Long-term follow-up of TREK-1 KO mice reveals the development of bladder hypertrophy and impaired bladder smooth muscle contractility with age

Abstract

Stretch-activated two-pore domain K⁺ (K_2P) channels play important roles in many visceral organs, including the urinary bladder. The TWIK-related K⁺ channel TREK-1 is the predominantly expressed K_2P channel in the urinary bladder of humans and rodents. Downregulation of TREK-1 channels was observed in the urinary bladder of patients with detrusor overactivity, suggesting their involvement in the pathogenesis of voiding dysfunction. This study aimed to characterize the long-term effects of TREK-1 on bladder function with global and smooth muscle-specific TREK-1 knockout (KO) mice. Bladder morphology, bladder smooth muscle (BSM) contractility, and voiding patterns were evaluated up to 12 mo of age. Both sexes were included in this study to probe the potential sex differences. Smooth muscle-specific TREK-1 KO mice were used to distinguish the effects of TREK-1 downregulation in BSM from the neural pathways involved in the control of bladder contraction and relaxation. TREK-1 KO mice developed enlarged urinary bladders (by 60.0% for males and by 45.1% for females at 6 mo; P < 0.001 compared with the age-matched control group) and had a significantly increased bladder capacity (by 137.7% at 12 mo; P < 0.0001) and compliance (by 73.4% at 12 mo; P < 0.0001). Bladder strips isolated from TREK-1 KO mice exhibited decreased contractility (peak force after KCl at 6 mo was 1.6 ± 0.7 N/g compared with 3.4 ± 2.0 N/g in the control group; P = 0.0005). The lack of TREK-1 channels exclusively in BSM did not replicate the bladder phenotype observed in TREK-1 KO mice, suggesting a strong neurogenic origin of TREK-1-related bladder dysfunction.

NEW & NOTEWORTHY This study compared voiding function and bladder phenotypes in global and smooth muscle-specific TREK-1 KO mice. We found significant age-related changes in bladder contractility, suggesting that the lack of TREK-1 channel activity might contribute to age-related changes in bladder smooth muscle physiology.

INTRODUCTION

Two-pore domain potassium (K_2P) channels regulate cell excitability by maintaining the resting membrane potential in excitable cells, such as neurons and smooth muscle cells (1). The TREK-1 (K₂P2.1, KCNK2) channel is the predominantly expressed K_2P channel in human bladder smooth muscle (BSM) (2). Its expression was also detected in the urinary bladders of mice, rats, guinea pigs, and monkeys (3–6). Under physiological conditions, TREK-1 activity can be elicited by mechanical stretch and is independent of intracellular calcium (Ca²⁺) (3). A TREK-1 channel opener, BL-1249, produced a concentration-dependent membrane hyperpolarization in cultured human BSM cells, inhibited KCl-induced contractions of rat bladder strips, and decreased the frequency of nonvoiding contractions (NVCs) in rats in vivo (7). Therefore, TREK-1-mediated K⁺ conductance is believed to play a physiological role in suppressing BSM cell excitability and muscle contractility in response to stretch, thereby facilitating bladder wall distension during the filling phase (1).

The significant downregulation of TREK-1 channels has been detected in pathological conditions. TREK-1 protein expression and TREK-1-mediated currents were significantly decreased in the urinary bladders of patients with detrusor overactivity (DO) (8). Inhibition of TREK-1 channel conductance also delayed human detrusor relaxation and triggered small-amplitude spontaneous contractions in response to stretch (2), further confirming the functional role of TREK-1 channels in the pathogenesis of DO. In mouse models of partial bladder outlet obstruction (PBOO), the bladder overactivity phenotype coincided with decreased TREK-1 channel expression in BSM (9). Systemic administration of methionine, which can inhibit TREK-1 currents and native stretch-dependent K⁺ (SDK) channels in bladder myocytes, also induced nonvoiding contractions during bladder filling (9). Blockade of TREK-1 channel conductance led to an increased detrusor cell contractility and longer distance of Ca²⁺ wave propagation through bladder tissue in response to stretch (4). Together, these data suggest that a diminished expression of TREK-1 channels in BSM may trigger detrusor instability.

Considering the established function of TREK-1 channels in governing smooth muscle excitability, we speculate that the lack of TREK-1 channels in BSM would lead to elevated detrusor contractility and DO symptoms in vivo. Our recent studies in young adult TREK-1 knockout (KO) mice showed that systemic deletion of TREK-1 gene caused increases in the baseline muscle tone, spontaneous BSM activity, and evoked contractions in in vitro assays using bladder strips. Additionally, we observed an increase in the number of nonvoiding contractions during urodynamic recordings in vivo (5). However, TREK-1 KO mice had a significantly larger bladder capacity during cystometry recordings, which contradicts the classic overactive bladder (OAB) symptoms (smaller bladder capacity, urgency, and frequency). Overall, TREK-1 KO mice developed a “mixed” bladder phenotype, which can be observed in patients with voiding dysfunction (10).

Here, we hypothesize that the loss of TREK-1 channels in the smooth muscle and bladder innervations differentially affect bladder function, leading to the mixed phenotypes observed in vivo in TREK-1 KO mice. Furthermore, the changes in bladder function due to the lack of TREK-1 channels might be age and/or sex dependent, which was not evaluated in our previous study, where we combined the data from animals of both sexes (5). The present study included a long-term follow-up of bladder function in TREK-1 KO mice and compared the effects of TREK-1 deficiency between global and smooth muscle-specific (sm)TREK-1 KO mice to distinguish the role of TREK-1 channels expressed in BSM from bladder innervation.

MATERIALS AND METHODS

Animals

Homozygous TREK-1 KO mice (3, 6, 9, and 12 m old) and age-matched C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME, strain no. 000664; RRID:IMSR_JAX:000664) were used in this study. Balanced numbers of males (total N = 102) and females (total N = 104) were included in each age/genotype group (Supplemental Table S1). Smooth muscle-specific TREK-1 KO (smTREK-1 KO) mice (N = 26) were obtained by crossing heterozygous SM22-Cre mice (The Jackson laboratory, Strain No. 017491; RRID:IMSR_JAX:017491) with homozygous TREK-1 conditional KO (cKO) mice (gift from Dr. H. A. Rockman, Duke University, Durham, NC) (11). The first-generation offspring were genotyped, and then the SM22-Cre^+/−:TREK-1^fl/WT mice were crossed again to homozygous TREK-1^fl/fl mice to produce experimental animals (SM22-Cre^+/−:TREK-1^Ex/Ex; smTREK-1 KO) as well as Cre-negative littermate control mice (SM22-Cre^−/−:TREK-1^fl/fl and SM22-Cre^−/−:TREK-1^fl/WT). All animals in the smTREK-1 KO group were genotyped by Transnetyx (Cordova, TN). The presence of Cre, fl, Ex, and WT sequences of the Kcnk2 gene was confirmed for each mouse.

Mice were housed in a temperature-controlled environment at the University of Colorado Anschutz Medical Campus (CU-AMC) vivarium (14:10-h light-dark cycle) with ad libitum access to food and water. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (No. 00517) of the CU-AMC. All experiments were performed in the CU-AMC in accordance with relevant guidelines and regulations, as well as in compliance with ARRIVE Guidelines 2.0 (https://arriveguidelines.org/arrive-guidelines). For survival surgeries (e.g., insertion of bladder catheters for cystometry), isoflurane (2%, continuous inhalation) was used as a primary anesthetic. Overdose exposure to CO₂ followed by a secondary method (cervical dislocation) was used as a method of euthanasia for mice used in terminal procedures.

Bladder Morphology and Histological Evaluation

The urinary bladders were isolated from TREK-1 KO (N = 36) and control (N = 33) mice, weighed, and fixed in paraformaldehyde. Paraffin sections (5 μm thickness) were prepared and processed for either hematoxylin and eosin (H&E) staining or second-harmonic generation (SHG) imaging for collagen content (12). The areas of the whole urinary bladder wall, the BSM layer, and collagen fibers (pseudocolored in gray) in SHG images of each section were measured. All measurements were taken in a blind fashion. Two sections taken at least 50 μm apart from each other from at least four animals per sex per group were analyzed for rigor and reproducibility.

Bladder Strip Contractility Recordings

Physiological evaluation of detrusor contractility and baseline spontaneous activity was conducted with freshly isolated bladder strips from TREK-1 KO (N = 33) and control (N = 32) mice at 6, 9, and 12 m, as previously described (13) (see Fig. 2 A, N = 5 or 6 per genotype per sex for TREK-1 KO and control mice, N = 6 or 7 per genotype per sex for smTREK-1 KO mice). The muscle contractility studies started with determination of the optimal length of an individual muscle strip (L_o) that produced maximal contraction in response to electric field stimulation (EFS, 32 Hz). Spontaneous contractions in each bladder strip were recorded for 2 min after initial equilibration for 30 min in Tyrode solution. Detrusor contractility was examined in responses to different stimuli including EFS (70 V, 2–32 Hz), the acetylcholine receptor agonist carbachol (CCh, 1–100 μM), substance P (SubP, 1 μM, sensory neuropeptide), αβ-methylene ATP (AMBA, 4.5 μM, a P2X receptor agonist), atropine (1 μM, a nonselective blocker of muscarinic receptors), and high KCl (125 mM). The peak force of the contractile response was calculated in newtons (N) of tension per weight of individual bladder strip. All pharmacological agents used are listed in Table 1. All data analyses were performed with PowerLab LabChart version 8.1.9 (AD Instruments, Colorado Springs, CO).

Figure 2.

TREK-1 knockout (KO) mice exhibit decreased bladder contractility compared with the age-matched control group. A: timeline of the experiments. ABMA, α,β-methylene ATP; CCh, carbachol; EFS, electric field stimulation; L_o. optimal length; SubP, substance P. B: bladder strip weight. m, Months. C: amplitude of spontaneous contractions. D and E: peak (D) and integral (E) contractile forces induced by KCl in TREK-1 KO and control bladder strips. F–I: peak contractile force in response to EFS (F), CCh (G), SubP (H), and ABMA (I). Open and filled bars represent control and TREK-1 KO groups, respectively (B–E, H and I). Open and filled squares in the line charts represent control and KO groups, respectively (F and G). J: component analysis in contraction evoked by EFS (32 Hz). Open, black, and gray bars represent the purinergic (ABMA sensitive), muscarinic (atropine sensitive), and nonpurinergic, nonmuscarinic (ABMA and atropine insensitive) component, respectively. Values are means ± SD. *P < 0.05 and #P < 0.005 (N = 5 or 6 per sex per group per age).

Table 1.

Pharmacological agents used in the bladder strip contractility assay

Name	Vendor	Working Concentration	Mechanism of Actions
Carbachol	Alfa Aesar (catalog no. L06674)	1, 10, 100 µM	Cholinergic receptor agonist
Substance P	Tocris (catalog no. 1156)	1 µM	Activating neurokinin-1 receptor, nociception, neuromodulator
α,β-Methylene ATP	Tocris (catalog no. 3209)	4.5 µM	P2X receptor agonist and desensitizer
Atropine	Sigma (catalog no. A0132)	1 µM	Nonselective muscarinic receptor blocker
KCl	Sigma (catalog no. P3911)	125 mM	Depolarizing muscle cells

Spontaneous Voiding Spot Assays

Void spot assay (VSA) was performed to reveal the potential impact of lack of TREK-1 channels on spontaneous voiding behavior (14). Clean cages equipped with elevated wire grid (1 cm above the cage floor, 0.64-cm² opening) were lined with a single layer of Whatman-grade filter paper (Fisher Scientific, Hampton, NH). Group-housed mice were transferred into individual testing cages with ad libitum access to water but no food during the 3-h testing period (9 AM to noon). Animals were returned to group housing after the assay. No acclimation was performed before testing. The urine spots on the filter paper were imaged with ultraviolet light on a transilluminator, and the number and area of the spots were analyzed with Adobe Photoshop CS6 (Adobe Systems, San Jose, CA; RRID:SCR_014199), as previously described (13). Urine spots were converted to volumes by construction of a standard curve that relates known volumes of mouse urine spotted on the filter paper with Adobe Photoshop. Urine spots of <0.3 cm² (corresponding to 10 μL of urine) were excluded from analysis, consistent with our previously described criteria (13, 14). The numbers of large (≥200 µL) and small (<200 µL) voiding spots were counted for each recording.

Cystometry in Unanesthetized Mice

For urodynamic evaluation of voiding patterns, mice underwent surgical catheter implantation in the bladder (see Fig. 2A), as previously described (14). Mice were allowed to recover single-housed in cages for 7 days before undergoing awake cystometry recordings (N = 3–7 per age group per sex for control mice, N = 4–9 per age per sex for TREK-1 KO mice). Briefly, an exteriorized bladder catheter (located at the base of the mouse neck) was connected to a pressure transducer and an infusion pump of the cystometry station (Catamount Research and Development, St. Albans, VT). Room temperature saline was infused into the bladder at the rate of 15 μL/min (see Fig. 4A). Each animal was observed for a minimum of four voiding cycles. Urodynamic values were recorded continuously during testing. Parameters including maximal micturition pressure (mmHg), bladder capacity (μL), voided volume (μL), threshold capacity (μL, the infusion volume at threshold pressure), contraction duration (s), intermicturition interval (min), and the number of nonvoiding contractions (NVCs) per voiding cycle were analyzed with Cystometry Analysis Software (SOF-552; Catamount Research and Development). The NVCs were defined as intravesical pressure spikes greater than one-third of average maximal voiding pressure in each animal without triggering micturition. Subsequently, micturition efficiency (voided volume/bladder capacity) and bladder compliance (bladder capacity/micturition pressure, μL/mmHg) were calculated. All measurements were based on the guidance published by the University of Wisconsin-Madison George M. O’Brien Center for Benign Urology Research (https://urology.wisc.edu/wp-content/uploads/2018/10/Analysis_of_Cystometry_Tracings.pdf) (15).

Figure 4.

TREK-1 knockout (KO) mice exhibited an increased bladder capacity compared with the age-matched control group. A: schematic representation of awake cystometry recordings. Image was created with BioRender software (Toronto, Canada, RRID:SCR_018361). B: intermicturition intervals. m, Months. C: bladder capacity. D: voided volume. E–H: micturition pressure. Values are means ± SD. *P < 0.05, #P < 0.005 between the TREK-1 KO group and the control group (N = 3–7 per age group per sex for control mice, N = 4–9 per age per sex for TREK-1 KO).

Data Analysis and Statistics

Statistical comparisons were performed between TREK-1 KO and control animals or between smTREK-1 KO and littermate control mice based on sex, genotype, and age groups. The study is exploratory in nature and was designed to test a biological hypothesis, not a prespecified statistical null hypothesis. Two-way ANOVA was performed to compare data among age and genotype groups between male and female sexes. Data found statistically significant by ANOVA were analyzed by Bonferroni test for multiple comparisons. A P value ≤ 0.05 was considered statistically significant. The calculated P levels can be interpreted only as descriptive values because of the lack of a prespecified statistical null hypothesis. All data are expressed as means ± standard deviation (SD). Statistical analyses were done with GraphPad Prism 9 (GraphPad Software, La Jolla, CA; RRID:SCR_002798). GraphPad outlier calculator (GraphPad Software, https://www.graphpad.com/quickcalcs/Grubbs1.cfm) was used to detect outliers. Figures were prepared with the most recent GraphPad Prism (RRID:SCR_002798) and Adobe Photoshop (RRID:SCR_014199). The schematic in Fig. 4A was created with BioRender (Toronto, Canada; RRID:SCR_018361).

RESULTS

TREK-1 KO Mice Developed Increased Bladder Weight and Size Compared With Age-Matched Control Mice

To reveal the physiological role of TREK-1 channels in bladder morphology, we compared bladder weight and histology between TREK-1 KO mice (6, 9, and 12 mo old) and age-matched control mice. Our previous study using 3-mo TREK-1 KO animals observed no significant changes in the bladder weight or histology (5). However, TREK-1 KO mice of 6 mo and older had larger and heavier urinary bladders compared with age-matched control mice (Fig. 1A). In male mice, the weight of the urinary bladder increased with age in both control (by 24.4% at 6 mo) and TREK-1 KO (by 26.9% at 6 mo) mice (Fig. 1A, left, P = 0.0003), which correlated with age-related increases in body weight in the same age groups (Fig. 1A, center). The bladder weight relative to the body weight showed no age-related changes in both groups (Fig. 1A, right). However, TREK-1 KO mice developed an increased bladder-to-body weight ratio (by 65.3%, 60.0%, and 86.3% at 6, 9, and 12 mo, respectively, all P < 0.0005) compared with age-matched control mice (Fig. 1A, right), suggesting the development of bladder hypertrophy in TREK-1 KO males.

Figure 1.

TREK-1 knockout (KO) mice showed bladder hypertrophy without signs of fibrosis. A: comparison of bladder weight (left), body weight (center), and bladder weight relative to body weight (right). m, Months. B and C: representative images of hematoxylin and eosin (H&E) staining (B) and collagen imaging (C) of the bladder wall from TREK-1 KO and control mice. Scale bars = 400 μm. D: area measurement of the transverse sections of the bladders from each group. Values are means ± SD (megapixels). E: area of the bladder smooth muscle (BSM) layer (left) or collagen (right) relative to that of the total bladder wall. The total cross-sectional area was taken as 100%. Values are means ± SD (in %) for the areas of total tissue on each section. *P < 0.05 and #P < 0.005 between the TREK-1 KO and control groups (A, D, and E). Blue and pink bars represent male and female mice, respectively. Filled and hatched bars represent control and TREK-1 KO mice, respectively.

In contrast, female animals did not show any age-related changes in bladder weight up to 12 mo of age in both control and TREK-1 KO groups (Fig. 1A, left). However, the body weight of the female animals increased over age (Fig. 1A, center, by 24.6% and 28.9% at 6 mo in the control and TREK-1 KO groups, respectively, P < 0.0001), which resulted in slight decreases in the bladder-to-body weight ratio over time (by 35.7% at 6 mo and 35.6% at 12 mo, P = 0.0142; Fig. 1A, right). In comparison to control female animals, TREK-1 KO females consistently had heavier bladders (Fig. 1A, left, 18.3–19.0 mg vs. 25.9–29.0 mg in control and TREK-1 KO groups, respectively, an increase of 40.1–52.4% between the age groups, P < 0.0001) as well as higher bladder-to-body weight ratio in 6 mo (by 35.7%) and 12 mo (35.6%) age groups (Fig. 1A, right, P < 0.0001). Together, our data indicate that the complete loss of TREK-1 channels led to bladder hypertrophy in adult mice as early as 6 mo of age.

We also found that bladder-to-body weight ratios were higher in TREK-1 KO male mice than in age-matched female mice [0.12% vs. 0.10% at 6 mo (P = 0.04), 0.13% vs. 0.08% at 9 mo, and 0.14% vs. 0.08% at 12 mo (P < 0.0001)]. Sex differences were also detected in 9- and 12-mo-old control mice but not in 6-mo-old mice [0.08% vs. 0.06% at 9 mo (P = 0.004) and 0.07% vs. 0.06% at 12 mo (P = 0.01) in males and females, respectively], suggesting that the loss of TREK-1 channels has an earlier impact on the changes in bladder morphology in males in comparison to females (16).

Bladders From TREK-1 KO Mice Had a Larger Size in All Layers of the Bladder Wall With No Signs of Fibrosis

To examine whether the increases in bladder weight and size in TREK-1 KO mice were due to BSM hypertrophy, we performed histological analysis of the urinary bladder wall using H&E staining. We found age-related increases in the bladder wall sections [Fig. 1, B and D, by 33.8% and 28.7% in control and TREK-1 KO male mice at 6 mo (P = 0.001) and by 36.6% and 11.0% in control and TREK-1 KO female mice at 12 mo (P = 0.006)], consistent with the pattern detected for bladder weight. Males showed larger bladder wall (measured in pixels) in all genotype/age groups in comparison to their female counterparts (by 8.8–28.8% in the control group and 16.1–34.6% in TREK-1 KO mice, P < 0.0001). The bladder wall thickness in TREK-1 KO mice was significantly bigger [Fig. 1, B and D, by 26.4–47.3% for male mice (P = 0.001) and by 18.5–45.9% for female mice (P = 0.006)] than in the age-matched control bladders from both sexes and all age groups, which is consistent with the changes found in bladder weight. When normalized to the bladder wall area, the relative BSM composition (%) was comparable between TREK-1 KO and control mice at all ages (Fig. 1E, left), indicating that the increases in BSM layer thickness were proportional to the increases in the total bladder wall thickness. These data suggest that a complete loss of TREK-1 channels did not affect the relative thickness of BSM layer in the urinary bladder.

We next examined the collagen content in the urinary bladders from TREK-1 KO mice and age-matched control mice with SHG imaging. The collagen content of the urinary bladders was comparable between the two genotypes and sexes and among all age groups (Fig. 1E, right), indicating that the loss of TREK-1 channel did not induce bladder fibrosis up to 12 mo of age. Taken together, our histology data suggest that increases in the size and weight of the urinary bladder in TREK-1 KO mice are not associated with bladder fibrosis or BSM-specific hypertrophy.

TREK-1 KO Mice Showed Decreased Bladder Contractility After 6 mo of Age

To determine the potential impact of TREK-1 downregulation on BSM contractility, in vitro contractility recordings were performed on isolated bladder strips (Fig. 2). First, the baseline activity of each bladder strip was recorded after initial calibration before electrical field stimulation (EFS) or any pharmacological agents were administrated (Fig. 2A). Evidence suggests that some TREK-1 channels may be active when the bladder is empty (3). The bladder strips from TREK-1 KO mice were significantly heavier than those from the control mice at all age groups, whereas there were no changes along age within a genotype (Fig. 2B, 8.5 mg vs. 13.6 mg in control and TREK-1 KO mice, respectively, P < 0.005). In our experiments, the bladder strips from control mice showed a statistically larger amplitude of spontaneous contractions than those from TREK-1 KO mice at 6 mo (37 ± 19 mN/g in control group and 22 ± 14 mN/g in TREK-1 KO mice, P = 0.014) and a trend of decrease with age, which was not observed in the TREK-1 KO group (Fig. 2C). These data echo our previous finding that the application of l-methionine (1 mM), a TREK-1 channel blocker, did not alter the spontaneous activity of bladder strips isolated from WT mice (2). Together, our data suggest that TREK-1 channel activity does not contribute to BSM spontaneous activity at the baseline (L_o).

Bladder strips from TREK-1 KO mice showed a smaller peak force [PF; a decrease by 53.6% at 6 mo, P < 0.001 and by 46.3% at 12 mo, P ≤ 0.05 vs. respective control) and integral force (IF; a decrease by 57.5% at 6 mo, P = 0.001, and by 44.7% at 12 mo, P = 0.007) of KCl-induced contractions (Fig. 2, D and E), suggesting a weaker detrusor contractility. Interestingly, there was a trend of age-related decrease in KCl-induced contractions in the bladder strips isolated from control mice, which was absent in the bladders from TREK-1 KO group. Taken together, the bladder strip contractility data confirmed that the loss of TREK-1 channels decreased BSM ability to contract, which coincided with BSM hypertrophy.

Lack of TREK-1 Channel Decreased Nerve- and Agonist-Evoked Bladder Contractility in Vitro

We next compared the EFS-induced and agonist-evoked bladder contractility in vitro between the TREK-1 KO and control groups (Fig. 2, F–J, and Table 1). TREK-1 KO bladder strips showed a reduced contractility induced by EFS compared with the age-matched control bladder strips (Fig. 2F). TREK-1 KO bladder strips also exhibited a reduced amplitude of contractions after CCh application (Fig. 2G), SubP (Fig. 2H), and ABMA (Fig. 2I, 1.3 ± 0.9 vs. 0.6 ± 0.4 N/g at 6 mo in control and TREK-1 KO mice, respectively, P = 0.003) treatments. There was a clear trend of age-related decline in the contraction evoked by EFS and all of the agonists in control bladder strips, which was absent in the TREK-1 KO group (Fig. 2, G–I), suggesting an association between TREK-1 channels and age-related changes in BSM contractility. Finally, the ABMA- or atropine-sensitive components for EFS (32 Hz)-evoked contractility were similar between control and KO groups (Fig. 2J), pointing toward similar contributions of purinergic and muscarinic signaling in the urinary bladders of TREK-1 KO mice. Overall, we found little evidence of TREK-1 downregulation-induced changes in muscarinic and purinergic receptor-mediated BSM contractility in the urinary bladders up to 12 mo of age.

Behavioral Assay Detected Voiding Pattern Shifts Toward Larger Voiding Spots in Female TREK-1 KO Mice and More Frequent Small Voiding Spots in Male TREK-1 KO Mice

We assessed spontaneous voiding behavior in mice with the VSA method. Mice typically produce very few (0 or 1) large (≥200 µL) voids corresponding to normal voiding of full bladders during the 3-h recording period. The presence of smaller urine spots indicates that animals void before the bladder becomes full and could suggest either the development of overactive bladder symptoms or the occurrence of overflow incontinence. TREK-1 KO males showed higher numbers of small voiding spots (<200 µL) than control males (Fig. 3A). In contrast, TREK-1 KO female animals mostly produced larger voiding spots (≥200 µL) than age-matched control mice (Fig. 3).

Figure 3.

Voiding spot assay (VSA) data. A: the volume distribution of all void spots. *P < 0.05, #P < 0.005 between the genotypes (red) or sexes in a genotype (green). Values are means ± SD. Each dot represents a single measured/counted void (n) for all animals in that sex/group. KO, knockout; m, months. B: the volume of large (≥200 µL) voided spots. No statistical analyses could be performed to compare male mice between the genotypes at 3 mo and 6 mo because of insufficient number of large voided spots. Values are means ± SD. *P < 0.05 (N = 10–19 per sex per group per age, N = 5 per sex in the control group at 3 mo).

The frequency of small and large voiding spots as well as the total voided volume were compared between TREK-1 KO mice and age-matched control mice (Table 2). Increased numbers of small voiding spots were observed in older (9 and 12 mo) TREK-1 KO males compared with age-matched control male mice (Table 2, 0.7 ± 0.9 vs. 5.4 ± 4.3 at 12 mo, P < 0.0001). Overall, the VSA revealed a trend toward larger voiding spot sizes in female TREK-1 mice (up to 6 mo) and more frequent small voiding in TREK-1 KO males, suggesting a sex-specific impact of the loss of TREK-1 channels on male and female voiding patterns.

Table 2.

Number of voiding spots and urine volume in TREK-1 KO mice of different ages

Age	Genotype	Male			Female
		Large	Small	Total, mL	Large	Small	Total, mL
3 mo	Ctrl	0.6 ± 1.0	2.9 ± 2.5	0.23 ± 0.11	0.6 ± 0.7	2.4 ± 1.6	0.27 ± 0.09
	TREK-1 KO	0.1 ± 0.3	2.6 ± 2.9	0.16 ± 0.14	1.0 ± 0.6	2.3 ± 2.0	0.51 ± 0.18
6 mo	Ctrl	0.2 ± 0.4	2.5 ± 2.7	0.40 ± 0.23	0.3 ± 0.5	2.9 ± 1.7	0.28 ± 0.17
	TREK-1 KO	0.2 ± 0.4	4.1 ± 3.4	0.23 ± 0.18	1.0 ± 0.7	2.0 ± 3.0	0.48 ± 0.26
9 mo	Ctrl	0.4 ± 0.5	2.5 ± 1.6	0.29 ± 0.19	0.7 ± 0.8	3.3 ± 3.8	0.48 ± 0.41
	TREK-1 KO	0.3 ± 0.7	5.2 ± 3.7	0.34 ± 0.27	0.5 ± 0.6	2.5 ± 3.2	0.33 ± 0.22
12 mo	Ctrl	0.5 ± 0.8	0.7 ± 0.9	0.35 ± 0.22	0.9 ± 1.0	1.7 ± 1.6	0.33 ± 0.12
	TREK-1 KO	0.9 ± 1.1	5.4 ± 4.3#	0.64 ± 0.34*	0.5 ± 0.7	1.3 ± 1.0	0.34 ± 0.24

Data are presented as mean ± SD number of large and small voiding spots and total urine volume. Ctrl, control; KO, knockout. *P ≤ 0.05 and #P ≤ 0.01 vs. the age-matched Ctrl group.

TREK-1 KO Mice Display Increased Bladder Capacity and Compliance During Urodynamic Evaluation of Voiding

To evaluate TREK-1-dependent effects on bladder function in vivo, we performed awake cystometry recordings in TREK-1 KO mice and age-matched control mice (Fig. 4A). Compared with the control groups, TREK-1 KO mice voided less frequently (intermicturition interval was increased from 6.1 ± 2.4 min in the control group to 12.8 ± 3.9 min in TREK-1 KO mice at 12 mo, P < 0.0001; Fig. 4B) and also had a significantly increased bladder capacity (90.9 ± 35.5 μL in the control group vs. 192.3 ± 58.2 μL in TREK-1 KO mice at 12 mo, P < 0.0001; Fig. 4C) and voided volume (71.3 ± 30.6 μL vs. 178.0 ± 53.0 μL at 12 mo in control and TREK-1 KO mice, respectively, P < 0.0001; Fig. 4D). Age-dependent changes were also detected in the same parameters (Fig. 4, B–D). As mice aged, the intermicturition intervals became significantly longer in both control and TREK-1 KO groups [5.4 ± 2.6 min at 3 mo vs. 8.7 ± 3.3 min in 9 mo control mice (P = 0.0452) and 7.5 ± 2.6 min at 3 mo vs. 12.8 ± 3.9 min at 12 mo in TREK-1 KO mice (P = 0.0050); Fig. 4B], and bladder capacity and voided volume [70.3 ± 40.4 μL at 3 mo vs. 119.1 ± 49.5 μL at 9 mo in control mice (P = 0.0487) and 108.0 ± 36.6 μL at 3 mo vs. 178.0 ± 53.0 at 12 mo in TREK-1 KO group (P = 0.0109); Fig. 4, C and D] were increased as well. At 12 mo of age, TREK-1 KO mice showed significant increases in all three parameters compared with age-matched control mice [intermicturition interval: 6.1 ± 2.4 min in the control group and 12.8 ± 3.9 min in TREK-1 KO mice (P ≤ 0.001); bladder capacity: 90.9 ± 35.5 μL in control mice vs. 192.3 ± 58.2 μL in TREK-1 KO mice (P ≤ 0.001); and voided volume: 71.3 ± 30.6 μL in the control group vs. 178.0 ± 53.0 μL in TREK-1 KO mice (P < 0.0001)]. These data are consistent with the larger bladder size in TREK-1 KO mice identified in histological experiments (Fig. 1). No sex differences were detected in any age group by two-way ANOVA (with genotype as the second variable).

No significant genotype differences were identified in bladder pressure at the time of micturition except for the 3-mo age group (39.2 ± 10.9 mmHg vs. 58.3 ± 9.7 mmHg in control and TREK-1 KO mice, respectively, P = 0.04; Fig. 4E). When the potential sex/genotyping differences were examined within the same age groups, a significant sex difference was identified in the 9-mo age group (55.4 ± 5.4 mmHg vs. 34.0 ± 11.4 mmHg in control female and male mice, respectively, and 56.9 ± 13.0 mmHg vs. 49.3 ± 11.7 mmHg in TREK-1 KO female and male mice, respectively, P = 0.0075) and only male mice showed significant genotyping differences (34.0 ± 11.4 mmHg vs. 49.3 ± 11.7 mmHg in control and TREK-1 KO mice, respectively, P = 0.03; Fig. 4F). Therefore, the data from female and male mice were analyzed separately, as shown in Fig. 4, G and H. Neither genotype nor age differences were detected in the micturition pressure in female mice (Fig. 4G). However, genotype differences were detected in the same parameter in males in the 3-mo age group (35.6 ± 13.0 mmHg vs. 64.2 ± 6.7 mmHg in control and TREK-1 KO mice, respectively, P = 0.02; Fig. 4H).

Consequently, bladder compliance (bladder capacity/micturition pressure) in TREK-1 KO animals was significantly higher than in their age-matched control counterparts, with the most significant changes detected in the 12-mo age group (2.2 ± 0.9 μL/mmHg vs. 4.0 ± 1.3 μL/mmHg in control and TREK-1 KO mice, respectively, P = 0.03; Fig. 5A). Age-dependent but not sex-dependent changes in bladder compliance were also detected (Fig. 5A). There were also no significant sex-, age-, or genotype-related differences in micturition efficiency (data not shown), suggesting preserved bladder emptying function despite the detected decreased BSM contractility observed in bladder strips from TREK-1 KO mice.

Figure 5.

TREK-1 knockout (KO) mice showed an increased bladder compliance, longer contraction duration, and more nonvoiding contractions (NVCs) compared with the age-matched control group. The threshold pressure remained unchanged between TREK-1 KO and control mice. A: bladder compliance. m, Months. B: contraction duration. C–F: threshold pressures. G and H: average numbers of NVCs per cycle (G) and frequency of occurrence (H). Values are means ± SD. *P < 0.05 and #P < 0.005 between the TREK-1 KO group and the control group (N = 3–7 per age group per sex for control mice, N = 4–9 per age per sex for TREK-1 KO).

K_2P channels can potentially contribute to BSM relaxation after micturition (1). We compared the duration of bladder contractions between TREK-1 KO mice and control mice (Fig. 5B). No sex- or age-related differences were detected regarding contraction duration. TREK-1 KO mice had a longer contraction duration than control mice (31.9 ± 23.7 s in the control group vs. 55.2 ± 32.3 s in the TREK-1 KO group at 6 mo, P = 0.01; Fig. 5B), suggesting slower contraction-relaxation period during micturition. Additionally, it has been shown that decreased TREK-1 channel activity leads to increased BSM excitability and contractions during mechanical stretch (1). Therefore, we analyzed the threshold pressure (bladder pressure at the micturition threshold) between TREK-1 KO and control mice (Fig. 5, C–F). No significant age or genotype differences were identified (Fig. 5C). Subsequently, significant sex differences in threshold pressure were determined in the 9-mo age group [30.9 ± 9.0 mmHg vs. 16.7 ± 9.2 mmHg in control female and male mice, respectively (P ≤ 0.05) and 31.2 ± 8.2 mmHg vs. 21.9 ± 6.2 mmHg in TREK-1 KO female and male mice, respectively (P = 0.002); Fig. 5D], with female and male data analyzed separately, as shown in Fig. 5, E and F. In either female or male mice, no significant genotype differences or age-related changes were identified (Fig. 5, E and F), suggesting that the loss of TREK-1 expression in the urinary bladder does not change threshold pressure in both male and female mice.

Increased numbers of NVCs were observed in our previous study using young TREK-1 KO animals (3 mo of age) (5). In this study, we quantified the average numbers of NVCs per voiding cycle in all age groups (Fig. 5G) as well as the occurrence of NVCs (percentage of voiding cycles in each animal with NVC present; Fig. 5H). Compared with age-matched control animals, TREK-1 KO mice showed a significantly elevated number of NVCs per voiding cycle (0.3 ± 0.3 vs. 1.1 ± 0.8 at 12 mo in control and TREK-1 KO mice, respectively, P = 0.02). The number of voiding cycles with NVCs also increased in TREK-1 KO mice from 19.2 ± 16.8% in the control group to 63.9 ± 38.1% in the TREK-1 KO group at 12 mo (P < 0.001 between age groups). There was no sex difference in either total numbers or occurrence of NVC among all age groups. Overall, the urodynamic results provided evidence of larger bladder capacity and increased frequency of NVC in TREK-1 KO mice without significant changes in pressure at micturition, suggesting the development of a mixed symptoms phenotype.

Smooth Muscle-Specific TREK-1 KO Mice Did Not Develop Altered Bladder Weight or Morphology

TREK-1 channels are expressed not only in BSM cells but also in the central and peripheral nervous systems. To compare the functional significance of TREK-1 channels in BSM versus the neural pathways involved in the control of bladder function, we performed a morphological analysis of the urinary bladder and contractility recordings in smooth muscle-specific TREK-1 KO mice (smTREK-1 KO) at 6 mo of age. First, bladder weights were compared among mice lacking TREK-1 in bladder detrusor cells (SM22-Cre^+/−:TREK-1^Ex/Ex) with two littermate control mice (SM22^−/−:TREK-1^fl/Wt and SM22-Cre⁻:TREK-1^fl/fl). No significant differences in bladder weight, body weight, or normalized bladder/body weight were detected among the three genotypes (Table 3). Since no significant differences were identified between the two littermate control genotypes, the data obtained from both control genotypes were combined in further analyses, including the contractility experiments (Fig. 6 and Fig. 7).

Table 3.

Body and bladder weight in smTREK-1 KO mice compared with littermate control mice (SM22-Cre^−/−) at 6 mo of age

	Body, g	Bladder, mg	Bladder, %
Females
SM22-Cre−/−::TREK-1fl/fl
Ctrl	24.6 ± 0.9	19.4 ± 3.8	0.09 ± 0.01
SM22-Cre−/−::TREK-1fl/WT
Ctrl	23.4 ± 1.3	15.8 ± 1.4	0.07 ± 0.00
SM22-Cre+/::TREK-1Ex/Ex
smKO	24.6 ± 0.9	16.0 ± 0.7	0.07 ± 0.00
Males
SM22-Cre−/−::TREK-1fl/fl
Ctrl	31.2 ± 0.7	25.4 ± 1.0	0.08 ± 0.00
SM22-Cre−/−::TREK-1fl/WT
Ctrl	31.5 ± 0.5	26.7 ± 0.6	0.08 ± 0.00
SM22-Cre+/::TREK-1Ex/Ex
smKO	33.7 ± 1.6	26.8 ± 0.9	0.08 ± 0.00

Data are presented as means ± SD. Ctrl, control; smKO, smooth muscle-specific (sm)TREK-1 knockout (KO).

Figure 6.

No significant changes were observed in bladder weight (A), body weight (B), and bladder/body weight (C) in smooth muscle-specific (sm)TREK-1 knockout (KO) mice in comparison to the control group. The control group includes littermate control mice with 2 genotypes (SM22-Cre^−/−). Values are means ± SD. *P < 0.05 and #P < 0.005 (N = 6 or 7 per sex per group).

Figure 7.

Smooth muscle-specific (sm)TREK-1 knockout (KO) mice showed a mild decrease in bladder strip contractility in response to electric field stimulation (EFS) at 6 mo. A: timeline of the experiment. ABMA, α,β-methylene ATP; CCh, carbachol; EFS, electric field stimulation; L_o. optimal length. B: bladder strip weight. C: amplitude of spontaneous contractions. D and E: peak (D) and integral (E) contractile force induced by KCl. F–H: peak force in response to EFS (F), CCh (G), and ABMA (H). I: component analysis in contraction evoked by EFS (32 Hz). Values are means ± SD. #P < 0.005 (N = 12 or 13 per group).

Male mice in all genotypes developed increased body and bladder weight compared with females. Specifically, body weight in males was increased by 131% in the control group (P = 0.04), by 137% in the smTREK-1 KO group (P = 0.001), and by 117% in the TREK-1 KO group (P = 0.03) compared with age-matched females. No significant differences in either body or bladder weight were detected between smTREK-1 KO and control mice (Fig. 6, A and B). In both sexes, TREK-1 KO mice at 6 mo had elevated bladder weight and bladder-to-body weight ratio [0.12 ± 0.02% vs. 0.08 ± 0.01% in control and smTREK-1 KO males, respectively (P < 0.0001), and 0.10 ± 0.03% vs. 0.07 ± 0.01% in females (P = 0.012); Fig. 6] compared with smTREK-1 KO mice (Fig. 6). These data indicate that the loss of TREK-1 channels in BSM is not responsible for the bladder hypertrophy observed in TREK-1 KO animals.

A Decrease in Nerve-Mediated Contractility in Bladder Strips From Smooth Muscle-Specific TREK-1 KO Mice

We assessed the potential changes in bladder strip contractility in smTREK-1 KO mice at 6 mo (Fig. 7A). The bladder strips isolated from smTREK-1 KO mice weighed similarly to those isolated from littermate control mice (Fig. 7B). No significant changes were detected in the amplitude of spontaneous contractions and maximal and integral contractile force induced by KCl compared with age-matched littermate control mice (Fig. 7, C–E). Unexpectedly, a decreased EFS-induced contractility was observed in smTREK-1 KO mice compared with littermate control mice (3.0 ± 2.0 N/g vs. 1.8 ± 1.2 N/g in the control and smTREK-1 KO mice, respectively, P = 0.001 at 32 Hz; Fig. 7F). There were no changes in contractility evoked by CCh and ABMA between the groups (Fig. 7, G and H). The different neural components of the EFS (32 Hz)-induced bladder strip contractions were also comparable between smTREK-1 KO mice and littermate control mice (Fig. 7I).

DISCUSSION

Loss of TREK-1 Channel Activity in Global KO Mice Leads to a Mixed Symptoms Bladder Phenotype

Bladder storage and voiding functions involve coordinated contraction-relaxation of the urinary bladder, urethra, and sphincter, which are controlled by the peripheral and central nervous systems (17). Changes in bladder afferent and efferent nerve activity as well as functional impairment of bladder and/or innervation to these tissues can cause a wide range of dysfunctions. In this study, we examined the physiological roles of TREK-1 channels in the urinary bladder, using global and smooth muscle-specific TREK-1 KO mice. Global TREK-1 KO animals exhibited significant increases in bladder weight and total bladder wall thickness but, however, had no detectable signs of fibrosis. TREK-1 KO bladders showed decreased contractile forces in vitro in responses to EFS and neurotransmitter agonists as well as to high-K⁺ solution when normalized to the weight of the bladder strips. Additionally, a substantially larger bladder capacity and less frequent voiding were observed in TREK-1 KO mice during urodynamic recordings. Nevertheless, micturition function was largely preserved in TREK-1 KO mice, including normal micturition pressure and efficient bladder emptying. These results suggest that the total contractile function of the bladder was maintained in TREK-1 KO mice at least partly because of the bladder hypertrophy compensating for a decrease in BSM contractility. Our data of bladder hypertrophy with preserved micturition functions are aligned with a recent study that demonstrated that global TREK-1 KO mice retained cardiac function with cardiac hypertrophy without developing fibrosis upon pressure overload (11).

Role of TREK-1 Channels in the BSM Contractility-Relaxation Cycle

The BSM relaxes during the filling phase to accommodate increasing volume of urine while maintaining low intravesical pressure (1). TREK-1 channel activation has been shown to produce membrane hyperpolarization in cultured human BSM cells as well as inhibiting KCl-induced contractions of rat BSM isolated strips and decreasing the frequency of nonvoiding contractions in mice in vivo (7). Therefore, the lack of TREK-1 channels could potentially lead to the changes in the pressure threshold before bladder contraction or peak pressure at the micturition. We found that TREK-1 KO mice, at any age or sex, exhibited similar threshold and micturition pressure compared with age- and sex-matched control mice. This might be related to potential compensatory changes in the expression/function of other K⁺ channels to accommodate physiological urine storage. Intermicturition interval was significantly increased in TREK-1 KO animals compared with control animals, suggesting functional alterations in BSM contraction-relaxation mechanisms due to the lack of TREK-1 conductance. We also observed a trend of increases in frequency of NVCs, especially in 12-mo TREK-1 KO animals (Fig. 5, G and H), suggesting a mild DO phenotype in the bladders that completely lacked TREK-1 channel expression. However, we observed no significant changes in the spontaneous activity in in vitro experiments, suggesting that the loss of TREK-1 channels leads to little if any change in the electrical activity of BSM cells.

One caveat of using TREK-1 KO mice is that TREK-1 channel expression and activity are lacking in all cells, including not only smooth muscle but also the nervous system (5). Therefore, the phenotypes we identified in TREK-1 KO mice could originate either from the lack of the channel in BSM or in bladder innervation or the combination of both factors. In the present study, we used a smTREK-1 KO mouse model to distinguish the contribution of TREK-1 to bladder function in BSM from that in the neurons. At 6 mo of age, TREK-1 KO mice showed consistent mixed bladder phenotypes in both sexes; however, smTREK-1 KO mice did not replicate the TREK-1 KO bladder phenotype based on the majority of measured parameters (Fig. 6 and Fig. 7 and Table 3). These data strongly suggest that the loss of TREK-1 channel activity in BSM alone is not sufficient to induce bladder hypertrophy, BSM wall thickening, or impaired BSM contractility.

Lack of TREK-1 Channels in the Bladder Causes Hypertrophy With Preserved Function

Overall, micturition pressure and voiding efficiency were completely normal in TREK-1 KO animals, even when BSM contractility was significantly impaired. A previous study focused on TREK-1 channel function in cardiac tissue showed that the loss of TREK-1 channels led to an exaggerated pressure overload-induced concentric hypertrophy (an increase in left ventricular myocardial mass) yet retained preserved systolic and diastolic cardiac function (11), similar to urinary bladder hypertrophy observed in TREK-1 KO mice. The absence of pressure overload-induced dysfunction in TREK-1 KO mice was associated with diminished cardiac fibrosis, which also echoes our observation of unchanged collagen content in the bladder. TREK-1 deletion in cardiomyocytes did not prevent cardiac dysfunction following pressure overload; however, TREK-1 deletion in fibroblasts prevented deterioration in cardiac function, suggesting a cell type-specific role of TREK-1 channels in the pathogenesis of pressure overload-induced cardiac dysfunction. Our data also confirm that downregulation of the TREK-1 KO channel in the BSM layer is not responsible for the bladder hypertrophy observed in global TREK-1 KO mice. The cellular origin of the absence of TREK-1 channels in significant bladder hypertrophy with preserved function warrants future investigations focused on cell-specific conditional deletion of TREK-1 in the urinary system.

Impact of Neuronal TREK-1 Deletion on Bladder Morphology and Function

Expression of TREK-1 channels has been detected in neurons (18, 19), glial cells (20), and immune cells (21) in the nervous system. TREK-1 channels play important roles in ischemic and epileptic neuroprotection, pain perception, and depression (18). In the peripheral sensory nervous system, TREK-1 channels are expressed in sensory neurons and participate in thermal and mechanical perception (22). Studies have shown altered expression of TREK-1 channel in dorsal root ganglion (DRG) neurons in pathological conditions, including colitis (23), PBOO (24), inflammatory pain (25), and chronic constriction injury-induced neuropathic pain (26), suggesting that TREK-1 channels play an active role in afferent signaling. Prior research on TREK-1 KO mice showed that these mice are more sensitive to painful heat stimulation with an intensity near the threshold between nonpainful and painful heat, with no changes in cold sensing (22). TREK-1 KO mice also displayed hypersensitivity to mechanical stimulation, providing evidence that the physiological role of the TREK-1 channel in DRG neurons is likely inhibitory (22). However, inflammation-induced mechanical and thermal hyperalgesia was lower in TREK-1 KO mice than in wild-type control mice, suggesting the opposite roles of TREK-1 channels in the peripheral sensitization of nociceptors during inflammation (22). In our study, the absence of TREK-1 expression in bladder afferent neurons could result in decreased sensitivity during the filling phase and, in turn, increase functional bladder capacity. It is possible that prolonged overextension of the urinary bladder resulted in a hypertrophic phenotype and increased bladder wall thickness in TREK-1 KO mice.

Future Directions: Promises and Challenges

This is the first study that directly compared voiding function and bladder phenotypes between TREK-1 and smTREK-1 KO mice. We also analyzed different age groups, including 3-, 6-, 9-, and 12-mo-old animals of both sexes, to follow up the changes in bladder structure and function in the long term. We found significant age-related changes in bladder contractility in control mice, whereas in TREK-1 KO mice similar declines were detected as early as 6 mo (Fig. 2). These data suggested that decreased TREK-1 channel activity might contribute to physiological aging and age-related changes in BSM physiology. We also identified several sex differences in bladder functions. Future studies are warranted to assess the sex-specific roles of the TREK-1 channel in bladder storage and micturition.

A major obstacle in studying K_2P channels has been the lack of selective channel agonists and inhibitors (1). TREK-1 channels are activated by heat, mechanical stretch, and unsaturated free fatty acids such as arachidonic acid and are inhibited by Gd³⁺, phorbol myristate (PMA), extracellular acidic pH, cAMP, forskolin, and l-methionine (2, 27–29). The use of global TREK-1 KO mice does not allow for spatial or temporal control of TREK-1 channel expression, presenting limitations in data interpretation as well. With accumulating evidence of TREK-1 expression in many cell types, a deletion of TREK-1 in specific cell types (e.g., fibroblasts, the urothelium, and sensory neurons) will be required to identify the functional contribution of TREK-1 to bladder physiology.

DATA AVAILABILITY

All original datasets and statistical analysis files were uploaded on the Figshare website (https://figshare.com) and are available at https://doi.org/10.6084/m9.figshare.25493176.v1.

GRANTS

This work was supported by National Institutes of Health Grant DK121506 (to A.P.M.). Collagen imaging experiments were performed in the University of Colorado Anschutz Medical Campus Advance Light Microscopy Core supported in part by the Rocky Mountain Neurological Disorders Core (P30NS048154) and by National Institutes of Health Grant UL1TR001082.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.X.X. and A.P.M. conceived and designed research; A.X.X. and N.I. performed experiments; A.X.X. and N.I. analyzed data; A.X.X., N.I., and A.P.M. interpreted results of experiments; A.X.X. and N.I. prepared figures; A.X.X. drafted manuscript; A.X.X., N.I., and A.P.M. edited and revised manuscript; A.X.X., N.I., and A.P.M. approved final version of manuscript.