Bimatoprost Increases Mechanosensitivity of Trigeminal Ganglion Neurons Innervating the Inner Walls of Rat Anterior Chambers via Activation of TRPA1

Abstract

Purpose: Our previous study found that some trigeminal ganglion (TG) nerve endings in the inner walls of rat anterior chambers were mechanosensitive, and transient receptor potential ankyrin 1 (TRPA1) was an essential mechanosensitive channel in the membrane. To address the effect of bimatoprost on the mechanosensitive TG nerve endings in the inner walls of rat anterior chambers, we investigated its effect on their cell bodies in vitro.

Methods: Rat TG neurons innervating the inner walls of the anterior chambers were labeled by anterior chamber injection of 1,1′-dilinoleyl-3,3,3′,3′-tetramethylindocarbocyanine, 4-chlorobenzenesulfonate (FAST DiI). Calcium imaging and whole cell patch clamp were used on neuronal cell bodies to detect the activation effect of TRPA1 channels. Whole cell patch clamp was performed to record the currents induced by drugs and mechanical stimulation. Mechanical stimulation was applied to the neurons by buffer ejection.

Results: Bimatoprost mimicked the effect of TRPA1 agonists, allyl isothiocyanate (AITC), and (R)-(+)-WIN55, 212-2 mesylate salt (WIN) in the TG neurons. Bimatoprost induced Ca²⁺ influx in HEK293 cells stably transfected with human TRPA1, but not in untransfected cells as AITC and WIN. Moreover, bimatoprost evoked inward currents via TRPA1 activation in FAST DiI–labeled TG neurons as WIN. Bimatoprost also enhanced mechanosensitivity of FAST DiI–labeled TG neurons via TRPA1 activation.

Conclusions: Our results indicate that bimatoprost is a novel agonist of TRPA1, and it can enhance mechanosensitivity of TG nerve endings in the inner walls of anterior eye chambers via TRPA1 activation in rats.

Latanoprost, travoprost, and bimatoprost are three efficacious prostanoid analogs used for the treatment of glaucoma. Their effects on aqueous humor outflow are similar. Despite the well-established efficacy of these drugs on intraocular pressure (IOP), the mechanism underlying the hypotensive effect is not fully understood. Latanoprost and travoprost are prodrugs of prostaglandin (PG) FP receptor agonists. A well-studied mechanism for their enhancement of aqueous outflow is the regulation of matrix metalloproteinases and remodeling of extracellular matrix via FP receptor activation.¹ However, experimental and clinical evidence suggests that bimatoprost and prostanoid FP receptor agonists stimulate different receptor populations.^2–4 Studies performed on a wide array of receptors, ion channels, and transporters have demonstrated that bimatoprost does not meaningfully interact with adrenergic, cholinergic, cannabinoid, dopaminergic, or any of the known prostaglandin receptors, indicating that these receptors are not involved in the mediation of bimatoprost-induced responses.^2–4 It has been shown that 17-phenyl prostaglandin F2alpha (PGF2α), an acid hydrolysis product of bimatoprost, is found in the aqueous humor after topical application of bimatoprost in humans.⁵ Based on the finding that 17-phenyl PGF2α is a potent FP prostanoid receptor agonist, it was proposed that bimatoprost may reduce IOP via its metabolite, 17-phenyl PGF2α.⁵ Nonetheless, controversy regarding this hypothesis still exists. Low levels of 17-phenyl PGF2α were detected in aqueous humor samples from patients with cataract after a single dose of bimatoprost; however, these low levels do not sufficiently account for the ocular hypotensive efficacy of bimatoprost.⁶ Moreover, bimatoprost effectively reduced the IOP of patients who are nonresponders/unresponsive to latanoprost, a potent and selective prostanoid FP receptor agonist in the form of an isopropyl ester prodrug.^7,8 Interestingly, a patient hypersensitive to topical bimatoprost, but with no responsiveness to both latanoprost and travoprost, has been previously reported. Further investigation revealed that this patient had allele C and A of rs3753380 and rs3766355, respectively, in the FP gene, and −224T>C variation of the myocilin gene.⁹ This evidence indicates that FP prostanoid receptor may not be involved in the mediation of bimatoprost-induced hypotension of IOP. 

Bimatoprost is pharmacologically unique and appears to mimic prostamide, a member of the fatty acid amide family.^1,2 One biosynthetic route to prostamide involves anandamide, an endogenous cannabinoid.² Thus, bimatoprost and cannabinoid may reduce the IOP via a similar mechanism. Cannabinoids reduce IOP when administered topically and systemically, and have been proposed to lower IOP by exerting either central or peripheral effects.^10–12 To date, a specific mechanism for this effect has not been elucidated.^13,14 Cannabinoids have two different receptor subtypes, CB1 and CB2. Based on the expression of CB1 in both structures of the eye where aqueous humor was produced and removed, it was proposed that cannabinoids induced IOP reduction via CB1 receptor activation.¹⁴ Besides CB receptors, cannabinoids can activate transient receptor potential ankyrin 1 (TRPA1) channels.^15,16 In addition to cannabinoids, prostaglandin A2 (PGA2) and 2,6-diisopropylphenol (propofol) also perform dual functions of both IOP reduction and TRPA1 excitation.^17–20 The existence of these kinds of drugs suggests that there may be an unknown relationship between IOP and TRPA1 channels. 

Accumulating evidence has shown that the trigeminal ganglion (TG) nerve may be the primary afferent nerve in neural regulation of the IOP.^21–25 Our previous study found that some TG nerve endings in the inner walls of rat anterior chambers were mechanosensitive (MS), and TRPA1 was an essential MS channel in the membrane.²⁶ These results also gave a possible answer as to how IOP variation triggered action potentials of the TG nerve. We therefore proposed that mechanosensitive TG nerve endings in the inner walls of anterior chambers might act as IOP baroreceptors.²⁶ 

Based on the above-mentioned studies, we hypothesized that TRPA1 might be the common target of bimatoprost and cannabinoid, and the mechanosensitivity of TG nerve endings in the inner walls of anterior chambers might be enhanced via TRPA1 activation. In the present study, we conducted a preliminary investigation to validate this hypothesis. It has been previously shown that activation of TRPA1 induces Ca²⁺ influx.²⁷ Therefore, in this study, we determined the agonist effect of bimatoprost on the TRPA1 channel using Ca²⁺ imaging. Inaccessibility of sensory nerve endings and its smaller size have hampered the direct investigation of mechanical transduction processes on the nerve endings. Based on the fact that specific receptors and ion channels present on the nerve endings are also present on the soma of cultured neurons, strategies using isolated neurons as a model to uncover the transduction mechanisms of the sensory nerve endings have been developed.^28,29 We harvested TG neurons innervating the inner walls of anterior chambers after 1,1′-dilinoleyl-3,3,3′,3′-tetramethylindocarbocyanine, 4-chlorobenzenesulfonate (FAST DiI; Invitrogen, Carlsbad, CA, USA) labeling as previously reported,²⁶ and performed a whole cell patch clamp assay to verify that bimatoprost targeted TRPA1 channels in the membrane. Finally, we investigated the effect of bimatoprost on the mechanosensitivity of FAST DiI–labeled TG neurons. 

Materials and Methods

Reagents

We dissolved FAST DiI (Invitrogen) to a stock solution of 5 mg/mL in dimethyl sulfoxide (DMSO), and diluted to a final concentration of 2.5 mg/mL with deionized water for FAST DiI retrograde labeling experiments. Cell permeant (Fura-2 AM; Invitrogen) was dissolved in DMSO to 1 μM for calcium imaging. Bimatoprost, (R)-(+)-WIN55,212-2 mesylate salt (WIN), ruthenium red (RR), allyl isothiocyanate (AITC), and HC-030031 were purchased from Sigma Aldrich Corp. (St. Louis, MO, USA). All the reagents were prepared as stock solutions and stored at −20°C or 4°C according to the manufacturers' instructions. These stock solutions were diluted to their final concentrations with the extracellular solution (147 mM NaCl, 5 mM KCl, 1 mM MgCl₂ • 6H₂O, 2 mM CaCl₂, 10 mM D-glucose, and 10 mM HEPES) before application. Bimatoprost (25 ng/mL), AITC (100 μM), KCl (K⁺, 75 mM), WIN (25 μM), RR (30 μM) and HC-030031 (25 μM) were applied by directly dissolving in the extracellular solution for calcium imaging. We diluted WIN to a concentration of 25 μM, bimatoprost to 25 ng/mL with extracellular solution before application, and applied via tube drug application system in whole cell patch clamp experiments. The intracellular solution (118 mM K-Asp, 20 mM KCl, 2 mM MgCl₂ • 6H₂O, 1 mM CaCl₂, 10 mM EGTA, 10 mM HEPES, 5 mM Na₂ATP) was used in whole cell patch clamp experiments. The final DMSO concentration was 0.05% for the patch clamp experiments. Dulbecco's modified Eagle's medium (DMEM/F-12) and fetal bovine serum (FBS) were purchased from Gibco (Carlsbad, CA, USA); Papain from Roche (Basel, Switzerland); Collagenase (type Xl-S) and DNase I (type IV) from Sigma Aldrich Corp. and penicillin-streptomycin from HyClone (Logen, UT, USA). 

Animals

Adult Sprague-Dawley (SD) rats (180–200 g) were individually housed in plastic cages with free access to food and water, and maintained in climate-controlled (23 ± 1°C) and light-controlled (12/12-hour dark/light cycle with light on at 8:00 AM) protected units for at least 10 days before the experiments. The Institutional Animal Care and Use Committee of Tongji Medical College, Huazhong University of Science and Technology, approved the experimental protocols. Handling and treatment of animals, and all other animal use procedures were carried out according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. In total, 127 rats were used in this study. 

FAST DiI Retrograde Labeling

FAST DiI retrograde labeling was performed to identify the cell bodies of TG neurons innervating the inner walls of anterior chambers, as previously described.^26,30 Briefly, after the rats were deeply anesthetized with 10% chloral hydrate (4 mL/kg) and topically anesthetized with 0.4% oxybuprocaine hydrochloride eye drops, a borosilicate glass was inserted through the cornea into the anterior chamber of the eye. Caution was used to avoid injuring the iris and lens, as well as to avoid leakage of aqueous humor from the cornea. The borosilicate glass pipette was connected to a 10-μL Hamilton syringe. Three microliters of FAST DiI solution (2.5 mg/mL in DMSO) was injected slowly into the anterior chamber (3 μL/minute) to avoid induction of acute intraocular hypertension. The animals were declined access to water for 12 hours after revival. 

Cultures of TG neurons

Four days after FAST DiI injection, the rats were decapitated under anesthesia with chloral hydrate. As previously described, the ipsilateral TG was dissected and TG neurons were prepared rapidly to minimize exposure of the cells to light for further experiments within 24 hours after plating.²⁶ Briefly, the ipsilateral TG was dissected in D-Hanks' balanced salt solution (HBSS) at 4°C and then incubated in 0.1% collagenase and 20 U/mL of papain in modified HBSS (pH 7.4). After incubation for 30 to 50 minutes at 37°C, individual cells were dissociated by triturating the tissue through a fire-polished glass pipette, followed by incubation for 10 minutes at 37°C in 10 mg/mL of DNase I in modified Hanks' solution. Cells were washed three times with modified HBSS, and then cultured in DMEM/F12 supplemented with 10% FBS and 100 units/mL of penicillin-streptomycin. Cells were plated on glass coverslips and cultured for 1 to 3 hours at 37°C in a humidified atmosphere with 20% O₂ and 5% CO₂. Cells were used for further experiments within 24 hours after plating. 

Cultures of Human Embryonic Kidney (HEK) Cells

We purchased the HEK293 cells stably transfected with human TRPA1 (hTRPA1-HEK) and nontransfected human TRPA1 HEK 293 cells (nt-TRPA1-HEK) from Genechem Co., Ltd. (Shanghai, China). The cells hTRPA1-HEK and nt-TRPA1-HEK were plated on glass coverslips, and cultured for 16 to 30 hours in DMEM/F-12 supplemented with 10% FBS and 100 units/mL of penicillin-streptomycin at 37°C in a humidified atmosphere with 20% O₂ and 5% CO₂. Cells were used for further experiments after plating. 

Calcium Imaging

Trigeminal ganglion neuron-covered or HEK293-covered coverslips were loaded with Fura-2 AM (1 μM) in extracellular solution and incubated for 40 minutes at 37°C, 5% CO₂. After rinsing in HBSS thrice, the cell-covered coverslips were placed in a mobile metal chamber of the inverted microscope, and then set aside for 10 minutes in the extracellular solution. Cell permeant (Invitrogen) fluorescence, which was dually excited sequentially at 340 and 380 nm, was recorded and measured by charge-coupled device imaging system (Apogee Imaging System, Roseville, CA, USA) and software (TillVision; TILL Photonics GmbH, Munich BioRegio, Germany). The obtained fluorescence ratio (ratio = F340/F380) indirectly reflects changes in the intracellular free calcium concentration ([Ca²⁺]_free). The ratio images were acquired every second. Baseline detection was performed 2 minutes before drug application. Total detection time was 5 to 15 minutes. We applied K⁺ (75 mM) to the cells to confirm their ion voltage sensitivity after each treatment. A total of 42 rats were used in experiments for calcium imaging. 

Electrophysiological Experiments

Whole cell patch clamp experiments were performed at room temperature (25°C) on FAST DiI–labeled TG neurons as previously described²⁶ using a HEKA EPC-10 patch clamp amplifier (HEKA, Lambrecht/Pfalz, Germany) connected to a compatible computer via a D/A and A/D converter. A Nikon fluorescence microscope was used to identify FAST DiI–labeled TG neurons. Voltage-clamp experiments were performed at a holding potential of −80 mV, and the sampling and filter rates were 10 and 2 kHz, respectively. The glass pipettes were made by a two-stage puller (PC-10; Narishige Instruments, Tokyo, Japan) from star-bore capillary tubes and had a resistance (Rs) of 2 to 4 MΩ after the pipette solution was perfused. The resistance series was electrically compensated for by at least 75% in most experiments to minimize the capacitive surge on the current recording. The results were excluded if the uncompensated Rs series caused voltage-clamp errors of more than 15 M. The pipette solution used for the whole cell recordings constituted 118 mM of K-Asp, 20 mM of KCl, 2 mM of MgCl₂ • 6H₂O, 10 mM of EGTA, and 5 mM of Na₂ATP. The osmolarity of the pipette solution was adjusted to 300 mosm/kg, and the pH was adjusted to 7.2 with N-methyl glucamine. The osmolarity of the extracellular solution was adjusted to 300 mosm/kg, and the pH was adjusted to 7.4. RR (5 μM) and HC-030031 (25 μM) were administered via bath application, while bimatoprost (25 ng/mL) and WIN (25 μM) were given via a tube drug application system. Data acquisition and analysis were carried out using acquisition and analysis software (Pulse + PulseFit 8.11; HEKA Elektronik Dr. Schulze GmbH, Lambrecht/Pfalz, Germany). The current magnitude was quantified using the peak current amplitude, and the current density (pA/pF) was analyzed. A cutoff of 20 pA was selected as the minimum amplitude for response in all of the experiments based on preliminary studies. A total of 85 rats were used in electrophysiological experiments. 

Mechanical Stimulation

Mechanical stimulation was applied to the cell bodies of FAST DiI–labeled TG neurons using extracellular buffer ejected from a micropipette by an automatic air pump (Suction Control Pro; Nanion Technologies GmbH, Munich, Germany) as previously described.²⁶ The automatic air pump provided graded pressure (10–40 mm Hg) for 200 ms each time to eject bath solution. The tip of the micropipette (2 μm in diameter) was placed 12 μm from the cell surface. 

Data Analysis

Cellular experiments were defined as follows: [Ca²⁺]_x was the apparent [Ca²⁺]_free of the cell at a given time point; [Ca2+]_b was the cellular mean baseline apparent [Ca²⁺]_free measured over 120s. Ratio (340/380 nm) = [Ca²⁺]_x/[Ca²⁺]_b. [Ca²⁺]_max was the cellular mean response apparent [Ca²⁺]_free measured over 120 seconds; ΔRatio = ([Ca²⁺]_max–[Ca²⁺]_b)/[Ca²⁺]_b. For the neuronal experiments, neurons were defined as “responders” to a given compound if the mean response was greater than the mean baseline plus 2 × the SD. Only neurons that responded to K⁺ were included in the analysis. Data were expressed as the mean ± SEM. Pairwise comparisons between the two groups were done using the Wilcoxon rank sum test with P < 0.05 as the significance threshold. We used R-3.0.2 software for the data analysis. 

The current magnitude was quantified using the peak current amplitude, and the current density (pA/pF) was analyzed. The quantitative data were expressed as the mean ± SEM. Differences between the two groups were compared using the paired Wilcoxon rank sum test with P < 0.05 as the significance threshold. We used R-3.0.2 software for the data analysis. 

Results

Dose-Response of Bimatoprost on Calcium Influx in TG Neurons

Since TRPA1 channels are associated with calcium influx, we explored the effects of bimatoprost on calcium influx in dissociated TG neurons to study the possible relationship between bimatoprost and TRPA1 channels. Activity of TG neurons was tested by application of K⁺ (75 mM) bath solution before the end of each calcium imaging. Responses of Ca²⁺ of TG neurons to bimatoprost (vehicle, 5, 25, 50, 100, and 200 ng/mL) bath solution are shown in Figure 1A. As shown in Figure 1B, TG neurons in calcium imaging assays were activated with bimatoprost at concentrations of 5 ng/mL (0.21 ± 0.03, n = 42, 42 neurons from 3 rats); 25 ng/mL (0.39 ± 0.08, n = 49, 49 neurons from 3 rats); 50 ng/mL (0.30 ± 0.10, n = 23, 23 neurons from 3 rats); 100 ng/mL (0.18 ± 0.04, n = 29, 29 neurons from 3 rats); and 200 ng/mL (0.10 ± 0.01, n = 23, 23 neurons from 3 rats), while 0 ng/mL bimatoprost (vehicle) failed to activate the TG neurons (0.05 ± 0.01, n = 15, 15 neurons from 3 rats). Since maximum calcium influx was produced by the application of 25 ng/mL bimatoprost bath solution, this concentration was selected for the subsequent experiments. 

Figure 1

 

Bimatoprost induces calcium influx in TG neurons. (A) Representative Ca2+ responses of TG neurons to application of bimatoprost bath solution (vehicle, 5, 25, 50, 100, and 200 ng/mL). Bimatoprost was applied for 120 seconds (blocked line). All neurons responded to K+ (75 mM) bath solution application. (B) The ordinate shows the magnitude of the Ca2+ responses of the neurons (Fura-2 ΔRatio [340/380 nm]). Data for Ca2+ responses of neurons that responded to bimatoprost (control [n = 15], 5 ng/mL [n = 42], 25 ng/mL [n = 49], 50 ng/mL [n = 23], 100 ng/mL [n = 29], 200 ng/mL [n = 23]) are shown as mean ± SEM. Bimatoprost at 25 ng/mL induced the maximum calcium influx in TG neurons.

Bimatoprost Mimics the Effects of AITC and WIN in TG Neurons

We further used calcium imaging to compare the effects of bimatoprost with TRPA1 agonists (e.g., AITC, WIN) on dissociated rat TG neurons. Pretreatment with either 25 μM HC-030031 (specific TRPA1 blocker)³¹ or 30 μM RR (nonspecific TRP family blocker)³² blocked 25 ng/mL bimatoprost-induced Ca²⁺ influx in TG neurons (p_{vehicle-Bimatoprost versus HC-Bimatoprost} < 0.01, p_{vehicle-Bimatoprost versus RR-Bimatoprost} < 0.01, Figs. 2A, 2B). Similarly, 25 μM HC-030031 and 30 μM RR blocked TRPA1 agonists (e.g., 100 μM AITC, 25 μM WIN) induced Ca²⁺ influx in TG neurons (p_{vehicle-AITC versus HC-AITC} < 0.01, p_{vehicle-AITC versus RR-AITC} < 0.01, Figs. 2C, 2D; p_{vehicle-WIN versus HC-WIN} < 0.01, p_{vehicle-WIN versus RR-WIN} < 0.01, Figs. 2E, 2F). All the neurons responded to K⁺ (75 mM) applied immediately before the end of the Ca²⁺ imaging. These results confirmed that bimatoprost mimics the effects of TRPA1 agonists, such as AITC and WIN, in TG neurons. 

Figure 2

 

Bimatoprost mimics the effect of AITC and WIN in TG neurons. (A) Ca2+ responses of dissociated TG neurons responding to bimatoprost (25 ng/mL) in the presence of vehicle, the TRPA1-specific inhibitor HC-030031 (HC, 25 μM), and the nonselective TRP inhibitor ruthenium red (RR, 30 μM). (B) The ordinate shows the magnitude of Ca2+ responses of the neurons. Data for Ca2+ responses of neurons that responded to bimatoprost after bath application of vehicle; HC and RR are shown as mean ± SEM. **P < 0.01: significant difference between each column. (C) Ca2+ responses of dissociated trigeminal neurons responding to AITC (100 μM) in the presence of vehicle, HC (25 μM), or RR (30 μM). (D) The ordinate shows the magnitude of Ca2+ responses of the neurons. Data for Ca2+ responses of the neurons that responded to AITC after bath application of vehicle; HC and RR are shown as mean ± SEM. **P < 0.01: significant difference between each column. (E) Ca2+ responses of dissociated trigeminal neurons responding to WIN (25 μM), in the presence of vehicle, HC (25 μM), or RR (30 μM). (F) The ordinate shows the magnitude of Ca2+ responses of the neurons. Data for Ca2+ responses of the neurons that responded to WIN after bath application of vehicle; HC and RR are shown as mean ± SEM. **P < 0.01: significant difference between each column. Blocked line denotes the duration of agonist treatment: bimatoprost, AITC, and WIN were applied for 120 seconds. All the neurons responded to K+ (75 mM) applied before the end of Ca2+ imaging.

Bimatoprost Induces Ca²⁺ Influx in hTRPA1-HEK, but not nt-TRPA1-HEK Cells

Because TRPA1 can be stimulated indirectly by increased levels of intracellular calcium,^33,34 it is possible that bimatoprost activates TRPA1 via an uncharacterized mechanism which leads to elevated intracellular calcium. We examined bimatoprost-induced Ca²⁺ response in hTRPA1-HEK and nt-TRPA1-HEK cells to conform the direct activation of TRPA1 by bimatoprost. As shown in Figure 3A, hTRPA1-HEK cells in calcium imaging assays responded robustly to bath solution application of bimatoprost (25 ng/mL), and responded again to bath solution application of 100 μM AITC,³⁵ a specific agonist of TRPA1, after bimatoprost washout. Furthermore, after AITC washout, hTRPA1-HEK cells responded again to bath solution application of 25 μM WIN, a synthetic and selective CB1 receptor agonist and a nonspecific agonist of TRPA1.¹⁶ In contrast, nt-TRPA1-HEK cells failed to responded to 25 ng/mL bimatoprost, 100 μM AITC and 25 μM WIN, but responded to K⁺ (75 mM; Fig. 3B). The magnitudes of HEK cells Ca²⁺ responses (ΔRatio [340/380 nm]) are shown in Figure 3C. The magnitudes of the Ca²⁺ responses of hTRPA1-HEK cells to bimatoprost (25 ng/mL) was 0.27 ± 0.021 (n = 75), while that of nt-TRPA1-HEK cells was 0.06 ± 0.002 (n = 140). The difference between the Ca²⁺ responses of hTRPA1-HEK and nt-TRPA1-HEK cells was significant (P < 0.01). Similarly, the difference between Ca²⁺ responses to AITC (100 μM) of hTRPA1-HEK (0.53 ± 0.033, n = 126) and nt-TRPA1-HEK (0.10 ± 0.03, n = 50) was significant (P < 0.01). Compared with the hTRPA1-HEK–mediated Ca²⁺ responses to WIN (25 μM; 0.22 ± 0.055, n = 25), Ca²⁺ responses to WIN in nt-TRPA1-HEK (0.08 ± 0.005, n = 49) were significantly different (P < 0.01). Our assays did not detect any bimatoprost-mediated increase in intracellular calcium in nt-TRPA1-HEK. The results confirm that bimatoprost induced calcium influx via direct activation of TRPA1. 

Figure 3

 

Bimatoprost mimics the TRPA1 activation effect of AITC and WIN in hTRPA1-HEK cells. (A) Representative Fura-2 AM ratiometric image of Ca2+ responses of hTRPA1-HEK cells to bimatoprost (25 ng/mL), AITC (100 μM), and WIN (25 μM). (B) Ca2+ responses of nt-TRPA1-HEK cells to bath solution application of bimatoprost (25 ng/mL), AITC (100 μM), WIN (25 μM), and K+ (75 mM). (C) The ordinate shows the magnitude of the Ca2+ responses of HEK cells (ΔRatio [340/380 nm]). Data for Ca2+ responses of hTRPA1-HEK and nt-TRPA1-HEK cells that responded to bimatoprost (25 ng/mL), AITC (100 μM), and WIN (25 μM) are shown as mean ± SEM. **P < 0.01: significant difference between each column. Blocked line denotes the duration of agonist treatment: bimatoprost, AITC, and WIN were applied for 120 seconds. All cells responded to K+ (75 mM).

Bimatoprost and WIN Induced Inward Currents via TRPA1 Activation in FAST DiI–Labeled TG Neurons

Besides the calcium imaging assay, we also performed the whole cell patch clamp recordings to confirm the TRPA1 activation effect of bimatoprost and WIN on FAST DiI–labeled TG neurons. Bimatoprost (25 ng/mL) led to an inward current in FAST DiI–labeled TG neurons (Fig. 4A), and induced the current density to 29.11 ± 11.83 pA/pF (n = 8; eight neurons from four rats). The bath application of HC-030031 (25 μM, the specific blocker of TRPA1^31,32,36) caused a 71% ± 9% decrease in the amplitude of the inward currents from 29.11 ± 11.83 pA/pF to 8.52 ± 4.00 pA/pF (Figs. 4A, 4B; n = 8, eight neurons from four rats, P < 0.01). The bath application of HC-030031 (25 μM), resulted in a 55% ± 12% reduction in the amplitude of the currents induced by WIN from 44.88 ± 16.66 pA/pF to 15.46 ± 6.59 pA/pF (Figs. 4C, 4D; n = 8, eight neurons from five rats, P < 0.01) in FAST DiI–labeled TG neurons. These results suggest that bimatoprost and WIN induced inward currents via TRPA1 channels in FAST DiI–labeled TG neurons. 

Figure 4

 

The inhibitory effect of HC-030031 on bimatoprost and WIN induced currents in FAST DiI-labeled TG neurons. (A) Inhibition of the bimatoprost activated whole cell currents by HC-030031. Currents elicited by bimatoprost in the same neuron before (left) and after (right) bath application of 25 μM HC-030031. (B) The ordinate shows the current density. Data for current density before and after bath application of 25 μM HC-030031 are shown as mean ± SEM (n = 8, eight neurons from four rats). **P < 0.01: significant difference between each column. (C) Inhibition of the WIN-activated whole cell currents by HC-030031. Currents elicited by WIN in the same neuron before (left) and after (right) bath application of 25 μM HC-030031. (D) The ordinate shows the current density. Data for current density before and after bath application of 25 μM HC-030031 are shown as mean ± SEM (n = 8, eight neurons from five rats). **P < 0.01: significant difference between each column.

Bimatoprost Enhanced Mechanosensitivity of FAST DiI–Labeled TG Neurons via TRPA1 Activation

Mechanosensitive currents at a clamp voltage of −80 mV were obtained via application of bath solution to the soma at a pressure of 30 mm Hg provided by the automatic air pump (Fig. 5A). Bimatoprost (25 ng/mL) caused 55% ± 12% increase in the amplitude of 30 mm Hg-induced MS currents from 7.09 ± 1.45 pA/pF to 10.30 ± 1.97 pA/pF (Figs. 5A, 5B; n = 11, 11 neurons from 8 rats, P < 0.01). 

Figure 5

 

Bimatoprost enhanced the mechanically induced currents in FAST DiI–labeled TG neurons via TRPA1 activation. (A) The currents shown were elicited via 200-ms jets of bath solution ejected from the pipette at pressures of 30 mm Hg in the same neuron before, after bath application of 25 ng/mL bimatoprost, after bath application of 25 μM HC-030031 + 25 ng/mL bimatoprost, and after washout (from left to right). Similar results were obtained in 10 other neurons. (B) The ordinate shows the current density. Data for current density before and after bath application of 25 ng/mL bimatoprost (n = 11, 11 neurons from 8 rats) are shown as mean ± SEM. **P < 0.01: significant difference between each column. (C) The ordinate shows the current density. Data for current density after bath application of 25 ng/mL bimatoprost and after bath application of 25 μM HC-030031 + 25 ng/mL bimatoprost (n = 5, five neurons from four rats) are shown as mean ± SEM. **P < 0.01: significant difference between each column.

To determine whether bimatoprost enhanced MS currents on FAST DiI–labeled TG neurons via TRPA1 activation, we pretreated the same TG neurons with HC-030031, a specific blocker of TRPA1, before treatment with bimatoprost. Although the gigaseal was ruptured in 6 out of 11 cells, the data from the other five cells showed that HC-030031 (25 μM) caused a 46% ± 7% decrease in the amplitude of MS currents from 9.38 ± 1.61 pA/pF to 4.96 ± 1.08 pA/pF (Figs. 5A, 5C, n = 5, five neurons from four rats, P < 0.01). The inhibited MS currents recovered partially after washout (Fig. 5A). The results indicate that bimatoprost enhanced mechanosensitivity of TG neurons via TRPA1 activation. 

Discussion

Our present study demonstrated that bimatoprost had an agonistic effect on the TRPA1 channel, and it enhanced the mechanosensitivity of TG nerve endings in the inner walls of anterior chambers via TRPA1 activation in rats. 

To explore the mechanism underlying the hypotensive effect, bimatoprost pharmacology has been extensively characterized by binding and functional studies at more than 100 drug targets, which comprise a diverse variety of receptors, ion channels, and transporters. However, bimatoprost exhibited no meaningful activity at receptors known to include antiglaucoma drug targets as follows: adenosine (A1–3), adrenergic (α1, α2, β1, β2), cannabinoid (CB1, CB2), dopamine (D1–5), muscarinic (M1–5), prostanoid (DP, EP1–4, FP, IP, TP), and serotonin (5HT1–7).² In addition, studies showing some effects of bimatoprost could not be explained by hydrolytic conversion of bimatoprost to 17-phenyl PGF2α. For example, a study on isolated iridial cells demonstrated that bimatoprost stimulated an entirely different cell population to those sensitive to PGF2α and 17-phenyl PGF2α.³ Another study using isolated uterine cells showed that bimatoprost and 17-phenyl PGF2α had different activity profiles, and the potent contractile effects of bimatoprost in the rabbit uterus could not be explained by its interaction with known prostanoid FP receptors. The study proposed the presence of a novel receptor population that preferentially recognizes bimatoprost in the uterus.⁴ Recent data showed that intrauterine application of HC-030031, the specific blocker of TRPA1, attenuated the AITC-induced uterine irritation and colonic hypermotility.³⁷ In the present study, our data show bimatoprost is an agonist of TRPA1 channel as WIN and AITC. Taken together, it is suggested that the potent contractile effects of bimatoprost in the rabbit uterus may attribute to TRPA1 channel activation. 

Many compounds with TRPA1 activation effect have been found. Some are electrophilic activators, such as AITC, 15-d-PGJ2, △¹²-PGJ2, isoprostane 8-iso-PGA2, and PGA2, which may induce covalent modifications of TRPA1 channel proteins via reactive electrophilic moieties.^17,38 Electrophilic nature of agonists is critical for TRPA1 activity.¹⁷ Bimatoprost also possesses electrophilic moieties. In our present study, data show that the effect of AITC on calcium efflux in hTRPA1-HEK cell is almost double that of bimatoprost. Possible reasons behind their different effects on TRPA1 may be attributed to their different electrophilic natures. However, some others like propofol, an intravenous anesthetic drug, modulate TRPA1,²⁷ but not via induced covalent modifications of the channel proteins. Among the activators of TRPA1, cannabinoid, prostaglandin A2 (PGA2) and propofol exhibit IOP decreasing effects.^15,17–20 The result from our present study show bimatoprost as a novel TRPA1 agonist with IOP hypotension effect. The data indicate toward an unknown relevance between IOP decreasing effect and TRPA1 channel activation. 

Although tremendous efforts have been made, how IOP variation triggers action potential and the mystery of the IOP baroreceptor remains elusive. In the past 2 decades, the mechanism of mechanoreceptors, such as blood pressure baroreceptors, inner ear hair cells, skin Pacinian corpuscles, were uncovered following the identification of MS channels.²⁹ All known mechanoreceptors of an organism are mechanosensitive nerve endings with special structures.²⁹ The ability of mechanoreceptors to detect mechanical stimulus relies on MS channels in the membrane.²⁹ The opening of MS channels can trigger the process of mechanotransduction.^28,29,39 Previous reports have shown that TG nerve endings in the scleral septa and the trabecular meshwork of human and monkey eyes exhibit typical morphologic characteristics of mechanoreceptive nerve endings, contain numerous mitochondria, and are in close contact with extracellular matrix components, especially the elastic-like fibers.^24,25 Taken together with our previous findings of the mechanotransduction process in TG nerve endings,²⁶ mechanosensitive TG nerve endings in the inner walls of rat anterior chambers may act as IOP baroreceptors. 

Recent developments have been made in identifying the molecules of mechanotransduction. Some members of the 2P-domain K⁺ channel (K2P), ASICs, and TRP channel superfamily proteins and Piezo proteins are considered as candidate channel proteins for mechanotransducers.^29,40–42 Transient receptor potential ankyrin 1 is a member of the large TRP family of ion channels, and functions as a Ca²⁺ permeable nonselective cation channel.²⁷ Numerous studies have demonstrated that the TRPA1 channel protein can affect the mechanosensation process in tissues.^27,43–45 Our previous study found that TRPA1 is an important mechanotransducer in the membrane of mechanosensitive TG nerve endings in the inner walls of rat anterior chambers, a potential IOP baroreceptor.²⁶ 

It has been shown that TRPA1 activation enhances the mechanosensitivity of visceral sensory neurons.^44,46,47 The respiratory tract is innervated by primary sensory afferent nerves, which are activated by mechanical and chemical stimuli. Data suggested that activation of TRPA1 on these vagal sensory afferents leads to central reflexes, including dyspnea, changes in the breathing pattern, and cough, which contribute to the symptoms and pathophysiology of respiratory diseases.⁴⁸ The channel TRPA1 has been reported to play a role in the stimulation of mechanosensitive afferent nerve activities of both Ad- and C-fibers of the rat bladder.⁴⁴ It has also been reported that TRPA1 tunes DRG neurons mechanosensitivity based on its degree of activation or expression.⁴⁵ Intriguingly, our present data showed that bimatoprost increased the mechanosensitivity of TG neurons innervating the inner walls of rat anterior chambers via direct TRPA1 activation. Based on the fact that the same receptors and channels exist in both soma and nerve ending of a neuron, the conclusion from the soma can be applied to the terminal.²⁹ Therefore, we propose that bimatoprost has the capacity to enhance the mechanosensitivity of TG nerve endings in the inner walls of rat anterior chambers via TRPA1 activation. 

Our current study has two major limitations. First, we did not directly measure the high IOP induced currents on the TG nerve endings in vivo due to small inaccessible size of the nerve endings. However, since most receptor molecules are expressed both on the membranes of a soma and nerve ending of a neuron, studies performed on the soma of a sensory neuron in vitro are substantial in understanding the mechanotransduction process. Second, we did not perform in vivo animal studies to prove that the hypotensive property of bimatoprost is mediated through TRPA1. However, IOP hypotensive effect of some other TRPA1 agonists (cannabinoid, PGA2 and propofol)^12,17–20 may provide indirect in vivo evidence. In our future studies, we aim to use TRPA1 gene knockout rats to prove that the hypotensive property of bimatoprost is mediated through TRPA1. 

In summary, our results provided evidence that TRPA1 is a novel receptor of bimatoprost, and bimatoprost may stimulate mechanosensitive TG nerve endings in the inner walls of the rat anterior chamber via TRPA1. However, it remains to be determined whether the mechanosensitive TG nerve endings in the inner walls of the rat anterior chambers act as IOP baroreceptors, and whether TRPA1 plays a role in IOP regulation. The results from our study present a valuable clue to explore further the mechanism leading to the hypotensive effect of bimatoprost. 

Acknowledgments

Supported by a grant from the National Natural Science Foundation of China (No. 81070727). 

Disclosure: Y. Ling, None; Z. Hu, None; Q. Meng, None; P. Fang, None; H. Liu, P