The role of focal adhesion kinase in bladder cancer: translation from in vitro to ex vivo human urothelial carcinomas

Normal porcine urothelial (NPU) cells were established as previously described.^11,20,21 In this study, we used both undifferentiated and differentiated in vitro models of NPU cells, which serve as a morphological and functional surrogate for normal human bladder urothelium. NPU cells were seeded at a density of 1 × 10⁵ viable cells/cm² and grown for 2 days in culture medium containing equal amounts of MCDB153 (Sigma-Aldrich, St. Louis, MO, USA) and A-DMEM (Gibco, Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 15 μg/ml adenine (Sigma-Aldrich, St. Louis, MO, USA), 0.1 mM phosphoethanolamine (Sigma-Aldrich, St. Louis, MO, USA), 5 μg/ml insulin (Sigma-Aldrich, St. Louis, MO, USA), 0.5 μg/ml hydrocortisone (Sigma-Aldrich, St. Louis, MO, USA), 2 mM glutamax (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and 2.5% FBS (Invitrogen, Carlsbad, CA, USA) to establish the undifferentiated normal urothelial model (2-day NPU cells). To establish the differentiated normal urothelial model, NPU cells, after reaching confluence (at day 7), were cultured for another three weeks in the same culture medium as described above but without fetal bovine serum (FBS) and with 2.5 mM CaCl₂. The differentiated normal urothelial model is a suitable in vitro model for investigating urothelial drug candidates.^20–23 The preparation of primary urothelial cells from porcine urinary bladders was approved by the Veterinary Administration of the Slovenian Ministry of Agriculture and Forestry in accordance with the Animal Health Protection Act and the Instructions for Granting Permits for Animal Experiments for Scientific Purposes (Decree No. U34453-15/2013/2).

Other in vitro models include the low-grade (LG) human non-invasive papilloma urothelial cell line RT4 for NMIBC (Ta) and the high-grade (HG) human muscle-invasive cancer urothelial cell line T24 for MIBC (T2), both from ATCC (Manassas, VA). RT4 and T24 cells were seeded at a density of 1 × 10⁵ viable cells/cm² and grown for 2 days in A-DMEM (Gibco) with Ham’s F12 (Gibco) to represent conditions early after tumour resection with few remaining urothelial cancer cells (2-day RT4 and 2-day T24 cells) or seeded at a density of 5 × 10⁴ viable cells/cm² and grown for 7 days to represent the later stages of regeneration after tumour resection (7-day RT4 and 7-day T24 cells). The selected time points (2-day and 7-day models of RT4 and T24 cells) were chosen based on our previous studies indicating significant changes in cellular activity and biological processes at these time intervals. The 2-day models represent the early post-surgery phase, while the 7-day models represent later time points, where changes associated with tumour progression become more evident. This approach allows us to better understand the mechanisms involved in bladder cancer progression.

Normal urothelial and urothelial cancer in vitro models were cultured at 37°C and 5% CO₂. All urothelial models were repeatedly tested negative for mycoplasma infection using MycoAlert mycoplasma detection kit (Lonza, Basel, Switzerland).

The study population consisted of 12 patients with either bladder cancer or normal urothelium who underwent TURB. Informed consent was obtained from all patients whose biopsies were included in this study. Tissue samples were collected from both cancerous and normal areas using cold cup biopsy, resulting in a total of 17 biopsies analysed. Participant recruitment and study design were at the discretion of the operating urologist, who determined the number of biopsy samples for each patient after intraoperative evaluation of tumour location, size, multiplicity, and the risk of acute bleeding. During a cold cup biopsy, a tissue sample is obtained using biopsy forceps. These forceps are at room temperature and feature two half-spherical jaws with sharp edges for sectioning. The forceps are introduced into the bladder via a cysto-scope. When the forceps close around the tumour tissue, a spherical tissue sample is cut between the jaws and retrieved by pulling the sample through the cystoscope. The diameter of the jaws, and thus the size of the tissue sample, is 4 mm. The biopsies captured the urothelium and part of the lamina propria. Samples were processed for western blot (from 9 patients) and paraffin embedding (from 8 patients). For paraffin sections, samples were fixed with 4% formaldehyde (FA) in phosphate buffered saline (PBS) overnight at 4°C and embedded in paraffin. Paraffin sections were 5 μm thick, cut from at least two different parts of each sample and stained with haematoxylin and eosin. The tissue samples were initially collected without prior knowledge of their nature. It was only after pathological evaluation that some samples were confirmed to represent normal urothelial tissue. Tissue samples were classified as normal if no signs of hyperplasia or dysplasia were detected. WHO classification of tumours of the urinary system was used for pathological staging and classification.²⁴ Urothelial carcinomas were diagnosed as invasive papillary urothelial carcinoma low-grade with invasion into the lamina propria (pT1 LG), invasive papillary urothelial carcinoma high-grade with invasion into the lamina propria (pT1 HG) and invasive papillary urothelial carcinoma highgrade with invasion into the muscularis propria (pT2 HG). Among 12 patients, 10 (83%) were male (aged 62–79) and 2 (17%) were female (aged 80 and 98). Of the 10 male patients, 3 were diagnosed with pT1 LG, however, biopsies taken 1 cm posterior to the interureteric ridge revealed normal urothelium. The biopsy-based diagnoses for the male patients were therefore distributed as follows: pT1 LG (3 patients), pT1 HG (3 patients), and pT2 HG (3 patients). Among the 2 female patients, one was diagnosed with normal urothelium and the other one with pT1 LG.

Human samples were homogenised in ice-cold buffer (0.8 M Tris-HCl, 7.5% sodium dodecyl sulphate (SDS), 1 mM phenylmethylsulphonyl fluoride). Treated and untreated NPU, RT4 and T24 cells were scraped, pelleted and lysed in RIPA buffer (EMD Millipore, Darmstadt, Germany) supplemented with protease and phosphatase inhibitors (100× Halt Cocktail, Thermo Scientific, 78441). Protein concentration was determined using the BCA Protein Assay Kit (Thermo Fisher Scientific Waltham, MA, USA). 10 μg of proteins per lane were loaded onto 4–20% Novex Tris-Glycine gels (Thermo Fisher Scientific, Waltham, MA, USA), separated and transferred onto nitrocellulose membranes (Amersham Biosciences, Amersham, UK). The membranes were blocked in 5% non-fat dry milk in PBS supplemented with 0.1% Tween-20 (T-PBS) for one 1 h at room temperature, and then incubated overnight at 4°C with mouse monoclonal antibodies against E-cadherin (610182, 1:1,000, BD Transduction), rabbit polyclonal antibodies against N-cadherin (13116, 1:1,000, Cell Signaling), rabbit polyclonal antibodies against FAK (3285, 1:1,000, Cell Signaling), mouse monoclonal antibodies against p-FAK (sc-81493, 1:500, Santa Cruz), rabbit monoclonal antibodies against the heavy chain of p-FAK (70-025-5, 1:1,000, Thermo Fisher Scientific, Waltham, MA, USA) and rabbit polyclonal antibodies against 32 kDa proenzyme and the 17 kDa active form of caspase-3 (ab4051, 1:200, Abcam). To confirm equal protein loading, the blots were stripped with Restore Western Blot Stripping Buffer (Pierce, Rockford, IL) and reprobed with mouse monoclonal antibody against β-actin (A2228, 1:2,000, Sigma) or mouse monoclonal antibody against GAPDH (sc-47724, 1:1,000, Santa Cruz). HRP-conjugated secondary antibodies (anti-mouse and anti-rabbit IgG HRP-linked total antibodies, Sigma, 1:1,000) were used and detected with SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher Scientific, Massachusetts, USA). Chemiluminescence signals were visualised using LAS-4000 CCD camera (Fujifilm, Tokyo, Japan) or iBright CL1500 (Thermo Fisher Scientific, Massachusetts, USA).

Paraffin sections were deparaffinised and hydrated, and endogenous peroxidase activity was blocked with 3% H₂O₂ in methanol. Sections were heated in the microwave. Non-specific labelling was blocked with 5% bovine serum albumin (BSA). Sections were incubated overnight at 4°C with rabbit polyclonal antibody against FAK (3285, 1:100, Cell Signaling) and rabbit polyclonal antibody against p-FAK (70-025-5, 1:50, Thermo Fisher Scientific, Waltham, MA, USA). For negative controls, incubation with the primary antibody was omitted or the specific primary antibody was replaced by a non-relevant antibody. For secondary antibodies, biotinylated porcine anti-rabbit IgG (E353, 1:200, Dako) were applied for 1 h at room temperature, followed by incubation with ABC/HRP complex (Vector Laboratories, Burlingame, CA, USA). After the standard DAB (Sigma, Taufkirchen, Germany) development procedure, sections were counterstained with haematoxylin, and examined with a light microscope (Eclipse TE300, Nikon, Japan).

Three plasmid DNA encoding different miRNA against FAK (anti-FAK plasmids) were constructed using the pcDNATM6.2-GW/EmGFP-miR plasmid backbone (Invitrogen). For each plasmid two complementary single-stranded DNA oligonucleotides with the engineered pre-miRNA were used, precisely; containing a 4-nucleotide 5’ overhang complementary to the vector, followed by 5′G + 21-nucleotide pre-miRNA sequence (Table 3), a short 19-nucleotide spacer to form a terminal loop, a short sense target sequence with 2 nucleotides removed to create an internal loop and followed again by a 4-nucleotide 5’ overhang complementary to the vector.

Standard molecular biology techniques of annealing, ligation and transformation were then performed in competent E. coli TOP10 (Thermo Fisher Scientific, Waltham, MA, USA) using the BLOCK-iT Pol II miR RNAi Expression Vector Kit with EmGFP (Thermo Fisher Scientific, Waltham, MA, USA). Expression of the miRNA molecules from the plasmid DNA in the cells was coupled to EmGFP expression under the control of the Pol II promoter, which allowed identification of the percentage of transfected cells with the plasmid DNA. Plasmid DNA was amplified in E. coli TOP10 and isolated using the EndoFree Plasmid Mega Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The amount of isolated plasmid DNA was determined by a spectrophotometer at 260 nm (Epoch Microplate Spectrophotometer, BioTek, Winooski, VT, USA) and the quality was determined by measuring the ratio of absorbance at A260 nm/280 nm and by agarose gel electrophoresis. The working concentration of 1 mg/ml was prepared with endotoxin-free water. Plasmids were sequenced to confirm the presence, correct orientation and sequence of the double-stranded DNA oligonucleotide (ds oligo) insert.

RT4 and T24 cells were harvested when they reached approximately 80% confluence. A suspension of 25 × 10⁶ cells/ml was prepared in cold electroporation buffer (125 mM sucrose, 10 mM K₂HPO₄, 2.5 mM KH₂PO₄, 2 mM MgCl₂×6H₂O). 44 μL of the cell suspension was mixed with 11 μL of plasmids (1 mg/ml). Then, 50 μL of the suspension was electroporated using an electric pulse generator (GT-01, Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia). Six square pulses with a voltage/distance ratio of 1300 V/cm, a pulse duration of 100 μs and a frequency of 4 Hz were used, delivered via two parallel stainless-steel plate electrodes with a spacing of 1.9 mm. After GET, the cells were transferred to 24-well plates and incubated for 5 min before 1 ml of the appropriate culture medium was added. The cells were then grown until further analysis.

Cells were collected, and 1 × 10⁶ cells were prepared for staining. Cells were first washed twice in PBS and then 1 μL of FVD viability dye (eBioscience™ Fixable Viability Dye eFluor™ 780, 650865, Thermo Fisher Scientific) was added per 1 ml of cells and immediately vortexed. Cells were then incubated at 2–8°C for 30 min, protected from light and then washed twice more with PBS. After the last wash, the supernatant was discarded and samples were vortexed with a pulse vortex to completely dissociate the pellet. Cells were fixed by adding 100 μl of the Intracellular (IC) fixation buffer (eBioscience™, Thermo Fisher Scientific, Waltham, MA, USA), mixed and incubated again for 30 min at room temperature protected from light. 2 ml of 1× permeabilisation buffer (eBioscience™, Thermo Fisher Scientific, Waltham, MA, USA) was added and centrifuged at 400–600 × g for 5 min at room temperature. The supernatant was discarded, and this step was repeated once more. Cells were resuspended in 100 μl of 1× permeabilisation buffer and a 1:100 dilution of recombinant anti-FAK antibody [EP695Y] (ab40794, Abcam) was added and incubated for 30 min at room temperature, protected from light. Cells were then washed twice in PBS and incubated for 30 min in 100 μl of 1× permeabilisation buffer containing secondary Cy3 AffiniPure Donkey Anti-Rabbit IgG (H+L) (Jackson Immunoresearch, Ely, UK 711-165-152) antibody at a dilution of 1:100. 2 ml of 1× permeabilisation buffer was added and centrifuged at 400–600 × g for 5 min at room temperature. The supernatant was discarded, and this step was repeated once more. Stained cells were resuspended in 400 μl of IC fixation buffer and measured using the FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA) using a 488-nm laser (air-cooled, 20 mW solid state) for excitation and both 530- and 650-nm bandpass filters were used for detection of green (GFP, transfection efficiency) and red (FAK silencing) fluorescence. To eliminate debris, cells were first gated and afterwards a histogram of gated cells against their fluorescence intensity was recorded. The number of fluorescent cells and their median fluorescence intensity were determined for each of the plasmids (software: BD FACSDiva V6.1.2).

Mechanisms of cell death of RT4 and T24 cells were determined using the Annexin V (647) Apoptosis Detection Kit with 7-AAD (BioLegend, San Diego, CA, USA) according to the manufacturer’s instructions 24 h after GET of anti-FAK plasmids. Measurements were performed using the FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA). A 488-nm laser (air-cooled, 20 mW solid state) was used for excitation and both 530- and 650-nm bandpass filters were used for green and red fluorescence detection. At least 40,000 events were measured for each cell sample. Data were analysed using BD FACSDiva software version 8.0.1 (BD Biosciences, San Jose, CA, USA). All debris was excluded for further analysis using the FSC/SSC scatter plot. The experiments were performed twice in three parallels. Live cells were identified as the subset of cells negative for both Annexin V and 7-AAD staining (Annexin V^–/7-AAD^–). Dead cells were defined as the subset negative for Annexin V and positive for 7-AAD staining (Annexin V^–/7-AAD⁺). Apoptotic cells were determined as the sum of two subsets: early apoptotic cells, characterized by positive Annexin V and negative 7-AAD staining (Annexin V+/7-AAD^–), and late apoptotic cells, characterized by dual positivity for Annexin V and 7-AAD (Annexin V+/7-AAD+).

For the FAK inhibitor experiments, we selected three inhibitors, PND-1186, PF-573228, and defactinib, based on published efficacy data (all from SelleckChem, Houston, TX, USA). We prepared stock solutions following manufacturer’s instructions by dissolving the inhibitors in the solvent a-dimethylsulfoxide at room temperature (23°C). The stock solutions were then aliquoted into microcentrifuge tubes and stored at –80°C for up to six months. Before each experiment, we thawed the required number of tubes to room temperature. The stock solutions were then diluted in prewarmed growth medium to the desired final concentrations. For each inhibitor, we selected two concentrations: a lower concentration recommended for cell line experiments (1 μM for PND-1186, 10 μM for PF-573228 and defactinib) and a tenfold higher concentration to study and determine the maximum effect of the inhibitors. Any remaining stock solutions were stored in the refrigerator at 4°C for up to two weeks. The cells were treated as shown in Figure 1.

FIGURE 1.

Schematic representation of focal adhesion kinase (FAK) inhibition. We used in vitro models of a) the primary model of differentiated normal urothelium (diff NPU), in which normal porcine urothelial (NPU) cells were maintained in culture for 4 weeks, 3 weeks of which were in medium UroM (+Ca²⁺ -S_FBS), b) the bladder cancer urothelium early after resection of the bladder tumour, i.e. 2-day RT4 and 2-day T24 cells (RT4 (2 d) and T24 (2 d)), and of c) the bladder cancer urothelium late after resection of the bladder tumour, i.e. 7-day RT4 and 7-day T24 cells (RT4 (7 d) and T24 (7 d)). Cells were treated with FAK inhibitors PND-1186 (1 μM and 10 μM), PF-573228 (10 μM and 100 μM) and defactinib (10 μM and 100 μM) for 2 h per day for 3 consecutive days (3 × 2 h), after which cell viability was measured.

Prior to the experiments, cells were seeded on black 96-well polystyrene plates (Costar, Kennebunk, ME, USA) as described above. Subsequently, the cells were treated with FAK inhibitors for 2 h per day for 3 consecutive days, which we labelled 1 × 2 h for the first day, 2 × 2 h for the second day and 3 × 2 h for the third day. The culture medium containing FAK inhibitors was replaced every 24 h. At the end of the 3-day treatment, cell viability was determined using the CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Luminescence was measured using a microplate reader (Safire2, Tecan, Mannedorf, Switzerland). Experiments were carried out in triplicate across at least three independent repetitions. For each cell line the data were presented as a percentage of the cell viability, calculated according to the following formula: Cell viability= Average luminiscence intensity of treated cells Average luminiscence intensity of untreated cells x100%{\rm{ }}Cell viability{\rm{ }} = {{{\rm{ Average luminiscence intensity of treated cells }}} \over {{\rm{ Average luminiscence intensity of untreated cells }}}}x100\%

Cells grown on glass coverslips were untreated or treated with FAK inhibitors. Then the cells were fixed with 4% FA (in PBS) for 10 min at room temperature. The fixed cells were washed in PBS and then incubated with blocking solution (0.5% bovine serum albumin, 0.1% saponin, 0.1% gelatin, 50 mM NH₄Cl, 0.02% NaN₃) for 30 min at 37°C. After incubation, cells were washed in PBS and incubated overnight at 4°C with primary antibodies. We used the following primary antibodies: FAK (rabbit polyclonal antibody, 3285, 1:50, Cell Signaling), p-FAK (Tyr-397) (rabbit monoclonal antibody, 700255; 1:250, Thermo Fisher Scientific, Waltham, MA, USA) and cleaved caspase-3 (rabbit polyclonal antibody, ab2302, 1:100, Abcam). Cells were then washed in PBS and incubated for 90 min at 37°C with secondary goat anti-rabbit Alexa Flour555 antibody (1:500, Thermo Fisher Scientific, Waltham, MA, USA). To label actin, cells were then washed in PBS and additionally fixed for 15 min at 37°C in 4% FA. The cells were then incubated with 16.7 μg/ml phalloidin-FITC (Sigma) for 1 h at room temperature. Cells were then washed in PBS and embedded in Vectashield with DAPI (Vector Laboratories). Imaging was performed using the Axio Imager Z.1 fluorescence microscope (Zeiss, Germany).

Cell viability values were statistically analysed using SPSS (version 23; IBM Corporation, USA). The data are average values from at least three cell viability experiments, each performed in a triplicate. Levene’s test was used to test for homogeneity of variances. If the variances were homogeneous, ANOVA, followed by Tukey’s post-hoc test was performed to determine statistical significance. If the variances were not homogeneous, Welch’s test was performed followed by Games-Howell’s post-hoc test. When comparing a pair of data sets, the twotailed Student’s t-test was performed. Differences were considered statistically significant at P < 0.05.

Relative quantification of gene expression was calculated using the 2^−ΔΔCt method.²⁵ The reference GAPDH was used as an endogenous control for normalisation of the data. SPSS (version 24; IBM Corporation, USA) was used for statistical analyses. Changes in relative gene expression profiles between different cell lines were assessed using the Mann-Whitney U test where appropriate. In all cases, P ≤ 0.05 was considered to indicate a statistically significant difference.

The study was conducted in accordance with the Declaration of Helsinki and approved by the National Medical Ethics Committee of the Ministry of Health of the Republic of Slovenia under approval number 76/10/10. Informed consent was obtained from all patients whose biopsies were used in this study.

To gain insights into the molecular characteristics of normal bladder tissue and bladder cancer tissues, we performed a comprehensive analysis of protein expression in ex vivo samples of human bladder cancer of varying grades and stages. In particular, we focused on studying the expression of E-cadherin, N-cadherin, FAK and p-FAK. Our results showed that their expression levels were higher in bladder cancer tissues compared to normal bladder tissues, which was confirmed by western blot analysis (Figures 2A, B and Supplementary Figure 1). Specifically, we observed the highest expression levels of E-cadherin and N-cadherin in pT1 HG and pT2 HG bladder cancer tissues, while the expression levels of FAK and p-FAK remained consistent across different grades and stages of bladder cancer tissue and were higher than in normal bladder tissue (Figures 2A, B and Supplementary Figure 1). In addition, immunohistochemistry showed a prevalence of FAK and p-FAK in pT1 LG, pT1 HG and pT2 HG bladder cancer tissue compared to adjacent normal bladder tissue and not in the lamina propria (Figure 2C).

FIGURE 2.

Western blot and immunohistochemistry analysis of ex vivo human bladder cancer samples. (A) Western blot analysis showing the relative expression levels of FAK, p-FAK, E-cadherin, and N-cadherin in pT1 low-grade (LG) and highgrade (HG) cancer tissues, pT2 HG cancer tissues, and the corresponding normal tissues (labelled as “normal tissue (pT1 LG)” for example), collected from the tumour-adjacent regions, which were not deemed normal at biopsy, but only after pathological evaluation. (B) Relative protein expression levels from biopsies of 2–4 patients ± the standard error of the mean (SEM) normalised to β-actin and presented as median with interquartile range. (C) Immunohistochemistry of FAK and p-FAK in normal tissue, pT1 LG and HG cancer tissues and pT2 HG cancer tissues.

FAK = focal adhesion kinase; L = lumen; LP = lamina propria; U = urothelium. Scale bar: 50 μm.

In our study, we examined the relative expression levels of genes UPK1B (uroplakin 1b), UPK3A (uroplakin 3a), CDH1 (E-cadherin), CDH2 (N-cadherin) and PTK2 (FAK) in different in vitro models of urothelial cells (Figure 3). Notably, the differentiated urothelial cells, obtained by maintaining NPU cells in culture for 4 weeks (3 weeks of which in UroM medium with Ca²⁺ and without FBS) represented normal human bladder urothelium, while the 2-day RT4 and 2-day T24 cells represented an early post-tumour resection state with few residual cancer cells. The 7-day RT4 and 7-day T24 cells represented later stages after tumour resection. These models were used in subsequent experiments to investigate the effects of FAK silencing and inhibition on the survival of normal urothelial and urothelial cancer cells.

The relative expression levels of UPK1B and UPK3A were significantly higher in differentiated NPU cells than in 7-day RT4 and 7-day T24 cells (Figure 3A). In contrast, the expression level of CDH1 was higher in 7-day RT4 and differentiated NPU cells, while no expression was detected in 7-day T24 cells. Conversely, 7-day T24 cells had a significantly higher expression level of CDH2 than 7-day RT4 and differentiated NPU cells. PTK2 expression was not significantly different among differentiated NPU, RT4 and T24 cells.

FIGURE 3.

Molecular characterisation of normal urothelial cells (NPU) and bladder cancer cells (RT4 and T24) in vitro.(A) Relative expression levels of UPK1B, UPK3A, CDH1, CDH2 and PTK2 were determined using qRT-PCR in differentiated NPU cells (diff NPU), 7-day RT4 and 7-day T24 cells (RT4 (7 d) and T24 (7 d)). (B) Expression of E-cadherin, N-cadherin, FAK and p-FAK in 2-day NPU cells (NPU (2 d)), diff NPU, 2-day RT4 cells (RT4 (2 d)), RT4 (7 d), 2-day T24 cells (T24 (2 d)) and T24 (7 d), determined by western blot. (C) The relative protein expression of E-cadherin, N-cadherin, FAK and p-FAK in NPU (2 d), diff NPU, RT4 (2 d), RT4 (7 d), T24 (2 d) and T24 (7 d) normalised to the expression of β-actin. The results are presented as median with interquartile range. (D) Quantification of FAK-positive RT4 (7 d) and T24 (7 d) by flow cytometry. Data are presented as the mean ± the standard error of the mean (SEM). *P < 0.05.

In addition, we performed western blot analysis to determine the relative protein expression levels of E-cadherin, N-cadherin, FAK and p-FAK in the in vitro models (Figures 3B, C and Supplementary Figure 2). The relative protein expression level of E-cadherin was significantly higher in 2-day RT4 than in 7-day RT4 cells, and both had significantly higher protein expression levels of E-cadherin than NPU and T24 cells, regardless of the number of cultivation days (Figure 3C). Notably, 2- and 7-day T24 cells did not express E-cadherin but showed a significantly higher protein expression level of N-cadherin than 2-day NPU, differentiated NPU, 2-day RT4 and 7-day RT4 cells. 7-day T24 cells showed significantly higher levels of N-cadherin than 2-day T24 cells. In addition, the expression of FAK and p-FAK was significantly higher in RT4 and T24 cells compared to NPU cells (Figures 3B, 3C and Supplementary Figure 2).

To validate the western blot data on FAK expression in urothelial cancer cell lines RT4 and T24, we quantified the percentage of FAK-positive cells by flow cytometry. Our results showed no significant differences in FAK expression between 7-day RT4 and 7-day T24 cells (Figures 3D and Supplementary Figures 3-6).

To investigate the role of FAK in urothelial cancer cells, we performed gene electrotransfer (GET) of three anti-FAK plasmids to induce FAK silencing in 2-day RT4 and 2-day T24 cells (Figure 4). Plasmids p44, p45 and p46 encode both anti-FAK miRNA and green fluorescent protein (GFP) mR-NA, which share the same promoter region. All anti-FAK plasmids were successfully transfected into RT4 and T24 cells, as shown by the GFP-positive cells (Figures 4A, D). FAK silencing was not significantly different for all plasmids tested in 2-day RT4 cells and was only about 10% lower than that of the control (Figure 4B). However, all three plasmids induced a significant increase in apoptosis, but not necrosis, in 2-day RT4 cells (Figure 4C). All three plasmids induced a statistically significant silencing of FAK in 2-day T24 cells, especially plasmid p45 (Figure 4E). All three anti-FAK plasmids caused a significant increase in apoptosis of 2-day T24 cells, whereas only plasmids p44 and p45 caused a statistically significant increase in necrosis (Figure 4F).

FIGURE 4.

Effect of focal adhesion kinase (FAK) silencing on the survival of 2-day RT4 and 2-day T24 cells (RT4 (2 d) and T24 (2 d)). (A–C) Gene electrotransfer (GET) of anti-FAK plasmids p44, p45 and p46 in RT4 (2 d) did not result in statistically significant FAK silencing in RT4 cells but did result in their apoptosis. (D–F) GET of anti-FAK plasmids p44, p45 and p46 caused statistically significant FAK silencing and led to apoptosis of T24 (2 d). Plasmids p44 and p45 caused a statistically significant increase in necrosis of T24 (2 d). *P < 0.05, **P < 0.001. (G) T24 cells were stably transfected with anti-FAK plasmid p45 after 14 days. However, growth of these cells was limited to a single colony, as seen under fluorescence and phase contrast microscope. Data are presented as the mean ± the standard error of the mean (SEM).

To achieve stable silencing of FAK, a stably transfected T24 cell line was produced using GET with the anti-FAK plasmid p45. The growth of these cells was inhibited and limited to the size of a colony that declined after 14 days, demonstrating a crucial role of FAK in the growth and survival of T24 cells (Figure 4G).

The effects of FAK inhibition on normal urothelial and urothelial cancer cells have not been thoroughly investigated. We analysed the role of FAK inhibition on the survival of normal urothelial and urothelial cancer cells with the aim of determining a possible treatment modality in the form of intravesical instillations after bladder tumour resection.

We measured the viabilities of different in vitro models following treatment with the FAK inhibitors PND-1186 (1 μM and 10 μM), PF-573228 (10 μM and 100 μM) and defactinib (10 μM and 100 μM). Incubation with 1 μM PND-1186 caused a significantly lower viability of 2-day T24 cells (96.1 ± 2.6%) compared to differentiated NPU cells (100.4 ± 2.3%) after treatment for 2 h per day for 3 consecutive days (3 × 2 h). Incubation with 10 μM PND-1186 for 3 × 2 h resulted in a greater decrease in the viabilities of both cancer cells lines (91.0 ± 15.4% for 2-day RT4 and 90.0 ± 3.6% for 2-day T24 cells) compared with differentiated NPU cells (94.5 ± 8.7%), but the difference was not significant. The decrease in the viability was also statistically significant for 2-day T24 cells after the 3 × 2 h treatment compared to T24 control cells (Figure 5A). Treatment with PND-1186 for 1 × 2 h did not lead to a significant decrease in the viabilities of any cells treated with inhibitors compared to their respective controls, nor among the different in vitro models.

FIGURE 5.

Effect of focal adhesion kinase (FAK) inhibitors on viabilities of differentiated normal urothelial (NPU) cells (diff NPU), 2-day RT4 and 2-day T24 cells (RT4 (2 d) and T24 (2 d)). (A) Treatment with FAK inhibitor 1 μM PND-1186 for 2 h per day for 3 consecutive days (3 × 2 h) led to a significantly lower viability of T24 (2 d) than diff NPU. (B) After 3 × 2 h treatment with 100 μM PF-573228, the viability of diff NPU was significantly higher than the viabilities of RT4 (2 d) and T24 (2 d). (C) Treatment with 100 μM defactinib for 3 × 2 h resulted in the greatest decrease in the viabilities of RT4 (2 d) and T24 (2 d) compared to the viability of diff NPU. Data are presented as the mean ± the standard error of the mean (SEM). To differentiate the statistical analysis, we compare the viability between different in vitro models (*, **, ***) and the viability between the control and the treated sample within each cell type (#). #P < 0.05. *P < 0.05. **P < 0.005. ***P < 0.001.

Incubation with 10 μM PF-573228 only caused a significant difference between the viabilities of 2-day T24 cells (114.4 ± 9.0%) and differentiated NPU cells (96.7 ± 7.7%) after treatment for 2 × 2 h. On the other hand, treatment with 100 μM PF-573228 for 3 × 2 h resulted in significantly lower viabilities of both 2-day RT4 and 2-day T24 (51.6 ± 16.1% and 66.1 ± 9.9%, respectively) compared to differentiated NPU cels (99.5 ± 18.1%) (Figure 5B). Treatment with 100 μM PF-573228 for 2 × 2 h resulted in a significant decrease in the viability of differentiated NPU and 2-day RT4 cells compared to their respective controls (87.0 ± 7.1% and 71.6 ± 14.9%, respectively). Similarly, treatment with PF-573228 for 1 × 2 h did not result in significant differences in the reduction of cell viabilities compared to controls, nor among the cell viabilities of the different in vitro models.

Treatment with 10 μM defactinib did not lead to a significant difference between the viabilities of any of the different in vitro models. On the other hand, treatment with 100 μM defactinib significantly reduced the viability of urothelial cancer cells, while the viability of normal urothelial cells was unaffected (Figure 5C). It caused statistically significant differences between the viabilities of normal urothelial compared to urothelial cancer cells at each treatment interval. The 1 × 2 h treatment caused a statistically significant difference between the viabilities of differentiated NPU (108.6 ± 17.7%) compared to 2-day RT4 and 2-day T24 cells (55.6 ± 18.0% and 58.9 ± 14.2%, respectively). The 2 × 2 h treatment caused an even greater decrease in the viabilities of 2-day RT4 and 2-day T24 (15.1 ± 9.6% and 12.4 ± 3.9%, respectively) compared to differentiated NPU cells (90.4 ± 10.1%). Our study showed that differentiated NPU cells remained highly viable after FAK inhibition with 100 μM defactinib for 3 × 2 h (108.4 ± 17.1%), whereas 2-day RT4 and 2-day T24 cells were almost completely eliminated with viabilities of 4.1 ± 2.7% and 7.6 ± 2.9%, respectively. These findings could pave the way for clinical applications, such as a treatment involving the instillation of defactinib into the bladder for 2 h at a time. Such adjuvant therapy could be beneficial in destroying the residual urothelial cancer cells from papillary and muscle invasive carcinomas after resection of the bladder tumour, while leaving normal urothelial cells unaffected.

The treatment of in vitro models representing the differentiated urothelium and later stages following tumour resection showed similar trends in results to those observed in the early post-surgery phase in vitro models. Treatment with the FAK inhibitor PND-1186 at both tested concentrations resulted in a statistically significant increase in the viabilities of 7-day T24 cells after the 1 × 2 h treatment (105.4 ± 9.4% and 108.3 ± 6.3%). Incubation with 10 μM PND-1186 caused a statistically significant decrease in the viabilities of differentiated NPU, 7-day RT4 and 7-day T24 cells compared to their untreated controls after the 2 × 2 h treatment (88.4 ± 5.8%, 95.5 ± 2.0% and 89.2 ± 7.4%, respectively). A 3 × 2 h treatment did not result in significant changes in cell viabilities. (Figure 6A).

FIGURE 6.

Effect of focal adhesion kinase (FAK) inhibitors on viabilities of differentiated normal urothelial (NPU) cells (diff NPU), 7-day RT4 and 7-day T24 cells (RT4 (7 d) and T24 (7 d)). (A) After treatment with 10 μM PND-1186 for 2 h per day for 3 consecutive days (3 × 2 h), no statistically significant differences in cell viability were observed among diff NPU, RT4 (7 d) and T24 (7 d). (B) Treatment with 100 μM PF-573228 for 3 × 2 h resulted in a significant decrease in the viabilities of both RT4 (7 d) and T24 (7 d) compared to the viability of diff NPU. (C) Treatment with 100 μM defactinib for 3 × 2 h caused the greatest difference between the viabilities of diff NPU, RT4 (7 d) and T24 (7 d). (D) Western blot of p-FAK after the 3 × 2 h treatment with 100 μM defactinib in 7-day NPU cells (NPU (7 d)), RT4 (7 d) and T24 (7 d) with GAPDH-loading control. Data are presented as the mean ± the standard error of the mean (SEM). To differentiate the statistical analysis, we compare the viability between different cell types (*, **, ***) and the viability between the control and the treated sample within each cell type (#). #P < 0.05. *P < 0.05. **P < 0.005. ***P < 0.001.

Treatment with 10 μM PF-573228 for 1 × 2 h led to a significantly higher viability of 7-day T24 (114.1 ± 7.0%) compared to the viability of 7-day RT4 cells (88.2 ± 5.6%). However, it led to a significantly lower viability of 7-day T24 (90.0 ± 11.4%) compared to differentiated NPU cells (108.9 ± 11.3%) after treatment for 3 × 2 h. Incubation with 100 μM PF-573228 caused a decrease in the viabilities of differentiated NPU, 7-day RT4 and 7-day T24 cells compared to their untreated controls after the 2 × 2 h treatment, but no significant differences between the viabilities of the different in vitro models (87.0 ±7.1%, 82.7 ± 2.8% and 81.6 ± 6.9%). A 3 × 2 h treatment resulted in a statistically significant reduction in the viabilities of 7-day RT4 and 7-day T24 cells (63.3 ± 6.1% and 64.2 ± 8.5%, respectively). In contrast, the viability of differentiated NPU cells remained high and was not significantly lower than that of the untreated control (99.5 ± 18.1%) (Figure 6B).

Again, defactinib was the most potent FAK inhibitor. Treatment with 10 μM defactinib caused, similarly to 10 μM PF-573228, a significantly higher viability of 7-day T24 (117.9 ± 8.9%) compared to 7-day RT4 cells (97.2 ± 4.7%) after 1 × 2 h of treatment and a significantly lower viability of 7-day T24 (91.2 ± 9.9%) compared to differentiated NPU cells (116.0 ± 13.4%) after 3 × 2 h of treatment. Incubation with 100 μM defactinib caused the greatest decrease in the viabilities of 7-day RT4 and 7-day T24 cells (45.5 ± 8.5% and 36.0 ± 6.2%, respectively) after the 3 × 2 h treatment, while differentiated NPU cells remained intact (108.4 ± 17.1%) (Figures 6C and Supplementary Figure 7). We additionally performed a western blot, which showed that p-FAK was present in 7-day RT4 and 7-day T24 cells without treatment but was absent after 3 × 2 h of treatment with 100 μM defactinib (Figure 6D). Supplementary Figure 8 illustrates the effect of FAK inhibitors on FAK and p-FAK expression in 7-day T24 cells after 24-hour treatment with FAK inhibitors. The p-FAK was not detected in either control or treated 7-day NPU cells, consistent with the resistance of NPU cells and the susceptibility of urothelial cancer cells to FAK inhibition observed in the cell viability assays.

To visualise the effects of defactinib on urothelial cancer cells, we immunolabelled 7-day RT4, 7-day T24 and their control cells after the treatment with 100 μM defactinib for 2 × 2 h (this treatment regimen was chosen to obtain sufficient numbers of cells for analysis). After the treatment of 7-day RT4 cells for 2 × 2 h, we observed fewer p-FAK-labelled cells and no significant changes in the actin filament organisation compared with untreated cells (Figure 7A). After the treatment of 7-day T24 cells for 2 × 2 h, the clusters of p-FAK were observed, however no significant changes in the organisation of actin filaments were observed in comparison to untreated cells (Figure 7B).

FIGURE 7.

Focal adhesion kinase (FAK) inhibition with 100 μM defactinib for 2 h per day for 2 consecutive days (2 × 2 h) leads to caspase-3-mediated apoptosis. (A) Treated 7-day RT4 cells (RT4 (7 d)) showed lower p-FAK-labelling than controls. White arrows indicate gaps in the distribution of actin filaments in treated RT4 cells. (B) 7-day T24 cells (T24 (7 d)) contained p-FAK labelling in clusters after treatment with 100 μM defactinib for 2 × 2 h (white arrowheads). No significant changes in actin filament localisation were observed in treated and control T24 cells. Scale bars: 50 μm (white), 10 μm (black). (C) Western blot analysis showed that cleaved caspase-3 was present only in RT4 (7 d) and T24 (7 d) after treatment with 100 μM defactinib for 2 h per day for 3 consecutive days (3 × 2 h).

To define the mechanism of action of FAK inhibitors, we analysed the type of cell death of normal urothelial and urothelial cancer cells. We performed a western blot of cleaved caspase-3, i.e. the large fragment (17/19 kDa) of activated cas-pase-3 produced by cleavage near Asp175 during apoptosis. Analysis showed cleaved caspase-3 in 7-day RT4 and 7-day T24 cells after 3 × 2 h of treatment with 100 μM defactinib. Cleaved caspase-3 was absent in all other treated in vitro models, including differentiated NPU cells, after 3 × 2 h of treatment with 100 μM defactinib, consistent with the results of cell viability assays (Figures 7C and Supplementary Figure 9).

The main limitation of our study arises from the limited number of human biological samples used in certain analyses (2–4 biopsies), which may introduce bias into the statistical outcomes. Larger sample sizes would improve the statistical power and robustness of the findings. Specifically, the analysis included three samples of pT1 tumours and two samples of pT2 tumours, which may explain the differences in E-cadherin levels. We hypothesise that, as T1 represents an earlier stage of tumour development, it might still benefit from E-cadherinmediated cell-to-cell junctions to facilitate tumour survival and angiogenesis. On the other hand, as the tumour advances and tight junctions disintegrate, E-cadherin may still be expressed and detected by antibodies, although it may no longer be functional at the plasma membrane. This could potentially explain the high levels of E-cadherin observed in pT1 tumours.

A limitation of our study is the absence of a control miRNA, which could help to delineate the non-specific effects of plasmid DNA delivery. It is known that plasmid DNA introduced into cells can activate DNA sensing pathways, potentially triggering immune responses or leading to cell death.^63,64 Therefore, some portion of the cell death observed following gene electrotransfer (GET) of anti-FAK plasmids may be attributed to these non-specific effects. Despite this, our findings are supported by the induction of caspase-3-mediated apoptosis observed with FAK inhibitors, which directly target and suppress FAK activity. The similarity in outcomes between these approaches strongly suggests that the primary mechanism driving apoptosis in our study is the inhibition of FAK rather than the non-specific activation of DNA sensing pathways. To further validate this conclusion, future studies could incorporate a control miRNA to better distinguish specific effects from potential artefacts of plasmid delivery. Additionally, experiments using scrambled miRNA or an empty plasmid would further help confirm that the observed effects are due to FAK inhibition rather than off-target or plasmid-related artefacts.

Although we identified caspase-3-mediated apoptosis after FAK inhibition, more comprehensive analysis of the apoptotic pathway, including upstream and downstream regulators, would provide a deeper understanding of the molecular events leading to cell death.