Luteolin suppresses bladder cancer growth via regulation of mechanistic target of rapamycin pathway

2.1 Cell culture and reagents

The human bladder cancer cell lines, T24, 5637 with a p53 mutation and RT-4 with wild-type p53, were acquired from the ATCC and maintained in DMEM and RPMI 1640, respectively (Thermo Fisher Scientific), supplemented with 10% FBS. All cells were authenticated by DNA profiling using short tandem repeat analysis. A rat bladder cancer cell line, BC31 with a p53 mutation, was derived from a BBN-induced rat bladder cancer and cultured as previously described.22-24 All cells were cultured at 37°C in a humidified atmosphere of 5% CO₂. Luteolin (3,4,5,7-tetrahydroxy flavone) and 1-methylpropyl 2-imidazolyl disulfide (PX-12) were purchased from Tokyo Chemical Industry and Sigma–Aldrich, respectively.

2.2 Cell viability assay

T24 and 5637 cells were seeded in 96-well plates at a density of 3 × 10³ cells/well. The cells were grown for 24 hours and treated with luteolin for 48 hours. Cell viability was evaluated using Cell Counting Kit-8 (Dojindo Laboratories). All experiments were performed in triplicate.

2.3 Flow cytometry of apoptosis and cell-cycle analysis

Cells were treated with luteolin for 48 hours and collected in 6-well plates. A FACSCalibur system (BD) was used for both apoptosis and cell-cycle assays. We used a phycoerythrin Annexin V Apoptosis Detection Kit with 7-aminoactinomycin D (BioLegend) for apoptosis assays, and a Cell Cycle Phase Detection Kit (Cayman Chemical) to evaluate the cell cycle by propidium iodide according to the manufacturer’s instructions. All experiments were performed in triplicate.

2.4 Western blot analysis

All proteins were collected from cells at 70%-80% confluency in 6-well plates. A total of 30 µg protein per lane was resolved on 12% acrylamide gels and transferred onto Hybond ECL membranes (GE Healthcare UK). Primary antibodies and their dilution ratios are listed in Table S1. All experiments were performed in triplicate.

2.5 In vitro siRNA transfection

Small interfering (si)RNAs targeting human TP53 sequences were obtained from Sigma-Aldrich. RT4 cells (1.5 × 10⁵/well) were seeded in 6-well plates and transfected with 20 nmol/L siRNA using Lipofectamine RNAiMAX (Thermo Fischer Scientific) according to the manufacturer’s protocol.8 Non–targeting siRNA, with no significant homology to any known rat and human genes, was also used as a negative control (NC). Confirmation of TP53 knockdown and treatment with luteolin were performed 24 hours after transfection. All experiments were performed in triplicate.

2.6 TRX1 gene expression analysis

The Cancer Genome Atlas (TCGA) bladder cancer datasets (http://cancergenome.nih.gov/) were analyzed as previously described.11 The unit is Fragments Per Kilobase of transcript per Million mapped reads Upper Quartile (FPKM-UQ).

2.7 TRX1 and DCFH-DA assay

TRX1 levels were evaluated using a Human Thioredoxin Assay Kit (IBL). Briefly, T24 cells were seeded into 6-well plates followed by a 48-hour exposure to luteolin. After treating the medium of each plate according to the manufacturer's instructions, we measured absorbance at 450 nm using a microplate reader. A dichloro-dihydro-fluorescein diacetate (DCFH-DA) assay was performed to evaluate intracellular ROS production, as previously described.22 All experiments were performed in triplicate.

2.8 Cancer patient tissue samples

A total of 166 patients who underwent radical cystectomy or transurethral resection for bladder cancer and whose pathology was UC or its variants, as diagnosed by experienced pathologists, were eligible for this study. This study was approved by the institutional review board at Nagoya City University Hospital and the approval number was NCU-893. All enrolled patients provided written informed consent.

2.9 Animals

Three or four experimental rats or mice were housed in each plastic cage on wood-chip bedding in an air-conditioned specific pathogen-free animal room at 22 ± 2°C and 55% ± 5% humidity with a 12-hour light/dark cycle. Food and tap water were available ad libitum. The present experiments were performed under protocols approved by the Institutional Animal Care and Use Committee of Nagoya City University School of Medical Sciences.

2.10 Xenograft mouse model

Seven-week-old male KSN nude mice (Japan SLC) were inoculated subcutaneously in the back with 5 × 10⁴ BC31 cells. Two days after injection, 24 mice were randomly divided into two treatment groups and fed either an AIN76A control diet (Oriental Bio Science) or AIN76A with 100 ppm luteolin for 5 weeks. Tumor size was monitored weekly. Tumor volume was calculated using the following formula: 1/2 × Length ×Width × Height.

2.11 N-butyl-N-(4-hydroxybutyl) nitrosamine-induced rat bladder carcinogenesis model

Six-week-old male F344 rats (Charles River Laboratories Japan) were given 0.05% BBN (Tokyo Chemical Industry) in drinking water for 10 weeks. A total of 60 rats were then randomly divided into three groups and received either a control diet, or 20 or 100 ppm luteolin for 20 weeks. A 24-hour specimen of urine was collected in the 29th week. All rats were killed, and blood, bladders, livers and kidneys were collected in the 30th week. Bladder tissues were divided equally into four portions and fixed by formalin, and H&E stains of all specimens were evaluated by experienced pathologists (N.A-I, SS) according to previously described instructions.25 Each tumor dimension was quantified using a BZ-9000 multifunctional microscope and associated analysis software (Keyence).

2.12 DNA extraction and direct sequencing for p53 gene in N-butyl-N-(4-hydroxybutyl) nitrosamine-induced rat bladder cancers

Genomic DNA was extracted from BBN-induced rat bladder tumors histologically diagnosed as cancer from the control group (n = 10) and the luteolin 100 ppm group (n = 10). The extraction from formalin-fixed paraffin-embedded tissues was performed using NucleoSpin DNA FFPEXS (MACHEREY-NAGEL). Exons 5, 6-7 and 8 of rat p53 gene were analyzed as previously reported.26 The primer sequences were 5′-CGCTGACCTTTGATTCTTTC-3′ and 5′-AGTTCTAACCCCACAGCAGT-3′ for exon 5, 5′-GTTAGAACTGGTTGTCCAGGG-3′ and 5′-CCCAACCTGGCACACAGCTTCCT-3′ for exon 6-7, and 5′-CTGTGCTCCTCTTGTCCCG-3′ and 5′-CCTCCACCTTCTTTGTCCTG-3′ for exon 8. PCR products were purified with NucleoSpin Gel and PCR Clean-up (MACHEREY-NAGEL). Direct sequencing was performed using 2 µL of aliquots by Applied Biosystems 3130xl Genetic Analyzer (Thermo Fischer Scientific) according to the manufacture’s instructions. Sequencing primers for exon 5 were 5′-GATTCTTTCTCCTCTCCTACAG-3′ and 5′-AGTTCTAACCCCACAGCAGT-3′, for exon 6-7 were 5′-GCCTCTGACTTATTCTTGCT-3′ and 5′-AACCTGGCACACAGCTTCCT-3′, and for exon 8 were 5′-TGCCTCCTCTTGTCCCGGGT-3′ and 5′-CACCTTCTTTGTCCTGCCTG-3′.

2.13 Immunohistochemistry and histological analysis

Deparaffinized formalin-fixed human and rat tissues were incubated with primary antibodies (listed in Table S2). A TUNEL assay was performed using an In situ Apoptosis Detection Kit from Takara, as previously described.22 Immunostaining was done according to the manufacturer’s instructions.27

In the analysis of positive cell rates for Ki67, TUNEL and p21, five microscopic fields were randomly selected and quantified with the same threshold nuclear intensity at a ×400 magnification according to analysis software (Keyence). The cytoplasmic intensities of p-S6 and cytokeratin 5/6 (CK5/6) were randomly selected in five cancerous areas at ×400 magnification and quantified with analysis software.

2.14 Metabolism of luteolin

Plasma and urine luteolin levels in F344 rats (control: n = 10, luteolin 100 ppm: n = 18) were analyzed by HPLC-tandem mass spectrometry at the Biodynamic Plant Institute. All samples were analyzed using the API 3200 System (ABSciex) with an HPLC system (Shimazu, Kyoto, Japan). Chromatographic separation was achieved on a YMC-Pack ProC18 (5 µm, 3.0 mm × 150 mm; YMC) and maintained at 40°C at a flow rate of 0.3 mL/min. The concentrations of luteolin-aglycone, luteolin-7-glucoside and luteolin-3ʹ-glucuronide were measured.

2.15 Statistical analysis

Data were analyzed using EZR software (Saitama Medical Center, Jichi Medical University) and Fisher’s exact test, Student’s t test or ANOVA, where appropriate. Spearman’s rank correlation coefficient was used to evaluate the correlation of two variables. P < 0.05 was considered a statistically significant difference.

3.1 Luteolin has a cancer suppressive effect and upregulates p21^Waf1/Cip1 in bladder cancer cells

As shown in Figure 1A, luteolin had an inhibitory effect on the cell viability in a dose-dependent manner against human bladder cancer cell lines. Figure 1B,C and Figure S1A show that luteolin clearly induced apoptosis in bladder cancer cells. Next, we found that luteolin induced G2/M arrest in a dose-dependent manner (Figure 1D,1 and Figure S1B). We further explored changes in cell-cycle-related protein expression in response to treatment with luteolin. We found upregulated p21^Waf1/Cip1 and p27^Kip1, and downregulated cyclin A and D1 in luteolin-treated T24 cells, which were derived from a high-grade human bladder cancer (Figure 1F). The upregulated expression of p21^Waf1/Cip1 and decreased expression of cyclin D1 with luteolin treatment were also evident in a low-grade human bladder cancer cell line, 5637 (Figure 1F). To explore the effect of p53 function on p21 ^Waf1/Cip1 regulation by luteolin, we further demonstrated an effect on a p53 wild-type human bladder cancer cell line, RT4. The expression level of p21 ^Waf1/Cip1 was higher in RT4 cells compared to T24 and 5637 cells. Luteolin increased the expression of p21^Waf1/Cip1 in RT4 cells (Figure 1G). TP53 knockdown clearly downregulated p21^Waf1/Cip1, indicating that p21^Waf1/Cip1 expression was p53-dependent. Luteolin still upregulated p21^Waf1/Cip1 in TP53-silenced RT4 cells (Figure 1H). Taken together, luteolin induces p21^Waf1/Cip1 in bladder cancer cells with both p53 wild-type and mutation.

3.2 Luteolin inhibits the mechanistic target of rapamycin pathway in bladder cancer

The mTOR regulates cell growth, proliferation, metabolism and autophagy. S6K is a major substrate of the mTOR kinase. The activity of S6K1, an isoform of S6K, is controlled with phosphorylation by mTOR, with the position of Thr389 being the most critical site of phosphorylation for kinase activity. S6 is a substrate of S6K regulated by phosphorylation and is involved in cell-cycle progression, ribosomal proteins, and elongation factors necessary for translation.14, 28, 29

Western blot analysis showed that levels of phospho(p)-Akt, phospho(p)-p70S6K and phospho(p)-S6 in bladder cancer cell lines were decreased by luteolin in a dose-dependent manner (Figure 2A,B). These effects were obtained 12 and 24 hours after exposure to luteolin (Figure 2C). We then compared the effect of luteolin to the mTOR inhibitor, Torin1. The inhibition of mTOR significantly reduced the survival rate at a concentration of more than 10 nmol/L, and induced G1/S arrest and apoptosis in T24 cells (Figure S2A,C-F). Western blot analysis indicated that Torin1 treatment of T24 cells did not induce p21 expression, in contrast to luteolin (Figure S2B). These results suggest that the anti–tumor effect of luteolin was not phenocopied by an mTOR inhibitor and was affected by p21 in bladder cancer. (Figure S2A-D). Next, we evaluated the role of the mTOR pathway in patients with bladder cancer by analyzing the immunohistochemistry of p-S6 and Ki67. Although no statistical differences were found in the relative intensity of p-S6 between pTa low-grade and T1 plus Ta high-grade tumors, significant differences were found between more than T2 and pTa low-grade or T1 plus Ta high-grade tumors (P < 0.001, respectively; Figure 2D,E). A statistical correlation between the Ki67-labeling index and the relative intensity of p-S6 was evident (P < 0.001, r = 0.46; Figure 2F). These results indicate that mTOR activity may be related to cell proliferation in bladder cancer.

3.3 The upregulation of TRX1 by luteolin suppresses cell proliferation in bladder cancer cells

We next evaluated whether luteolin affected the TRX1 level or the relationship between mTOR and TRX1. Intracellular ROS decreased and TRX1 increased in a concentration-dependent manner in luteolin-treated T24 cells (Figure 3A,B). To explore whether the suppressive effect of luteolin on cell viability can be attributed to TRX1 upregulation and whether TRX1 is associated with the mTOR pathway, we treated cells with the TRX1 inhibitor, PX-12.11 PX-12 decreased T24 cell proliferation in a dose-dependent manner. This suggests that an imbalance in oxidative stress occurred that adversely affected cell viability (Figure 3C). T24 cells were harvested after 48-hour exposure to luteolin, followed by another 48-hour exposure to luteolin, DMSO control or 2 µmol/L PX-12 (nearly 40% decreased cell proliferation). Cells treated with luteolin followed by DMSO recovered from the decrease in cell survival induced by luteolin (P < 0.001; Figure 3D). PX-12 treatment also led to an increase in cell viability, despite the growth-suppressive concentration (Figure 3D). This indicates that TRX1 played a role in the inhibition of ROS production and cell proliferation in bladder cancer cells. Western blot analysis revealed that levels of p-p70S6K and p-S6 recovered with PX-12 treatment after luteolin-induced mTOR suppression quicker than those after DMSO treatment (exposure post–3-hour to PX-12 vs 12-24-hour to DMSO; Figure 3E). We further analyzed TRX1 expression and also prognosis by using a TCGA dataset of 338 patients with bladder cancer. Figure 3F,G shows no difference in the TRX1 level between normal and cancer tissues, and between low-grade and high-grade tumors. Furthermore, a significant difference in overall survival between high and low TRX1 expression groups was not observed (Figure 3H). These data indicate that TRX1 is not necessarily a negative prognostic factor in bladder cancer, unlike for other cancers,9-11 and luteolin prevents the accumulation of intracellular ROS levels and mTOR signaling by upregulating the TRX1 level in bladder cancer.

3.4 Anti–tumor effect of luteolin in a subcutaneous BC31 xenograft model

We next examined the anti–tumor effect of dietary luteolin in a subcutaneously transplanted BC31 xenograft model. Of 24 mice, the tumors of 10 out of 12 mice in the control group and 9 out of 12 mice in the luteolin 100 ppm group were successfully engrafted; and these 19 mice were subsequently used for analyzing the effect of luteolin. Significant differences in the amount of dietary intake, and body and organ weights between the two groups were not noted (Figure S3A-C). Histological analyses revealed that no toxic effect was induced by luteolin. The tumor volumes of mice of the luteolin 100 ppm group were significantly smaller than those of the control group of mice (P < 0.05; Figure 4A,B). Ki67-labeling index was significantly decreased and rates of TUNEL-positive cells were significantly increased by luteolin (P < 0.001 for both; Figure 4C). These results suggest that luteolin inhibited proliferation and induced apoptosis in bladder cancer cell xenografts (Figure 4C). The percentage of p21-positive cells was significantly higher (P < 0.01) and p-S6 intensity was significantly lower (P < 0.001) in the luteolin 100 ppm group compared to the control group, which was consistent with in vitro results (Figure 4C).

3.5 Luteolin metabolites decrease mechanistic target of rapamycin signaling and cell proliferation in an N-butyl-N-(4-hydroxybutyl) nitrosamine-induced rat bladder cancer model

Finally, we investigated the anti–tumor effect of the oral intake of luteolin using a BBN-induced rat bladder cancer model. A rat in the 100-ppm group died because of bladder cancer without metastasis or hydronephrosis the day before it was due to be killed. Significant differences in the amount of BBN water, dietary intake, and body and organ weights among the three groups were not observed (Figure S4A-D). No histological change was observed in livers and kidneys.

All rats developed superficial bladder cancer without invasion. DNA sequencing analysis revealed that point mutations of p53 gene were observed in rat bladder cancers (Table S3). Mutations in codon 195 (GTG → ATG) and codon 251 (ACC → ATC) were similar to those reported in a previous study.26 Bladder weight and tumor dimension tended to be reduced by luteolin treatment (Figure 5A-C). Immunohistochemical analyses revealed a significantly decreased Ki67-labeling index (P < 0.01) and p-S6 level (P < 0.001) in the high-dose luteolin group, but differences in positive rates of TUNEL and p21 staining were not observed (Figure 5D). We further analyzed any change in squamous differentiation, which relates to worse recurrence and progression rates30 with luteolin, because we previously found that the oxidative stress-regulated gene glutathione peroxidase 2 (GPX2) was upregulated in several carcinomas, including UC,22, 27, 31, 32 and promoted squamous differentiation.22 Moreover, we recently found that luteolin downregulated GPX2 in the rat prostate tumor and in a human prostate cancer cell line.33 The frequency of squamous differentiation in the tumors was significantly decreased in the luteolin-treated groups (P < 0.05; Figure 5E,5). Consistent with this result, the level of CK5/6, a marker of squamous differentiation,30, 34 was decreased by luteolin treatment (P < 0.01; Figure 5D,E). Immunohistochemistry revealed that the expression of GPX2 was induced in bladder cancer with squamous differentiation, compared to pure UC, and was significantly reduced by luteolin in rat BBN-induced bladder cancer (P < 0.01; Figure 5G,H). Moreover, the expression level of GPX2 and the incidence of squamous differentiation were significantly correlated in the model (P < 0.001; Figure 5I). GPX2 expression was shown in RT4 and BC31 cells but not in T24 and 5637 cells, and was not affected by luteolin treatment in vitro (Figure S5). These results indicate that luteolin reduces squamous differentiation through the suppression of GPX2 in rat bladder cancer in vivo.

Next, we analyzed the metabolism of luteolin and whether it is involved in the anti–cancer effect. Previous reports described the oral administration of luteolin-aglycone or glucoside and their conjugation to glucuronide in rats occurring mainly in plasma and tissues.35, 36 Consistent with this, luteolin-aglycone and glucoside were not detected in plasma. Instead, concentrations of luteolin-3ʹ-glucuronide in plasma (1.66 ± 1.30 ng/mL) and urine (25.1 ± 18.5 ng/mL), respectively, were significantly increased after luteolin treatment (Figure 6A). Concentrations of luteolin-3ʹ-glucuronide showed a mild correlation between urine and plasma (r = 0.25, P = 0.20; Figure 6B).

The luteolin-3ʹ-glucuronide concentration in urine (r = −0.41, P = 0.03), as well as that in plasma (r = −0.31, P = 0.11), was significantly correlated with the Ki67-labeling index of the tumors (Figure 6C), even though it did not correlate with tumor dimension (Figure 6D). Moreover, the relative p-S6 intensity of the tumors and luteolin-3ʹ-glucuronide concentration in plasma, as well as that in urine, showed a strong correlation (r = −0.60, P < 0.001 and r = −0.62, P < 0.001, respectively; Figure 6E). With respect to the preventive effect of luteolin on squamous differentiation, luteolin-3ʹ-glucuronide concentrations in plasma were significantly correlated with a lack of squamous differentiation and low GPX2 intensity (P < 0.01; Figure 6F, P < 0.01; Figure 6G).