Wnt/β-catenin signaling activation by TIMP1 confers cisplatin-resistant gastric cancer cells to malignant behaviors and epithelial–mesenchymal transition

Gene expression profiling interactive analysis (GEPIA)

GEPIA (http://gepia.cancer-pku.cn/) [18], a web-based tool to deliver fast and customizable functionalities based on TCGA and GTEx data, was applied to analyze the predictive significance of the TIMP1 expression level on tumor stage and overall survival in GC.

GEO dataset analysis

GSE56807 dataset [19] was utilized to analyze the relative expression level of TIMP1 mRNA in five paired GC and adjacent normal tissues.

Cell culture

AGS cell line was obtained from ATCC, and BGC-823 cell line was kindly provided by Cell Bank, Chinese Academy of Sciences. DDP-resistant GC cell lines (DPP/R) were developed by continuously exposing the parental ones to gradient DDP in a previous study [20]. To maintain DDP resistance, DPP/R cell lines were continuously exposed to 1 μg/mL DDP. DDP-resistant AGS cells, denoted as AGS/DDP, were used to perform further analysis.

During the biological functional experiments in this study, the concentrations of DDP and AR-A014418 (Cat. no A3230, Sigma-Aldrich, MO, USA) were 1 μg/mL and 2.5 μM, respectively.

Reverse-transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted as per standard protocols. Afterward, on the 7,500 Real-Time PCR System, RT-PCR was performed with SYBR green reagent (Cat. No. 4309155, Applied Biosystems, CA, USA). This study considered GAPDH an internal reference gene to quantitate the mRNA expression of TIMP1 based on the 2^−ΔΔCt method [21]. The primer sequences were as follows: TIMP1 forward primer (F): 5′ -GGGGCCATTCCCTAAATCGT-3′, (R) reverse primer: 5′-CTGGTCCGTCCACAAGCAAT-3′; GAPDH (F): 5′ -AGTTAATGCCGCCCCTTACC-3′, (R): 5′-CAGGGCTGACTACAAACCCA-3′.

Cell transfection

The cells were divided into six groups to receive different transfections as follows: si-NC group, cells transfected with an irrelevant nucleotide (si-NC (F): 5′-UUCUCCGAACGUGUCACGUTT-3′, (R): 5′-ACGUGACACGUUCGGAGAATT-3′) to act as a negative control; (2) si-TIMP1 groups (#1, #2, and #3),cells transfected with a short interfering RNA (siRNA) that specifically targeted TIMP1 (si-TIMP1#1 (F): 5′-GGAACGGAAAUUUGCACAUTT-3′, (R): 5′- AUGUGCAAAUUUCCGUUCCTT-3′; si-TIMP1#2 (F): 5′-GCACAGUGUUUCCCUGUUU-3′, (R): 5′-AAACAGGGAAACACUGUGC-3′; si-TIMP1#3 (F): 5′-CCACCCUUAUACCAGCGUUA-3′, (R): 5′-UAACGCUGGUAUAAGGUGG-3′); (3) vector group, cells transfected with empty vector; (4) TIMP1 OE group, cells transfected with the overexpression vector of TMIP1 (TIMP1 OE) that was synthesized by GenePharma (Shanghai, China). Briefly, both AGS cells were transfected using Lipofectamine^® 2000 (Invitrogen, CA, USA) following the manufacturer’s manuals, and then harvested 48 h post-transfection to evaluate the transfection efficiency using RT-PCR and Western blot. The siRNA with the highest efficiency was chosen for the subsequent experiment.

Western blot

Western blot is often used in research to separate and identify proteins [22]. After finishing the treatments according to the study design, cells were harvested and subsequently lysed with RIPA buffer on ice for 20 min at 4 °C to acquire the total protein. Nuclear fractions were prepared using a nuclear extraction kit (Cat. No. ab113474; Abcam, MA, USA). Equivalent amounts of protein were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels and subsequently electrotransferred onto polyvinylidene fluoride membranes (Millipore Sigma, MA, USA). Then, the membranes were blocked and then incubated with primary antibodies against TIMP1 (Cat. No. ab211926, Abcam, MA, USA), CD44 (Cat. No. ab19857, Abcam, MA, USA), CD47 (Cat. No. AF6423, Beyotime, Beijing, China), Nanog (Cat. No. AF1912, Beyotime, Beijing, China), Oct4 (cat.no ab189524, Abcam, MA, USA), E-cadherin (cat.no AF6759, Beyotime, Beijing, China), N-cadherin (Cat. No. AF5237, Beyotime), GSK3β (Cat. No. AF1543, Beyotime, Beijing, China), phospho-GSK3β (Cat. No. AF1531, Beyotime, Beijing, China), β-catenin (Cat. No. AF1189, Beyotime, Beijing, China), and GAPDH (Cat. no. ab8245, Abcam, MA, USA). Then, the membranes were rinsed thrice with a secondary antibody before incubation for 1 h (Cat. no ab6721, Abcam, MA, USA). The protein bands were visualized using an ECL kit (Cat. no. ab65623, Abcam, MA, USA).

CCK-8 assay

CCK-8 assay was exploited to determine cell viability as previously reported [23]. Briefly, GC cells were seeded in 96-well plates after being digested and resuspended. After complete adherence, cells were treated with DMEM (Gibco, CA, USA) containing diverse concentrations of DDP for three time points (24, 48, and 72 h). The cells added with 0.1% dimethyl sulfoxide were used as the experimental control, and wells containing only DMEM were set as the blank group. After the incubation, the medium was replaced with the CCK-8 reagent (Dojindo, Tokyo, Japan) (10%, v/v, dissolved in DMEM). After incubating for 2 h, the absorbance was examined with a microplate reader at 450 nm to determine the cell viability. To determine the resistance of cells to DDP, the median inhibitory concentration (IC50) of DDP for 24, 48, and 72 h on cells was calculated based on the cell viability curve.

Sphere formation assay

The sphere formation assay provides a useful tool to assess the stem cell population residing in tumors or cancerous cell lines [24]. Cells were plated on 6-well ultra-low attachment plates at a density of 1,000/well and grown in a medium supplemented with 5 μg/mL insulin, 20 ng/mL basic fibroblast growth factor, and 20 ng/mL epidermal growth factor. The plates were further incubated for 15 days. Finally, under a light microscope (Olympus, Tokyo, Japan), the number of spheroids with a diameter of >50 μm in each well was counted.

Wound healing assay

To explore the directional cell migration of cells from diverse groups, the wound healing assay (also called scratch assay), a well-developed and simple laboratory technique [25], was performed. In brief, cells were seeded into six-well plates to form a monolayer. The 200-μL pipette tip was utilized to create a scratch in the cell layer. After scratched cells were removed by rinsing with PBS, the remaining cells were cultured in DMEM in a 5% CO₂ incubator at 37 °C for 24 h. A BX51 microscope (Olympus, Tokyo, Japan) was utilized to photograph the same position of scratches at 0 and 24 h after scratching.

Transwell assay

The cell invasion ability was determined using transwell chambers with Matrigel-coated membrane (Corning, NY, USA) [26]. Briefly, DMEM-containing serum and cell suspension were added in the lower and upper chambers, respectively. Twenty-four hours later, cells that adhered to the membrane of the upper chamber were wiped out, whereas cells crossing the membrane were fixed before staining with 0.1% crystal violet. Finally, under a light microscope, the stained cells were imaged and counted.

Statistical analysis

Data were expressed as mean ± standard deviation. The differences were analyzed by analysis of variance using GraphPad Prism 8.0.1 software (GraphPad Software, CA, USA). Graphpad Prism was also applied to calculate IC50 values based on the cell viability curve. p <0.05 was considered statistically significant.

TIMP1 upregulation is closely related to the tumor stage, poor survival, and DDP chemoresistance in GC

Based on GEPIA, STAD, which means stomach adenocarcinoma, represents GC. TIMP1 was highly expressed in GC (p <0.05) (Figure 1A). In the analysis of the GSE56807 dataset, the expression of TIMP1 was significantly higher in gastric cancerous tissues than in adjacent normal tissues, (p <0.05) (Figure 1B). GEPIA showed that the high TIMP1 expression was positively associated with tumor stage (p <0.001) and poor survival (p =0.017) in GC (Figure 1C and D). To investigate whether TIMP1 expression also correlated with the DDP chemoresistance of GC, two DPP/R GC cell lines (AGS/DDP and BGC-803/DDP) were established. RT-PCR analysis showed that TIMP1 expressions in two DPP/R GC cell lines were both higher than that in the parental ones (Figure 1E), which suggested a potential role of TIMP1 in DDP-resistant GC. Among the two paired GC cell lines, parental AGS and AGS/DDP exhibited the most difference in TIMP1 expression (8.70-fold, p <0.001). Hence, we chose AGS and AGS/DDP cells for further analysis on the role of TIMP1 in GC.

Figure 1:

TIMP1 is highly expressed and closely related to the tumor stage, poor survival, and DDP chemoresistance in GC. (A) The expression analysis for TIMP1 between GC (n=408) and normal samples (n=211) based on the GEPIA database. (B) Expression analysis for TIMP1 on five pairs of GC and adjacent normal tissue samples from the GSE56807 dataset. (C) Expression analysis for TIMP1 on different stages of GC based on the GEPIA database. (D) By using GEPIA data, survival analysis for patients with GC (n=192) was performed based on TIMP1 expression. (E) Relative expression levels of TIMP1 in two GC cell lines (normal) and their DDP-resistant ones (DDP/R) were detected using RT-PCR. ^* p<0.05, ^** p<0.01, and ^*** p<0.001 vs. the normal group.

Resistance to DDP and stemness in DDP-resistant AGS cells is mediated by TIMP1

Initially, siRNAs against TIMP1 were transfected into AGS/DDP cells. Then, the efficiency of transfection was measured by RT-PCR and Western blot (Figure 2A and B). Notably, si-TIMP1#2 led to the most efficient suppression of TIMP1 expression in cells. Thus, si-TIMP1#2 (hereafter defined as “si-TIMP1”) was chosen for subsequent experiments. The CCK-8 assay was conducted to evaluate cell viability after treatment with a series of DDP concentrations for 24, 48, and 72 h. The viability curve showed that AGS/DDP cells exerted significantly stronger resistance against DDP than AGS cells (Figure 2C), as evidenced by the increased IC50 value (Table 1). AGS/DDP transfecting si-TIMP1 showed significantly enhanced sensitivity to DDP in comparison with AGS/DDP cells or AGS/DDP cells delivered with si-NC (Figure 2C and Table 1). These results suggested a critical role of TIMP1 in regulating the response of GC cells to DDP.

Figure 2:

Resistance to DDP and stemness of DDP-resistant AGS cells are restrained by TIMP1 silencing. AGS/DDP cells were transfected with/without si-NC or si-TIMP1; parental AGS cells were defined as the positive control. The protein and mRNA expression levels of TIMP1 were examined by (A) Western blot and (B) RT-PCR after the transfection of si-NC, si-TIMP1#1, si-TIMP1#2, and si-TIMP1#3; si-RNA with the highest silencing effect on TIMP1 expression was chosen for further analysis. After treating with DDP for 24, 48, and 72 h, (C) cell viability in different groups was evaluated by the CCK-8 assay. (D) The self-renewal of cells was determined by the sphere formation assay. (E) Western blot detected the expression levels of stem cell markers in cells. ^** p<0.01 and ^*** p<0.001 vs. the AGS group; ^## p<0.01 and ^### p<0.001 vs. the AGS/DDP group.

Sphere formation assay showed that AGS/DDP cells formed approximately 2.57-fold (p=0.002) greater numbers of spheres per field than AGS cells (Figure 2D), suggesting a stronger stemness in AGS/DDP cells. Western blot detecting stem cell markers further confirmed this point, which showed that CD44, CD47, Nanog, and Oct4 expression levels in AGS/DDP cells were obviously elevated compared with those in AGS cells (all p<0.001) (Figure 2E). Moreover, TIMP1 silencing significantly restrained the stem properties in AGS/DDP cells (Figure 2D and E), which signposted that the stemness of AGS/DDP cells was mediated by TIMP1. The numbers of formed spheres from the AGS/DDP + si-TIMP1 group were markedly less than those from the AGS/DDP group (p=0.003) (Figure 2D). Additionally, the expression levels of CD44, CD47, Nanog, and Oct4 in AGS/DDP cells were significantly decreased after transfecting with si-TIMP1 (all p<0.001) (Figure 2E).

TIMP1 is involved in migration, invasion, EMT process, and Wnt/β-catenin signaling pathway in DDP-resistant AGS cells

We further evaluated cell migration and invasive capacities in the AGS cell line and its DDP-resistant counterpart (AGS/DDP). Results revealed that both the migration and invasive abilities of AGS/DDP cells were markedly stronger than those of AGS cells (all p<0.001), which were significantly restrained by knocking down TIMP1 (p<0.001) (Figure 3A and B). As known, EMT activation frequently occurred in drug resistance, stemness, and invasiveness of cancer cells [27]. Compared with AGS cells, E-cadherin upregulation and N-cadherin downregulation (both p<0.001) were observed in AGS/DDP cells (Figure 3C), which suggested EMT activation occurred in AGS/DDP cells. Then, our study further examined whether TIMP1 participates in the EMT process in AGS/DDP cells. Si-TIMP1 transfection suppressed the epithelial converse to mesenchymal phenotype in AGS/DDP cells, as evidenced by the reduction of E-cadherin (p<0.001) and elevation of N-cadherin (p<0.001) (Figure 3C).

Figure 3:

TIMP1 is involved in the migration, invasion, and EMT of DDP-resistant GC cells. AGS/DDP cells were transfected with si-NC or si-TIMP1; AGS/DDP cells without transfection were considered the control, whereas parental AGS cells were defined as the positive control. (A) The wound healing assay evaluated cell migration; scale bar=100 μm. (B) The transwell assay detected cell invasion; scale bar=100 μm. (C) The expression of EMT-related markers was detected by Western blot. (D) The expression and phosphorylation levels of GSK3β and expression of nuclear β-catenin were detected by Western blot. ^*** p<0.001 vs. the AGS group; ^### p<0.001 vs. the AGS/DDP group.

To further explore the mechanisms involved in the effect of TIMP1 on DDP-resistant GC, the expressions of GSK3β and β-catenin were analyzed by Western blot. The phosphorylation level of GSK3β and the accumulation of nuclear β-catenin were significantly increased in AGS/DDP cells compared with AGS cells (both p<0.001) (Figure 3D). In the meantime, the transfection of si-TIMP1 led to a significant decrease in GSK3β phosphorylation and nuclear accumulation of β-catenin in AGS/DDP (both p<0.001) (Figure 3D), which suggested that TIMP1 contributes to activating Wnt/β-catenin in DDP-resistant GC. Taken together, TIMP1 knockdown impaired the cell migration and invasion capacities and blocked the EMT process in AGS/DDP cells, which might relies on the repression of Wnt/β-catenin signaling.

TIMP1 contributes to the malignant behaviors and EMT of DDP-resistant AGS cells via Wnt/β-catenin signaling

To determine whether the effect of TIMP1 on the malignant behaviors and EMT of DDP-resistant AGS cells relies on its role in Wnt/β-catenin signaling, AGS/DDP cells were treated with 2.5 μM AR-A014418 (a selective GSK-3 inhibitor) after being transfected with TIMP1 overexpression vector (TIMP1 OE). The result of the Western blot confirmed that TIMP1 OE transfection successfully overexpressed TIMP1 in AGS/DDP cells (Figure 4A). As expected, overexpressing TIMP1 led to aggravation in the malignant phenotypes of AGS/DDP cells (Figure 4B–F), which is revealed by the significant elevation in the IC50 value for DDP (17.48–28.54 μg/mL), numbers of formed spheres, stemness-related marker expression, and capacities of migration and invasion. In the meantime, E-cadherin expression was markedly reduced, and N-cadherin expression was dramatically increased in TIMP1-overexpressing AGS/DDP cells (Figure 4G), revealing that TIMP1 promotes the EMT process of AGS/DDP cells. After treatment with AR-A014418, the effects induced by TIMP1 OE transfection on the malignant behaviors and EMT of AGS/DDP cells were almost blocked (Figure 4B–G), which demonstrated that TIMP1 contributes to the malignant behaviors and EMT of DDP-resistant AGS cells by suppressing Wnt/β-catenin signaling.

Figure 4:

TIMP1 regulated the Wnt/β-catenin signaling pathway in DDP-resistant GC cells. After being transfected with a vector or TIMP1 OE, AGS/DDP cells were treated with or without 2.5 μM AR-A014418 (a selective GSK-3 inhibitor). (A) The protein expression levels of TIMP1 were examined by Western blot after the transfection of a vector or TIMP1 OE. (B) Cell viability in different groups was evaluated by the CCK-8 assay after treatment with DDP for 24 h. (C) The self-renewal of cells was determined by the sphere formation assay. (D) Western blot detected the expression levels of stem cell markers in cells. (E) The wound healing assay evaluated cell migration; scale bar=100 μm. (F) The transwell assay detected cell invasion; scale bar=100 μm. (G) The EMT-related marker expression was detected by Western blot. ^** p<0.01 and ^*** p<0.001 vs. the AGS/DDP group; ^### p<0.001 vs. the AGS/DDP + TIMP1 OE group.