We have reported that hyperhomocysteinemia (HHcy) evoked by folate depletion increases arterial permeability and stiffness in rats and that low folate without HHcy increases arterial permeability in mice. In this study, we hypothesized that HHcy independently increases arterial permeability and stiffness in mice. C57BL/6J mice that received rodent chow and water [control (Con), n = 12] or water supplemented with 0.5% l-methionine (HHcy, n = 12) for 18 ± 3 wk had plasma homocysteine concentrations of 8 ± 1 and 41 ± 1 μM, respectively (P < 0.05), and similar liver folate (∼12 ± 2 μg folate/g liver). Carotid arterial permeability, assessed as dextran accumulation using quantitative fluorescence microscopy, was greater in HHcy (3.95 ± 0.4 ng·min−1·cm−2) versus Con (2.87 ± 0.41 ng·min−1·cm−2) mice (P < 0.05). Stress versus strain curves generated using an elastigraph indicated that 1) maximal stress (N/mm2), 2) physiological stiffness (low-strain Young's modulus, mN/mm), and 3) maximal stiffness (high-strain Young's modulus, N/mm) were higher (P < 0.05) in aortas from HHcy versus Con mice. Thus, chronic HHcy increases arterial permeability and stiffness. Carotid arterial permeability also was assessed in age-matched C57BL/6J mice before and after incubation with 1) xanthine (0.4 mg/ml)/xanthine oxidase (0.2 mg/ml; X/XO) to generate superoxide anion (O2−) or 50 μM dl-homocysteine in the presence of 2) vehicle, 3) 300 μM diethylamine-NONOate (DEANO; a nitric oxide donor), or 4) 10−3 M 4,5-dihydroxy-1,3-benzene disulfonic acid (tiron; a nonenzymatic intracellular O2− scavenger). Compared with preincubation values, X/XO and dl-homocysteine increased (P < 0.05) permeability by 66 ± 11% and 123 ± 8%, respectively. dl-Homocysteine-induced increases in dextran accumulation were blunted (P < 0.05) by simultaneous incubation with DEANO or tiron. Thus, acute HHcy increases arterial permeability by generating O2− to an extent whereby nitric oxide bioavailability is reduced.
- cardiovascular risk
- vascular mechanics
hyperhomocysteinemia (HHcy) is a cardiovascular risk factor that may be caused by deficiencies of folate or B vitamins (e.g., cobalamin and pyridoxine) and/or by genetic defects in folate or methionine metabolism. We (37) have reported that HHcy evoked by folate depletion increases arterial permeability and stiffness in rats. HHcy, low folate, and/or their combination could have contributed to these responses. More recently, we (38, 39) found that a low folate status produces arterial dysfunction in a manner that is independent of its ability to elevate homocysteine. Specifically, conductance and resistance vessel function were impaired in rats (38), and carotid arterial permeability was greater in mice (39) that possessed low liver folate concentrations in the absence of HHcy.
In the present investigation, we tested the hypothesis that HHcy independently increases arterial permeability and stiffness in mice. Homocysteine is produced from dietary methionine through the intermediates S-adenosylmethionine and S-adenosylhomocysteine (Fig. 1). To produce HHcy without influencing folate status, l-methionine was added to the drinking water of mice. We chose to assess arterial permeability and stiffness because of their potential relevance to cardiovascular disease. In this regard, greater endothelial cell layer permeability facilitates arterial lipoprotein accumulation and contributes to lesion development and/or severity, and decreased arterial compliance increases afterload to an extent whereby myocardial oxygen demand may be elevated inappropriately (2, 30).
All protocols used in this study were approved by Animal Use and Care Committees (University of California, Davis, CA, and University of Utah, Salt Lake City, UT) and conformed to guidelines set forth by the American Physiological Society and Animal Welfare Act.
Experimental animals and diets.
C57BL/6J mice were housed individually under controlled temperature (23°C) and light conditions (12:12-h light-dark cycle). At the time of weaning, mice received commercially available rodent chow (Harlan Teklad, Madison, WI) and consumed standard water [control (Con) mice, n = 12] or water supplemented with 0.5% l-methionine (HHcy mice, n = 12) for 18 ± 3 wk. The diet contained 0.75 mg folate/100 g chow. Furthermore, the antibiotic succinylsulfathiazole (1%) was included by the manufacturer to eradicate intestinal microflora that are capable of synthesizing folate endogenously.
After an overnight fast, mice were anesthetized (0.05 mg/g pentobarbital sodium ip), the chest was opened, and blood was collected from the right ventricle to measure plasma total homocysteine (tHcy). Plasma tHcy was measured using HPLC with fluorescence detection (7, 37) and refers to the combination of free reduced homocysteine (∼1% of total), mixed disulfides (20–30% of total), and protein-bound homocysteine (70–80% of total) (22). Next, samples of the liver (for folate analysis), thoracic aorta (to measure arterial elasticity), and abdominal aorta (for hydroxyproline quantification) were removed, snap frozen in liquid nitrogen, and stored at −80°C. The right and left common carotid arteries were excised and prepared immediately to assess vascular permeability.
Liver folate was measured using a conventional microbiological assay (37, 40) because it is more indicative of long-term folate status and is less susceptible to fluctuations in metabolism than serum folate (4, 6, 37). Hydroxyproline was quantified to determine collagen content using a mini-Woessner microtiter plate method (36). The concentration of hydroxyproline was normalized to vessel length (2, 37).
Measurement of arterial permeability.
Carotid arteries were placed in a microscope viewing chamber containing Krebs-Henseleit (KH) buffer (pH 7.4, 37°C). The proximal and distal ends of each artery were cannulated and perfused with 1% BSA at 2 ml/min (pH 7.4, 37°C) for 15 min using a peristaltic pump. During this equilibration period, both arteries were viewed under a fluorescence microscope connected to a photometer and video camera while fluorescence intensity was recorded on a computer and chart recorder. The assessment of vascular permeability using quantitative fluorescence microscopy involved three perfusion phases. In the first phase, the artery was perfused (2 ml/min, 100 mmHg, 37°C, pH 7.4, 8 min) with clear, nonfluorescent KH buffer + 1% BSA to measure baseline fluorescence intensity. Second, 4,400-molecular-mass dextran molecules (estimated Stokes diameter, 1.4 nm; 42 μg/ml in perfusate) labeled with tetramethylrhodamine isothiocyanate (TRITC; 0.0833 mg/ml TRITC-dextran in perfusate, 494-nm excitation maximum and 518-nm emission maximum) were perfused through the arterial lumen for 5 min and viewed/recorded through an inverted light microscope. Dextran was used as the reference molecule because this nonlipid particle does not bind specifically to the artery wall. During the second phase, a rapid increase in intraluminal fluorescence intensity occurred as TRITC-labeled dextran filled the artery lumen. In the third phase, the artery was perfused for 8 min with nonfluorescent buffer to wash the TRITC-labeled dextran out of the lumen. The three phases were collectively termed as a perfusion “run.”
The washout phase (i.e., phase 3) was analyzed as two distinct processes. The first, rapid washout represents dextran exiting the vessel lumen, whereas the second, slower washout represents dextran exiting the vessel wall. Arterial permeability was estimated by the amount of TRITC-labeled dextran that accumulated in the arterial wall (If accumulation). Calculating If accumulation involved finding the intersection of tangents drawn to approximate the rapid and slow washout phases. To determine If accumulation rate, If accumulation was divided by the time of perfusion. Fluorescence values were then converted from millivolts per minute to nanograms of TRITC-dextran per centimeter squared per minute with the knowledge of the 1) surface area of the vessel in the photometric window; 2) arterial lumen volume of the vessel in the photometric window; 3) If at time 0, which occurred at the beginning of the TRITC perfusion; and 4) concentration of dextran in the perfusate. If accumulation rates were performed in triplicate for each vessel, and the values were averaged (5, 24–26, 37, 39, 46, 47).
In addition to perfusion studies performed on mice after the dietary interventions, five acute experiments were performed using age-matched C57BL/6J mice. First, carotid arteries were prepared for quantitative fluorescence microscopy, three perfusion runs were performed, and the results were averaged. Next, vessels were incubated for either 120 min with 1) 50 μM dl-homocysteine (n = 10); 2) 50 μM dl-homocysteine plus the nitric oxide (NO) donor diethylamine-NONOate (DEANO; 300 μM, n = 7); 3) 50 μM dl-homocysteine plus the nonenzymatic intracellular superoxide anion (O2−) scavenger 4,5-dihydroxy-1,3-benzene disulfonic acid (tiron; 10−3 M, n = 2); or 4) 50 μM dl-homocysteine plus vehicle/volume (n = 4) or for 5 min with 5) xanthine (0.4 mg/ml)/xanthine oxidase (X/XO; 0.2 mg/ml, n = 7) to generate O2− (17). After the respective incubation period, three perfusion runs were performed, and the results were averaged.
Measurement of vascular stiffening.
Vascular stiffening was estimated using a modified vessel myograph, termed an elastigraph (37). After vessels were thawed overnight, two stainless steel rods were inserted in a parallel manner through the lumen of a 1-mm segment of the thoracic aorta while the vessel was immersed in KH buffer. One rod was fixed to a force transducer, whereas the other was attached to a motorized controller. The elastigraph allowed the vessel to be stretched radially at a constant rate until breakage while vessel tension was recorded via a force transducer. In preparation for each stretch, aortic segments were preconditioned three times at ∼10% of their maximal load (maximal tension at the vessel breaking point). Stress (vessel tension development/vessel area, N/mm2) versus strain [(vessel width at breakage − vessel width at start)/vessel width at start, %] curves were generated using three 1-mm aortic segments from each animal, and the results were averaged. From these curves, the following parameters were calculated: 1) maximal stress, which was the stress at vessel breakage (N/mm2); 2) maximal strain, which was the strain at vessel breakage (ultimate extensibility, %); 3) physiological stiffness, which was the slope of the linear region between 30% and 40% strain of a stress versus strain curve (low-strain Young's modulus, mN/mm); and 4) maximal stiffness, which was the slope of the linear region at high load values of a stress versus strain curve (high-strain Young's modulus, N/mm) (2, 34, 37, 41).
Drugs and solutions.
Unless noted otherwise, all chemicals were purchased from Sigma Chemical (St. Louis, MO). KH solution contained (in mmol/l) 116 NaCl, 5 KCl, 2.4 CaCl2·H2O, 1.2 MgCl2, 1.2 NH2PO4, and 11 glucose with 1% BSA.
Animal and vessel characteristics, dextran accumulation rate, and indexes of vascular stiffening were compared between Con and HHcy mice using an unpaired Student's t-test. For acute studies of arterial permeability, dextran accumulation rates before and after each respective incubation period were compared using a paired t-test. Results are presented as means ± SE. Statistical significance was accepted at P < 0.05.
Data from Con mice have been published previously (39) but were collected and analyzed concurrently with HHcy animals. Body weights were similar between Con (24 ± 1 g) and HHcy (25 ± 1 g) mice. Plasma tHcy was elevated by approximately fivefold in HHcy versus Con animals, but liver folate was not different between groups (Fig. 2, A and B). These results allowed us to evaluate the contribution from HHcy to arterial permeability and stiffness in the absence of folate deficiency.
Carotid arteries from HHcy mice were more permeable than those from Con animals. In this regard, dextran accumulation was ∼38% greater in HHcy versus Con mice (Fig. 3). This experiment indicates that HHcy has a primary effect to injure endothelial layer function.
Stress versus strain curves generated using aortas indicated that vessels from HHcy mice were stiffer than those from Con animals. Specifically, whereas the maximal strain (ultimate extensibility) was similar between groups, arterial breakage occurred at approximately fourfold higher maximum stress in HHcy versus Con mice (Fig. 4, A and B). Increased maximum stress indicates that vessels produced more force at their break point, as can happen with increased intermolecular cross-links in elastic fibers (32). Furthermore, physiological and maximal stiffness were markedly increased in arteries from HHcy versus Con mice, consistent with their increased maximal stress (Fig. 4, C and D). Collagen content was elevated in aortas from HHcy (49 ± 3 hydroxyproline/cm) versus Con (37 ± 4 hydroxyproline/cm) mice.
Dextran accumulation in carotid arteries increased from 2.94 ± 0.36 to 4.87 ± 0.48 ng·min−1·cm−2 after the exposure to X/XO and from 1.80 ± 0.19 to 4.02 ± 0.20 ng·min−1·cm−2 after an incubation with dl-homocysteine. When xanthine or xanthine oxidase were added independently to the vessel perfusate, they increased neither arterial permeability nor O2− production (46). Both the NO donor DEANO and the O2− scavenger tiron significantly limited increases in arterial permeability resulting from the incubation with dl-homocysteine. Specifically, the simultaneous incubation of dl-homocysteine plus DEANO blunted increases in dextran accumulation (3.29 ± 0.3 ng·min−1·cm−2) versus Con (2.39 ± 0.29 ng·min−1·cm−2). Likewise, a concomitant incubation of dl-homocysteine + tiron attenuated increases in dextran accumulation (2.04 ± 0.44 ng·min−1·cm−2) versus Con (1.56 ± 0.34 ng·min−1·cm−2). Figure 5 shows the percent increase from preincubation values after an incubation with X/XO, dl-homocysteine, dl-homocysteine plus DEANO, or dl-homocysteine plus tiron.
Internal control experiments measuring the time and volume effects of the perfusion system demonstrated no changes in arterial permeability. For example, the average dextran accumulation from three perfusion runs before (2.48 ± 0.22 ng·min−1·cm−2) and after (2.41 ± 0.14 ng·min−1·cm−2) a 120-min incubation with a vehicle/volume varied by 3 ± 7%.
Findings from this study support the hypothesis that arterial permeability and stiffness increase in response to chronic elevations of tHcy, even in the absence of folate deficiency. Results from acute experiments suggest that HHcy exerts its deleterious effects on arterial permeability by generating O2− to an extent whereby NO bioavailability is reduced.
Previously, we (37–39) observed that HHcy evoked by folate depletion produced vascular oxidant stress and increased arterial permeability. Subsequently, Kamath et al. (11) compared vascular permeability between heterozygous cystathionine β-synthase (Cbs)-deficient (Cbs+/−) mice that consumed methionine-enriched, low-folate chow for 8 wk (tHcy ∼98 μM) and wild-type (WT) mice fed standard chow (tHcy ∼4 μM). The authors reported greater permeability in the brain but not peripheral (e.g., skin, liver, and spleen) vasculature from Cbs+/− versus WT mice. Because HHcy and low folate coexisted in both studies (11, 37), the relative contribution from each to increasing arterial permeability cannot be discerned. Recently, however, we (39) showed that low folate increased carotid arterial permeability in the absence of HHcy.
To determine the independent contribution from HHcy to increasing arterial permeability, we supplemented the drinking water of mice with l-methionine. Arterial permeability was assessed using real-time measurements of dextran accumulation in the vessel wall via methods whereby flow rate, hydrostatic pressure, pH, temperature, and superfusate and perfusate compositions were controlled to simulate physiological conditions (5, 24–26, 37, 46, 47). We found that arterial permeability increased in response to high physiological/pathophysiological concentrations of tHcy (19, 33), thus suggesting an independent effect of HHcy.
HHcy-induced increases in arterial permeability may involve oxidant stress produced by tumor necrosis factor (TNF)-α. TNF is a proinflammatory cytokine (20) that elevates O2− in endothelial cells (27, 48) and vascular smooth muscle cells (21). Previously, we (46) showed that acute exposure to TNF increased the permeability of isolated perfused vessels and that this effect was attenuated by the antioxidants estradiol or α-tocopherol. Recently, Ungvari et al. (43) reported greater TNF expression in arteries from rats that consumed methionine-supplemented drinking water to produce HHcy. In that study (43), TNF increased O2− production in vessels from HHcy animals by upregulating nox1-based NAD(P)H oxidase in arterial smooth muscle.
Results from our acute experiments have also implicated O2− as a potential mechanism for elevated arterial permeability in response to HHcy. First, dextran accumulation was ∼70% greater in carotid arteries after the incubation with X/XO, a well-known system for O2− generation (17, 46). Second, the incubation with 50 μM homocysteine, a dose similar to the concentration (i.e., 41 ± 5 μM) experienced chronically by HHcy mice, elevated dextran accumulation by ∼120% versus Con values obtained from the same vessel. Third, when carotid arteries were incubated with homocysteine in the presence of a potent, nonenzymatic intracellular O2− scavenger (e.g., tiron) or a NO donor (e.g., DEANO), increases in arterial permeability were attenuated compared with results obtained in response to homocysteine alone. These findings demonstrate that acute (i.e., 120 min) elevations of homocysteine to levels experienced chronically (i.e., ∼18 wk) in HHcy mice are capable of generating O2− to an extent that reduces NO bioavailability and increases arterial permeability. Although these are acute experiments, they do provide a rationale for future studies to test whether increased vascular permeability produced by chronic HHcy can be attenuated by a simultaneous administration of antioxidants or NO donors, or be exacerbated by reducing NO bioavailability.
Although the precise mechanism(s) whereby O2− limits NO bioavailability was not examined in the present study, several possibilities exist. For example, HHcy-induced O2− could react with NO to produce peroxynitrite in the vascular wall (42). Not only does this reaction inactivate NO, but the resultant increase in vascular peroxynitrite has multiple adverse effects. Of these, peroxynitrite can cause 1) tyrosine nitration of the major mitochondrial isoform of SOD and 2) oxidation of tetrahydrobiopterin. The former compromises the antioxidant capacity of the cellular environment, and the latter reduces the activity of endothelial NO synthase (eNOS) and/or promotes eNOS uncoupling, which is a circumstance wherein the normal flow of electrons within the eNOS enzyme is diverted so that superoxide rather than NO is produced (3, 10). Any and/or all of these mechanisms could be operative in the present study.
Arterial stiffness is a risk factor for cardiovascular disease (14) and is observed in patients with atherosclerosis, diabetes, and hypertension (12). tHcy has been found to correlate strongly (1, 29), marginally (35), or not at all (13, 44, 45, 49) with arterial stiffness in experiments that have used a variety of methods to evaluate stiffness and increase tHcy. We evaluated the passive elastic properties of thoracic aortas using a vessel elastigraph modified to measure arterial elasticity (2, 37, 41). This is a sensitive procedure compared with traditional compliance/distensibility methods (2, 31, 34). Maximal vascular stress and maximal and physiological stiffness increased markedly in aortas from HHcy versus Con mice.
The increased collagen content we observed in aortas from HHcy versus Con mice could have contributed to our findings because collagen is ∼1,000 times stiffer than elastin (9, 23). Indeed, Majors et al. (18) reported that homocysteine enhanced collagen synthesis in cultured aortic smooth muscle cells, and Neves et al. (29) showed greater collagen deposition and vascular stiffness in mesenteric arteries from mice with mild elevations (∼8 μmol/l) versus normal concentrations (∼4 μmol/l) of tHcy (29). Furthermore, Mujumdar et al. (23) demonstrated severe elastin breakdown and impaired distensibility in rat aortic rings that were incubated with 100 μM homocysteine for 14 days versus Con arterial segments (23). Increased arterial stiffening could also result from greater collagen deposition combined with the accumulation of advanced glycation end products (AGEs) that occurs during nonenzymatic glycation of elastin or collagen within the vascular wall (32, 34, 50). During this process, oxidant stress stimulates the incorporation of glucose-derived cross-links such as pentosidine between collagen fibers. Collagen cross-linking is one mechanism thought to be responsible for the reduced vascular distensibility in diabetes and atherosclerosis and could be operative in response to HHcy. In this regard, others (8) have reported increased AGE receptor expression in mice with HHcy, and we (37) have shown that pentosidine is elevated 60-fold in arterial tissue from rats with HHcy that possess local and global indexes of increased oxidant stress.
Elevated tHcy is associated with increased risk for cardiovascular disease, but the underlying mechanisms are not well understood (for reviews, see Refs. 15 and 16). In the present study, high physiological/pathophysiological concentrations of tHcy (19, 33) evoked by methionine supplementation increased arterial permeability and stiffness in mice with normal folate status. Because one of the initial steps in the development of atherosclerosis is the accumulation of low-density lipoprotein in the artery wall (28, 30), HHcy could increase permeability, facilitate lipoprotein accumulation, and contribute to the enhanced lesion development that has been reported to occur in vessels from HHcy mice (30). Furthermore, aortic stiffening produced by HHcy could play a role in elevating cardiac afterload to an extent whereby ventricular wall stress and myocardial oxygen demand are elevated inappropriately. Taken together, our studies indicate that HHcy might contribute to cardiovascular disease, at least in part, by increasing arterial permeability and stiffness.
This work was supported by American Heart Association (AHA)-National Affiliate Scientist Development Grant 0130099N; University of California-Davis Clinical Nutrition Research Unit National Institutes of Health (NIH) Pilot and Feasibility Grant DK-35747 (to C. H. Halsted as the Principal Investigator; to J. D. Symons); funds provided by the University of Utah College of Health (to J. D. Symons); NIH Grant HL-55667 (to J. C. Rutledge); NIH Grant HL-63943 and a United States Department of Veterans Affairs grant (to S. R. Lentz); and a University of California-Davis President's Undergraduate Fellowship and an AHA-Western States Affiliate Summer Research Fellowship (to U. B. Zaid).
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.
- Copyright © 2006 the American Physiological Society