hyperproliferation at least at 5 days. The apoptosis marker single stranded DNA (ssDNA)‐positive nuclei (arrowheads in upper panel of b) of SCC cells are presented in brown. At 5, 10 and 15 days in culture, there is no significant difference of the rate of ssDNA‐positive nuclei of SCC cells between with irradiated and non‐irradiated fibroblasts (P > 0.05). As the number of irradiated fibroblasts decreased to about 30% of that of non‐irradiated fibroblasts (see Fig. 4), apoptosis of SCC cells with 30 × 10⁵ irradiated fibroblasts was compared with that with 10 × 10⁵ non‐irradiated fibroblasts. However, apoptosis of SCC cells had no significant difference between both conditions above (P > 0.05).

Invasion of T3M‐1 SCC cells.  To elucidate effects of NIH 3T3 fibroblasts with or without irradiation on the invasion of T3M‐1 SCC cells, we examined the invasion frequency and depth of SCC cells. SCC cells with irradiated or non‐irradiated fibroblasts underwent their invasion into the gel (2, 3). The invasion degree of both conditions gradually increased along with culture time. However, the invasion frequency (Fig. 6a) and depth (Fig. 6b) of SCC cells with irradiated fibroblasts was always greater than those of SCC cells with non‐irradiated fibroblasts.

Figure 6.

Effect of NIH 3T3 fibroblasts with or without irradiation on both invasion frequency (a) and depth (b) of T3M‐1 squamous cell carcimona (SCC) cells into gel. Invasion frequency (arrowheads in upper panel of a) and depth (vertical line in upper panel of b) of SCC cells are determined, as explained in Materials and Methods. At 5, 10 and 15 days in culture, there are significant differences of both invasion frequency and depth of SCC cells between with irradiated and non‐irradiated fibroblasts (P < 0.05 at 5 days; P < 0.01 at 10 and 15 days), indicating that irradiated fibroblasts accelerate invasion of SCC cells more greatly than non‐irradiated fibroblasts.

Invasion‐, motility‐ and growth‐related molecule expression of T3M‐1 SCC cells.  At 10 days in culture, we examined the expression of cellular invasion‐, motility‐ and growth‐associated molecules in T3M‐1 SCC cells with irradiated or non‐irradiated NIH 3T3 fibroblasts. Immunohistochemistry showed that SCC cells with irradiated fibroblasts expressed MMP‐1, MMP‐9, laminin 5, filamin A, Raf‐1, MEK‐1 and ERK‐1/2 more greatly than those of SCC cells with non‐irradiated fibroblasts (Fig. 7). These results were supported by Western blotting (Fig. 8).

Figure 7.

Effect of NIH 3T3 fibroblasts with (b, d, f, h, j, l and n) or without irradiation (a, c, e, g, i, k and m) on immunohistochemical expression of matrix metalloproteinase‐1 (MMP‐1) (a and b), MMP‐9 (c and d), laminin 5 (e and f), filamin A (g and h), Raf‐1 (i and j), MEK‐1 (k and l) and ERK‐1/2 (m and n) of T3M‐1 squamous cell carcimona (SCC) cells. SCC cells with irradiated fibroblasts express all of these molecules above more strongly than those with non‐irradiated fibroblasts. MMP‐1, MMP‐9, laminin 5, filamin A, Raf‐1 and MEK‐1 are displayed in the cytoplasm of SCC cells, while ERK‐1/2 is expressed mainly in their cytoplasm and partly in the nucleus. Note that both laminin 5 and filamin A are more strongly displayed in invasive fronts of SCC cells with irradiated fibroblasts.

Figure 8.

A representative case of Western blot of matrix metalloproteinase‐1 (MMP‐1; 47 KDa) and ‐9 (92 KDa), laminin 5 (150 KDa), filamin A (250 KDa), Raf‐1 (74 KDa), MEK‐1 (43 KDa) and ERK1/2 (44/42 KDa) in T3M‐1 squamous cell carcimona (SCC) cells with irradiated or non‐irradiated NIH 3T3 fibroblasts. (a, b) SCC cells with irradiated fibroblasts express all molecules above more prominently than those with non‐irradiated fibroblasts, supporting immunohistochemical results (Fig. 7). (b) Densitometry analyses.

Focus formation of the genomic instability marker 53BP1 in T3M‐1 SCC cells.  To evaluate a genomic instability, we examined IRIF formation of 53BP1. Irradiated NIH 3T3 fibroblasts themselves formed IRIF of 53BP1 protein (Fig. 9b,d), whereas non‐irradiated fibroblasts did not organize the IRIF (Fig. 9a,c). Irradiated fibroblasts, in turn, enabled non‐irradiated T3M‐1 SCC cells to enhance IRIF formation of 53BP1 (Fig. 9b,d). At 0.5, 1 and 24 h and 5 days in culture, the rate of 53BP1 IRIF‐positive SCC cells with irradiated fibroblasts was significantly greater than that with non‐irradiated fibroblasts (Fig. 9e), although there is no significant difference of the rates of 53BP1 IRIF formation in SCC cells between both situations at 15 days in culture (Fig. 9e).

Figure 9.

Effect of NIH 3T3 fibroblasts with (b and d) or without irradiation (a and c) on p53‐binding protein‐1 (53BP1) irradiation‐induced foci (IRIF) formation of fibroblasts and T3M‐1 squamous cell carcimona (SCC) cells by immunofluorescence. At 24 h and 5 days in culture, irradiated fibroblasts show 53BP1 IRIF (arrows in b and d) in their nuclei, whereas non‐irradiated fibroblasts have no. 53BP1IRIF (a and c). Formation of 53BP1 IRIF (arrowheads in insets of b and d) within nuclei of SCC cells with irradiated fibroblasts is presented in yellow green. Irradiated fibroblasts promote 53BP1 IRIF in SCC cells (b and d) more prominently than non‐irradiated fibroblasts (a and c). At 0.5, 1, and 24 h, and 5 days in culture, there is significant difference of the rate of 53BP1 IRIF‐positive nuclei of SCC cells between with irradiated and non‐irradiated fibroblasts (P < 0.05), although SCC cells with irradiated and non‐irradiated fibroblasts have no significant difference of values at 15 days in culture. This suggests that irradiated fibroblasts enable non‐irradiated SCC cells to enhance 53BP1 IRIF formation in the nuclei more greatly than those with non‐irradiated fibroblasts.

Radiation‐induced bystander effects on behavior of HEp‐2 cells as another SCC cell type.  To estimate whether irradiated fibroblasts affect the biological behavior of SCC cell type other than T3M‐1 SCC cell type, we examined the bystander effects above, using HEp‐2 SCC cells. At 10 days, irradiated NIH 3T3 fibroblasts promoted the growth and expression of ERK‐1/2 and 53BP1 IRIF in HEp‐2 cells more extensively than non‐irradiated counterparts, but irradiated fibroblasts did not affect the invasion, apoptosis and MMP‐1 expression of HEp‐2 cells (Fig. 10).

Figure 10.

Effects of NIH3T3 fibroblasts with or without irradiation on invasion, growth, apoptosis and matrix metalloproteinase‐1 (MMP‐1), ERK‐1/2 and p53‐binding protein‐1 (53BP1) irradiation‐induced foci (IRIF) expression of HEp‐2 squamous cell carcimona (SCC) cells at 10 days in culture. Irradiated fibroblasts promote cellular stratification (b), 53BP1 IRIF formation (d and e), growth (f) and ERK‐1/2 expression (h) of SCC cells more prominently than non‐irradiated fibroblasts (a, c, e, f and h), but fibroblasts with or without irradiation do not affect invasion (a and b), apoptosis (g) and MMP‐1 expression (h) of SCC cells. *P < 0.05.

Possible mediators of radiation‐induced bystander effects.  In culture assemblies of T3M‐1 cells at 10 days, only irradiated fibroblasts, but not non‐irradiated ones, expressed TGF‐β1 (Fig. 11c,d), whereas TNF‐α were not detected in any fibroblasts (Fig. 11a, b). In the case of HEp‐2 cells, similar results to above were detected (data not shown). This suggests that TGF‐β1 may be a mediator responsible for the radiation‐induced bystander effects.

Figure 11.

TNF‐α and TGF‐β1 expression of NIH 3T3 fibroblasts with or without irradiation in their co‐culture with T3M‐1 squamous cell carcimona (SCC) cells by immunohistochemistry at 10 days in culture. Fibroblasts with (b) or without irradiation (a) do not display TNF‐α. Irradiated fibroblasts express TGF‐β1 (d), but non‐irradiated fibroblasts (c) do not express it.

Are radiation‐induced bystander effects reversible or irreversible?  The invasion and 53BP1 IRIF formation of irradiated fibroblast‐affected T3M‐1 cells on collagen gel alone were similar to those of non‐irradiated fibroblast‐affected T3M‐1 cells on collagen gel alone (Fig. 12). In the case of HEp‐2 cells, similar results to above were detected (data not shown). This suggests that the radiation‐induced bystander effects may be reversible at least in our culture system.

Figure 12.

Reversibility of radiation‐induced bystander effects of irradiated NIH 3T3 fibroblasts on invasion and p53‐binding protein‐1 (53BP1) irradiation‐induced foci (IRIF) formation of T3M‐1 squamous cell carcimona (SCC) cells at 10 days in culture. The reversibility of the bystander effects is examined according to the way described in Materials and Methods. The invasion (b) and 53BP1 IRIF formation (d and e) of irradiated fibroblast‐affected T3M‐1 cells on collagen gel alone were similar to those (a, c and e) of non‐irradiated fibroblast‐affected T3M‐1 cells on collagen gel alone. Arrow, invasion.

Discussion

Here we have shown for the first time that irradiated fibroblasts, even with their drastically decreased number, promote the growth and invasion of non‐irradiated T3M‐1 SCC cells more prominently than non‐irradiated fibroblasts, supporting other studies that found that irradiated fibroblasts accelerate the invasive growth of non‐irradiated adenocarcinoma cells of the breast and pancreas.^{( 5} , ^{7 )} We have also demonstrated that irradiated fibroblasts enhance expression of some growth‐, invasion‐ and motility‐related molecules in the cancer cells more greatly than non‐irradiated fibroblasts. The irradiated fibroblast‐induced phenomena above are elicited through bystander mechanism under cancer–stromal cell interaction, because SCC cells themselves are not irradiated. This suggests that effects of irradiation on not only cancer cells, but also the stromal cell types should be taken into account for cancer radiotherapy. Finally, we also have shown that irradiated fibroblasts promote the growth of the laryngeal SCC cell type HEp‐2 cells more prominently than non‐irradiated fibroblasts, but the irradiated fibroblasts did not affect their invasion and apoptosis. Thus, the bystander effects seem critical for the biological behavior of at least two SCC cell types. It is also likely that the characteristics of radiation‐induced bystander effects may be SCC cell type‐dependent.

In the present study, the precise mechanistic basis of irradiated fibroblast‐induced phenomena above remains unclear. Irradiation is suggested to induce 53BP1 IRIF formation that indicates cellular DNA damage and genomic instability.^{( 10} , ¹² , ¹⁴ , ^{15 )} As described here, irradiated fibroblasts indeed form IRIF of 53BP1 protein, whereas non‐irradiated fibroblasts do not organize the IRIF. Furthermore, SCC cells with irradiated fibroblasts gain more increased number of IRIF of 53BP1 in a bystander manner than those with non‐irradiated fibroblasts.^{( 15 )} This suggests that irradiated fibroblasts under cancer–stromal cell interaction may cause SCC cells to enhance their own non‐direct DNA damage responses that lead to invasive growth of SCC cells. Although the number of irradiated fibroblasts is prominently lower than that of non‐irradiated fibroblasts, irradiated fibroblasts do induce more drastic changes of SCC cells than non‐irradiated fibroblasts. This suggests that the irradiated fibroblast‐induced bystander effect is a powerful player in the biological behaviors of SCC cells. In addition, some radiation‐induced bystander effects have been recently shown to be mediated by TNF‐α and TGF‐β1 produced by irradiated cells.^{( 9 )} Our current study has shown that only irradiated fibroblasts expressed TGF‐β1. Taken together, these results suggest that irradiated fibroblast‐induced bystander effects on the invasive growth of SCC cells may be mediated by TGF‐β1 produced by irradiated fibroblasts. To address the mediators and their actions in the bystander effects in more detail, further studies are needed.

As shown in this study, the invasion and 53BP1 IRIF formation of irradiated fibroblast‐affected SCC cells on collagen gel alone were similar to those of non‐irradiated fibroblast‐affected SCC cells on collagen gel alone. This suggests that irradiated fibroblasts may induce reversible effects in SCC cells at least in our culture system.

Recently, the MAPK cascade is suggested to be involved in the growth‐signaling pathway of various cell types of normal or neoplastic origin in response to many extracellular stimuli.^{( 31 )} As described in that study, irradiated fibroblasts enable SCC cells to increase Raf‐1, MEK‐1 and ERK‐1/2 expression of the MAPK cascade along with their hyperproliferation. In the radiation‐induced bystander effect, MAPK pathways are also suggested to play a critical role.^{( 31} , ^{32 )} Irradiated fibroblasts also enhance c‐Met and Ras expression of the tumor cells.^{( 7 )} Some studies have demonstrated that c‐Met and Ras signals are closely linked to the activation of MAPK cascade.^{( 20 )} Taken together, these results suggest that irradiated fibroblasts promote the growth of SCC cells through the activation of the c‐Met, Ras/MAPK pathway in a bystander manner, in which effects of paracrine and cell–cell contact may be involved.

The invasion and metastatic spread of cancer cells are regulated by various factors such as proteolytic, locomotive and angiogenic agents.^{( 33} , ³⁴ , ^{35 )} In our current study, irradiated fibroblasts enhance the expression of the proteolytic enzyme MMP‐1, MMP‐9^{( 21 )} and the motility factors laminin 5^{( 22 )} and filamin A^{( 23 )} in T3M‐1 SCC cells more extensively than non‐irradiated fibroblasts. The invasive degree of the cancer cells with irradiated fibroblasts is greater than that with non‐irradiated fibroblasts. This suggests that irradiated fibroblasts accelerate the invasion of T3M‐1 SCC cells through the increased expression of MMP‐1 and 9 and filamin A in a bystander manner under cancer–stromal cell interaction.

As described here, apoptosis of SCC cells had no significant difference between irradiated and non‐irradiated fibroblasts. Even with same number of irradiated and non‐irradiated fibroblasts, apoptosis of SCC cells had no change. This suggests that the irradiated fibroblast‐induced bystander effect is not involved in the apoptotic pathway of SCC cells.