Heart and Circulatory Physiology

Diminished arteriolar responses in nitrate tolerance involve ROS and angiotensin II

Mary D. Frame, Randall J. Fox, Dongsoo Kim, Amy Mohan, Bradford C. Berk, Chen Yan


Our purpose was to evaluate hyporesponsivity to nitric oxide (NO)-induced dilation in small arterioles during nitrate tolerance. An Alza osmotic pump was implanted in the left flank of adult rats (n = 56) for continuous administration of nitroglycerin (140 μg/h) or vehicle (propylene glycol). On postoperative day 3, arcade (∼50-μm diameter) and terminal (∼20 μm) arterioles were observed in the cremaster preparation with in vivo video microscopy. Local vascular responses were obtained with micropipette-applied NO donors, with and without superoxide dismutase (SOD), Mn(III) tetrakis(4-benzoic acid) porphyrin chloride (MnTBAP), or losartan. On day 3, NO-mediated dilation was significantly attenuated in nitroglycerin-treated rats. Attenuation was greater in the terminal arterioles compared with the arcades. Control responses were restored by SOD, MnTBAP, or losartan, suggesting a role for elevated angiotensin II and reactive oxygen species (ROS) as mediators of the attenuated NO dilation (nitrate tolerance). Addition of losartan to the drinking water likewise prevented nitrate tolerance. In summary, terminal arterioles are affected by nitrates to a greater extent than the arcade arterioles that feed them, in a process dependent on angiotensin II and ROS.

  • microvascular responses
  • steady vs. pulsatile flow
  • superoxide

clinical use of nitroglycerin (NTG) to improve cardiac function in heart failure was proposed by Murrell in 1879 (28) and continues to be a treatment of choice to improve cardiac output through its peripheral dilatory effects, decreasing preload and afterload on the heart. Nitrate tolerance is a long-standing detrimental side effect of nitrate therapy involving vascular hyporesponsivity to continued nitrate treatment. In fact, cross-tolerance occurs, in which there is a failure of blood vessels to dilate to all nitric oxide (NO) donors; this is associated with an increase in constriction to many agents, including angiotensin II.

In the systemic circulation, the mechanism of nitrate tolerance includes early activation of the renin-angiotensin system with elevation of vascular superoxide anion concentration (6, 17, 21,27). The cellular mechanism involves upregulation of angiotensin II and NADPH oxidase, with reduced expression and activity of superoxide dismutase (SOD) (26, 27). There is evidence that the diminished dilation to NO occurs via a process after cGMP synthesis (36). Recently, we showed (20) that cyclic nucleotide phosphodiesterase (PDE) 1A1 is upregulated in nitrate tolerance, suggesting that this cGMP-hydrolyzing PDE is involved in diminishing NO/cGMP-mediated vasodilation in tolerance. Thus our understanding of the cellular mechanism is improving. However, our understanding of nitrate tolerance in the whole animal is not yet complete.

Most studies have examined nitrate tolerance with in vivo or in vitro models of the systemic circulation. Although the involvement of the peripheral circulation has been suggested and documented for many years, especially with regard to differences between vascular beds (see, e.g., Refs. 1 and 4), direct investigation of nitrate tolerance within the peripheral circulation has not been done. In our study, two classifications of small arterioles were investigated in vivo with the rat model of nitrate tolerance. In the rat cremaster muscle microcirculation, we studied the larger arcading arterioles (maximal diameter ∼55 μm) and the smaller terminal arterioles (maximal diameter ∼25 μm) directly fed by the arcades. The flow characteristics of these two groups of vessels are very different, with oscillatory flow patterns in the arcade arterioles and unidirectional flow in the terminal arteriolar feeds. Several recent studies clearly show that endothelium-dependent dilatory capability is enhanced at the transcription level in endothelial cells exposed to unidirectional shear stress (as is found in terminal arterioles) compared with oscillatory shear stress (as would be found in the arcading arterioles) (30, 33, 37). We therefore studied both groups of arterioles in the present work. We hypothesized that nitrate tolerance would diminish arteriolar responsivity to NO to a greater extent in the smaller terminal arterioles than in the larger arcading vessels. We also examined the role of reactive oxygen species (ROS) and the renin-angiotensin system on nitrate tolerance in small arterioles.


Model of nitrate tolerance.

With approval of the University of Rochester, adult male rats [Harlan Sprague-Dawley (HSD), 315 ± 27 g (mean ± SD);n = 70] were anesthetized with intraperitoneal ketamine (4 mg/100 g), xylazine (0.5 mg/100 g), and acepromazine (0.5 mg/100 g). An Alza osmotic pump (model 1003D) was implanted in the left flank and was used for continuous administration of NTG (140 μg/h; dissolved in propylene glycol) or vehicle (control; infusion of propylene glycol alone). Buprenorphine (0.5 mg/kg sc) was administered for immediate postsurgical analgesia and again 6–12 h later. Two protocols were completed. Protocol 1(n = 56) consisted of two treatments: vehicle or NTG (140 μg/h) administered via osmotic pump. Protocol 1animals were used in acute experimentation on postsurgical day 1 or day 3. Protocol 2 (n = 14) consisted of four treatments: vehicle or NTG (140 μg/h) administered via osmotic pump either with or without losartan (30 μM) in the drinking water. Protocol 2 animals were used in acute experimentation on postsurgical day 3. Plasma nitrate was determined for animals of protocol 2 with a Sievers nitric oxide analyzer (model 280).

Acute microvascular experimentation.

With University of Rochester approval, adult male rats (HSD, 315 ± 27 g; n = 70) were anesthetized with pentobarbital sodium (50 mg/kg ip), tracheostomized, and maintained on a constant infusion of pentobarbital sodium (10 mg/ml at 0.56 ml/h ip). Systemic hematocrit (44 ± 3%) was unaffected by treatments. Body temperature was maintained (37–38°C) by conductive and convective heat sources. Mean arterial pressure was monitored in some animals via a left femoral arterial catheter; cardiac output was measured in these animals by indicator dilution of fluorescein isothiocyanate conjugated to bovine serum albumin (FITC-BSA) injected via a right jugular venous catheter (31). The right cremaster muscle was prepared for in vivo microcirculatory observations (14, 34). The preparation was continuously superfused with bicarbonate-buffered saline containing (in mM) 132 NaCl, 4.7 KCl, 2.0 CaCl2, 1.2 MgSO4, and 20 NaHCO3(equilibrated with gas containing 5% CO2 and 95% N2; pH 7.4 at 34°C). All chemicals were obtained from Sigma (St. Louis, MO) unless otherwise noted.

The microcirculation was observed with transillumination with a modified Nikon upright microscope (Nikon, Tokyo, Japan) with a ×25 (Nikon) objective. Epi-illumination was used to visualize the FITC-BSA dye for cardiac output with a Chroma B1E filter (Chroma, Brattleboro, VT). Video images were produced with a charge-coupled device 72s video camera with a Gen/Sys II video intensifier (Dage-MTI, Michigan City, IN). During a 60-min stabilization period, arteriolar tone was verified by dilation to topically applied 10−3 M adenosine (maximum dilation) and constriction to 10% O2 gas added to the superfusate. At the end of this 60-min period, cardiac output was determined. Cardiac output was measured from VHS playback of the light intensity changes due to fluorescent dye passing through a 20-μm-diameter arteriole. From the change in light intensity, and an established calibration of light intensity to dye concentration, cardiac output was calculated as previously described (31).

Two functionally distinct vessel types were studied: arcading arterioles [maximum diameter 56 ± 17 μm (mean ± SD)] and terminal arterioles (27 ± 6 μm). Flow characteristics in the arcades involve significant and frequent flow reversal (oscillatory flow), whereas terminal arterioles maintain unidirectional flow. Under a strict Wiedeman classification (35), the arcades were A3 or A4 vessels and the terminal arterioles were A4 or A5 vessels. Each observation site was the branching point of the terminal arteriole as it arose from the arcade; the arcade and terminal arteriolar responses were paired measurements obtained simultaneously.

Acute local vascular responses were tested by micropipette application of drugs to the observed site for 60 s at 0.5 psi (pneumatic injection). Vascular responses to three NO donors were tested: 3-morpholinosydnonimine (SIN-1; 10−6–10−3 M), sodium nitroprusside (SNP; 10−9–10−3 M), and (Z)-1-[2-(aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2- diolate (DETA-NO; 10−8–10−3 M); additionally, dilation to micropipette-applied adenosine (10−9–10−3 M) and to papaverine (10−9–10−3 M) was tested. Because of the literature showing that micropipette-applied papaverine is not as efficacious in dilating as papaverine in the tissue bath, two concentrations of papaverine (10−7 and 10−4M) were tested in the tissue bath at the end of the experiment. At micromolar concentrations in vascular smooth muscle, papaverine induces dilation via action as a nonselective PDE inhibitor (2, 3, 19,24). The involvement of ROS in the attenuated dilation to NO donors was tested with two superoxide scavengers (micropipette applied): SOD (100 U/ml), and Mn(III) tetrakis(4-benzoic acid) porphyrin chloride (MnTBAP, 10 μM; also a scavenger of hydrogen peroxide and peroxynitrite). The role of angiotensin type 1 receptors (AT1R) in the attenuated dilation to NO donors was tested acutely with losartan (AT1R antagonist; 100 nM) directly applied via micropipette to the vessels of rats in protocol 1 and separately tested systemically by administration of losartan (30 μM) in the drinking water of rats in protocol 2. Constriction to micropipette-applied angiotensin II (10−12–10−6 M) was tested for rats inprotocol 2. Remote dilation was tested with SNP (10−3 M) or LM-609 (10 μg/ml; agonist to αvβ3-integrin signal transduction; Ref.22). For remote dilation responses, drugs were applied via micropipette for 60 s at a location 700 ± 120 μm downstream from the observation point where data were taken, along the terminal arteriole. For all diameter responses, peak diameter change during exposure is reported; in no cases were there biphasic responses.

All diameter changes were obtained on-line with a video caliper system (Microvascular Research Institute, College Station, TX) and a software data acquisition system (Strawberry Tree; Workbench, Sunnyvale, CA) calibrated with a micrometer. The diameter change was calculated as fractional change from baseline: (peak − initial)/(maximum − initial), where the maximal diameter is with topical 10−3 M adenosine. In Figs. 2 and 3, lines are the sigmoid fit to the data; EC50 values could not be reliably estimated in most cases because of the absence of plateau responses even at high concentrations. Comparisons were made between groups, as a function of concentration, by ANOVA (repeated measures). For all statistics, differences were considered significant whenP < 0.05.


Systemic changes with nitrate tolerance.

There was no significant weight loss for any treatment in any test group protocol. The initial weight at the time of implantation of the osmotic pump was 309 ± 27 g (n = 21), and the weight at the time of acute microvascular experimentation was 313 ± 22 g (day 3 animals). Neither mean arterial pressure (163 ± 41 mmHg) nor heart rate (377 ± 48 beats/min) was different by treatment group at the time of acute experimentation (anesthetized with pentobarbital). Cardiac index (cardiac output/animal weight) was significantly elevated in day 1 NTG-treated rats from protocol 1 (Fig.1). Hence, as expected, the day 1 NTG-treated rats had a decreased total peripheral resistance (34% of control animals) and an increased stroke volume (600 μl/beat compared with 240–260 μl/beat in control animals). As expected, the beneficial effect of NTG on cardiac index was absent in day 3 animals, indicating that they were nitrate tolerant. Systemic losartan treatment (protocol 2) did not alter cardiac index, either alone or with NTG [day 3: control, 337 ± 24 (n = 4); losartan alone, 372 ± 34 (n = 3); NTG, 364 ± 58 (n = 3); losartan + NTG, 355 ± 48 ml · min−1 · kg−1(n = 3)]. To verify NTG exposure in losartan-fed animals, plasma nitrate was measured in 14 animals from protocol 1 and in all protocol 2 animals. Plasma nitrate in animals from protocol 1 was significantly elevated in NTG-treated rats compared with day 3 controls [13 ± 3 (n = 9) vs. 8 ± 2 (n = 5) μm]. Likewise, for animals from protocol 2, plasma nitrate was significantly elevated in NTG-treated rats [without losartan, 15 ± 1 μM (n = 3); with losartan, 17 ± 2 μM (n = 3)] compared with day 3 controls [without losartan, 9 ± 1 μM (n = 4); with losartan, 11 ± 2 μM (n = 3)].

Fig. 1.

Cardiac index (CI) for protocol 1 rats implanted with osmotic pumps to deliver nitroglycerin (NTG; 140 μg/h) or vehicle (control; propylene glycol). CI is significantly elevated inday 1 NTG-treated animals. n, No. of animals. * P < 0.05 vs. other groups.

Microvascular responses to three NO donors.

Each of the three NO donors tested showed a decreased dilatory response in day 3 NTG-treated animals compared with day 3controls. [For day 1 animals, there were no significant changes in dilatory responses for NTG-treated compared with control animals (day 1 data not shown).] Day 3 responses are shown in Fig. 2. The key data in support of our primary hypothesis show that the terminal arterioles from NTG-treated animals displayed a greater reduction in maximal response than the arcades. For 10−3 M SNP, dilation was diminished from 99 ± 7% to 41 ± 17% (P < 0.05) of maximal dilation for the terminal arterioles and from 90 ± 7% to 64 ± 2% for the arcades [mean ± SE; 100 × (peak − initial)/(max − initial)]. Importantly, this effect lowers the dilatory capacity of the terminal arterioles by ∼20% compared with the arcades (41% vs. 64%; P < 0.05) despite the fact that the terminal arterioles were initially ∼9% more sensitive to the NO donors (99% vs. 90%). Thus the terminal arterioles are affected by nitrate tolerance to a much greater extent than the arcades.

Fig. 2.

Dilation responses (peak − initial/max − initial) for terminal arterioles (A, C, E) and arcade arterioles (B, D, F) with 3-morpholinosydnonimine (SIN-1; A and B), sodium nitroprusside (SNP; C and D), or (Z)-1-[2-(aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2- diolate (DETA-NO; E and F). Responses are for day 3 control and NTG-treated animals. Maximal dilation is to adenosine (10−3 M) determined before stimulation with NO donors. Lines are the sigmoid fit to the data. * Different from control at the same dose (P < 0.05).

One way to explain why the terminal arterioles showed a diminished dilation to NO donors to a greater extent than the arcades would be a difference in the level of arteriolar tone. Arteriolar tone was not different for the arcades and terminal arterioles for any treatment group (defined as initial baseline diameter/maximal diameter with 10−3 M adenosine). Furthermore, initial tone was not different for terminal arterioles (0.49 ± 0.03; n= 42) compared with arcade arterioles (0.55 ± 0.03;P = 0.31). Thus neither tone nor maximal dilation to adenosine was affected by nitrate therapy—only dilation to NO donors was suppressed.

Importantly, NTG treatment suppressed the maximal dilation to each NO donor for both groups of arterioles while retaining cAMP-mediated dilation to adenosine and retaining dilation mediated by PDE inhibition with papaverine. Figure 3 shows the concentration response curve for micropipette-applied vasodilators. Tissue bath papaverine has been shown to be more efficacious in dilating microvessels than locally applied papaverine (see, e.g., Refs.2, 3, 24); therefore, we tested 5-min exposure to 10−7 and 10−4 M papaverine in the tissue bath of these animals. At the lower dose, there was no difference between micropipette-applied and tissue bath-applied papaverine (tissue bath 10−7 M; arcades: control 9 ± 4%, NTG 5 ± 5%; terminal: control 15 ± 3%, NTG 1 ± 11%). However, at the higher concentration (10−4 M) of papaverine, dilation was 20–30% greater in all groups with tissue bath-applied compared with micropipette-applied papaverine, consistent with the long-standing literature. Interestingly, at the high concentration, the arcade arterioles from NTG-treated rats dilated more than all other groups (tissue bath 10−4 M; arcades: control 58 ± 6%, NTG 86 ± 10%; terminal: control 57 ± 9%, NTG 59 ± 6%). Hence, PDE-mediated tone was enhanced in the arcades but not in the terminal arterioles with nitrate tolerance.

Fig. 3.

Concentration-response relationship for SNP (A andB), adenosine (C and D), and papaverine (E and F) in day 3 control and NTG-treated rats. The diameter response is calculated as the peak change from the baseline (initial), and the maximum diameter is in response to topical adenosine (10−3 M), obtained before other agents. * Different from initial baseline diameter (P < 0.05).

Effect of ROS/angiotensin II on nitrate tolerance.

To delineate further differences for animals of protocol 1, we studied the effect on diameter of ROS and the renin-angiotensin system. There was a significant shift toward angiotensin-linked constrictor tone in day 3 NTG-treated animals despite the finding that the level of total tone was unchanged. In day 3control animals, diameter change to losartan (100 nM) was not significant, whereas in day 3 NTG-exposed animals, there was a significant dilation in both the arcades and the terminal arterioles (Fig. 4 A). The contribution of AT1R activation to overall tone was greater in the arcades compared with the terminal arterioles in the NTG-treated animals. The level of tone contributed by ROS (predominantly superoxide) was also evaluated with both MnTBAP and SOD (Fig. 4 B). In control animals of protocol 1, the arcades, but not the terminal arterioles, exhibited baseline tone maintained by superoxide, as evidenced by the dilation with SOD alone. This was unchanged in NTG-treated animals for the arcades; however, for the terminal arterioles, NTG treatment enhanced superoxide-linked constrictor tone in day 3 animals. Thus data with SOD suggest that NTG treatment enhanced superoxide-linked tone for the terminal arterioles and not the arcades. The paradoxical constriction with MnTBAP alone is consistent with findings of other studies (see, e.g., Ref.20), as noted in discussion. There is a clear difference between vessel groups regarding constrictor tone of nitrate-tolerant rats. Nitrate tolerance enhances both superoxide- and angiotensin II-mediated tone in terminal arterioles but only angiotensin II-mediated tone in arcades.

Fig. 4.

Peak diameter change from baseline (initial) for terminal arterioles and arcade arterioles during the first 5-min application of losartan (100 nM) [angiotensin type 1 receptors (AT1R); A] or Mn(III) tetrakis(4-benzoic acid) porphyrin chloride (MnTBAP; 10 μM) or superoxide dismutase (SOD; 100 U/ml) [reactive oxygen species (ROS); B] in day 3 control or NTG-treated animals. * Different from initial baseline diameter (P < 0.05).

To determine whether the elevated AT1R tone was responsible for the diminished dilatory response to NO donors, we tested dilation to SNP in the presence of 100 nM losartan. Figure 5shows that losartan restored the maximal dilation for both the arcades and terminal arterioles; hence, the elevated AT1R constrictor tone contributed to the diminished dilatory response to NO in both groups of vessels.

Fig. 5.

Dilation responses to SNP (10−3 M) ± losartan (10 nM) for terminal arterioles and arcade arterioles inday 3 control and NTG animals. * Different from control, ▴different from SNP alone (P < 0.05).

To determine whether the elevated ROS-linked tone was involved in the diminished dilatory response to NO donors, we tested dilation to DETA-NO with MnTBAP and dilation to SNP with SOD. Figure6 shows that only the terminal arteriolar responses were significantly restored with these superoxide scavengers. The paradoxical attenuation of the control responses with these agents is addressed in discussion.

Fig. 6.

Dilation responses to DETA-NO (10−3 M) with MnTBAP (A) and SNP (10−3 M) with SOD (B) for terminal arterioles and arcade arterioles inday 3 control and NTG-treated animals. * Different from control, ▴different from NO donor alone (P < 0.05).

Because of the shift toward angiotensin-mediated constrictor tone and the finding that losartan restored the dilation to SNP for each arteriolar group, protocol 2 was included in this study. As expected, losartan treatment in the drinking water (protocol 2) prevented all constriction to angiotensin II in both arcades and terminal arterioles at all concentrations tested for both control and NTG-exposed animals (data not shown). In day 3NTG-treated animals, the dose response to angiotensin II was significantly shifted to the left, but only for the terminal arterioles and not the arcades. Thus a maximal constriction response was observed at 10−10 M angiotensin for the terminal arterioles (−32 ± 6%, day 3 NTG; P < 0.05) but not for arcades (−1 ± 5%), whereas equivalent constriction was obtained at higher concentrations of angiotensin II [day 3NTG 10−6 M: terminal, −35 ± 5% (P< 0.05); arcade, −27 ± 7% (P < 0.05)]. The data reinforce the idea that nitrate tolerance significantly increased the angiotensin II tone in both types of arterioles yet enhanced the angiotensin II constrictor capability more in the terminal arterioles than in the arcades.

The key finding for protocol 2 animals is that day 3 NTG-treated animals fed losartan showed normal dilatory responses to SNP (Fig. 7). Additionally, the data provide further evidence that superoxide-linked responses are contributing to diminished responsivity in both groups of arterioles.

Fig. 7.

Dilation responses to SNP (10−3 M) with and without SOD for terminal arterioles (A) and arcade arterioles (B) in day 3 NTG-treated animals with or without losartan (30 μM) in the drinking water (protocol 2).Day 3 control data are not shown. * Different from other groups at same concentration (P < 0.05).

Remote vascular responses.

Remote dilatory responses have not been assessed before in nitrate tolerance. Two distinct types of remote dilatory responses were tested here, each linked to endogenous NO metabolism. SNP remote dilations likely occur via gap junctional communication (11). The remote dilation to SNP is significantly attenuated by nitrate tolerance (Fig. 8). Animals receiving losartan in the drinking water (plus NTG) show normal responses, thus supporting a role for angiotensin-linked tone in modulation of vascular communication events. LM-609, although a blocking agent for angiogenesis during chronic exposure, acts acutely in a pharmacological manner as an agonist for the αvβ3-integrin receptor (22). Stimulation of this receptor elicits a remote dilation; all evidence to date supports the idea that this remote dilation is not transmitted by gap junctional communication but instead is an ascending flow-dependent dilation in response to an initial acute elevation in wall shear stress (13, 14). Thus the remote flow-dependent dilation to LM-609 is dependent on NO production. Surprisingly, the remote dilation to LM-609 is not compromised in nitrate tolerance (Fig. 8). Importantly, these data show that, although direct vasoactive responses to NO donors are compromised in nitrate tolerance, the flow-dependent dilatory pathways are intact.

Fig. 8.

Remote dilation responses to SNP (10−3 M;A) or LM-609 (10 μg/ml; B) for terminal arterioles in day 3 animals of protocol 2. * Different from control (P < 0.05).


This study demonstrates that NO-mediated dilation of terminal arterioles is inhibited by nitrate tolerance to a greater extent than in arcading arterioles that feed them. In nitrate tolerance, both angiotensin II- and superoxide-mediated tone are altered in the terminal arterioles, whereas angiotensin II- and PDE-mediated tone are altered for the arcades. Furthermore, both direct and remote dilation to NO donors are attenuated by nitrate tolerance; however, flow-dependent dilatory responses remain intact.

The key finding of this study is that terminal arterioles show a greater net sensitivity to nitrate tolerance than the arcade vessels. Table 1 summarizes the findings of this study. It is important to note that the observation sites for the arcade vessel and terminal arteriole were within 200 μm of each other and represent paired measurements taken during the same exposure. Despite their proximity, the ongoing flow conditions in these two groups of vessels are very different. The terminal arterioles of skeletal muscle are the inflow to a defined arteriolar network with prescribed flow control (10, 13, 15, 29), structure (12), and innervation (32). The larger arcade arterioles studied here serve primarily as conduits within the tissue and not as controllers of capillary perfusion per se (10,29). Thus the difference in response for the terminal and arcading arterioles to nitrate tolerance reported in the present study represents another functional distinction between the two groups of vessels.

View this table:
Table 1.

Summary of findings for nitrate tolerance vs. control

Why is there a difference in response for these two vessel groups? The responses of individual small arterioles in nitrate-tolerant animals have not been reported previously. We have ruled out diminished dilatory capability in general, because maximal dilation to adenosine is unchanged (Fig. 3). There remain several potential explanations for the finding that terminal arterioles are affected to a greater degree than arcades by nitrate tolerance. The explanations center around the difference in the baseline flow conditions for these two groups of vessels and the cellular basis for maintaining vascular tone in the terminal vs. arcade arterioles. Terminal arterioles have steady laminar flow with constant high wall shear stress, whereas the arcades have oscillatory flow and, hence, nonuniform shear. Endothelial cell nitric oxide synthase activity is increased by high laminar shear and decreased by oscillatory shear (18, 30). Thus one possibility is that the difference in flow conditions is responsible for a difference in NO-dependent dilatory capability in general and, consequently, has allowed a larger effect by nitrate tolerance in terminal arterioles compared with arcade arterioles. This difference in baseline flow conditions is additionally a potential reason for differences in the cellular basis in maintaining constrictor tone. In control conditions, arcade arterioles displayed a significant ROS constrictor tone, whereas the terminal arterioles did not, consistent with findings by others that oscillatory flow stimulates ROS production chronically, whereas maintained steady flow conditions do not (9,18). With nitrate tolerance, the shift in tone involves both angiotensin II and ROS for the terminal arterioles, consistent with the idea that the induction of O 2 levels by elevated angiotensin II (or other mechanisms) is a major molecular mechanism in these vessels. It is important to consider that the absence of evidence for upregulation of ROS-linked tone in the arcades may be due to a more complex interaction between tolerance and the flow conditions, as discussed above.

In the arcades, angiotensin II- and PDE-mediated tone are altered, suggesting a different underlying mechanism. On the basis of our prior study (20), it is clear that nitrate therapy upregulates the calcium/calmodulin-dependent cGMP-hydrolyzing PDE 1A1. We propose the following scenario. Elevated NO increases cGMP levels. Elevated cGMP then provides a stimulus for upregulation of the PDE 1A1 as a means to hydrolyze the excess cGMP. Simultaneously, elevated NO provides a stimulus for elevated renin-angiotensin system (and other vasoconstrictor pathways). Angiotensin II activates PDE 1A1, which then hydrolyzes cGMP, decreases dilator capability, and directly constricts the vessels. Losartan prevents the AT1R-mediated constriction and prevents activation of PDE 1A1, thus permitting cGMP levels to rise in response to SNP (i.e., stimulus for dilation), and the subsequent dilation is unhindered by constrictor tone due to AT1R activation. This scenario is reinforced by data showing that nitrate tolerance in the rat (aorta) results in activation of the renin-angiotensin system with upregulated expression of AT1R; losartan prevented development of nitrate tolerance in that study (21). The shift to PDE-mediated tone is likewise consistent with this proposed mechanism. In the present study, plasma nitrate levels remained high because of ongoing nitrate therapy and hence are part of the systemic treatment.

Aside from the mechanism of nitrate tolerance, the data show a difference in NO dilatory capability for these NO donors, with SNP and DETA-NO being more efficacious than SIN-1. SIN-1 releases superoxide along with NO, preferentially forming peroxynitrite (see, e.g., Ref.7); this alone explains the lower maximal dilation with SIN-1 for both vessel groups. Additionally, in control conditions, arcade arterioles displayed basal tone that could be inhibited by SOD, whereas the terminal arterioles did not. It is possible that diminished response to SIN-1 in the arcades is related to exogenous superoxide and not a true difference in NO dilatory capability between these vessel types, despite evidence that the superoxide formed from SIN-1 is not available for reaction with SOD (8). Together this would support our speculation that there is a higher basal release of NO from the terminal arterioles.

Some responses with MnTBAP are inconsistent with its action as a pure free radical scavenger. MnTBAP alone constricted the arcade arterioles and not the terminal arterioles (data in Fig. 5). This was unrelated to NTG treatment. Although the manganese porphyrin compounds are free radical scavengers that have been reported to augment NO effects (25), the same laboratory has reported that these compounds can induce constriction through peroxynitrate formation (23). Hence, the paradoxical constriction to MnTBAP reported in the present study may be due to reduction of basal NO activity. Constriction of only the arcades and not the terminal arterioles remains a puzzle, unless the terminal arterioles have a higher basal NO production rate that was not overwhelmed by this dose of MnTBAP.

A puzzling and consistent finding in control animals was attenuated dilation to DETA-NO with MnTBAP and to SNP with SOD for all vessel types (data in Fig. 6). If NO released from DETA-NO is destroyed by MnTBAP (as discussed above), then we would anticipate a diminished dilatory response in the presence of MnTBAP. If NO is consumed by MnTBAP, this suggests that the recovery for NO-induced dilation with MnTBAP in NTG-treated animals is, in fact, greater than the data demonstrate. Likewise, with SOD there may be an overriding chemical reaction to explain the diminished dilation to SNP with SOD for the control animals, but, unlike the MnTBAP case, the scenario with SOD does not explain the data with NTG-treated animals. This scenario is as follows: SNP contains five cyanide groups, each of which must be released before NO release from the compound (5). Cyanide is a potent inhibitor of endothelial cell Cu/Zn SOD (16). Thus it follows that for the control animals, endogenous and exogenous SOD would be blocked by cyanide, resulting in a net increase in superoxide and diminished dilation to a uniform concentration of SNP. However, with nitrate tolerance in vivo, SOD expression and activity are decreased (26), leaving less to be blocked by cyanide. Thus, for the present study, restoration of a significant dilation to SNP by SOD in day 3 NTG-treated animals could not be explained in this manner. We would expect, instead, that the total SOD would be decreased in NTG-treated animals and further inhibition of SOD by cyanide from SNP would not permit recovery with SOD. It is clear that the chemistry of oxygen free radical scavengers is complex, and the mechanism for the attenuated dilation in control animals remains to be determined. The mechanistic interpretation of the restored dilation in terminal arterioles of NTG-treated animals is considered to be via removal of superoxide, which is elevated because of nitrate tolerance. This is demonstrated with two separate pairs of NO donors and free radical scavengers and cannot presently be explained by a consistent mechanism involving purely chemical reactions for each pair tested.

In summary, this study suggests that the flow conditions found in these two distinct vessel types may play a role in determining the extent of hyporesponsivity to NO dilation in nitrate tolerance. The net result is that terminal arterioles, which control capillary perfusion, are detrimentally affected to a greater extent by nitrate tolerance than the arcading arterioles.


We gratefully acknowledge the support of American Heart Association (AHA) Scientist Development Grant 0030302T (to C. Yan), AHA Grant EI-0040197N, and National Heart, Lung, and Blood Institute Grants HL-55492 (to M. D. Frame), HL-63462, and HL-49192 (to B. C. Berk).


  • Address for reprint requests and other correspondence: M. D. Frame, Dept. of Anesthesiology, 601 Elmwood Ave, Univ. of Rochester, Rochester, NY 14642 (molly_frame{at}urmc.rochester.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • First published February 14, 2002;10.1152/ajpheart.00429.2001


View Abstract