TGF-β receptor inhibitor LY2109761 enhances the radiosensitivity of gastric cancer by inactivating the TGF-β/SMAD4 signaling pathway

High levels of TGF-β were associated with poor overall survival (OS) of GC patients

Analyzing the Cancer Genome Atlas database (Figure 1A) showed that GC patients with low TGF-β1 expression generally had a prolonged OS (median survival time undefined) and that upregulated TGF-β1 was significantly correlated with poor OS (median survival time: 19.32 months, n=203, P=0.023). Figure 1B shows that GC patients with low TGF-β2 expression had a prolonged OS (median survival time 68.99 months) compared with that of patients with high TGF-β2 expression (median survival time 26.45 months, n=205, P=0.0409). These data showed that TGF-β may be a better prognostic marker than grading or tumor necrosis metastasis (TNM) staging (not significantly related to GC patient prognoses; Supplementary Figure 1A).

Figure 1. Relationship between the TGF-β expressions with the overall survival and radioresistance in GC patients. (A) Kaplan-Meier survival curves for high and low mRNA TGF-β1 expression level in GC patients. (B) Kaplan-Meier survival curves for high and low mRNA TGF-β2 expression level in GC patients. (C) Immunohistochemical analysis: representative immunohistochemistry images of the TGF-β1, TGF-β2, SMAD4, E-CA and N-CA expressions in biopsy samples of GC patients. (a) Specimen from a patient whose tumor sample showed a complete response to radiotherapy (radiosensitive GC patient). (b) Specimen from a patient whose tumor sample showed no change after radiotherapy (radioresistant GC patient). Original magnification, ×400. P<0.05.

Dysregulation of TGF-β and EMT markers in GC patients with radioresistance

Twenty-four cancer tissue samples were harvested from GC patients before radiotherapy (diagnostic biopsies). Computed tomography images from before and two months after radiotherapy were reviewed to examine the radiosensitivity of these samples. Sixteen specimens responded completely to radiotherapy (radiosensitive tissue), while 8 were nonreactive after radiotherapy (radioresistant tissue). Table 1 lists the patients’ characteristics. TGF-β1 and TGF-β2 were strongly upregulated in radioresistant GC patients compared with those in radiosensitive patients (Figure 1C). The downstream molecule, SMAD4, was downregulated, and EMT makers were dysregulated in radioresistant GC patients (E-cadherin was downregulated, and N-cadherin was upregulated). The staining indices differed statistically (Supplementary Figure 1B).

Establishment of radioresistant GC cell line RR

The radioresistant GC cell line RR (SGC-7901-R and AGS-R) were established by exposing parental GC cells to fractioned irradiation for 6 months at a total dose of 60 Gy and validated via colony-formation assays [25]. Colony formation and survival fractions were significantly increased in the RR cells compared with those of the parental GC cells (Figure 2A; P<0.05). The reactive oxygen species (ROS) levels in RR cells were significantly increased compared with those of the controls (Figure 2B; P<0.05). Real-time PCR (Supplementary Figure 2A) and western blot (Figure 2C) showed that TGF-β1 and TGF-β2 were overexpressed, SMAD4 and E-cadherin were downregulated, and N-cadherin was upregulated in RR cells. In addition, inflammatory factors IL-1β and IL-6 and immune checkpoint PD-L1 were significantly overexpressed, and the DNA double-strand break biomarker, γ-H2AX, was significantly lower in RR cells. The sphere-formation assay showed that RR cells generated sphere cells and showed self-renewal potential compared with the parental GC cells (Figure 2D). Flow cytometry showed that SGC-7901-R cells possessed considerably enhanced expression of the putative CSC markers, CD24 (P<0.0001) and CD133 (P=0.0018; Figure 2E). Cell apoptosis analysis showed that irradiation induced fewer SGC-7901-R cells into apoptosis (P=0.0006; Figure 2F). Wound-healing assays indicated that the wound-healing percentages were significantly increased in SGC-7901-R cells (mean ± SEM: 60.43% ± 0.9563% vs 26.73% ± 1.099%, n=3, P<0.0001; Figure 2G). Transwell assays showed that the numbers of invaded cells were significantly increased in SGC-7901-R cells (449 vs 335, n=3, P=0.0006; Figure 2H). The same phenotype was also observed in AGS-R cells (Supplementary Figure 2B–2E).

Figure 2. Different biological characteristics in radioresistant RR and parental GC cells. (A) Clonogenic survival assay: the ability of colony formation between RR and parental GC cells after a range of radiation doses. (B) Flow cytometry analysis: ROS levels between RR and parental GC cells. (C) Western blot analysis: representative results of the TGF-β1, TGF-β2, SMAD4, E-CA, N-CA, IL-1β, IL-6, PD-L1 and γ-H2AX expressions in RR and parental GC cells. (D) Sphere formation assays: captured images of sphere formation assays in RR and parental GC cells. (E) Flow cytometry analysis: the expressions of the CSC markers CD24 and CD133 between SGC-7901 and SGC-7901-R cells. (F) Flow cytometry analysis: cell apoptosis between SGC-7901 and SGC-7901-R cells after irradiation treatment. (G) Cell migration assay: captured images of wound healing assay of SGC-7901 and SGC-7901-R cells, columns indicated the percentage of wound healing width of SGC-7901 and SGC-7901-R cells. (H) Cell invasion assay: captured images of transwell assay of SGC-7901 and SGC-7901-R cells, columns indicated the invaded cell percentage of SGC-7901 and SGC-7901-R cells. All data represent three independent experiments, mean ± SEM, **P<0.01, ***P<0.001, ****P<0.0001.

TGF-β inhibitor LY enhanced GC cell sensitivity to irradiation

Since TGF-β was upregulated in RR, we examined whether the TGF-β inhibitor LY could lead to radiosensitivity in GC cells. CCK-8 assays (Figure 3A, 3B) indicated that RR clones were more sensitive to LY, and LY pretreatment inhibited GC cell proliferation dose- dependently following irradiation (P<0.01). Flow cytometry showed that LY combined with IR increased GC cell apoptosis compared with that of IR alone (P<0.0001; Figure 3C). Clonogenic assays showed that LY significantly reduced the clonogenic ability of GC cells when combined with IR (P=0.0021 and P=0.0005; Figure 3D). The wound-healing (Figure 4A) and transwell assays (Figure 4B) confirmed that LY significantly decreased GC cell migration and invasion abilities after irradiation treatment (P<0.05).

Figure 3. Effects of LY on radiosensitivity of GC cell lines. (A) CCK-8 assay: the proliferation of RR and parental GC cells influenced by LY. (B) CCK-8 assay: the proliferation of GC cells influenced by LY with or without irradiation treatment. (C) Flow cytometry analysis: the cell apoptosis in GC cells influenced by LY with or without irradiation treatment. (D) Clonogenic survival analysis: the clonogenic ability of GC cells influenced by LY with or without irradiation treatment. All data represent three independent experiments, mean ± SEM, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 4. Effects of LY on the migration and invasion abilities of irradiation induced GC cell lines. (A) Cell migration assay: captured images of wound healing assay in GC cells influenced by LY with or without irradiation treatment, columns indicated the percentage of wound healing width in four groups. (B) Cell invasion assay: captured images of transwell assay in GC cells influenced by LY with or without irradiation treatment, columns indicated the invaded cell percentage in four groups. All data represent three independent experiments, mean ± SEM, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

LY inhibited the TGF-β/SMAD4 pathway and altered the expression of other downstream molecules in GC cells

To determine the potential mechanism of LY in GC radioresistance, we detected several TGF-β-related signaling pathways in GC cell lines. PCR (Supplementary Figure 2F), western blotting (Figure 5A) and enzyme-linked immunosorbent assay (ELISA; Figure 5B) revealed that LY decreased TGF-β1 and TGF-β2 overexpression after irradiation treatment. LY enhanced SMAD4 expression, reversed the irradiation-induced EMT in GC cells and decreased the expressions of inflammatory-related factors IL-1β and IL-6 and immune checkpoint PD-L1. Pretreating SGC-7901 cells with cycloheximide to inhibit protein synthesis showed that TGF-β protein stability was significantly increased in the IR group compared with that in the control group (Figure 5C). Immunofluorescence images showed that LY inhibited phospho-SMAD2 expression after IR treatment (Figure 6A). Flow cytometry assays showed that LY significantly reduced the overexpression of CSC markers CD24 and CD133 in irradiated GC cell lines (P<0.05; Figure 6B and Supplementary Figure 3A).

Figure 5. Effect of LY with or without irradiation on the gene expression of GC cell lines. (A) Western blot analysis: representative results of the TGF-β1, TGF-β2, SMAD4,E-CA, N-CA, IL-1β, IL-6, PD-L1 and γ-H2AX expressions in GC cells after LY with or without irradiation treatment. (B) ELISA: the expression of TGF-β1, TGF-β2 and IL-6 in culture medium of GC cells after LY with or without irradiation treatment. (C) Western blot analysis: the protein stability of TGF-β1 and TGF-β2 regulated in response to irradiation. All data represent three independent experiments, mean ± SEM, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 6. Effect of LY with or without irradiation on the expression of p-Smad2 and CSC markers in GC cell lines. (A) Immunofluorescence assay: representative results of the p-Smad2 expressions in GC cells after LY with or without irradiation treatment. (B) Flow cytometry analysis: the representative results of the expressions of the putative stem cell markers CD24 and CD133 in GC cells with or without LY and irradiation treatment.

LY enhanced the radiosensitivity of a GC xenograft tumor model in vivo

To verify the radiosensitization effect of LY on GC in vivo, we established SGC-7901 subcutaneous xenograft tumor models in BALB/c nude mice and examined the effect of LY alone, IR alone, and a combination of LY with IR on the growth of subcutaneous xenograft tumors. The tumor volume was measured twice weekly, and the tumor volumes of the control, LY, and IR-alone groups were larger than those of the LY with IR group (Figures 7A and 7B). Tumor growth was inhibited in mice treated with LY and IR combination (approximately 85% inhibition at day 18) and tumor growth curves were delayed in IR and LY with IR group, (the tumor doubling time of control and LY group was 9 days and 15 days, respectively, and the tumor stopped growing and the tumor volume gradually decreased in IR and LY with IR group 3 days after treatment; P<0.0001; Figures 7C). The tumor weight in the LY with IR group was significantly reduced compared with that of the IR group (P<0.0001; Figures 7D).

Figure 7. Effects of LY on radiosensitivity of gastric cancer in vivo. (A and B) Changes in tumor sizes: images of BALB/c nude mice with tumors and resected tumors after LY with or without irradiation treatment. (C) Changes in tumor volumes: the statistical curves of gastric tumor volumes influenced by LY with or without irradiation treatment. (D) Changes in tumor weights: the final tumor weight after LY with or without irradiation treatment for 18 days. (E) TUNEL assays: representative images of the apoptostic cells in GC xenografts influenced by LY with or without irradiation treatment. (F) Changes in genetic expression in GC xenografts: the immunohistochemical analysis of the TGF-β1, TGF-β2, SMAD4, E-CA, N-CA and Ki-67 expressions in GC xenografts influenced by LY with or without irradiation treatment. Original magnification, ×400. *P<0.05, ****P<0.0001.

LY affected the radiation-induced genetic expression in vivo

TUNEL assay indicated that apoptotic cells were increased in the LY with IR group compared with those of the IR-alone group (Figure 7E). The expressions of TGF-β1, TGF-β2, SMAD4, E-cadherin, N-cadherin and Ki-67 in tumor xenografts were detected via immune- histochemical staining. Figure 7F from left to right shows the control, LY-alone, IR-alone and LY with IR groups. TGF-β1, TGF-β2, and N-cadherin expressions were significantly upregulated in tumor xenografts treated with IR alone but were decreased after LY with IR treatment. The expressions of SMAD4 and E-cadherin were significantly downregulated in tumor xenografts with IR treatment alone and were restored in the LY pretreatment group. The LY with IR group showed decreased Ki-67 expression. The staining indices differed statistically (Supplementary Figure 3B). These results were consistent with the in vitro results and confirmed that LY enhanced GC radiosensitivity via TGF-β/SMAD4 signaling pathways and promoted EMT reversal in vivo.

Patient data sources and clinical patient specimens

TGF-β mRNA expression and related clinical data on 416 GC patients were obtained from TCGA database. Matching analysis was conducted to select 100 GC patients with low TGF-β1 expression, 103 GC patients with high TGF-β1 expression, 102 GC patients with low TGF-β2 expression and 103 GC patients with high TGF-β2 expression. Survival curves were drawn according to the patients’ survival information.

The study population consisted of 24 patients with advanced GC diagnosed at the Department of Radiation and Medical Oncology in Zhongnan Hospital of Wuhan University from 2012 to 2018. The study was conducted in accordance with the recommendations of the Zhongnan Hospital Ethics and Scientific Committee, and written informed consent was received from each patient before tissue acquisition. We determined 8 radioresistant GC patients (no tumor reduction after radiotherapy and no response to radiotherapy) and 16 radiosensitive GC patients (obvious tumor reduction after radiotherapy or complete response to radiotherapy) according to computed tomography imaging results 2 months after radiotherapy. Tissue samples were obtained from gastroscopic diagnostic biopsies before radiotherapy.

Immunohistochemistry analysis of tumor tissues in GC patients

Immunohistochemistry staining was used to assess gene expression in radioresistant and radiosensitive GC patients. Paraffin-embedded patient tissue sections were subjected to antigen retrieval using 0.01 M citrate buffer (pH 6.0). After blocking the endogenous peroxidase, the tissue sections were incubated with primary antibody overnight in a 4 °C wet box. The antibodies used were as follows: TGF-β1 (#ab179695, Abcam, 1:200), TGF-β2 (#P61812, CUSABIO, 1:100), SMAD4 (#46535, CST, 1:200), E-CA (#3195, CST, 1:200), and N-CA (#13116, CST, 1:200). After washing the sections with phosphate-buffered saline (PBS), horseradish peroxidase (HRP)-labeled secondary antibody was added to incubate at 37 °C for 20 minutes. Color development, counterstaining, dehydration and sealing were performed in sequence, and the results were observed under an optical microscope and judged by three pathologists.

Cell culture and treatments

The human GC cell lines, AGS and SGC-7901, were purchased from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The radioresistant GC cell line, RR, was obtained via sublethal dose of irradiation using the Small Animal Radiation Research Platform (PXI X-RAD 225Cx, CT, USA). Cells were cultured in RPMI-1640 (HyClone, USA) supplemented with 10% fetal bovine serum (Gibco, USA) and 1% penicillin-streptomycin sulfate (Invitrogen, USA). TGF-β inhibitor LY2109761 was purchased from Selleckchem (Houston, TX, USA), constituted in dimethyl sulfoxide and stored at -80 °C. All cells were planted and irradiated with 12-Gy X-rays (pretreated with LY 10 μmol/L or not), and cells were exposed to LY 2 h prior to irradiation.

Establishment and validation of radioresistant GC cell line

The GC cell lines were irradiated with 2-Gy X-rays at 300 cGy/min and an irradiation range of 20 cm × 20 cm to establish the radioresistant GC cell line, RR. The total exposure time was up to 6 months, and the total exposure dose was 60 Gy. After stable passage, a colony-formation assay was used to measure the radioresistance of the RR cells. Cells were seeded into six-well culture plates and exposed to a range of radiation doses (2–10 Gy). Ten days after irradiation, cell colonies were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The colonies were counted (up to 50 cells were counted as one colony). The plating efficiency (PE) and survival fraction (SF) were calculated as follows: PE = (colony number/number of inoculated cells) × 100%; SF = colonies counted/(cells seeded × [PE/100]).

Detection of ROS levels

The collected parental GC cells and radioresistant GC cell lines were incubated with dichlorofluorescein diacetate (DCFH-DA) dye (Beyotime, China) for 25 min at 37 °C, and the ROS levels were detected via flow cytometry (BD Biosciences, USA).

Sphere formation assays

Sphere-formation assays were conducted to detect GC cell self-renewal. Briefly, 2×10⁴ cells were inoculated in a 10-cm ultra-low-adhesive culture dish, and 10 mL of Dulbecco’s modified Eagle’s medium (DMEM)/F12 containing B27 (1×) and 5 ng/mL epidermal growth factor was added and incubated for 5–7 days to determine spheroid cell growth.

Cell proliferation assay

A Cell Counting Kit-8 (CCK-8) assay (Dojindo, Japan) was used to measure cell proliferation. GC cells were plated into 96-well plates and treated with an increasing concentration gradient of LY or irradiation (pretreated with LY at 10 μmol/L or not) with a range of radiation doses. After 48 h of incubation, cell viability was measured with CCK-8 following the manufacturer’s instructions. Data are expressed as the percentage of cell proliferation relative to the nonradiated control group and the percentage of cell activity reduced by LY.

Cell apoptosis assay

Cells were treated as described above. After 24 h of incubation, apoptosis was measured via flow cytometry with 5 μL annexin V and 10 μL propidium iodide cell stains (Multi Sciences, China). Data are expressed as the percentage of apoptotic cells.

Colony-formation assay

Cells were planted in 6-well plates (600/well). After treatment, the colony-formation ability was detected as described above. Data are expressed as the percentage of the nonradiated control group.

Cell migration and invasion assessment

Cell migration ability was examined using wound-healing assays. Cells were cultured for 24 h and grown until confluency. Confluent cell monolayers were disrupted by standardized wound scratching with a sterile 200-μL pipette tip, then treated as described above and incubated in 1% FBS medium for 48 h. Cell images of the wound area were captured with an optical microscope (Olympus, Japan). Data are expressed as the percentage of the wound-healing width. Cell invasion ability was detected via transwell assay. At 48 h after treatment, 1 × 10⁴ cells were collected and seeded into the upper chamber, which was precoated with Matrigel (BD Bioscience, USA). Medium containing 10% FBS was added to the lower chamber. After 24 h of incubation, cells that migrated to the lower surface were fixed in methanol, stained and counted under a microscope (Olympus, Japan). Data are expressed as the percentage of the control.

Flow cytometry assay for putative stem cell markers

Expression of putative stem cell markers CD24 and CD133 in GC cell lines were evaluated by fluorescence-activated cell sorting (FACS) analysis. After treatment, cells were washed with cooled PBS, resuspended, and stained with CD24 and CD133 antibodies (Biolegend, USA). The putative stem cell markers, CD24 and CD133, were detected using a flow cytometer (BD Biosciences, USA). Data are expressed as the percentage of the CD24⁺ or CD133⁺ cells.

Western blot

Cells were washed with PBS, lysed in 200 μL of protein lysate (containing 100 mg/L phenylmethanesulfonyl fluoride) at 4 °C for 30 min, then the supernatant was centrifuged. The protein was quantified using a BCA Protein Assay kit (Beyotime Institute of Biotechnology, China). Protein (20 μg) was thoroughly mixed with loading buffer, boiled for 5 min, separated by 7.5% sodium dodecyl sulfate-polyacrylamide gel and transferred onto a PVDF membrane (0.22 μm; Millipore). After blocking with 5% skim milk powder for 1 h, the membranes were incubated with the following primary antibodies: TGF-β1 (#ab179695, Abcam, 1:1000), TGF-β2 (#P61812, CUSABIO, 1:600), SMAD4 (#46535, CST, 1:1000), E-CA (#3195, CST, 1:1000), N-CA (#13116, CST, 1:1000), PD-L1 (#MA878942A1m, CUSABIO, 1:600), IL-1β (#PA003023, CUSABIO, 1:600), IL-6 (#PA06757A0Rb, CUSABIO, 1:600), γ-H2AX (#155226, mlbio, 1:600) and β-actin (#60008-1-Ig, Proteintech, 1:5000).

ELISA

The amounts of TGF-β1, TGF-β2 and IL-6 secreted in the culture medium were detected with ELISA kits (DAKEWE, China) following the manufacturer’s instructions. Data are expressed as the quantity of secreted protein in the culture medium.

Immunofluorescence

phospho-SMAD2 expressions were detected via immunofluorescence assays. Cells were collected, fixed with 4% paraformaldehyde, permeabilized with Triton X-100 and washed with PBS. Primary antibody (bioss, China) was added and incubated in a wet box at 37 °C for 2 h. Secondary antibodies to the corresponding species of the primary antibody were added and incubated for 50 min in the dark. The cells were stained with DAPI for 5 min, then viewed under an inverted fluorescence microscope to observe and collect images.

Animal model establishment and treatment

Four- to six-week-old BALB/c male nude mice were fed at the animal experiment center of Zhongnan Hospital, and all procedures were approved by the Animal Care and Use Committee of Wuhan University. A total of 1×10⁶ SGC-7901 cells were injected subcutaneously in the left hind limbs of the mice. Twenty mice were randomly divided into 4 groups: the control, LY, IR, and LY plus IR at n=5/group. Treatments started after 20 days when the tumors had reached at least 200 mm³. Tumors in the IR and LY plus IR groups were irradiated with fractionated radiotherapy (4×6 Gy, twice weekly); the LY and LY plus IR groups were treated with LY orally twice daily (50 mg/kg, on days 1–5 each week) until the animals were sacrificed. Tumor volume was directly measured twice weekly using a Vernier caliper (volume=length × width × width × 0.5).

TUNEL assays and tumor immunohistochemistry in the mice

All animals were sacrificed via carbon dioxide inhalation after 18 days of treatment, and the tumor tissues were removed and measured. The tissues were sectioned, TUNEL assays (Roche, China) were used to detect cell apoptosis in vivo following the manufacturer’s instructions, and gene expression at the protein level was detected via immunohistochemistry as described above.

Statistical analysis

Each experiment was performed at least three times to ensure the repeatability of the results. Statistical analysis was performed using GraphPad Prism 7.0 (GraphPad Software). The results were analyzed using Student’s t-test when two groups were compared or one-way analysis of variance when more than two groups were compared. OS curves were plotted using the Kaplan-Meier method and compared using a log-rank test. Differences were considered statistically significant at P<0.05.

Ethics approval

The study was conducted in accordance with the recommendations of Zhongnan Hospital Ethics and Scientific Committee and written informed consent was received from each patient before tissue acquisition (Ethical approval number: 2019165). All the procedures in animal experiments were approved by the Animal Care and Use Committee of Wuhan University.