Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Furchgott and Zawadski (1980) were the first to report the phenomenon of endothelium-dependent relaxation of blood vessels. It has been suggested that this action results from the release of an EDRF that was likely to be NO or a closely related entity (Palmer et al., 1988). Since these initial reports, it has become apparent that the endothelium-dependent relaxation of certain blood vessels to certain stimulants is not entirely mediated by EDRF/NO but may also include other relaxation factors such as EDHF or prostanoids (Cohen and Vanhoutte, 1995).

Arginine analogs that are competitive antagonists to NOS have been used to assess the contribution of NO to relaxation induced by a variety of endothelium-dependent agonists. These compounds do not antagonize NO directly but only block NO synthesis via NOS. A different approach to assess the role of NO is by using pharmacological probes that directly react with or scavenge NO. Indeed, we have previously reported that the iron-thiol containing NO scavenger, MGDFe, significantly attenuated ACH-stimulated relaxation in both control and diabetic rat aortic rings; however, a significant component of relaxation was resistant to MGDFe antagonism (Pieper and Lai, 1997). These observations were interesting because ACH-mediated relaxation in both experimental groups is virtually abolished by L-NA and relaxation to ACH is not modified by indomethacin (Pieper et al., 1997). Taken together, these studies raise some question whether endothelium-dependent relation in this model is, in fact, entirely mediated by NO radical (·NO) per se.

To further understand this discrepancy, we used an agent belonging to a uniquely different class of NO antagonist known as nitronyl nitroxides. One of these nitronyl nitroxides, CPTIO, has been shown to be a valuable probe to discount the specific role of ·NO in antimicrobial action (Yoshida et al., 1993). CPTIO has also been used to distinguish the role of ·NO in relaxation of rat aorta compared to nerve-stimulated relaxation in anoccygeous muscle and gastric fundus strips (Rand and Li, 1995). CPTIO completely inhibited relaxation induced by NO gas in rat aorta (Rand and Li, 1995) and inhibited by >90% the relaxation elicited by authentic EDRF using ACH in rabbit aorta (Akaike et al., 1993). Thus, we used CPTIO to discriminate a role of NO-dependent vs. NOS-dependent, endothelium-dependent relaxation to ACH in control and diabetic rat aorta.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Male Sprague-Dawley rats were made diabetic by an i.v. injection of 55 mg/kg streptozotocin as previously described (Pieper and Peltier, 1995). Diabetes was verified at 1 wk using an Exac Tech glucometer and test strips. Diabetic and age-matched control rats were housed for 8 wk before evaluation of vascular function. After this duration, animals were anesthetized with 65 mg/kg sodium pentobarbital. The descending thoracic aorta was carefully isolated, cleaned and sectioned into 3-mm long aortic rings. Extreme care was taken to avoid damage during the isolation process. In a few of the rings, the endothelium was intentionally removed by rubbing the lumen with a forceps.

Rings were mounted between parallel wires in isolated tissue baths at 37°C. The bath medium contained Krebs bicarbonate buffer which was oxygenated at 95% O₂:5% CO₂ to maintain pH at 7.4. The buffer contained (in mM): NaCl 118, KCl 4.7, CaCl₂ 2.5, MgSO₄ 1.2, NaHCO₃ 24 and glucose 10.5. Rings were equilibrated under optimal tensions of 2.0 g for both control and diabetic rings before performing measurements of contractile reactivity to PE. Isometric tension were recorded on a Gould (Valley View, OH) TA6000 recorder using Radnoti (Monrovia, CA) model 159901 force-displacement transducers.

Each ring was contracted with cumulative concentrations of PE. After washing and equilibrium, rings were contracted with a submaximal concentration of PE (i.e., 1 µM) before evaluating relaxation produced by increasing concentrations of the endothelium-dependent vasodilator, ACH. In a few rings, nitroglycerin was used as a vasodilator for endothelium-independent relaxation.

Previously, we showed that endothelium-dependent relaxation to ACH is impaired in diabetic aortic rings that are contracted with either norepinephrine or PE (Pieper et al., 1997). In these studies as in many other investigations, the concentration was varied, if necessary, to provide equipotent contraction based on concentration-response curves to each agonist. In our study, an equimolar PE concentration was used in both control and diabetic preparations. This protocol was designed to prevent variability in the proportion of NO vs. EDHF that contributes to total ACH-induced relaxation. This proportion is known to vary at different concentrations of vasoconstrictor (Hatake et al., 1995) and might confound the interpretation of the efficacy of CPTIO in control versus diabetic rings.

After the initial ACH challenge to verify similar baseline responses between rings from the same animal, individual rings were washed, equilibrated and contracted a second time to PE. At the peak of PE-induced contraction, various concentrations of CPTIO were added. At the peak of response to CPTIO, the rings were then challenged a second time to increasing ACH concentrations. Previous studies indicated that relaxation between the first and second challenges to ACH in both control and diabetic rings are similar (Pieper and Peltier, 1995, Pieper et al., 1996a). In a select few experiments, we also used a related nitronyl nitroxide derivative known as PTIO. In a few other experiments, we challenged rings with a single ACH concentration of 30 µM to evaluate whether drug interventions altered either the initial rapid phase vs. maximum sustained phase of relaxation to ACH since the factors or molecular species responsible for hyperpolarization and relaxation has been suggested to be different at each phase (Vanheel et al., 1994; Rubanyi et al., 1985).

To evaluate the CPTIO-resistant component of ACH-mediated relaxation, we also performed several experiments using specific drug interventions. These included 10 µM indomethacin (to inhibit cyclooxygenase); 100 µM L-NA (to inhibit NOS); 1 mM TEA (to inhibit K⁺ channels); 1 mM L-cysteine (to scavenge nitroxyl anion and/or peroxynitrite anion); 1000 U/ml catalase (to decompose H₂O₂) and 3 µM of ODQ to inhibit soluble guanylate cyclase (Garthwaite et al., 1995; Moro et al., 1996).

Data are presented as the mean ± S.E.M. where n equal the number of individual animals. Data were analyzed by analysis of variance or repeated analysis of variance followed by Fishers PLSD test for multiple mean comparisons, where appropriate; or by unpaired t test and paired t test for comparison of two group means, where appropriate. A P < .05 was considered to designate statistical significance.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Contractile function. The average maximum tension development of all aortic rings in response to increasing PE concentrations was reduced by diabetes (table 1); however, tension normalized for cross-sectional area was not different. The sensitivity (pD₂) to PE was modestly different between the two groups. Removal of the endothelium increased the tension development in both control and diabetic rings; however, the increase was greater in control vs. diabetic rings.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Effects of 300 µM CPTIO on tension development in control (upper panel) and diabetic (lower panel) aortic rings contracted with 1 µM PE. Rings are untreated (control: n = 23; diabetic: n = 23), treated with 10 µM indomethacin (control: n = 7; diabetic: n = 9), 1 mM TEA (control: n = 10; diabetic: n = 11), 1000 U/ml catalase (control: n = 9; diabetic: n = 10), 1 mM L-cysteine (control: n = 4; diabetic: n = 7), 3 µM ODQ (control: n = 8; diabetic: n = 15) or ODQ in combination with either TEA, indomethacin and catalase (n = 3-4 each). P < .01 vs. untreated

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Example showing acetylcholine (ACH)-induced relaxation in control rings (tracings a and b) or diabetic rings (tracings c and d) and inhibition of relaxation following pretreatment with 300 µM CPTIO (tracings a and c), or reversal by CPTIO given after increasing ACH concentrations (tracings b and d). ACH concentrations are noted by dots and listed as the -log M. Boxes indicate when ACH or CPTIO is present. PE (phenylephrine: 1 µM).

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Tension development in rings contracted with PE after addition of 100 µM CPTIO (control: n = 23; diabetic: n = 32) compared to rings without CPTIO but after pretreatment with the NOS inhibitor, L-NA (control: n = 12; diabetic: n = 14), and in rings without endothelium (control: n = 12; diabetic: n = 11). P < .01 vs. corresponding control rings

Vascular relaxation analysis. For all rings, the 1 µM PE concentration used for the relaxation studies produced 66 ± 2 and 58 ± 2% maximum tension for control vs. diabetic rings, respectively. ACH fully relaxed both control and diabetic rings (examples shown in fig. 7, b and d) and this relaxation was significantly inhibited by L-NA, ODQ or by removal of the endothelium in both groups (fig. 3). Although blockade of ACH -mediated relaxation by L-NA and ODQ was complete in control rings, a significant residual component of ACH -stimulated relaxation was present in diabetic rings equivalent to 15 ± 5 and 27 ± 6% in L-NA-treated and ODQ-treated rings, respectively. In contrast, ODQ caused complete and equivalent inhibition of nitroglycerin-induced relaxation in both control and diabetic rings (not shown).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effects of removal of endothelium (n = 11 each) or incubation with either 100 µM L-NA (control: n = 4; diabetic: n = 7) or 3 µM ODQ (control: n = 4; diabetic: n = 7) on ACH-induced relaxation compared to untreated control (total n = 13) or untreated diabetic (total n = 20) rings. *P < .05 and P < .01 vs. corresponding untreated rings

Effects of CPTIO on ACH-induced relaxation. Addition of 30 to 300 µM CPTIO to control rings caused a concentration-dependent inhibition in the ACH-induced relaxation (fig. 4). Addition of 500 µM CPTIO caused no further inhibition (not shown). CPTIO also caused a concentration-dependent inhibition of relaxation in diabetic rings (not shown). We used 300 µM CPTIO to further evaluate the role of NO in ACH -mediated relaxation of control and diabetic rings. The % residual ACH -induced relaxation that was resistant to CPTIO was greater in diabetic rings than in control rings (fig. 5). Addition of either indomethacin (not shown), catalase or TEA or L-cysteine did not alter the CPTIO-resistant component of ACH-induced relaxation in both control and diabetic rings (fig. 6). In contrast, the CPTIO-resistant component of ACH-mediated relaxation was inhibited completely by preincubation with ODQ in control rings but left a small residual portion of relaxation in diabetic rings (fig. 6). Although pretreatment with CPTIO alone before addition of increasing concentrations of ACH partially inhibited the relaxation in control and diabetic rings, the addition of CPTIO after addition of the highest concentration of ACH completely restored tension in both control and diabetic rings to levels greater than that observed prior to the addition of ACH (fig. 7).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Concentration-dependent inhibition of acetylcholine-induced relaxation in control aortic rings by 30 to 300 µM CPTIO (n = 4-11 each). *P < .05; P < .01 vs. relaxation in the same rings before treatment with CPTIO

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Acetylcholine-mediated relaxation which is insensitive to antagonism by 300 µM CPTIO in control (n = 16) and diabetic (n = 19) rings. P < .01 vs. control rings

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Failure of incubation with 1 mM TEA (control: n = 5; diabetic: n = 6), 1 mM L-cysteine (control: n = 4; diabetic: n = 5), or 1,000 U/ml catalase (control: n = 5; diabetic: n = 6) but not 3 µM ODQ (control: n = 3; diabetic: n = 6), to alter the ACH-induced relaxation component resistant to antagonism by 300 µM CPTIO in control and diabetic rings. TEA. *P < .05 vs. pair-matched, untreated rings with CPTIO alone

View larger version (27K): [in this window] [in a new window]   Fig. 8.   Residual relaxation to a single concentration of acetylcholine (30 µM) that is resistant to inhibition by 300 µM CPTIO in control or diabetic rings. The CPTIO-insensitive component of relaxation is inhibited by 3 µM ODQ (control: n = 4; diabetic: n = 9) but not by either 10 µM indomethacin (control: n = 4; diabetic: n = 6), 1 mM TEA (n = 5 each) or 1000 U/ml catalase (n = 4 each). P < .01 vs. pair-matched untreated rings with CPTIO only (total control: n = 7; total diabetic: n = 12)

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Failure of substitution with 300 µM PTIO or addition of PTIO plus 300 µM CPTIO to alter the CPTIO-resistant portion of acetylcholine-induced relaxation in control rings (n = 5 or 8) but not diabetic rings (n = 6-7 each). *P < .05 vs. pair-matched rings treated with CPTIO alone

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In our study, we have characterized the contribution of NO to basal tone and ACH-stimulated endothelium-dependent relaxation of control and diabetic rat aortic rings. Our studies suggest that the nitronyl nitroxide, CPTIO, antagonized all of the NO released under basal conditions in both control and diabetic rings. In addition, CPTIO inhibited a significant portion of ACH-stimulated relaxation in both control and diabetic rings; however, a residual component of relaxation remained which was larger in diabetic rings. This effect occurred despite the observation that relaxation was virtually eliminated in both groups using the NOS inhibitor, L-NA. Although our study is the first known application of the nitronyl nitroxide class of NO antagonists to provide valuable insight into the nature of endothelium-dependent relaxation in diseased blood vessels, it is clear that these agents give important information about NO-like activity that would not be achieved using NOS inhibitors alone.

Basal NO tone. Our studies suggest that ·NO per se is the likely product released by both control and diabetic rings under control unstimulated conditions and that the nitronyl nitroxide, CPTIO, is effective in counteracting all of the ·NO activity released under these conditions. This conclusion is based on several observations including: 1) the CPTIO-sensitive incremental increase of tension development in PE-contracted rings was insensitive to indomethacin, cysteine, TEA and catalase but was inhibited by ODQ; 2) the CPTIO-sensitive component of tension development was equivalent to that achieved by removal of the endothelium or by pretreatment with L-NA and 3) the addition of CPTIO to rings without endothelium or L-NA-treated rings with endothelium did not produce any significant change in tension.

Agonist-stimulated endothelium-dependent relaxation in diabetes. Endothelium-dependent relaxation to ACH and other agonists is impaired in a variety of conduit and resistance blood vessels taken from experimental diabetic animals. Because endothelium-dependent relaxation is also impaired in both type I (Johnstone et al., 1993) and type II (McVeigh et al., 1992) diabetes mellitus in humans, information derived from experimental diabetic models may provide insight to the defect in human diabetes. Several studies using a variety of vessels from different species show that indomethacin does not shift the relaxation responses to ACH (Oyama et al., 1986; Hattori et al., 1991; Cameron and Cotter, 1992; Chang and Stevens, 1992; Pieper et al., 1992, 1997; Mayhan, 1992; Dai et al., 1993; Taylor et al., 1994, Diederich et al., 1994; Kamata and Kobayashi, 1996). These collective observations suggest that prostanoid factors cannot account for this dysfunction and that non-prostanoid factors such as deficits in NO per se can readily account for endothelial dysfunction in diabetes.

Role of NO in diminished endothelium-dependent relaxation in diabetes. In support of deficits in NO activity, cGMP concentration in response to ACH was diminished (Kamata et al., 1989; Abiru et al., 1990 Pieper et al., 1996a; Abiru et al., 1993). The assumption that the ACH-stimulated cGMP generation and relaxation is entirely mediated by free NO radical, ·NO, is based on the observation that the cGMP generated is essentially blocked by L-NA. This assumption may not be entirely valid because NO-independent molecular species arising from the NOS reaction (see discussion below) or molecular rearrangement of ·NO in equilibrium with other redox-dependent forms such as the nitrosonium cation (NO⁺) or the nitroxyl anion (NO) may occur. Indeed, both NO⁺ and NO have been claimed to be vasodilators (Fukoto et al., 1992; Pino and Feelisch, 1994; Stamler et al., 1992), although others have argued that only ·NO can activate purified guanylate cyclase (Dierks and Burstyn, 1996). Nevertheless, the biological evidence of any of these transformed species under physiological or pathophysiological condition has yet to be documented.

Potential role of an aberrant NOS reaction/reaction products. In the presence of an arginine deficiency or limited cofactor, purified NOS enzyme can reduce molecular oxygen to H₂O₂ or superoxide anion radical accompanied by diminished production of ·NO (Heinzel et al., 1992; Pou et al., 1992). This may have some implications in diabetic blood vessels. Indeed, the concentration of tetrahydrobiopterin has been shown to be diminished in brain tissue of diabetic animals (Hamon et al., 1989) although supplementation with a tetrahydrobiopterin derivative in vitro restores endothelium-dependent relaxation to ACH in diabetic rat aorta (Pieper, 1997). In addition, we have shown decreases in plasma arginine concentration (Pieper and Peltier, 1995) and in arginine content in vascular tissue of diabetic rats (Pieper and Dondlinger, 1996). Furthermore, supplementation with L-arginine in vitro or in vivo restores endothelium-dependent relaxation to ACH in diabetic rat aortic rings and improves cGMP generation (Pieper and Peltier, 1995; Pieper and Dondlinger, 1996; Pieper et al., 1996a). We have also shown that superoxide dismutase plus catalase restores ACH-induced relaxation to normal in diabetic aortic rings (Pieper et al., 1997) which would be consistent with increased basal production of superoxide anion radicals, ·O₂ (Chang et al., 1993; Pieper, 1995) and H₂O₂ (Pieper, 1995) by diabetic aorta. Thus, it is theoretically possible that an aberrant NOS may be a source of increased reactive oxygen production from diabetic endothelium.

Comparison of the effects of nitronyl nitroxide vs. NOS inhibitors. The CPTIO-resistant component of relaxation to ACH that was greater in diabetic vs. control rings is not unique to CPTIO because a similar observation was shown using MGDFe (Pieper and Lai, 1997), an agent that acts to interact with NO in a manner quite different than that achieved by CPTIO. Our use of a nitronyl nitroxide compared with L-NA appears be a useful candidate probe to discriminate the effects of ·NO vs. other products of the NOS reaction on endothelium-dependent relaxation in control vs. diabetic blood vessels. Reliance alone on the efficacy of L-NA to inhibit relaxation to ACH in both control and diabetic rings does not prove that relaxations are, in fact, entirely mediated via ·NO. Rather, it simply suggests that relaxation arises from a NOS pathway. Indeed, L-NA might also eliminate the reactive oxygen by-products of the NOS reaction as well. We suggest, therefore, that a portion of the L-NA-sensitive, ACH -mediated relaxation that was insensitive to CPTIO and larger in diabetic rings might result from increased generation of products arising from an aberrant NOS reaction in diseased blood vessels.

Potential role of EDHF. We considered the possibility that the CTPIO-resistant component of relaxation to ACH is related to release of an EDHF that may act via cGMP-dependent or cGMP-independent pathways (Cohen and Vanhoutte, 1995). Previous studies discounted a role of alterations in EDHF to account for impaired endothelium-dependent relaxation in diabetic rat aorta (Endo et al., 1995). In our study, our endothelium-dependent relaxation protocols using CPTIO in the presence of indomethacin, TEA or catalase suggest that any putative alteration in prostanoid production, activation of calcium-activated K⁺ channel-sensitive hyperpolarizing factor or H₂O₂ generation in diabetic rings is unlikely to account for this larger CPTIO-resistant component of relaxation to ACH in diabetic compared to control rings.

Potential role of nitrosothiols. We also considered the possibility that a portion of the relaxation produced by ACH arises from release of a thiol-bound derivative such as S-nitrosocysteine (Myers et al., 1990; Rubanyi et al., 1991) because CPTIO is unable to interact chemically with nitrosothiols (Akaike and Maeda, 1996). Reaction of ·NO with cysteine produces S-nitrosocysteine that increases the potency of authentic EDRF and increases its half-life. In this regard, L-cysteine has been shown to potentiate ·NO-dependent relaxation but inhibit relaxation mediated by NO   (Pino and Feelisch, 1994) and might also interact with ONOO (Pfeiffer et al., 1997). Because L-cysteine did not change the CPTIO-resistant portion of relaxation to ACH in either control or diabetic rings, we conclude that the CPTIO-resistant portion of relaxation to ACH in diabetic rings is likely not mediated by ONOO.

Role of sequestration of ·NO in the lipid phase. Previous cell-free studies using nitroxide compounds (rather than nitronyl nitroxides) suggest that ·NO might partition in lipid environments (Singh et al., 1994). In fact, at least two reports suggest that ·NO might be sequestered in large quantitities in lipid bilayers (Lancaster, 1996; Denicola et al., 1996). Thus, we considered the possibility that some of the ·NO escapes reaction with CPTIO (as with our previous studies using MGDFe) due to the fact that both are predominately water-soluble agents and, therefore, might be unable to effectively counteract the high diffusibility of ·NO under agonist-stimulated conditions.