METHODS

Methods for superfusion of drugs on the cerebral cortex, induction of hypercapnia, monitoring of CBF by laser-Doppler flowmetry, and the NO synthase (NOS) assay have been described previously (14, 16) and are summarized below.

General Surgical Procedures

Studies were performed in 25 male Sprague-Dawley rats (Sasco, Omaha, NE) weighing 300–380 g. Rats were anesthetized with 5% halothane in 100% oxygen. After induction of anesthesia the concentration of halothane was reduced to 1–2%. Cannulas were inserted in the left femoral artery and in the trachea. Animals were then placed on a stereotaxic frame (model 1404, Kopf Instruments, Tujunga, CA) and artificially ventilated with a nitrogen-oxygen mixture by a mechanical ventilator (model 638, Harvard Apparatus, S. Natick, MA). The proportion of oxygen in the mixture was adjusted to maintain arterial$P{O}_2$ at ∼150 mmHg [153 ± 23 (SD) mmHg]. Body temperature was maintained at 37 ± 0.5°C by use of a heating lamp controlled by a rectal probe (model 73A-TA, Yellow Springs Instrument, Yellow Springs, OH). The arterial catheter was used for blood sampling and continuous recording of arterial pressure and heart rate on a chart recorder (model 716P, Grass, Quincy, MA). Arterial $P{\text{CO}}_2$(${\text{Pa}}_{{CO}_2}$),$P{O}_2$, and pH were measured at multiple times on 50–100 μl of blood with use of a blood-gas analyzer (model 178, Ciba-Corning, Medfield, MA). After completion of the surgical procedures the halothane concentration was reduced to 1%. The level of anesthesia was tested by verifying the absence of corneal reflexes and motor responses to tail pinch. Rats were then paralyzed with tubocurarine. After paralysis the depth of anesthesia was monitored by testing the cardiovascular responses to tail pinch and by the pattern of the electrocorticogram (16). The electrocorticogram was recorded monopolarly using a metal screw inserted through the bone at a site near the cranial window. As described in detail elsewhere (14,16), a 3 × 3-mm cranial window was drilled at a site 2–3 mm lateral and 1–2 mm rostral to bregma. The dura was carefully removed, and the craniotomy site was continuously superfused at a rate of 0.33 ml/min with Ringer solution warmed to 37°C and aerated with 95% oxygen-5% carbon dioxide (pH 7.3–7.4).

Monitoring of CBF by Laser-Doppler Flowmetry

Laser-Doppler flowmetry was performed using a Vasamedic BPM403 instrument, the output of which was displayed on a chart recorder (14,16). The probe (tip 0.8 mm) was mounted on a micromanipulator and placed ∼0.5 mm above the pial surface within the cranial window. Probe position and reactivity of the preparation were tested at each site by determining the cerebrovascular reactivity to changes in${\text{Pa}}_{{CO}_2}$. Once a suitable placement was obtained, the probe was left at that site for the duration of the experiment.

NOS Assay

Methods for measuring brain Ca²⁺-dependent NOS catalytic activity by the isotopic conversion assay of Bredt and Snyder have been described in detail previously (16) and are briefly summarized. The entire forebrain of four rats was homogenized in 20 mM HEPES containing 0.5 mM EGTA, 1 mM dithiothreitol, and 0.32 M sucrose (Polytron/PT3000, Brinkmann). The homogenate was centrifuged at 20,000 rpm for 15 min, and triplicates of aliquots of the supernatant (150 μg protein) were incubated for 45 min (37°C) with a buffer containing 20 mM HEPES, 0.5 mM EGTA, 1 mM dithiothreitol, 0.32 M sucrose, 0.5 mM Ca²⁺ (1 μM free Ca²⁺), 200 μM NADPH, 1 μMl-arginine, and 1 μCi/mll-[³H]arginine. The reaction was stopped by addition of 2 ml of 20 mM cold HEPES containing 2 mM EDTA (pH 5.5). Samples were applied to Dowex AG50W-X8 (Na⁺ form) columns to removel-[³H]arginine. Columns were then washed with 2 ml of water, andl-[³H]citrulline was quantified in the flow-through fraction by use of a liquid scintillation spectrophotometer (model LS 6000, Beckman). The level ofl-[³H]citrulline was computed after subtraction of the blank value, which represents nonspecific radioactivity in the absence of enzyme activity. In studies in which the effect of homocysteine on NOS activity was tested, homocysteine (0, 1, 10, 100, 1,000 μM) was added to aliquots of homogenate before the 45-min incubation period. Aliquots run in parallel were treated with comparable concentrations of the NOS inhibitorN-monomethyl-l-arginine. Enzyme inhibition was expressed as a percentage of NOS catalytic activity of vehicle-treated aliquots.

Experimental Protocol

The cranial window was superfused with normal Ringer solution, and blood gases were adjusted (Table 1). The laser-Doppler probe was positioned on the cerebral cortex for continuous monitoring of CBF. Experiments commenced once arterial pressure, blood gases, and CBF were in a steady state.

Effect of homocysteine on resting CBF.

The cranial window was superfused with Ringer solution for 30 min. The superfusion solution was sequentially changed to1) Ringer solution containing 40 μM Cu²⁺ for 60 min,2) Ringer solution containing 40 μM Cu²⁺ and 1 mM homocysteine for 120 min, and 3) normal Ringer solution for 90 min (Ringer 2; Fig.1). The changes in CBF associated with these treatments were continuously recorded. These experiments were performed in six rats.

Effect of homocysteine on the increase in CBF produced by ACh, hypercapnia, SNAP, or papaverine.

The increase in CBF produced by hypercapnia, the NO-dependent vasodilator ACh (10 μM), the NO donor SNAP (500 μM), or the NO-independent vasodilator papaverine (1 mM) was studied. Hypercapnia (${\text{Pa}}_{{CO}_2}$ = 50–60 mmHg) was produced by adding carbon dioxide to the circuit of the ventilator. The elevation in ${\text{Pa}}_{{CO}_2}$ was maintained until the resulting flow increase was in a steady state (usually 3 min) (14,16). Agents were dissolved in Ringer solution and topically superfused until the CBF elevation reached a steady state. Concentrations yielding ∼50% of maximal response were used. Similarly, we used levels of hypercapnia ($P{\text{CO}}_2$ = 50–60 mmHg) that produce ∼50% of the maximal CBF elevation (17). CBF responses to hypercapnia, ACh, SNAP, and papaverine were tested in random sequence during superfusion with normal Ringer solution (Ringer 1; Fig.1), 40 μM Cu²⁺, or Cu²⁺-homocysteine (1 mM). Responses were tested 60–90 min after the Cu²⁺-homocysteine superfusion was started. The time interval between responses was 10–15 min. At the end of the superfusion with Cu²⁺-homocysteine, the solution was changed to normal Ringer (Ringer 2), and responses were tested again 60 min later. These experiments were performed in five rats.

Dose-response relationships.

CBF responses to hypercapnia, ACh, and SNAP were tested during superfusion with normal Ringer solution or Ringer solution containing 40 μM Cu²⁺ and increasing concentrations of homocysteine (100, 500, 1,000 μM). Each concentration was applied for 60–90 min before the responses were tested. The time interval between responses was 10–15 min. These experiments were performed in five rats.

Effect of SOD on homocysteine-induced attenuation of vasodilator responses.

Responses to hypercapnia, ACh, and SNAP were tested in random sequence during superfusion with normal Ringer solution, Ringer solution containing 1,000 U/ml of SOD, or Ringer solution containing 40 μM Cu²⁺, SOD (1,000 U/ml), and 1 mM homocysteine. Vasodilator responses were tested after each solution was superfused for 60–90 min. The time interval between responses was 10–15 min. These experiments were performed in five rats.

Data Analysis

Values are means ± SE. Multiple comparisons were evaluated by ANOVA and Tukey’s test (Systat, Evanston, IL). Differences were considered statistically significant for P < 0.05.

RESULTS

Effect of Homocysteine on Resting CBF

First, we studied whether Cu²⁺-homocysteine influences resting CBF. Superfusion with 40 μM Cu²⁺ alone did not affect resting CBF. Superfusion with Cu²⁺ and homocysteine (1 mM) attenuated resting CBF, an effect that was maximal after 90 min of superfusion (Fig. 1). At this time, CBF was reduced by 28 ± 4% of baseline (P < 0.05 by ANOVA, n = 6). After the superfusion solution was switched back to Ringer solution, CBF was not different from values before Cu²⁺-homocysteine application (Fig. 1; P > 0.05). Cu²⁺-homocysteine superfusion did not produce electrocorticographic evidence of seizures.

Effect of Homocysteine on CBF Responses to ACh, Hypercapnia, SNAP, and Papaverine

In these experiments we sought to determine whether homocysteine affects selected vasodilator responses of the cerebral circulation. During Ringer solution superfusion, 10 μM ACh, hypercapnia ($P{\text{CO}}_2$ = 50–60 mmHg; Table 1), 500 μM SNAP, and 1 mM papaverine increased CBF (Figs.2 and 3). The response was not affected by superfusion with Cu²⁺(P > 0.05; Figs. 2 and 3). Superfusion with Cu²⁺-homocysteine attenuated the increase in CBF produced by ACh (−73 ± 6%), hypercapnia (−31 ± 9%), and SNAP (−48 ± 4%;P < 0.05,n = 5), but not that evoked by papaverine (P > 0.05,n = 5; Figs. 2 and 3). After the superfusion solution was switched back to normal Ringer solution, responses to ACh, hypercapnia, SNAP, and papaverine were not different from those observed before Cu²⁺-homocysteine (P > 0.05; Figs. 2 and 3).

The relationship between the concentration of homocysteine and its cerebrovascular effects is illustrated in Figs.4 and 5. The reduction in resting CBF and the attenuation of vasodilator responses to ACh, hypercapnia, and SNAP were observed at 1 mM homocysteine (n = 5,P < 0.05). The response to ACh was slightly reduced at 500 μM (Fig. 4). However, this reduction did not reach statistical significance (P > 0.05).

Effect of SOD on the Attenuation of Cerebrovascular Dilator Responses

We then investigated whether the cerebrovascular effects of Cu²⁺-homocysteine were mediated by production of superoxide. Topical superfusion of the superoxide scavenger SOD (1,000 U/ml) did not affect resting CBF or the vasodilation produced by ACh, hypercapnia, or SNAP (Figs.6 and 7;P > 0.05,n = 5). However, SOD prevented the reduction of resting CBF and the attenuation of the response to ACh, hypercapnia, and SNAP produced by Cu²⁺-homocysteine (Figs. 6 and 7;P > 0.05 vs. Ringer).

Effect of Homocysteine on Brain NOS Enzymatic Activity

To rule out the possibility that the cerebrovascular effects of homocysteine were related to inhibition of brain NOS, we investigated the effects of homocysteine on NOS activity in forebrain homogenates (n = 4). Homocysteine (1–1,000 μM) had no effect on NOS activity, whereas comparable concentrations ofN-monomethyl-l-arginine markedly reduced NOS activity (Fig. 8). These data suggest that homocysteine at a concentration effective in reducing resting CBF and vascular reactivity does not inhibit brain NOS catalytic activity.

DISCUSSION

We have demonstrated that homocysteine, in the presence of Cu²⁺, reduces resting CBF and attenuates the increases in CBF produced by ACh, hypercapnia, and SNAP, whereas the response to papaverine is not affected. The cerebrovascular effects of Cu²⁺-homocysteine are prevented by pretreatment with SOD. The data suggest that homocysteine exerts profound cerebrovascular effects that are mediated via Cu²⁺-catalyzed production of superoxide anion.

The cerebrovascular actions of Cu²⁺-homocysteine cannot be attributed to differences in arterial pressure or blood gases, because these parameters were closely monitored and controlled and did not differ between groups. The effect of homocysteine is also unlikely to result from a nonselective impairment of smooth muscle relaxation, because the increase in CBF produced by papaverine was not affected. Similarly, the reduced cerebrovascular reactivity to ACh, hypercapnia, or SNAP cannot be attributed to a time-dependent deterioration of the preparation, because responses returned to normal after Cu²⁺-homocysteine was discontinued. In addition, in preliminary experiments, we found that superfusion with homocysteine alone for 5 h affected neither resting CBF nor vasodilator responses to ACh, hypercapnia, or SNAP (F. Zhang and C. Iadecola, unpublished observations). It is also important to point out that the effect of Cu²⁺-homocysteine is not related to endothelial injury, because the increase in CBF produced by ACh, a response mediated by production of NO in endothelial cells (25, 35), is fully reestablished after discontinuation of homocysteine superfusion. Homocysteine has been reported to induce seizures (1). However, it is unlikely that the cerebrovascular effects of homocysteine are due to seizures or to a postictal state after seizures, because homocysteine superfusion did not induce electrocorticographic evidence of seizures. The concentration of SOD used in the present study (1,000 U/ml) was higher than that used by others (50–100 U/ml) (34). This concentration was chosen to ensure that maximal scavenging of superoxide was achieved. However, SOD at 1,000/U ml did not affect resting CBF or the reactivity to hypercapnia, ACh, and SNAP. Therefore, the effect of SOD cannot be a consequence of nonspecific actions of this enzyme on resting CBF or baseline vasodilator responses.

Homocysteine is emerging as an important risk factor for ischemic stroke (see Ref. 3 for review). Patients with inherited deficiency of cystathionine β-synthase, an enzyme involved in homocysteine metabolism, have markedly elevated levels of plasma homocysteine and are more susceptible to ischemic stroke (22). More recently, it has been proposed that homocysteine is a stroke risk factor in patients with moderate elevations of plasma homocysteine (3). Patients with folate deficiency, a condition associated with hyperhomocysteinemia, have an increased risk of ischemic stroke (11). Furthermore, case-control studies, including a prospective study (32), have demonstrated that patients with cerebrovascular diseases tend to have increased levels of plasma homocysteine (2, 4, 7). Therefore, hyperhomocysteinemia might be a risk factor for ischemic stroke in the general population.

The mechanisms by which increased levels of homocysteine might increase the incidence of cerebral ischemia have not been elucidated. One possibility is that homocysteine predisposes to ischemic injury by its well-known ability to induce vascular damage, atherosclerosis, and thrombosis (12, 13, 28, 33; see Ref. 24 for review). Another possibility is that homocysteine impairs cerebrovascular regulation, thereby decreasing the ability of the brain circulation to compensate for the reduction in blood flow that occurs during ischemia. One potential mechanism by which homocysteine could compromise vascular regulation is by interfering with the action of endothelial vasodilators such as NO. This possibility is supported by our observation that the vasodilation produced by ACh or hypercapnia, responses that depend on NO synthesis (10, 14, 35), is attenuated by Cu²⁺-homocysteine, whereas the vasodilation produced by papaverine, a response that is independent of NO (14), is not affected. The attenuation of NO-dependent responses is not related to inhibition of NOS activity, because homocysteine does not affect Ca²⁺-dependent citrulline production in brain homogenates ex vivo. Rather, the fact that the impairment of NO-dependent responses is prevented by SOD is consistent with the hypothesis that NO is scavenged by superoxide anion generated by Cu²⁺-homocysteine (23). The observation that, in cranial window preparations, generation of superoxide by other means attenuates NO-dependent cerebrovascular responses supports this conclusion (18, 34). Cu²⁺-homocysteine attenuates also the vasodilation produced by SNAP, a compound that generates NO nonenzymatically, and by hypercapnia, a response that depends on the availability of endogenous NO (16). It would, therefore, seem that superoxide produced by Cu²⁺-homocysteine scavenges NO independently of its source. NO reacts with superoxide extremely rapidly, leading to production of peroxynitrite (8). The mechanisms of superoxide production by autoxidation of thiol are well described and involve reduction of a transitional metal followed by reaction of the metal with molecular oxygen and formation of superoxide (6, 21, 23). Thiol autoxidation can also generate hydrogen peroxide and hydroxyl radicals (28). The role of these reactive oxygen species in our model has not been established and needs to be investigated in future studies.

Homocysteine has also been reported to react directly with NO to formS-nitrosohomocysteine, a product that retains the vasoactive properties of NO but is less cytotoxic than homocysteine (27). AlthoughS-nitrosohomocysteine could be formed in our model, it is unlikely that this process is the predominant mechanism responsible for the attenuation of NO-dependent responses. This is because the reaction of NO with superoxide is faster thanS-nitrosation (8). Moreover,S-nitrosylated species do not block vasodilation; rather, they retain the vasodilatory effect of NO (27). Homocysteine can also react with adenosine to formS-adenosylhomocysteine, a reaction catalyzed by a specific hydrolase present in the central nervous system (5). This reaction can, therefore, lower the extracellular concentration of adenosine, a potent cerebrovasodilator, and decrease resting CBF (26). However, it is unlikely that the findings of the present study are related to this mechanism. Adenosine is not involved in the increase in CBF produced by ACh or NO donors (9, 20) and, consequently, lack of adenosine could not explain the reduction in these responses induced by homocysteine. Furthermore, the reaction of adenosine with homocysteine should not be superoxide dependent and, as such, the effect of homocysteine should not be prevented by SOD.

The cerebrovascular effects of homocysteine were observed at 1 mM, a concentration similar to that used in previous studies in which other vascular effects of exogenous homocysteine were investigated (26-29). These concentrations are higher than those observed in patients with inherited or acquired hyperhomocysteinemia (31). Therefore, findings obtained with high concentrations of homocysteine should be interpreted with caution. However, reduction in endothelium-dependent vasodilation has been reported in the carotid artery of monkeys in which the plasma concentrations of homocysteine were increased only to ∼11 μM by dietary measures (19). This finding may suggest that chronic exposure to moderately elevated levels of homocysteine may be more effective than acute exposure. In addition, because there is compartmentation of homocysteine in tissues (30), it is conceivable that intracellular and interstitial concentrations are higher than plasma concentrations. Alternatively, other factors may play a role in the reduction of the vascular responses in diet-induced hyperhomocysteinemia (see Ref. 19 for discussion). These hypotheses remain to be tested experimentally. Studies in chronic models of hyperhomocysteinemia with moderate elevation of plasma homocysteine would be needed to address this issue.

The present results provide an additional insight into the mechanisms of the increased susceptibility of cerebral ischemia observed in hyperhomocysteinemia. There is substantial evidence that the vascular-hemodynamic effects of NO are beneficial in the early stages of cerebral ischemia (15). Immediately after ischemia, NO may facilitate collateral flow to the ischemic territory and maintain microvascular patency by inhibiting platelet aggregation (15). In hyperhomocysteinemia, NO scavenging by superoxide may limit the amount of NO available to exert these beneficial hemodynamic effects, resulting in worsening of cerebral perfusion. Increased platelet aggregation resulting from removal of NO may promote thombus extension and microvascular occlusions. Additional studies, however, are required to determine whether hyperhomocysteinemia worsens the outcome of experimental cerebral ischemia and, if so, whether the effect is related to a greater flow reduction in the ischemic territory and to increased platelet aggregation.

In conclusion, we have demonstrated that 1 mM homocysteine, in the presence of Cu²⁺, reduces resting CBF and selectively attenuates the increase in CBF produced by ACh, hypercapnia, and SNAP, responses entirely, or in part, mediated by NO. The cerebrovascular effects of homocysteine are prevented by the superoxide scavenger SOD. The data are consistent with the hypothesis that superoxide, generated from metal-catalyzed oxidation of homocysteine’s thiol group, scavenges NO and decreases the availability of NO for vasodilation. Because the cerebrovascular and antiplatelet effects of NO are beneficial in the early stages of cerebral ischemia, homocysteine-induced scavenging of NO may be one of the mechanisms by which this amino acid worsens the outcome of cerebral ischemia. However, future studies using more clinically relevant models of hyperhomocysteinemia are required to better assess the impact of these cerebrovascular effects on the cerebrovascular complication of this condition.