5-hydroxytryptamine (5-HT) reduces total peripheral resistance during chronic infusion: direct arterial mesenteric relaxation is not involved

Serotonin (5-hydroxytryptamine; 5-HT) delivered over 1 week results in a sustained fall in blood pressure in the sham and deoxycorticosterone acetate (DOCA)-salt rat. We hypothesized 5-HT lowers blood pressure through direct receptor-mediated vascular relaxation. In vivo, 5-HT reduced mean arterial pressure (MAP), increased heart rate, stroke volume, cardiac index, and reduced total peripheral resistance during a 1 week infusion of 5-HT (25 µg/kg/min) in the normotensive Sprague Dawley rat. The mesenteric vasculature was chosen as an ideal candidate for the site of 5-HT receptor mediated vascular relaxation given the high percentage of cardiac output the site receives. Real-time RT-PCR demonstrated that mRNA transcripts for the 5-HT2B, 5-HT1B, and 5-HT7 receptors are present in sham and DOCA-salt superior mesenteric arteries. Immunohistochemistry and Western blot validated the presence of the 5-HT2B, 5- HT1B and 5-HT7 receptor protein in sham and DOCA-salt superior mesenteric artery. Isometric contractile force was measured in endothelium-intact superior mesenteric artery and mesenteric resistance arteries in which the contractile 5- HT2A receptor was antagonized. Maximum concentrations of BW-723C86 (5- HT2B agonist), CP 93129 (5-HT1B agonist) or LP-44 (5-HT7 agonist) did not relax the superior mesenteric artery from DOCA-salt rats vs. vehicle. Additionally, 5-HT (10–9 M to 10–5 M) did not cause relaxation in either contracted mesenteric resistance arteries or superior mesenteric arteries from normotensive Sprague- Dawley rats. Thus, although 5-HT receptors known to mediate vascular relaxation are present in the superior mesenteric artery, they are not functional, and are therefore not likely involved in a 5-HT-induced fall in total peripheral resistance and MAP.

Background 5-HT is a vasoactive amine synthesized in the enterochromaffin cells of the gastrointestinal tract and the raphe nucleus of the central nervous system [1,2]. The physiological actions of 5-HT are mediated by 7 major receptor subtypes (5-HT 1 -5-HT 7 ) [3]. Initially, 5-HT was identified as a potent constrictor of isolated blood vessels in the rat [4]. The primary contractile receptor in the peripheral vasculature of the rat is the 5-HT 2A receptor [5]. Additionally, activation of 5-HT receptors contracts human blood vessels, including: coronary arteries [6], saphenous vein, internal mammary artery [7], cutaneous hand vein [8], and pulmonary artery [9].
The constrictor effects of 5-HT initially led investigators to hypothesize that elevated levels of free plasma 5-HT, present in human and experimental models of hypertension [10,11], were acting to constrict the peripheral vasculature, resulting in elevated blood pressure. However, Diaz et al. demonstrated that chronic 5-HT infusion produced a sustained fall in blood pressure in the deoxycorticosterone acetate (DOCA)-salt and sham rat [12]. This suggests elevated levels of free plasma 5-HT may instead lower blood pressure. The mechanism by which chronic 5-HT infusion lowers blood pressure is not yet known. However, Diaz et al. further demonstrated that inhibition of nitric oxide synthase (NOS) prevented the chronic 5-HT-induced fall in blood pressure in the sham and DOCA-salt rat. This suggested an important role for NOS, and potentially nitric oxide (NO), in enabling a 5-HT-induced fall in blood pressure.
Given that a 5-HT-induced fall in blood pressure may be mediated by either a reduction in total peripheral resistance (TPR), or a reduction in cardiac output (CO), we initially sought to establish that 5-HT lowers blood pressure through a reduction in TPR. We targeted the mesenteric vasculature, which receives a large portion of cardiac output (~20%) as a potential site at which 5-HT may be acting to reduce TPR [13].
In addition to its well known function as a vasoconstrictor, 5-HT can also cause dilation of the peripheral vasculature. Thus, one mechanism by which 5-HT may reduce blood pressure is through direct receptor mediated vascular relaxation. Specifically, the 5-HT 2B , 5-HT 1B , and 5-HT 7 receptors have been linked to dilation of peripheral arteries [14][15][16]. Therefore, we tested the hypothesis that 5-HT lowers blood pressure through direct receptor-mediated vascular relaxation.
We studied the 5-HT 2B , 5-HT 1B , and 5-HT 7 receptors at the level of mRNA (real time PCR), protein (Western blot and immunohistochemical detection), and function (isolated contractile studies) in isolated mesenteric arteries to determine whether selective activation of these receptors was capable of direct receptor mediated vascular relaxation. We then investigated whether 5-HT was capable of relaxing a true mesenteric resistance artery. Further, we conducted experiments in both sham and DOCA-salt rats to assess potential mechanistic differences in how 5-HT lowers blood pressure.
This study provides the first evidence that chronic infusion of 5-HT reduces TPR, and that 5-HT does not act directly on the mesenteric vasculature to stimulate direct receptor mediated vascular relaxation, enabling a 5-HT-induced fall in blood pressure.

Effects of 5-HT on CO and TPR
Whole animals were used to determine whether TPR (rather than cardiac output) was reduced to facilitate a 5-HT-induced fall in blood pressure. There were no between-group difference in blood pressure prior to 5-HT or vehicle infusion (C1-C2). Continuous 5-HT infusion resulted in a significant fall in MAP on the first 2 days of infusion (designated S1 and S2) ( Figure 1A). Starting on day S3, MAP tended to normalize yet remained below baseline levels. 5-HT produced a significant increase in HR, cardiac index (CI) and stroke Figure 1 Effect of chronic 5-HT or vehicle infusion on A) mean arterial pressure (MAP), B) heart rate (HR), C) cardiac index (CI), D) stroke volume, and E) total peripheral resistance (TPRI). Data points are 1-hr group means +/− SEM. Number of animals per group is shown in parentheses. The vertical line denotes 5-HT or vehicle osmotic pump implant. C1 and C2 -control recordings. S1 -S7 days since initiation of 5-HT/vehicle infusion. * P < 0.05 vs. vehicle. volume (SV) compared to vehicle infusion. HR returned to pre-administration levels after 1 day of infusion, while CI and SV remained significantly elevated throughout the experiment ( Figure 1). Rats receiving 5-HT demonstrated a significant reduction in total peripheral resistance (TPRI; 0.31 ± 0.03 mmHg/ml/min/kg; S1) compared sto vehicletreated animals (0.51 ± 0.01 mmHg/ml/min/kg). This effect persisted throughout the remainder of the study.
Real time RT-PCR 5-HT receptor transcripts were quantified using real time RT-PCR (2 -delta Ct x1000) in the SMA of sham and DOCA-salt rats using beta-2 microglobulin as a calibrator ( Figure 2). The 5-HT 2B receptor mRNA was expressed at high levels in both the SMA of sham and DOCA-salt rats (4.46 ± 0.44 vs. 10.41 ± 2.32 respectively). 5-HT 2B receptor mRNA was expressed at significantly greater levels in the SMA of the DOCA-salt rat compared to the sham rat (P < 0.05). mRNA transcripts for the 5-HT 1B (DOCA-salt 0.83 ± 0.24; Sham 0.81 ± 0.11), and 5-HT 7 receptor (DOCA-salt 0.78 ± Sham 0.62; 0.21 ± 0.04) were expressed at significantly lower levels compared to the 5-HT 2B receptor (P < 0.05) but were not different between sham and DOCA rats. The 5-HT 2B , 5-HT 1B , and 5-HT 7 receptors were considered the major products and followed by Western blot, immunohistochemical detection, and contractile studies.
Immunohistochemical detection of the 5-HT 2B , 5-HT 7 , and 5-HT 1B receptor(s) Immunohistochemical detection was conducted to determine the location of 5-HT receptors within the SMA of sham and DOCA-salt rats. We observed positive staining as black/brown precipitate. 5-HT 2B receptor staining was observed in the endothelial, smooth muscle, and adventitial layers of SMA sections from sham and DOCA-salt rats ( Figure 3B). Similar staining, corresponding to the 5-HT 1B receptor was observed in the media and adventitial layers ( Figure 5B). Negligible 5-HT 7 staining was observed in the adventitia. ( Figure 4B). Staining was observed in the rat brain (positive control) with the 5-HT 2B and 5-HT 1B receptor antibodies, but not the 5-HT 7 receptor antibody.

5-HT does not relax mesenteric resistance vessels or the superior mesenteric artery in normotensive rats
Because the above studies were negative, we next investigated whether 5-HT was capable of relaxing a true mesenteric resistance artery. Endothelial-intact mesenteric resistance vessels (200 micron diameter) were mounted in a wire myograph chamber for measurement of isometric contractile force. 5-HT (10 -9 M -10 -5 M) contracted the isolated mesenteric resistance vessel from baseline and when contracted with prostaglandin F 2 alpha (PGF2α) and incubated with ketanserin ( Figure 6A). 5-HT (10 -9 M -10 -5 M) also contracted the isolated SMA when contracted with U46619 and incubated with ketanserin ( Figure 6B). These same resistance vessels relaxed over 50% to ACh (10 -6 M) and forskolin (1 μM). Relaxation to 5-HT was not observed under any condition. These findings suggest that 5-HT alone does not relax the superior mesenteric resistance vessels.

Discussion
The main objective(s) of this study were to: 1) determine whether 5-HT lowers blood pressure through a reduction in total peripheral resistance (TPRI) or cardiac output (CO), 2) identify the primary 5-HT receptors expressed in the superior mesenteric artery (SMA) of sham and DOCA-salt rats, and 3) determine whether activation of these receptors relaxes SMA of sham or DOCA-salt rats. 5-HT reduces total peripheral resistance in sham and DOCA-salt rats 5-HT infusion produced a significant reduction in MAP and TPRI, supporting our original hypothesis that 5-HT lowers blood pressure through direct receptor mediated vascular relaxation. These data confirm and extend our previous findings that chronic elevations in free circulating 5-HT reduce MAP, and increase HR over the course of a 1 week 5-HT infusion [12]. TPR is reduced throughout the duration of the infusion, despite a recovery in blood pressure, a novel finding. The recovery of blood pressure may be due to the persistent elevation in SV and CI. HR is only transiently increased and therefore raises the question of whether 5-HT is also acting directly on the heart to increase SV. 8-OH-DPAT (5-HT 1A and 5-HT7 receptor agonist) produces a sustained increase in cardiac output (CO) following hemorrhagic shock [17]. Thus, activation of the 5-HT 1A receptor may explain why blood pressure recovers despite a persistent reduction in TPR. Ultimately, by defining these important hemodynamic parameters (HR, SV, CI and TPR), we were able to pursue a careful investigation of the vasculature as a potential site of action for a 5-HT-induced fall in blood pressure. The mesenteric vasculature was selected as the potential site of action for 5-HT due to the large portion of cardiac output it receives (~20%) [13]. We hypothesized that 5-HT may be acting on mesenteric vasculature to reduce TPR and lower blood pressure.

5-HT does not directly relax the mesenteric arterial vasculature
The 5-HT 2B receptor was a leading candidate in our investigation due to its location on the endothelial cell, reports of its ability to stimulate NO release [18], and the absolute dependence of the 5-HT-induced depressor response on nitric oxide synthase [12]. Bordoff et al. demonstrated that treatment with either a 5-HT 2B receptor antagonist or an inhibitor of NOS abolished a bypassinduced fall in blood pressure in the male normotensive Wistar rat [14]. Bordoff et al.  from mechanically disrupted platelets, contributed to falling blood pressure in patients placed on cardiopulmonary bypass. This suggested 5-HT may activate NOS, and potentially release NO. The 5-HT 1B and 5-HT 7 receptors were also considered leading candidates in our investigation. Calama et al.. demonstrated that intra-arterial administration of L-694, 247 (5-HT 1B/1D receptor agonist) produced vasodilation in the hindquarter of the anesthetized rat [15]. Further, activation of the 5-HT 7 receptor is associated with vascular smooth muscle relaxation, and a long-lasting depressor response in the anesthetized rat [16,19,20]. These studies suggest 5-HT may be acting at either the 5-HT 1B , or 5-HT 7 receptor(s) to lower blood pressure.
RT-PCR, Western blot, and IHC analysis identified expression of the 5-HT 2B , 5-HT 7 , and 5-HT 1B receptors in the SMA of sham and DOCA-salt rats. The 5-HT 7 receptor antibody recognized protein in SMA homogenates from sham and DOCA-salt rats, but the same antibody did not elicit a strong immunohistochemical stain in sections of SMA from either sham or DOCAsalt rats. Furthermore, we also failed to demonstrate adequate staining in brain sections using multiple other antibodies against the 5-HT 7 receptor. Nonetheless, RT-PCR and Western blot data strongly support the existence of the 5-HT 7 receptor in the SMA of sham and DOCA-salt rats.
Despite compelling evidence that 5-HT 2B , 5-HT 7 , and 5-HT 1B receptors are expressed in the mesenteric vascular bed, no significant relaxation to their respective 5-HT receptor agonists was observed in either the SMA of the sham or DOCA-salt rat. Instead, activation of the 5-HT 2B receptor (via BW723C86) and 5-HT 1B (via CP93129) caused contraction of the SMA of the DOCA-salt rat. These results are consistent with previous studies in the DOCA-salt aorta, which demonstrated that activation of the 5-HT 2B and 5-HT 1B receptor(s) in the DOCA-salt aorta elicits a contractile response [21,22]. Interestingly, a significant body of evidence suggests opposing roles for the 5-HT 2B receptor depending on its location within the vasculature (endothelial cell vs. smooth muscle), and/or underlying pathology (presence or absence of hypertension) [21,22]. For example, the 5-HT 2B receptor is up-regulated, and contractile in the aorta of the DOCA-salt rat vs. sham rat [22]. This evidence may explain why BW723C86 contracted the SMA of DOCAsalt rat, but did not contract the SMA of the sham rat. Finally, 5-HT itself did not relax either a true mesenteric resistance vessel or the SMA in the presence of ketanserin (a 5-HT 2A receptor antagonist). We hypothesized that the presence of ketanserin might reveal or unmask a relaxant response in the mesentery. Thus, we must conclude that the 5-HT 2B , 5-HT 7 , and 5-HT 1B are not involved in relaxing the SMA and 5-HT is not likely acting directly in the mesenteric vasculature to mediate a 5-HT-induced fall in TPR and blood pressure.

Potential mechanism(s) underlying a 5-HT-induced fall in blood pressure
Several possibilities may explain a 5-HT-induced fall in blood pressure. First, 5-HT may be acting in a different vascular bed, rather then the mesentery, to release nitric oxide (NO), reduce TPR, and lower blood pressure. Direct receptor-dependent vascular relaxation has been observed in the rat jugular vein, the porcine pulmonary artery, the canine coronary artery, and the human umbilical artery [23][24][25][26][27]. Further, previous studies also suggest an important role for the serotonin transporter (SERT) in mediating a 5-HT-induced fall in blood pressure [28]. Thus, we cannot rule out the possibility that 5-HT may be taken up via SERT at an alternate site (vascular), leading to nitric oxide synthase (NOS) activation and NO release.
The skeletal muscle vasculature, and the cutaneous vasculature are alternative sites at which 5-HT may be acting to reduce TPR, and lower blood pressure [28,29]. For example, fluoxetine (5-HT reuptake inhibitor) dilates isolated skeletal muscle arterioles [30]. Additionally, Alsip et al. demonstrated 5-HT is capable of dilating the rat skeletal muscle arteriole [31]. The skeletal muscle vasculature receives a large portion of cardiac output (~40%), such that changes in resistance could account for large changes in blood pressure [13]. The cutaneous vasculature is also a potential site at which 5-HT may be acting to reduce TPR, and lower blood pressure, but this is more controversial. 5-HT causes a significant increase in tail temperature in the rat, suggestive of heat loss through cutaneous vasodilation, and a fall in blood pressure [32,33]. Similarly, the 5-HT releaser fenfluramine is associated with cutaneous dilation in the conscious rat [34]. Low doses of intra-arterially infused 5-HT significantly increased human forearm blood flow in healthy volunteers [35]. Collectively, these data suggest 5-HT may relax the cutaneous vasculature, reduce peripheral resistance, and lower blood pressure. However, the association between changes in cutaneous blood flow and systemic change(s) in TPR needs to be studied further.
Second, 5-HT may reduce vascular tone by inhibiting sympathetic nerve activity (SNA) rather than directly causing relaxation. 5-HT 1B/1D receptors expressed on the presynaptic sympathetic nerve terminals inhibit norepinephrine release, an action that would decrease vascular tone [36]. 5-HT may also act centrally to inhibit SNA. Recent studies demonstrated 5-HT is capable of crossing the blood brain barrier via the serotonin transporter (SERT) [37]. At present, we cannot rule out the possibility that 5-HT acts centrally to lower blood pressure.

Conclusions
We recognize several limitations of the present study. First, it is possible that alternate 5-HT receptor(s) (identified by PCR) are expressed and functional in the SMA, causing direct receptor mediated vascular relaxation. However, none of these receptor(s) have been as strongly linked to vascular relaxation as those chosen for the current study. Finally, we also acknowledge that only one 5-HT receptor agonist was used to investigate each 5-HT receptor (5-HT 2B , 5-HT 1B , 5-HT 7 ) of interest. The fact that 5-HT itself did not relax the "unmasked", antagonized vessel, supports that the mechanism of direct relaxation does not exist in the mesenteric vasculature. In summary, this study provides the first evidence that 5-HT reduces blood pressure through a reduction in total peripheral resistance during a chronic infusion. However, the reduction in TPRI is not likely due to activation of the 5-HT 2B , 5-HT 7 , and 5-HT 1B receptors in the SMA of sham or DOCA-salt rat.

Animals
The Michigan State University and Loyola University Chicago, Stritch School of Medicine Institutional Animal Care and Use Committees (IACUC) approved all protocols. Male Sprague Dawley rats (225-250 g, Charles River Laboratories) were used. Hemodynamic studies [measurement(s) of mean arterial pressure (MAP), heart rate (HR), cardiac output (CO), stroke volume (SV), and calculations of cardiac index (CI) and total peripheral resistance] were completed in conscious freely moving Sprague Dawley rats. Isolated tissue experiments (RT-PCR, Western blot, immunohistochemistry, and isolated contractile experiments) were performed in the SMA from sham and DOCA-salt rats.

DOCA-salt hypertension
Sprague-Dawley rats underwent left uninephrectomy, and a DOCA pellet was implanted subcutaneously (200 mg/kg). Sham rats underwent left uninephrectomy, but were not implanted with DOCA pellet. Rats were given standard rat chow ad libitum. DOCA-salt water was supplemented with 1% NaCl and 0.2% KCl for the duration of the study. Sham rats received tap water for the duration of the study. Hypertension was established after 4 weeks of initiation of DOCA. Sham systolic blood pressure was < 140 mmHg, while DOCA-salt systolic blood pressure was > 140 mmHg measured by tail cuff method.

Anesthesia and analgesia
All rats were anesthetized with isoflurane (2% in 100% O 2 ) and ventilated mechanically. At the time of surgery, the incision site was treated with bupivicaine. Incisions were closed with silk suture. Rats were treated with amoxicillin (150 mg/kg/i.m.) following surgery and 3 days thereafter. All rats were treated with rimadyl (5 mg/kg) and buprenorphine (35 μg/kg) for 2 days for general analgesia.

Surgical methods: alzet osmotic pump
A small incision was made at the base of the neck. Blunt dissection was used to create a small subcutaneous pocket between the scapulae. Alzet Osmotic Pumps (Model 2ML1, Duret Corporation, Cupertino, CA, 10 μL/hr 7 days) were loaded with 5-HT creatinine complex (25 μg/kg/min) dissolved in 1 M HCl, (normalized to pH 7 by 4 M NaOH) and inserted under the skin.
Surgical methods: ascending aortic flow probe Rats were given an injection of antisialogogue, glycopyrrolate (4 mg/kg, s.c.), then anesthetized with isoflurane, intubated and mechanically ventilated. An ultrasonic transit-time flow probe (model 3SB; Transonic Systems Inc., Ithaca, NY) was placed around the ascending aorta through an incision in the third intercostal space. The thoracic incision was closed in layers, and the lungs were reinflated with negative pressure. The probe cable was tunneled s.c. and externalized at the nape of the neck and secured with a silastic cuff.

Surgical methods: blood pressure probe
During the same surgery, radiotelemetery probes (C50-PXT, Data Science International, St. Paul, MN, USA) were implanted to enable chronic blood pressure recording. The left femoral artery was exposed through a groin incision and ligated distally. The tip of the probe catheter was advanced through a hole made in the artery and advanced into the abdominal aorta. The body of the probe was placed subcutaneously on the left flank. All rats recovered for 10 days prior to the start of data collection. Mean arterial pressure (MAP), heart rate (HR), and ascending aortic blood flow (i.e. CO) were recorded daily, between 10:00 am and 2:00 pm while the rats rested unrestrained in their home cage. Recordings began at least 60 min for connection to the blood flow meter, after which MAP and HR were recorded continuously for 1 hr. Total peripheral resistance (mmHg/ml/ min/kg) and cardiac Index (CI) were normalized to body mass. Where CI = CO / (9.1 x g 0.66 ).
Where g = body mass in g, (9.1 x g 0.66 ) = body surface area in cm 2 , and TPRI = CI/MAP.

Tissue preparation
Rats were anesthetized with pentobarbital (60 mg/kg i.p.) and the superior mesenteric artery or mesenteric resistance artery was removed and placed in physiological salt solution PCR conditions were: 95°C for 10 minutes followed by 40 cycles of (95°C, 15 sec; 60°C, 60 sec). A standard dissociation curve was run following the above cycle conditions. Each sample was run in duplicate. No template controls (NTC) were run for each primer set as well as no Reverse Transcriptase (no RT) samples to verify that the primers themselves did not amplify and that no genomic contamination had occurred with RNA isolation, respectively.

Immunohistochemistry
Slides containing sections of paraffin-embedded rat superior mesenteric artery were dewaxed, unmasked using Unmasking Solution (Vector Laboratories, Burlingame, CA) and taken through a standard protocol. Slides were incubated with 5-HT 1B , 5-HT 2B or 5-HT 7 antibody (5 μg/ ml) (same antibodies as used for Western blots) in 1.5% blocking serum. Slides were washed in phosphate buffered saline and incubated with a peroxidase-conjugated secondary antibody in 1.5% blocking serum for 30 minutes, followed by a 30 min incubation in Vectastain Elite ABC Reagent (Vector Laboratories). 3,3-diaminobenzidine/H 2 O 2 was applied until staining appeared (1-4 minutes). The slides were counterstained with hematoxylin (Vector Laboratories, Burlingame, CA).

Isolated tissue bath
Endothelium-intact SMA were cleaned and cut into helical strips for measurement of isometric contractile force. SMA strips were mounted in warmed (37°C) and aerated (95% O 2 , 5% CO 2 ) tissue baths (30 ml PSS) on Grass isometric transducers (FT03; Grass instruments, Quincy, MA, USA), connected to an ADInstruments PowerLab (ADInstruments, Colorado Springs, CO). Tissues were placed under optimal resting tension (600mg; previously determined) and allowed to equilibrate for 1 hr before an initial challenge with a maximal concentration of phenylephrine (10 -5 M). After the initial challenge, tissues were washed until tone returned to baseline. Then, a half-maximal concentration of phenylephrine (10 -7 M) was added to the bath, followed by acetylcholine (ACh; 10 -6 M) to determine the integrity of the endothelial layer. Tissues that relaxed more than >50% in ACh were considered to have intact endothelium. Tissues were washed until they returned to baseline. Strips were incubated with ketanserin (50 nM; 5-HT 2A antagonist) for 15 minutes to prevent activation of the 5-HT 2A receptor. Tissues were then contracted half-maximally to U46619 (thromboxane A 2 agonist) (~30 min incubation with ketanserin). Cumulative concentration-response curves were generated for each of the following agonists: 5-HT vs. vehicle (H 2 O), BW-723C86 (5-HT 2B agonist) vs. vehicle (H 2 O), CP 93129 (5-HT 1B agonist) vs. vehicle (H 2 O), and LP-44 (5-HT 7 agonist) vs. vehicle (DMSO). At the end of each experiment, sodium nitroprusside (SNP) (10 -6 M) was added to demonstrate that each SMA was able to relax under experimental conditions set forth.

Wire myograph
Under a stereomicroscope with a calibrating eyepiece and in cold PSS, two tungsten wires were inserted through the lumen of a cleaned mesenteric resistance artery (~200-250 micron diameter). One wire was attached to an isometric force transducer with a detection range of 0.002-10g, while the other was connected to a micrometer-attached support. Force generation was recorded on a Grass Model 7D polygraph. The dual chamber, kept at 37°C and perfused via a peristaltic pump with warmed and oxygenated PSS, allowed mounting of two parallel vessels. A passive tension of~400 mg (appropriate for generating optimal force in a resistance artery) was applied. Vessels were allowed to equilibrate for 1 hr prior to initial challenge with PE (10 -5 M). After initial washout, experiments were performed as described above for isolated tissue bath.

Statistical analysis
For in vivo data analysis, between group differences were assessed by a two-way ANOVA with repeated measures followed by post hoc testing using Bonferroni's procedure (GraphPad Prism 5). For isometric contractile studies, relaxation is reported as a percentage of initial contraction to a half-maximal concentration of U46619 (thromboxane A 2 agonist). Repeated measures two-way ANOVA followed by the Bonferroni post hoc test was used to compare concentration-response curves. An unpaired Students t-test was used to compare differences in maximal response(s) between agonist vs. vehicle. In all cases, p < 0.05 was considered significant. All results are presented as the mean ± SEM.