Altered expression and modulation of the two-pore-domain (K_2P) mechanogated potassium channel TREK-1 in overactive human detrusor

Abstract

Detrusor overactivity (DO) is the abnormal response of the urinary bladder to physiological stretch during the filling phase of the micturition cycle. The mechanisms of bladder smooth muscle compliance upon the wall stretch are poorly understood. We previously reported that the function of normal detrusor is regulated by TREK-1, a member of the mechanogated subfamily of two-pore-domain potassium (K_2P) channels. In the present study, we aimed to identify the changes in expression and function of TREK-1 channels under pathological conditions associated with DO, evaluate the potential relationship between TREK-1 channels and cytoskeletal proteins in the human bladder, and test the possibility of modulation of TREK-1 channel expression by small RNAs. Expression of TREK-1 channels in DO specimens was 2.7-fold decreased compared with control bladders and was associated with a significant reduction of the recorded TREK-1 currents. Isolated DO muscle strips failed to relax when exposed to a TREK-1 channel opener. Immunocytochemical labeling revealed close association of TREK-1 channels with cell cytoskeletal proteins and caveolins, with caveolae microdomains being severely disrupted in DO specimens. Small activating RNA (saRNA) tested in vitro provided evidence that expression of TREK-1 protein could be partially upregulated. Our data confirmed a significant downregulation of TREK-1 expression in human DO specimens and provided evidence of close association between the channel, cell cytoskeleton, and caveolins. Upregulation of TREK-1 expression by saRNA could be a future step for the development of in vivo pharmacological and genetic approaches to treat DO in humans.

the physiological function of the urinary bladder includes the cyclical stretch of its wall during the storage phase of the micturition cycle. The extent of bladder wall expansion before the sensations of fullness and urgency develop defines bladder compliance; stretch of the bladder wall activates many mechanosensitive receptors expressed on the urothelial and detrusor smooth muscle cells followed by transduction of mechanosensitive stimuli to the peripheral afferents and, further, to the central nervous system (5). Pathological alterations in mechanotransduction often lead to the development of detrusor overactivity (DO), which is a cosymptom of many lower urinary tract dysfunctions including overactive bladder (OAB), obstructed bladder, diabetic overactivity, urinary incontinence, and bladder pain syndrome (2–4, 21).

Mechanosensitive ion channels play an important role in the physiology of mechanotransduction, as they are capable of detecting and transducing membrane mechanical forces into electrical and chemical intracellular signals (7). Previous animal (9, 10, 38) and human (8, 9, 12, 31) studies provided evidence that stretch-activated two-pore-domain (K_2P, KCNK) potassium channels are key regulators of bladder wall relaxation during physiological stretch. TREK-1 is the predominantly expressed member of the K_2P family in the human bladder and has been confirmed to play a significant role in bladder compliance via hyperpolarization of smooth muscle cell membrane and reduction in detrusor excitability (35). A decrease in functional expression of TREK-1 channel in the detrusor was shown to lead to the development of DO in animal models of partial bladder outlet obstruction (PBOO) (9). In humans, increased expression of TREK-1 was detected in pregnant women’s myometrium but was followed by a decline in expression in term laboring tissues (14, 15, 27). To date, no studies have evaluated the changes in expression and function of TREK-1 channels in the human urinary bladder in pathophysiological conditions such as DO.

Upon membrane stretch, mechanical force could be transmitted to the channel directly or via accessory proteins/macromolecular complexes tethering the channel to the cell matrix or cytoskeleton (6). Previous studies confirmed that when mechanical force is transmitted to the channel through the lipid bilayer, the cytoskeleton acts as a tonic repressor, thus limiting channel activation by membrane tension (34, 42, 43, 46). The transfer of force through connections between the actin cytoskeleton and mechanosensitive ion channels was previously observed in human gastrointestinal smooth muscle (52).

Expression of TREK-1 could be also modulated by the expression of caveolins, integral membrane proteins required for the formation and maintenance of caveolae (29, 51). These membrane invaginations are enriched with signaling molecules, ion channels, and regulatory proteins necessary for protein trafficking and signal transduction in different cell types (19, 54). There are three known members of the caveolin protein family (28). Caveolin-1 and -2 are expressed in most cells, whereas caveolin-3 is a muscle-predominant isoform (53). Caveolin-2 is often colocalized with caveolin-1, with this interaction required for membrane localization and stabilization (40). All three caveolins are expressed in the urinary bladder (29, 30, 53). Caveolin-1 knockout mice also lack caveolae in the bladder and exhibit several urological defects including bladder hypertrophy and decreased contractility during stimulation with carbachol (29, 32, 33, 55). In human studies, the expression of caveolin-1 was significantly downregulated in the obstructed bladders of men with benign prostatic hyperplasia (11). However, whether mechanogated properties of the TREK-1 channel in the human bladder are modulated by any cytoskeletal proteins is yet to be established. Therefore, in the present study, we tested the hypothesis that changes in expression and/or function of TREK-1 are associated with DO in humans. Further, we also evaluate the relationship between TREK-1 channels and cytoskeletal proteins, and the possible modulation of TREK-1 channel by siRNA and saRNAs.

MATERIALS AND METHODS

Human subjects and tissue processing.

Human full-thickness urinary bladder specimens (N = 48) were obtained from deidentified healthy donors and patients with muscle invasive bladder cancer who underwent cystectomy. Whole healthy bladders from organ donors were provided by the National Disease Research Interchange Network (NDRI) following ethical and research protocols approved by the NDRI and the Colorado Multiple Institutional Review Board (COMIRB). Bladders were delivered in human transplant medium to the laboratory within 24 h of collection. The protocols to use surgical waste tissue from the human bladders isolated during cystectomies were approved by the University of Pennsylvania Institutional Review Board (no. 812269) and the Colorado Multiple Institutional Review Board (COMIRB, no. 14–1557). Inclusion criteria comprised negative dipstick or urinalysis before surgery. Exclusion criteria included: chemotherapy; pelvic radiation; the use of narcotic analgesics for >3 days/wk; impaired liver and/or kidney function; presence of discrete anatomical urinary tract abnormalities; sacral neurostimulation; and the presence of neurodegenerative diseases. Evaluation of lower urinary tract symptoms (LUTS) was performed by the American Urological Association (AUA) symptom score questionnaire, and experimental groups were based on the AUA symptom score, with AUA < 8 being considered as a control group (N = 23) and with AUA > 20 being a group with idiopathic DO (N = 25). Healthy, full-thickness bladder tissue (~1–1.5 square inch) obtained no closer than 2–3 cm to the edge of the cancerous tissue was placed in ice-cold Ca²⁺ free Tyrode’s solution upon delivery to the laboratory as previously described (35). Detrusor smooth muscle tissue was divided into several pieces and used for whole cell patch-clamp recordings, immunohistochemical (IHC), and biochemical analysis.

Dissociation of primary single smooth muscle cells from the human detrusor.

Small strips of denuded detrusor tissue were cut into 2- to 3-mm pieces and placed in a 15-ml conical tube containing prewarmed (37°C) low-Ca²⁺ (10 µM) Tyrode’s solution supplemented with collagenase (1.3 mg/ml; Yakult, Tokyo), bovine serum albumin (BSA; 2.0 mg/ml; Sigma-Aldrich, St. Louis, MO), and trypsin (2.0 mg/ml; Sigma-Aldrich). The small tissue pieces were incubated in this solution at 37°C for 20 min under gentle constant agitation. After incubation, the tissue was left to settle for 2–3 min and the supernatant was aspirated. Next, prewarmed (37°C) low-Ca²⁺ Tyrode’s solution containing 1.3 mg/ml of collagenase and 2.0 mg/ml of BSA was added to the tissue and was incubated for an additional 10 min at 37°C under gentle constant agitation. Following this, the tissue was left to settle for 2–3 min, the supernatant was removed, and 3 ml of prewarmed (37°C) low-Ca²⁺ Tyrode’s solution containing 2.0 mg/ml of BSA was added to the tube. Individual human bladder smooth muscle cells (BSMC) were released by mechanical trituration through a series of blunt glass Pasteur pipettes with decreasing tip diameters. Dispersed cells were plated on poly-l-ornithine (Sigma-Aldrich)-coated, 25-mm glass coverslips (Carolina Biological Supply, Burlington, NC) and were allowed to settle down for 15–20 min at room temperature (RT). Freshly dissociated noncontracted BSMC were used for immunohistochemical labeling and electrophysiological recordings.

Immunocytochemistry and imaging.

Freshly dissociated BSMC and a human bladder smooth muscle cell line (hBSMC, lot no. 0000198879, LONZA, Basel, Switzerland) were used for immunocytochemical labeling. The cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PFA-PBS, in mM: 137.0 NaCl, 2.7 KCl, 10.0 Na₂HPO₄, 2.0 KH₂PO₄; Sigma Aldrich) for 15 min at RT, and then washed twice for 5 min each time with 0.1 M PBS. Nonspecific labeling was minimized by incubation in a blocking solution containing 1.0% BSA, 0.1% Triton-X (Sigma Aldrich), and 0.2% coldwater-fish gelatin (Sigma-Aldrich) in PBS for 30 min at RT. The cells were then incubated with the following primary antibodies for 12–24 h at 4°C: 1:500 rabbit anti-TREK-1 (H75, Santa Cruz Biotechnology, Dallas, TX); 1:200 rabbit anti-caveolin-1, anti-caveolin-2, and anti-caveolin-3 (Cell Signaling, Danvers, MA); and 1:500 mouse anti-tubulin and anti-actin (Abcam, Cambridge, MA). After primary antibody incubation, cells were washed three times for 5 min each in PBS at RT and then incubated with secondary fluorescent antibody conjugates [goat anti-rabbit or goat anti-mouse Texas Red (1:250) or Alexa 647 (1:500; Invitrogen, Carlsbad, CA)] for 1–2 h at RT. Coverslips were washed with PBS before being mounted on glass slides in Fluoroshield mounting media with DAPI (Abcam, Cambridge, MA). Imaging was performed on an Olympus FV1000 confocal microscope with 60× Plan-APo/1.4-NA objective and FV-viewer software (Olympus, Tokyo). For detection of fluorescence, three-dimensional Z-stack images of x-y sections with 1.0- to 1.5-µm steps were collected. Two-dimensional average-intensity projection images were generated for analysis. Quantification of fluorescent images was performed in Fiji [Fiji is ImageJ (48)].

Tissue immunohistolabeling.

To evaluate spatial distribution of TREK-1 channels and sarcoglycans, isolated detrusor specimens were fixed in paraformaldehyde (4%), cryoprotected in sucrose (20%) for 3 days, embedded in OCT, and sectioned on a cryotome at 10-μm increments. Tissue sections were incubated in blocking solution with mouse monoclonal antibodies against β-sarcoglycan (NCL-b-SARC, 1:200, Novocastra Laboratories, Newcastle-upon-Tyne, UK) or mouse monoclonal anti-α-sarcoglycan (NCL-α-SARC, 1:200, Novocastra Laboratories) and rabbit polyclonal TREK-1 (H-75, Santa Cruz Biotechnology) at 4°C overnight. The next day, sections were rinsed three times with PBS for 10 min each, and then incubated with biotinylated goat anti-mouse or anti-rabbit IgG (1:200) in blocking solution for 1 h at RT. For double-antibody localization, tissue sections were first stained with β- or α-sarcoglycans and Alexa-fluor 594 (1:400, Invitrogen) and then with antibody to TREK-1 and Alexa-fluor 488 (1:400, Invitrogen). A number of controls were employed for these studies: 1) the use of an unrelated antibody or PBS in place of the primary antibody, 2) the use of the secondary antibody or the fluorochrome-conjugated antibodies alone, and 3) the use of PBS in place of the secondary antibodies. In all instances, controls showed either no staining or no detectable background staining. Imaging data were obtained using a Bio-Rad Scanning Confocal Microscope equipped with three lasers and appropriate software for image analysis as previously described (6, 20).

Measurements of smooth muscle tone in the human detrusor strips in vitro.

Denuded smooth muscle strips (~40–50 mg, N = 4, n = 5–6 per group) were placed in a glass organ bath (Radnoti, Monrovia, CA) containing 10 ml of Tyrode’s buffer (in mM: 124.9 NaCl, 2.5 KCl, 23.8 NaHCO₃, 0.5 MgCl₂, 0.4 NaH₂PO₄, 1.8 CaCl₂, 5.0 glucose) equilibrated with 95% O₂-5% CO₂. One end of the strip was attached to a glass rod at the bottom of the organ chamber, while the other end was attached to a force displacement transducer connected to a computerized system for data acquisition and analysis (AD Instruments, Colorado Springs, CO). After a 1-h equilibration, the length of optimal force development (L_o) was determined by manually increasing the length of each strip by 1.5-mm increments until the maximal contractile force in response to electric field stimulation (EFS; 80 V, 1 ms, and 32 Hz) was achieved. The bath solution was then changed to fresh Tyrode’s buffer containing 1 μM of TTX (Sigma-Aldrich) to reduce the potential confounding effects of nerve stimulation on BSM contractility. The strips were incubated for 30 min, followed by the treatments with either arachidonic acid (AA; 10 μM) or l-methionine (1 mM) for 60 min. Changes in smooth muscle tone/baseline were recorded during drug applications.

Patch-clamp recordings of TREK-1 currents.

Electrophysiological recordings used whole cell configuration, voltage-clamp protocol (39). Extracellular solution contained (in mM) 135.0 NaCl, 5.4 KCl, 1.0 MgCl₂, 1.0 CaCl₂, 0.33 NaHPO₄, 5.0 HEPES, 5.5 glucose, pH 7.4) and was supplemented with TEA (5 mM), 4-AP (1 mM), nifedipine (1 µM), and ω-conotoxin GVIA (5 µM). All experiments were performed at room temperature. The recording patch electrodes were made from borosilicate glass capillary tubes (GC120F-10, Warner Instruments, Hamden, CT) and were filled with an intracellular solution containing (in mM) 30.0 KCl, 100.0 potassium aspartate, 2.0 MgCl₂, 1.0 EGTA, 5.0 HEPES, pH 7.2. The typical resistances of the electrodes when filled with this solution varied between 2.5 and 3.4 MΩ. The patch pipette’s Ag-Cl wire was connected to the input of an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA). pClamp software (Molecular Devices, Sunnyvale, CA) was used to generate voltage-clamp protocols, as well as for data acquisition and analysis. Recordings were subjected to analysis only when capacitative transients showed monoexponential decays, membrane resistance was equal or larger than 120 MΩ, access resistance was equal or lower than 12 MΩ and remaining stable during the course of the experiment, and the leak current was smaller than 150 pA. The series resistance was compensated (65–70%) and monitored throughout the course of the experiment. If the SR changed by >20%, the cell was eliminated from analysis. The input capacitance of the cells included in this study was 56.3 ± 4.2 pF. Steady-state current amplitude was measured as the average value of a 10-ms interval 55 ms after the depolarizing test voltage step began. Normalization of current amplitude to cell size was achieved by dividing the steady-state current amplitude by the membrane capacitance, which serves as an indicator of membrane surface area [1 pF/cm² (39)]. For I–V plots, we present current density as a function of voltage.

TREK-1 gene silencing and activating with small RNAs.

Experiments to test the effectiveness of gene silencing techniques were performed using the hBSMC cell line. Cells were cultured in SmGM-2 BulletKit Medium (LONZA; Basel, Switzerland) at 37°C in a 5% CO₂ humidified incubator (Sanyo IncuSafe, Sanyo Biomedical, Tokyo). The medium was replaced every 3 days until cells reached 80% confluence. Cells were then rinsed once with prewarmed (37°C) PBS, and 1 ml/25 cm² of a 1X trypsin/EDTA solution (ThermoFisher, Waltham, MA) was added to the culture. After 2–10 min, detached cells were resuspended in 3 ml of SmGM-2 BulletKit medium, and counted in a hemocytometer. Cells (1.8 × 10⁵) were transfected with 20 nM of Silencer Select Human TREK-1 siRNA (Ambion, ThermoFisher) or Silencer Select Negative Control #1 siRNA (Ambion, ThermoFisher) using Lipofectamine RNAiMAX Reagent (Invitrogen, ThermoFisher) in 12-well plates and incubated at 37°C for 48 h.

We have also designed in-house three sets of saRNA oligos: 1) saRNA-TREK1-#1 at −326 bp (sense GCGAAGAGGGGCUGAACGAuu; antisense uuCGCUUCUCCCCGACUUGCU); 2) saRNA-TREK1-#2 at −490 bp (sense AGAGGAGGGAGUUCCGAAAuu; antisense uuUCUCCUCCCUCAAGGCUUU); and 3) saRNA-TREK1-#3 at −639 bp (sense GAUCUUUAGAAGGGAGCAUuu; antisense uuCUAGAAAUCUUCCCUCGUA) using the human TREK-1 promoter sequence (NM_014217). hBSMC were cultured until 75% of confluency and then transfected with 20 nM of TREK-1-saRNAs from each set. Cells were harvested at 3 days posttransfection and prepared for qRT-PCR and Western blot analysis.

Real-time quantitative PCR (qRT-PCR) and Western blotting.

Total RNA was extracted and purified from full-thickness human bladder specimens, as well as from hBSMC cells after TREK-1-siRNA or TREK-1-saRNA transfections using RNeasy Plus Mini Kit (Qiagen, Valencia, CA) according to manufacturer’s instructions. cDNA was made using QuantiTect Reverse Transcription Kit (Qiagen). Real-time PCR was performed on a LightCycler 96 Real-time PCR system (Roche, Indianapolis, IN) using TaqMan Universal PCT master mix. TaqMan probes and primers for TREK-1 (Hs01005159-m1) and GAPDH (Hs02758991-g1) were purchased from Applied Biosystems (ThermoFisher). Samples were amplified in triplicate, and nontransfected cells were included as negative control. TREK-1 gene expression was normalized to GAPDH expression.

Total protein extraction was conducted by conventional cell lysis with RIPA buffer containing protease inhibitors (Roche protease inhibitor cocktail, Roche, Indianapolis, IN). Control and experimental protein samples (20 µg each) were size fractionated by SDS-PAGE using 4–20% polyacrylamide Mini-Protean gels (Bio-Rad, Hercules, CA). Proteins were electroblotted to polyvinylidene difluoride membranes (PVDF, Bio-Rad) which were incubated in Tris-buffered 5% nonfat milk in PBS blocking buffer for 1 h at RT and subsequently washed 3 times in PBS with 0.1% Tween-20 (PBST, Sigma-Aldrich) for 10 min each. Membranes were then incubated with anti-TREK-1 (H-75, 1:200; Santa Cruz Biotechnology) or anti-GAPDH (6C5, 1:1,000; Santa Cruz Biotechnology) primary antibodies in 1% BSA-PBS overnight at 4^oC under continuous shaking. Membranes were rinsed 3 times in PBST for 10 min each and incubated with anti-rabbit or anti-mouse horseradish peroxidase-conjugated antibodies (Promega, Madison, WI) at a 1:2,000 dilution in 1% BSA-PBS for 1 h at RT. Protein bands were visualized using Clarity ECL Western blotting substrate (Bio-Rad). Image densitometry analysis was performed with FluorS MultiImager and Image One Software (Bio-Rad). Volume integration signal intensity was normalized to GAPDH from the same membrane as a loading control.

Statistical analysis of the data.

Data are expressed as means ± SE. The differences between the groups were statistically analyzed using two-way repeated-measures ANOVA followed by comparison between the groups using Bonferroni's t-test method (Systat Software, San Jose, CA). Plots were made with OriginLab Data analysis and graphing software (OriginLab, Northampton, MA). Differences between the groups were considered statistically significant at P ≤ 0.05.

RESULTS

Expression of TREK-1 channel is downregulated in DO specimens.

Previous animal studies provided evidence that TREK-1 channel is downregulated under pathological conditions associated with overactive bladder (9). To determine if TREK-1 transcript levels were altered in humans with DO, we compared the expression level of TREK-1 proteins in bladder specimens obtained from patients with and without DO. Figure 1A shows representative Western blot data reflective of reduced levels of TREK-1 protein expression in smooth muscle tissue obtained from patients with DO compared with normal bladders (NB). When normalized to GAPDH, the level of TREK-1 expression in the NB was 2.7-fold higher than in the DO group (P ≤ 0.05, Fig. 1B). Decreased expression of TREK-1 protein was detected in parallel with downregulation of TREK-1 mRNA in DO specimens (P ≤ 0.05 compared with NB group, Fig. 1C). To evaluate whether a statistically significant decrease in TREK-1 channel expression in DO specimens could be also associated with altered current expression, we next examined freshly dissociated bladder smooth muscle cells (BSMC) by whole cell patch-clamp technique (Fig. 1D). In all recorded cells, the observed resting membrane potential was −54.0 ± 1.5 mV. However, the measured I_TREK-1 peak current in NB cells was ~10-fold larger than in DO cells (1,087.0 ± 99.3 vs. 98.6 ± 66.3 pA/pF at 60 mV, respectively, n = 6 in each group, P ≤ 0.001). Although no changes in voltage activation threshold were detected between the groups, a detailed voltage and kinetic analysis of I_TREK-1 was limited due to small current amplitude observed in DO cells. All together, these results confirmed a reduced expression of TREK-1 channels in human bladders with DO without detectable changes in channel kinetics.

Fig. 1.

Expression of TREK-1 in the normal and overactive human detrusor. A: TREK-1 protein expression in full-thickness bladder tissue obtained from normal bladders (NB1-3) was higher than in tissue obtained from detrusor overactivity bladders (DO1-3). BSMC lane represents human bladder smooth muscle cell line for comparison. B: expression of TREK-1 in NB and DO tissues normalized to GAPDH (*P ≤ 0.05). C: level of mRNA expression normalized to GAPDH (*P ≤ 0.05). D: representative whole cell recordings of TREK-1 currents obtained from freshly dissociated normal (NB, n = 6; top left) and DO BSMC (DO, n = 6, top right). Cells obtained from NB displayed large TREK-1 current (I_TREK-1) amplitude in comparison to those detected in DO cells. TREK-1 currents were evoked by a series of voltage steps from −40 to +60 mV in 10-mV increments from a holding potential of −50 mV. For illustrative purposes, current traces are shown in 20-mV increments. Current-voltage (I-V) relationships (bottom) revealed no changes in current activation threshold. Other kinetic parameters were not evaluated due the small size of TREK-1 current in DO specimens. E: denuded detrusor muscle strips isolated from NB (top trace, n = 8) and DO (n = 7, bottom trace) bladders did not display spontaneous contractile activity. Addition of arachidonic acid (AA; arrow) significantly decreased muscle tone in NB but only slightly in muscle strips from DO bladders. F: addition of l-methionine (pointed by arrow) to NB tissue strip (top trace) increased muscle tone without having any substantial effects on DO muscle strips (bottom trace). G: bar graph shows quantification of data shown in D and E (*P ≤ 0.05 compared with baseline).

We next tested the effects of reduced TREK-1 channel expression on detrusor tone in vitro. Isolated strips of denuded human detrusor (~1 × 4 mm) were incubated with either arachidonic acid, a TREK-1 channel opener (AA, 10 μM), or l-methionine, a TREK-1 channel inhibitor (1 mM), for 60–80 min. Whereas NB strips showed a decrease in tone after application of AA (due to an increase in TREK-1-mediated potassium conductance), detrusor strips from DO specimens failed to relax (n = 8, Fig. 1E). Application of l-methionine caused an increase in muscle tone in NB but had no effect on DO strips reflective of decreased expression of TREK-1 channels in DO specimens (n = 7, Fig. 1, F and G).

Structural relationship between TREK-1 and cell cytoskeleton.

Previous studies established that expression of TREK-1 proteins has profound effects on the cytoskeleton architecture of many cell types (46). Additionally, it was found that cell treatment with latrunculin A or patch excision produced an increase in channel activity suggesting that actin may act as a tonic repressor of TREK-1 channels. We used Lonza’s hBSMC cell line to analyze the distribution of TREK-1 channel transcripts and smooth muscle cytoskeletal proteins by immunocytochemical labeling and confocal fluorescence microscopy. Figure 2 shows immunocytochemical labeling with antibodies against F-actin (Fig. 2A), α-tubulin (Fig. 2B), α-actin (Fig. 2C), and TREK-1 (n = 5 per group, middle panels). Confocal image analysis revealed close association between the cytoskeletal proteins and TREK-1 channel as indicated by their colocalization index (Fig. 2D). Pearson’s correlation coefficient (PCC) was 0.75, 0.72, and 0.45 for F-actin, α-tubulin, and α-actin, respectively. The tension transfer apparatus (TTA) of smooth muscle cells includes cytoskeletal coupling proteins that connect the actin skeleton with transmembrane proteins (7). Previous animal studies established that sarcoglycans play an important role as part of the TTA in the urinary bladder during the physiological stretching upon bladder filling (4). These proteins are part of sarcoglycan/dystrophin complexes which couple the sarcolemma of smooth muscle cells with the intracellular F-actin. To test whether sarcoglycans participate in a macromolecular complex between TREK-1 and the cytoskeleton, we performed double immunolabeling against TREK-1 and either α- or β-sarcoglycans using frozen sections of detrusor tissue. Although we confirmed that both types of sarcoglycans are expressed in the human bladder smooth muscle, the absence of colocalization and the pattern of sarcoglycan staining suggested that these proteins are not in close proximity to one another (n = 3 per group, Fig. 3). Therefore, we conclude that neither α- nor β-sarcoglycans participate as direct connectors between TREK-1 and cytoskeletal complexes in the human bladder.

Fig. 2.

Colocalization of TREK-1 with cytoskeletal proteins in hBSMC. Representative confocal fluorescent images revealed close association of immunolabeled TREK-1 channels (middle columns, in green) with F-actin (A, in red), α-tubulin (B, in red), and α-actin (C, in red). Cell nuclei were stained with DAPI. For colocalization analysis, red and green images were pseudocolored to magenta and cyan, respectively. White color in merged images represents distribution of colocalized points (8 bits) between 2 fluorophores. D: bar graph shows colocalization coefficients for each of tested cytoskeletal proteins and TREK-1 channels. Data are presented as means ± SE. Three different regions in 3 different slides were analyzed per target. Scale bar, 20 µm.

Fig. 3.

Sarcoglycans do not colocalize with TREK-1 channels in the normal human detrusor. Analysis of confocal images of normal human bladder detrusor sections costained with TREK-1 (green, n = 3), and either α- (A, red, left panel) of β-sarcoglycan (B, red, right panel) showed no colocalization of sarcoglycans with TREK-1 channel. Scale bar, 50 µM.

Expression of TREK-1 and caveolins in the human detrusor and BSMC.

The caveolin protein family (caveolins 1–3) are the main structural proteins necessary for the formation of caveolae, membrane invaginations that function as protein trafficking, signal transduction, and lipid regulation centers that are particularly abundant in vascular, endothelial, and smooth muscle cells. Previous studies have shown that all three members of the caveolin family are present in bladder smooth muscle cells; however, their expression patterns at the cellular level have not been confirmed in the human bladder. We performed immunocytochemical labeling for the caveolin family in hBSMC. Confocal immunofluorescence analysis showed diffuse expression for all three caveolin proteins and TREK-1 channels in the cytoplasm of these cells with limited localization at the cell membrane (n = 5 per group, Fig. 4, A, B, and C, left and middle panels). Double immunofluorescence analysis in regions of membrane expression showed a statistically significant colocalization between Cav-1, Cav-3, and TREK-1 channels, whereas the overlap coefficient for Cav-2 and TREK-1 was fourfold lower (Fig. 4D; PCC was 0.81 ± 0.1, 0.20 ± 0.07, and 0.74 ± 0.2 for Cav-1, Cav-2, and Cav-3, respectively; P ≤ 0.001).

Fig. 4.

Cellular distribution of TREK-1 and caveolin-1, -2, and -3 proteins in hBSMC. Representative images of double immunolabeling in hBSMC line with TREK-1 (red, middle column) and caveolin-1 (A), caveolin-2 (B), and caveolin-3 (C). Nuclei were stained with DAPI. All caveolins (green) and TREK-1 (red) images were pseudocolored to magenta and cyan for colocalization analysis (right column). D: correlation coefficients calculated from double-labeled confocal images in regions of membrane protein expression showed a significant overlap of caveolin-1 and -3 with TREK-1. Data are presented as means ± SE. Three different regions from separate slides were analyzed per target (**P ≤ 0.001). Scale bar, 20 µm.

Next, we aimed to confirm if the patterns of caveolin expression detected in hBSMC hold true in freshly dissociated BSMC from the human detrusor. Immunofluorescent analysis of confocal images confirmed expression for all three members of the caveolin protein family in freshly dissociated BSMC from NB specimens (n = 5, Fig. 5). Interestingly, in contrast to the hBSMC cell line, membrane expression of Cav-1 and Cav-2 proteins were abundant while expression for Cav-3 appeared to be the weakest among the three isoforms, displaying diffuse cytoplasmic distribution throughout the cells and poor localization on the cell membrane. Furthermore, Cav-1 and -2 showed an organized dashed-like expression pattern aligned along the long axis of the smooth muscle cells which was absent in the cells immunolabeled with Cav-3 (Fig. 5, A, B, and C, and respective insets).

Fig. 5.

Freshly dissociated human BSMC show distinct membrane microdomain expression of caveolins-1 and -2. Confocal images of freshly dissociated BSMC were obtained from NB specimens immunolabeled with caveolin-1 (A), caveolin-2 (B), and caveolin-3 (C) antibodies. Nuclei were stained with DAPI. Caveolin-1 and -2 were expressed at the cell membrane in defined dashed-like pattern following the cell’s long axis, whereas caveolin-3 showed low and diffuse expression (scale bar, 20 µm). Insets show increased magnification of the regions marked with squares (scale bar, 10 µm).

Expression of caveolins and TREK-1 proteins is reduced in DO specimens at the single cell level.

We evaluated expression and localization of Cav-1, Cav-2, Cav-3, and TREK-1 in freshly dissociated BSMC obtained from human bladders with DO (Fig. 6). Cells obtained from DO tissue displayed a reduction in cell size and altered morphology. Whereas cells obtained from NB tissue where narrow and elongated, cells from DO tissues where shorter and did not maintain the elongated shape (Fig. 6A). In contrast with BSMC from NB bladders shown in Fig. 4, the dashed-like expression pattern for Cav-1 and Cav-2 proteins was reduced or absent in DO bladder smooth muscle cells (Fig. 6, A and B, left panels). Although Cav-3 protein expression in DO smooth muscle cells seemed to be similar to that observed in cells from NB as measured by fluorescence intensity (data not shown), protein distribution appeared to be condensed and localized in various regions of the cytoplasm rather than on the cell membrane. Further, expression of TREK-1 protein was also significantly reduced (Fig. 6, middle panels) and, in some cases, it was almost undetectable (data not shown). Colocalization analysis revealed a significant reduction in PCC between TREK-1, Cav-1, and Cav-2 when compared with NB data (0.20 ± 0.10, 0.30 ± 0.10, and 0.20 ± 0.02 for Cav-1, Cav-2, and Cav-3, respectively, n = 4 per group, P ≤ 0.001. Figure 6D).

Fig. 6.

Expression of caveolins is reduced in freshly dissociated BSMC from DO specimens. Colocalization analysis of caveolin-1, -2, and -3 (A, B, and C; left, in green) with TREK-1 (red, middle panels). The microdomain organization in BSMC was reduced for caveolin-1, and almost absent for caveolin-2. Similarly, a significant reduction in protein expression of TREK-1 was observed in DO cells. D: colocalization indexes were also reduced between caveolin-1, -2, -3 and TREK-1 (PCC = 0.16, 0.23, and 0.02, respectively). Overlap correlation coefficients were calculated from double-labeled confocal images. Bars and error bars represent means ± SE. Three different regions from separate slides were analyzed per target. Scale bar, 20 µm. **P ≤ 0.001.

siRNA silencing and saRNA activation of TREK-1 channel in hBSMC.

Considering low levels of TREK-1 expression in DO specimens, we wanted to explore ways of modulating expression of TREK-1 channels in the human detrusor. We chose small inhibiting (siRNA) and small activating (saRNA) RNAs as an experimental approach. Since gene silencing by siRNAs is much better established and studied than gene activating with saRNAs, we first evaluated the feasibility and efficiency of transfections with established and commercially available siRNA for TREK-1 (see materials and methods). hBSMC were transfected with commercially available siRNA for TREK-1, and both TREK-1 mRNA and protein expression were evaluated by qRT-PCR and Western blotting at 3 days posttransfection. As seen in Fig. 7, TREK-1-siRNA silencing resulted in a significant (78%) downregulation of TREK-1 mRNA expression in hBSMC cells by TREK-1-siRNA (Fig. 7A) followed by a 23% reduction in TREK-1 protein expression (Fig. 7B).

Fig. 7.

TREK-1 expression in hBSMC was modulated by both siRNA and saRNA approaches. A: TREK-1 mRNA transcripts were significantly reduced in hBSMC after transfection with TREK-1-siRNA (top panel). Bottom graph shows mRNA expression normalized to GAPDH in control siRNA and TREK-1-siRNA transfected cells (*P < 0.05). B: Western blot gel shows expression of TREK-1 protein downregulated by TREK-1-siRNA. Bottom graph presents TREK-1 expression in hBSMC normalized to smooth muscle α-actin after siRNA transfection (n = 5, *P ≤ 0.05 compared with control). C: two representative Western blots showing raw data of TREK-1 protein expression after transfection with 3 different sets of TREK-1-saRNAs. Expression of TREK-1 was not altered by transfection with TREK-1-saRNA-#1, but upregulated by ~20% after transfection with TREK-1-saRNA-#2 and TREK-1-saRNA-#3 [*P ≤ 0.05 compared with control (Ctrl)]. D: expression of TREK-1 protein after transfections with 3 sets of saRNAs.

We then performed similar experiments using three different sets of saRNAs designed in our laboratory. Analysis of qRT-PCR data indicated that mRNA levels of TREK1 were decreased up to 42% by TREK-1-saRNA-#3 compared with control group at both day 1 and day 3 postinfection (data not shown). However, protein expression was significantly increased by the same saTREK1-#3 as well as by saTREK-1-#2 (Fig. 7C). TREK-1 protein expression normalized by GAPDH showed some effectiveness of TREK-1-saRNA-#2 and TREK-1-saRNA-#3 in upregulation of channel expression (Fig. 7D), indicating that these saRNAs could indirectly increase the stability of TREK-1 protein at the posttranscriptional level.

DISCUSSION

Intrinsic regulation of myogenic excitability is an important aspect of urinary bladder control. This control allows bladder smooth muscle cells to adaptively relax in response to stretch to accommodate progressively increasing urine volume. We previously confirmed that response of the normal human detrusor to stretch is regulated by TREK-1, a predominantly expressed member of the mechanogated subfamily of two-pore-domain potassium (K_2P) channels. In the present study, we provided additional evidence that impaired mechanotransduction in DO could be associated with lower expression of TREK-1 channels in human BSMC. We also evaluated a potential relationship between TREK-1 channels and the cell cytoskeleton and established that coexpression of caveolins with TREK-1 was significantly altered in DO bladders.

Decreased expression of TREK-1 in DO specimens and increased muscle tone observed in our study correlate with previous animal and human data which confirmed that TREK-1 expression/activation is necessary for smooth muscle relaxation during stretch under different physiological conditions. Stretch-activated potassium channels are differentially regulated and contribute to the control of the membrane resting potential to counteract contractile stimuli (12). Membrane stretch was shown to increase TREK-1 expression in mouse cardiac muscle (58) consistent with the role for TREK-1 in the muscle response to stretch. Our results also directly correlate with the previous animal data, which determined that TREK-1 was downregulated in detrusor myocytes in a murine model of partial bladder outlet obstruction associated with bladder overactivity (9). Subsequent human studies detected higher expression of TREK-1 protein in pregnant vs. nonpregnant myometrium in both mice and humans (14, 45).

Deformation of the cell membrane upon stretch not only directly affects the function of mechanosensitive ion channels but also modulates channel activity by surrounding cytoskeletal networks, membrane-associated proteins, and lipids through several major mechanisms. One mechanism, transfer of force through connections between the actin cytoskeleton and the ion channel, was previously observed in human gastrointestinal smooth muscle (52). Previous studies suggested that there is a dynamic interaction between TREK-1 and the actin cytoskeleton (34). Mechanical force is usually transmitted to the channels through the membrane bilayer, with the cytoskeleton acting as a tonic repressor, limiting channel activation by membrane tension (34, 42, 43). Conversely, the expression of TREK-1 may markedly alter the cytoskeletal network and induce the formation of membrane filipodia-like structures (34). For instance, in the anterior eye chamber, TREK-1 was shown to be coexpressed with cochlin, a secreted extracellular matrix protein involved in mechanosensing of fluid shear (26). Cochlin-TREK-1 interaction resulted in changes in cell shape and motility with both proteins participating in filopodia formation (26).

We also tested whether TREK-1 is connected to the actin cytoskeleton via the TTA of the smooth muscle, which includes dystroglycan and sarcoglycan complexes and integral membrane proteins that span the sarcolemma (2, 3). In the skeletal muscle, disruption of the dystroglycan/sarcoglycan complexes results in the damage to the sarcolemma and ultimately cell death (3). Previous animal studies in bladder smooth muscle partially characterized these membrane structures and established their important role in the response of the detrusor muscle to stretch (2). However, our data from the human bladder did not show an overlap in IHC staining of TREK-1 with either of the tested sarcoglycans, suggesting that they do not directly connect TREK-1 with the actin cytoskeleton in the human BSM cells.

Activation of TREK-1 channels could be affected by changes in expression/function of caveolins. Caveolins are the principal structural component of caveolae, which are omega-shaped invaginations typically associated with membrane microdomains that regulate and facilitate signaling events by promoting the spatial sequestration of signaling molecules essential for endocytosis, intracellular signaling, and cholesterol homeostasis (23, 25, 41, 50). The caveolin family consists of three members: caveolin-1 (Cav-1), caveolin-2 (Cav-2), and caveolin-3 (Cav-3). Cav-1 and -2 are abundant in all cell types, whereas Cav-3 expression appears to be restricted to muscle cells. Bladder smooth muscle cells express all three caveolin isoforms (16, 18, 48). Kim et al. (30) reported that Cav-2 and Cav-3 were the most abundant members of the family in the rat bladder. In the present study, we confirmed that freshly dissociated human BSMC predominantly express Cav-1 and Cav-2 isoforms with very low membrane expression of Cav-3. Our results are similar to the reports by other groups which detected low expression of Cav-3 in the vascular (49), intestinal (16), and other types (57) of smooth muscle tissues.

Studies using caveolin-1 knockout mice provided evidence that loss of caveolin-1 impaired urinary bladder contractions (18, 33, 44). Polyak et al. (44) reported decreased numbers of caveolae in hypertrophied detrusor smooth muscle. Mice lacking caveolin-1 also lack caveolae in the bladder and exhibited several urological defects, including decreased contractility during stimulation with carbachol (32, 33, 55). Thus association of TREK-1 with caveolins may strongly influence the response of the human detrusor to stretch under both physio- and pathological conditions.

We acknowledge several limitations of our study. First, the presence of bladder cancer in patients with DO is a potential confounding factor. Although all specimens were closely evaluated by a pathologist, and the normal tissue was excised as far as possible from the malignant site, the presence of cancer in these patients may affect a variety of genetic, metabolic, and functional processes. Second, DO has not being confirmed urodynamically but was based on the results of AUA questionnaires. Although we acknowledge that bladder cancer itself can contribute to LUTS, urodynamic evaluation of bladder function cannot be clinically justified for, and, therefore performed on, bladder cancer patients.

The applicability of gene therapy for lower urinary tract dysfunction has been previously examined in many preclinical studies focused on overactive bladder, underactive bladder, decreased bladder sensation, bladder hyperalgesia, and urinary incontinence (17, 57). Therapeutics based on RNA interference (RNAi) proved to be powerful for the treatment of many diseases because of their high specificity, high efficacy, and low toxicity of the RNAi trigger—small dsRNA (13, 22). siRNA was successfully used to inhibit expression of the genes upregulated by certain pathological conditions. However, only recently small RNAs were shown to be effective in activation of gene expression. Thus Li and collaborators (36) reported that dsRNA molecules can induce sequence-specific transcriptional gene activation by targeting gene promoter regions and has termed this phenomenon as RNA-induced gene activation (RNAa) and such dsRNA molecules as small activating RNAs [saRNAs (36)]. In a quest for future therapeutic approaches to increase bladder capacity during filling, we tested if TREK-1 protein expression could be partly increased by saRNA approach. Previous reports established that saRNA-induced upregulation of p21WAF1/CIP1 (56) and E-cadherin (37) expression inhibits cell proliferation, migration, and invasion of human bladder cancer cells. In our study, although we observed a significant increase in TREK-1 protein levels using saRNA-TREK1-#2 and saRNA-TREK1-#3, there was an inverse correlation between TREK-1 mRNA and protein expression. This could be due to a feedback mechanism of increased protein stability in response to mRNA loss, or a nonspecific off-target effect of saRNAs (1, 47). The inverse correlation observed in our study is consistent with the previous reports showing that 46% of upregulated proteins correlated with unchanged or downregulated mRNA expression (24). We could not perform patch-clamp recordings from siRNA transfected cells due to a downregulation of functional TREK-1 channels in these cell, but were able to record from the cells transfected with saRNA. Unfortunately, we did not detect a statistically significant upregulation of TREK-1 current after saRNA using a patch-clamp approach. The reasons could include a significant variability in protein expression levels and efficiency of transfection between individual cells. Additionally, in the transfected cells, TREK-1 protein was mostly localized in the cytosol and was not transported to the cell membrane, thereby making patch-clamp recordings uninformative. Future studies are warranted to validate the applicability of this approach for the development of in vivo pharmacological and genetic interventions to treat DO in humans.

Conclusions.

Bladder overactivity in humans is associated with a significant reduction in the expression of mechanosensitive TREK-1 channels. Immunocytochemical labeling provided evidence of close association between the channel, cell cytoskeleton, and caveolins. Upregulation of TREK-1 expression by saRNA could be a future step for the development of in vivo pharmacological and genetic approaches to treat bladder overactivity in humans.

GRANTS

This study was funded by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK-095817 to A. P. Malykhina.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

R.H.P. and A.P.M. conception and design of the research; R.H.P., B.N., J.H., X.-Q.P. and A.P.M performed the experiments; R.H.P., B.N., X.-Q.P. and A.P.M analyzed the data; R.H.P., B.N., X.-Q.P. and A.P.M interpreted the results of the experiments; R.H.P., B.N., X.-Q.P. and A.P.M prepared the figures; R.H.P. and A.P.M drafted the manuscript; R.H.P., B.N., S.W., R.B.M. and A.P.M edited and revised the manuscript; R.H.P., B.N., X.-Q.P., S.W., R.B.M., and A.P.M. approved the final version of the manuscript.