Thrombin-Induced Responses Via Protease-Activated Receptor 1 Blocked by the Endothelium on Isolated Porcine Retinal Arterioles

Purpose: Thrombin, a serine protease, causes organ-specific responses to vessels. However, the mechanism by which thrombin affects the retinal microcirculation remains unclear. We examined the effects of thrombin on the retinal microvasculature and signaling mechanisms. Methods: Porcine retinal arterioles were isolated, cannulated, and pressurized (55 cmH 2 O) without flow in this in vitro study. Videomicroscopy techniques recorded changes in diameter in the retinal arterioles in response to thrombin at concentrations ranging from 0.001 to 20 mU/ml. Results: Extraluminal administration of thrombin induced concentration-dependent vascular responses, i.e., vasoconstriction at low concentrations less than 5 mU/ml and vasorelaxation with high concentrations greater than 5 mU/ml. However, intraluminal administration of thrombin (5 mU/m) did not constrict the retinal arterioles; in denuded vessels, intraluminal administration constricted the retinal arterioles. Thrombin-induced vasoconstriction was significantly ( p < 0.01) suppressed by pretreatment with a protein kinase C (PKC) inhibitor and a protease-activated receptor (PAR)-1 inhibitor but not by PAR-2 and PAR-4 inhibitors or denudation. A rho kinase (ROCK) inhibitor also


Introduction
Thrombin, a serine protease, plays an important role in hemostasis by converting fibrinogen into fibrin.Many studies have reported the responses of the vascular system to thrombin.Thrombin causes vasodilation of the coronary (pig, dog, and human) [1][2][3][4] and pulmonary (newborn pig) arteries 5 and vasoconstriction of the pulmonary (pig) 6 and cerebral (dog) arteries. 7,8In the porcine renal interlobar artery, thrombin also causes vasorelaxation at low concentrations and vasoconstriction at high concentrations. 9Taken together, thrombin causes various responses depending on the vessels in the various organs.
Moreover, thrombin mediates various vascular effects via protease-activated receptor (PAR). 10Since PAR is expressed in the vascular endothelium and smooth muscle, [11][12][13][14][15] the vascular reactivity of thrombin might be induced via PAR on the vessel wall.
Thrombin causes vasoconstriction of bovine retinal arterioles. 160][21][22] Based on these factors, elevated serum and intravitreal thrombin concentrations might contribute to the pathophysiology of ocular Takahashi et al.--5   disorders via activation of PAR.However, the role of PAR in thrombin-induced vasomotion of the retinal arterioles remains unclear.
Because blood only flows in the vascular lumen under normal physiologic conditions, thrombin can cause a direct response on the endothelium.In most studies, the effects of thrombin on the endothelium have been evaluated in cultured endothelial cells, i.e., thrombin increases production of nitric oxide (NO) and prostacyclin (PGI 2 ), potent vasodilators of retinal arterioles, in cultured endothelial cells. 3,23However, extraluminal administration of thrombin also causes vasoconstriction of the retinal arterioles. 16Thus, the intraluminal effects of thrombin might differ from the extraluminal effects of thrombin on the endothelium of the retinal arterioles.The current study examined the intraluminal and extraluminal effects of thrombin and investigated the signaling mechanisms via PAR involved in the vasomotor activity of the retinal microvasculature.We hypothesized that only extraluminal administration of thrombin can cause the response and not intraluminal administration, because the permeability of the inner blood-retinal barrier (BRB) to molecules is below 1 kDa 24 and the molecular weight (MW) of thrombin is 33.6 kDa.

Animal preparation
The Animal Care Committee of Asahikawa Medical University approved all animal protocols, which were conducted according to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.
Eyeballs from white pigs of either sex (age, 22-24 weeks; weight, 100-150 kg) were obtained from local slaughterhouses and transported to the laboratory immediately in a humid chamber on ice.

Isolation and cannulation of microvessels
[27][28] One retinal arteriole was isolated in each porcine eye and cannulated, and the pressure in the lumen was increased to 55 cmH 2 O without flow. 29After stable basal tone developed without any preconstrictors, the changes in the vascular diameter in response to pharmacologic agents were recorded using videomicroscopy techniques. 25hashi et al.--7
In previous studies using our methods regarding the intravitreal concentration of thrombin in ocular disorders, [19][20][21][22] we assessed the response of the retinal arterioles to extraluminally administered thrombin at concentrations ranging from 0.001 to 20 mU/ml within 15 minutes.In the first series of studies, we compared the thrombin-induced vasomotor activity between intraluminal and extraluminal administration.We performed intraluminal administration as described previously. 28In the second series of studies, to determine the participation of thrombin receptor subtypes, the vessels were pretreated with PAR-1, -2, and -4, and the specific thrombin antagonists RWJ 56110 (10 μM PAR-1), 30 FSLLRY-NH 2 (10 μM PAR-2), 31 tcY-NH 2 (1 μM PAR-4), 32 and hirudin (50 mATU/ml), the specific Takahashi et al.--8 thrombin antagonist (twice the thrombin concentration). 9The vessels were incubated with respective blockers for 30 minutes, and to determine the response of the retinal arterioles to extraluminally administered thrombin at concentrations ranging from 0.001 to 20 mU/ml.In the third series of studies, to determine the role of the endothelium in the thrombin-induced responses of the retinal vessels, we removed the endothelium using previously reported techniques. 33Briefly, the vessels were filled with nonionic detergent, CHAPS (0.4%) intraluminally, for 1 to 2 minutes to destroy the endothelial cells.We then assessed the response of the retinal arterioles to thrombin at concentrations ranging from 0.001 to 20 mU/ml.The specific inhibitors of NO synthase and hydrogen peroxide (10 μM NG-nitro-L-arginine methyl ester [L-NAME]), 25,26,34 cyclooxygenase and hydrogen peroxide (10 μM indomethacin), 26,35,36 and Ca2+-activated K+ (K Ca ) channels and endothelium-derived hyperpolarizing factor (EDHF) (combination of 0.1 μM apamin and 0.1 μM charybdotoxin [ChTX] 0.1 μM) [37][38][39][40] were used to assess the contributions of NO, PGI 2 , and K Ca .In the final series of studies, to assess the contributions of extracellular Ca 2+ , protein kinase C (PKC), and Rho-associated protein kinase (ROCK), factors important for vascular smooth muscle contractions, to thrombin-induced vasoactivity, Ca 2+ -free PSS, the Takahashi et al.--9 broad-spectrum PKC inhibitor Gö-6983 (3 μM), 41 or the non-selective ROCK inhibitor H1152 (10 μM) 41 was administered.Gö-6983 was administered using the same methods as in the second study.In the study of extracellular Ca 2+ and ROCK, after Ca 2+ -free PSS or H1152 was administered, the resting diameter changes over 20 minutes were recorded.
After establishing a new stable tone with Ca 2+ -free PSS or H1152, these vessels were exposed to thrombin (5 mU/ml) and the changes in the vascular diameter were monitored for another 15 minutes.Because Ca 2+ -free PSS and H1152 reduce vascular tone, in another set of experiments, for comparison, the vessels were incubated with the endothelium-independent vasodilator sodium nitroprusside (SNP) (30 µM) for 20 minutes before administration of thrombin.Arterioles with tone were exposed first to thrombin (5 mU/ml) for 15 minutes to establish stable vasoconstriction and then incubated with Ca 2+ -free PSS or H1152 (10 µM) for another 20 minutes.

Chemicals
Thrombin was obtained from Roche (Mannheim, Germany).RWJ 56110, FSLLRY-NH 2 , tcY-NH 2 , and H1152 were obtained from Tocris Bioscience (Bristol, UK).Other drugs Takahashi et al.--10 were obtained from Sigma-Aldrich (St. Louis, MO).Gö-6983 was in dimethylsulfoxide (DMSO).Other drugs were dissolved in PSS.The final concentrations of DMSO were less than 0.1% in the vessel bath. 26As a pilot study, vehicle control did not affect the diameter of the retinal arterioles.

Analysis
At the end of the functional experiment, the vessel was relaxed with EDTA (1 mM) Ca 2+ -free PSS to obtain the maximal diameter at 55 cmH 2 O. 25 The diametric changes observed during responses to thrombin were normalized to the resting vessel diameter after development of basal tone and are expressed as the percentage changes in diameter.The data are expressed as the mean ± standard error of the mean; n represents the number of vessels studied.Distribution was analyzed by the Kolmogorov-Smirnov test.The Student's t-test was used for the statistical comparisons of changes in the resting tone caused by the inhibitors.One-way analysis of variance followed by Turkey's test was applied to determine the significant differences in the controls (p< 0.05).

Responses of retinal arterioles to thrombin
All vessels (n = 138) had similar levels of basal tone, resting diameters, and maximal diameters, i.e., constricted to 61.4 ± 5.9% of the maximal diameter, 59.0 ± 8.1 micrometers, and 96.2 ± 10.3 micrometers, respectively, under the same conditions.Thrombin induced the maximal responses of the retinal arterioles within 2 to 4 minutes and a concentration-dependent vascular response, i.e., initial vasoconstriction at a low concentration (≦ 5 mU/ml) and then vasorelaxation at a high concentration (> 5 mU/ml) (Figure 1).The thrombin-induced response was elicited only once on one vessel, and a higher concentration (> 30 mU/ml) of thrombin did not result in any responses on retinal arterioles (data not shown).

Comparison of responses between intraluminal and extraluminal administration of thrombin
Extraluminal administration of thrombin (5 mU/ml) caused vasoconstriction, but intraluminal administration of thrombin (5 mU/ml) did not cause any changes in the vessel Takahashi et al.--12 diameter.After removing the endothelium, intraluminal administration of thrombin (5 mU/ml) resulted in similar vasoconstriction with extraluminal administration of thrombin (5 mU/ml) (Figure 2).

Contribution of PAR subtype to thrombin-induced vasoconstriction
Pretreatment with the PAR-1 inhibitor RWJ56110 and the specific thrombin antagonist hirudin significantly (p < 0.01) suppressed the constriction resulting from thrombin administration; the highest rate of vasoconstriction caused by thrombin (5 mU/ml) remained at 30% (Figure 3).However, the PAR-2 inhibitor FSLLRY-NH 2 and the PAR-4 inhibitor tcY-NH 2 did not affect the response to thrombin.

Role of the endothelium
Endothelial denudation and pretreatment with L-NAME significantly (p < 0.05) suppressed vasorelaxation caused by a high concentration of thrombin, but pretreatment with indomethacin and a combination of apamin and ChTX did not suppress the vasorelaxation (Figure 4A, B).To evaluate the thrombin-induced vasorelaxation, the diameter at the concentration of 20 mU/ml was normalized to the diameter at the concentration at 5 mU/ml Takahashi et al.--13 with basal tone and maximal vasorelaxation and expressed as the percentage of the maximal relaxation (Figure 4B).No procedures suppressed the thrombin-induced vasoconstriction (Figure 4A).

Role of Rho kinase
Thrombin-induced vasoconstriction after H1152 and Ca 2+ -free PSS were administered was suppressed significantly (p < 0.01) compared with thrombin-induced vasoconstriction after SNP was administered (Figure 6A).Thrombin (5 mU/ml) caused similar vasoconstriction in all groups of vessels (Figure 6B).The retinal arterioles constricted by thrombin (5 mU/ml) relaxed with 15-minute exposure to Ca 2+ -free PSS and ROCK inhibitor H1152.

Takahashi et al.--14
Many studies of the effect of thrombin on vascular reactivity have reported that thrombin induced vasodilation in the coronary [1][2][3][4] and pulmonary arteries. 5In contrast, other studies have reported that thrombin induced vasoconstriction in the pulmonary 6 and cerebral arteries 7,8 or caused a biphasic response, i.e., both vasorelaxation at a low concentration and vasoconstriction at a high concentration in the renal interlobar artery. 9These studies have indicated that thrombin can cause both vasoconstriction and vasorelaxation depending on the vascular bed.In bovine retinal arterioles (~200 µm) after preconstriction with prostaglandin F 2α , thrombin (40 mU/ml) caused vasoconstriction. 16The current study showed that thrombin caused a dose-dependent biphasic response, i.e., after vasoconstriction of the retinal arterioles at a low concentration (≦ 5 mU/ml) and vasorelaxation at a high concentration (> 5 mU/ml) (Figure 1).The current study also showed that intraluminal administration of thrombin does not cause vasoconstriction (Figure 2).These contradictory results might be caused by differences in species, vessel size, use of a preconstrictor, and methods of administration.A previous study 16 indicated that thrombin causes vasoconstriction via its catalytic activity, but the current study found that not only its catalytic activity but also PAR-1 were involved with vasoconstriction Takahashi et al.--15 (Figure 3).Intraluminal administration of thrombin did not cause vasoconstriction at 5 mU/ml (Figure 2).In the endothelium-denuded retinal arterioles, intraluminal administration of thrombin (5 mU/ml) showed vasoconstriction similar to the extraluminal administration of thrombin (5 mU/ml).Extraluminal administration of thrombin could cause the vascular responses via both the smooth muscles and endothelium.The permeability of the inner BRB to molecules is below 1 kDa. 24Because the MW of thrombin is 33.6 kDa, intraluminal administration of thrombin would not affect the smooth muscles.In addition, the normal blood concentration of thrombin is 5.5 ± 1.1 mU/ml in healthy subjects, 42 and the current study showed that intraluminal administration of thrombin (5 mU/ml) did not cause vasoconstriction on normal retinal arterioles (Figure 2).These results suggested that thrombin does not cause vasoconstriction on retinal arterioles of healthy subjects.However, thrombin can cause a vascular response on the outer side of the retinal arterioles.0][21][22] Regarding the Takahashi et al.--16 current study (Figure 1), thrombin might cause 50% of constriction in central retinal vein occlusion and 30% of constriction in other ocular disorders on the retinal arterioles, and the constriction might be related to circulatory disorders in those diseases.
Moreover, Ku and Zaleski detected differences in the thresholds at which thrombin receptor agonist peptide induced vasorelaxation and vasoconstriction, i.e., 0.03 and 0.3 µM on canine coronary arteries, respectively. 56Due to the different thresholds, thrombin might cause a biphasic response in the retinal arterioles as in the coronary artery.
8][59] In addition, hydrogen peroxide also causes endothelium-dependent vasorelaxation. 34In the current study, because thrombin-induced vasorelaxation (20 mU/ml) was abolished after denudation, the vasorelaxation depended on endothelium-derived relaxing factor.Moreover, blockade of NO and hydrogen peroxide by L-NAME also abolished the thrombin-induced vasorelaxation (20 mU/ml) but not indomethacin and the combination of apamin and ChTX (Figure 4A, B).Thrombin-induced vasorelaxation also was suppressed by L-NAME. 60Both L-NAME 34 and indomethacin 36 block hydrogen peroxide and only L-NAME abolished thrombin-induced vasorelaxation.
Overall, these results indicated that NO but not PGI 2 , EDHF, and hydrogen peroxide are Endothelin-1 (ET-1), a potent vasoconstrictor of the retinal arterioles, is produced by thrombin in the endothelial cells. 61,62In addition, long-term treatment (~24 hours) with thrombin up-regulates endothelin-converting enzyme-1, which is processed to bioactive ET-1, 63 in human umbilical vein endothelial cells. 64In the current study, we found that endothelial denudation did not affect the thrombin-induced vasoconstriction (Figure 4A).
ET-1 also caused constriction of the retinal arterioles in 20 minutes, 41,65 whereas thrombin-induced vasoconstriction occurred within 2 to 4 minutes in the current study.
Overall, thrombin might elicit vasoconstriction via PAR-1 in an endothelium-independent manner in the smooth muscles in the retinal arterioles without any contribution of ET-1 to vasoconstriction caused by thrombin.PKC activation might be involved in thrombin-induced vasoconstriction. 66In the current study, the PKC inhibitor Gö-6983 significantly suppressed the thrombin-induced vasoconstriction, whereas the combination pretreatment with both Gö-6983 and RWJ56110 did not intensify the suppression of the thrombin-induced vasoconstriction (Figure 5), which agreed with the previous finding. 668][69] To determine the association between thrombin-induced vasoconstriction and ROCK, thrombin (5 mU/ml) was administered to the retinal arterioles after extraluminal administration of H1152, SNP, or Ca 2+ -free PSS.Thrombin-induced vasoconstriction after H1152 and Ca 2+ -free PSS was suppressed significantly compared with thrombin-induced vasoconstriction after SNP (Figure 6A).In addition, pretreatment with H1152 or Ca 2+ -free PSS reversed the thrombin-induced vasoconstriction (Figure 6B).Our data indicated that ROCK and extracellular Ca 2+ were involved in thrombin-induced vasoconstriction.Although our data did not show an association between PKC and ROCK in the thrombin-induced responses, a previous study reported that PKC and ROCK interact with each other. 70 agents abolished the thrombin-induced vasoconstriction in the current study.We also performed additional experiments after pretreatment with higher concentrations of RWJ56110 (20 µM) and hirudin (100 mATU/ml) to exclude the possibility that the concentrations of RWJ56110 and hirudin were not sufficiently high to abolish the Takahashi et al.--20 thrombin-induced vasoconstriction.We did not find any difference in the vascular responses at higher concentrations.We also performed an additional study to investigate the possibility that thrombin causes responses to retinal arterioles before hirudin suppresses the effect of thrombin.In that study, pretreatment with hirudin in the microtubes containing thrombin did not cause any difference in the thrombin-induced vascular response.
Unsuppressed vasoconstriction might be caused by catalytic activity as reported previously. 16e current study had two limitations.First, extraluminal administration of thrombin causes vasorelaxation after vasoconstriction in a dose-dependent manner.If the thrombin-induced vasoconstriction is suppressed with inhibitors, i.e., RWJ56110 and hirudin, it is difficult to evaluate precisely the mechanism of vasorelaxation because thrombin caused vasorelaxation at a concentration higher than 5 mU/ml and the vessel diameter in the groups of thrombin with inhibitors differed from the control at 5 mU/ml.Second, our method evaluated the vessel responses to thrombin only based on changes in the vessel diameter.Thus, the current study can describe the responses of the retinal arterioles to thrombin as a phenomenon, but the mechanical evaluation is difficult.

Takahashi et al.--21
To the best of our knowledge, this is the first study to report that thrombin causes constriction of the retinal arterioles via PAR-1 and to report the different responses between intraluminal and extraluminal administration of thrombin on the retinal arterioles.We also found that a low concentration (≦5 mU/ml) of thrombin caused vasoconstriction via PKC, and ROCK on smooth muscle, and a high concentration (>5 mU/ml) of thrombin caused vasorelaxation via endothelial NO synthesis.This novel finding of thrombin-induced responses and an understanding of the complicated mechanism of thrombin-induced vasoactivity might have important clinical implications in the pathophysiology of ocular diseases.

Declaration of interest
The authors report no conflicts of interest.The authors alone are responsible for the content and writing of the paper.Supplemental material 1.To determine if hirudin is a specific blocker of thrombin, using the classic vasoconstrictor endothelin-1.As a result, a higher concentration of hirudin (100 mATU/ml) did not suppress the endothlein-1-induced vasoconstriction.

Legends
Takahashi et al.--18 produced in the endothelium, resulting in vasorelaxation of the retinal arterioles.

Figure 1 .
Figure 1.Response of isolated retinal arterioles to extraluminally administered thrombin

Figure 3 .
Figure 3. Roles of PARs in the vascular response to thrombin.Effects of pretreatment with

Figure 4 .
Figure 4. A) Effects of removing endothelium with 0.4% CHAPS and pretreatment with

Figure 5 .
Figure 5.The effect of pretreatment with the PKC inhibitor Gö-6983 for 30 minutes.After