Detailed evaluation of coronary function early in diabetes mellitus (DM)-associated coronary artery disease (CAD) development is difficult in patients. Therefore, we investigated coronary conduit and small artery function in a preatherosclerotic DM porcine model with type 2 characteristics. Streptozotocin-induced DM pigs on a saturated fat/cholesterol (SFC) diet (SFC + DM) were compared with control pigs on SFC and standard (control) diets. SFC + DM pigs showed DM-associated metabolic alterations and early atherosclerosis development in the aorta. Endothelium-dependent vasodilation to bradykinin (BK), with or without blockade of nitric oxide (NO) synthase, endothelium-independent vasodilation to an exogenous NO-donor (S-nitroso-N-acetylpenicillamine), and vasoconstriction to endothelin (ET)-1 with blockade of receptor subtypes, were assessed in vitro. Small coronary arteries, but not conduit vessels, showed functional alterations including impaired BK-induced vasodilatation due to loss of NO (P < 0.01 vs. SFC and control) and reduced vasoconstriction to ET-1 (P < 0.01 vs. SFC and control), due to a decreased ETA receptor dominance. Other vasomotor responses were unaltered. In conclusion, this model demonstrates specific coronary microvascular alterations with regard to NO and ET-1 systems in the process of early atherosclerosis in DM. In particular, the altered ET-1 system correlated with hyperglycemia in atherogenic conditions, emphasizing the importance of this system in DM-associated CAD development.
- coronary circulation
- diabetes mellitus
- endothelial function
diabetes mellitus (dm) is widespread in industrialized countries, and the incidence of type 2 is rapidly increasing worldwide. DM is an independent and strong predictor of coronary artery disease (CAD) (26), characterized by atherosclerosis of the conduit arteries (16, 23) and dysfunction of the microcirculation (51). Endothelial dysfunction is an important determinant of altered vascular reactivity and plays a major role in the genesis of DM-induced macro- and microvascular complications (51). Impaired coronary vasodilation, which is present well before the development of angiographically visible atherosclerosis in DM patients, is known to be at least partially due to impairment of the nitric oxide (NO) system (45). On the other hand, it is less clear whether and how the vasoconstricting endothelin (ET)-1 system is affected. Evidence of an altered ET-1 system in DM patients is predominantly coming from studies in the peripheral circulation in advanced stages of disease (30), and details on coronary alterations in relation to ET-1 in DM subjects are scarce (48, 62), in particular in early CAD development. Moreover, the pathological significance of the ET-1 system is still not completely established and is complicated by interactions with other vasoregulatory systems such as the NO system (57). Given that the regulation of endothelial function may vary in different vascular beds, additional research on alterations in NO and ET-1 systems in the coronary circulation during early CAD development is needed (20).
Knowledge of early coronary endothelial dysfunction is highly relevant, as it is an independent predictor of cardiovascular events (46), and its reversal might even result in the prevention of events. Such detailed evaluation is difficult to achieve in humans, since the precise onset of DM and CAD is usually unknown. Therefore, suitable animal models focusing on early disease development are necessary. Because of the highly similar anatomy and physiology of the cardiovascular system, porcine models with toxin-mediated pancreatic damage in combination with an atherogenic diet have been previously used to study DM and cardiovascular complications (16, 17, 23, 37, 41, 42, 63). However, most of these models have focused on advanced CAD with clear presence of atherosclerotic plaque.
In light of these considerations, the present study was undertaken to study coronary function early in the process of atherosclerosis development in a DM type 2-like porcine model (35) with focus on endothelial function via NO and ET-1 systems. For this purpose, we examined vasomotion in vitro of both coronary conduit and coronary small arteries of DM pigs after 10 wk on a saturated fat/cholesterol diet (SFC + DM) compared with control pigs on the same SFC diet (SFC) and to controls on a standard low fat diet (control). Because a previous study (34) indicated that there were minimal effects of DM combined with unsaturated fat or starch diets on plasma cholesterol and central atherosclerosis development after 10 wk, a group of DM pigs on the standard diet was omitted.
MATERIALS AND METHODS
All experiments were performed in accordance with the American Physiological Society's Guiding Principles in the Care and Use of Vertebrate Animals in Research and Training and with approval of the Local Animal Ethics Committee of Lelystad, The Netherlands.
Before the start of the study period of 10 wk, 16 male crossbred pigs (Yorkshire × Landrace) were housed in three groups. Initially, the first group (control group, n = 5) was fed a standard diet for growing pigs, and the second and third groups (SFC, n = 5; and SFC + DM, n = 6) were fed a diet complemented with 25% of saturated fats and 1% of cholesterol (SFC diet). All pigs started the prestudy phase at ∼30 kg of body wt and were fed ad libitum with free access to water. At ∼70 kg of body wt, the pigs were individually housed in metabolism cages and fed two meals a day, during which the pigs had ad libitum access to food for 1 h. The food in the study period consisted of control diet in the case of the control pigs and of SFC diet in the case of the SFC and the SFC + DM pigs. Details of the composition of the study diets are given in Table 1.
In the first study week, pigs were anesthetized with intramuscular 2 mg/kg azaperone (Stressnil; Janssen, Tilburg, The Netherlands), followed by intravenous 15 mg/kg thiopental (Nesdonal; Rhone Merieux, Lyon, France). A permanent blood vessel catheter (Secalon Seldy, 16-G polyurethane; Becton Dickinson, Franklin Lakes, NJ) was inserted in the ear vein and fixed firmly to the ear. DM was induced in pigs of the SFC + DM group by slow titrated injection of streptozotocin (STZ; 80 mg/kg; Pharmacia and Upjohn, Kalamazoo, MI) in the ear vein to produce subtotal destruction of pancreatic β-cells, as described previously (34, 35). The following STZ injections (respectively on days 3, 5, and 8 after the first dose) depended on the level of urinary glucose excretion (total STZ dose per pig ranged from 110 to 150 mg/kg), as we know from experience that urinary glucose excretion of ∼400 g/day corresponded with a fasting plasma glucose concentration of 15–20 mmol/l. In the week of DM induction, the SFC + DM pigs had free access to food all day to avoid hypoglycemia. After 1 wk of DM settlement, the SFC + DM pigs were also placed on the twice daily feeding regime, as outlined above, until the end of the study.
In the seventh week of the study period, all pigs were sedated with 2 mg/kg of intramuscular azaperone (Stressnil), and anesthetized with 15 mg/kg of intravenous thiopental (Nesdonal). Pigs were intubated and anesthesia was maintained by inhalation of 3% sevoflurane combined with 40% oxygen and room air. All pigs were equipped with two permanent polyethylene blood vessel catheters (Tygon; Norton, Akron, OH), which were placed in the external jugular vein and in the carotid artery for blood sampling procedures and hemodynamic measurements during the study. The catheters were fixed firmly at the site of insertion, tunneled subcutaneously to the back of the pig and exteriorized between the shoulder blades. The catheters were filled and sealed off with saline containing heparin and penicillin (Procpen, Cuijk, The Netherlands) and kept in a backpack and glued to the skin. In the first week after surgery, pigs were habituated to stress-free blood sampling. In the eighth week of the study period, fasting blood samples for determination of glucose, fructosamine, insulin, β-hydroxybutyrate, nonesterified fatty acids, triglycerides, total cholesterol, very-low-density lipoprotein, low-density lipoprotein, high-density lipoprotein, tumor necrosis factor-α, and markers of liver function (aspartate aminotransferase) and kidney function (creatinine) were collected and analyzed. Moreover, at this time point a meal tolerance test was performed to assess DM metabolic characteristics. Blood was sampled repeatedly before, during, and up to 8 h after a morning meal. Per sampling time point, small amounts of blood (5 ml) were collected to monitor the responses of glucose, insulin, and triglycerides to the meal. Subsequently, blood pressure and heart rate were determined 5 h postprandially via the carotid artery using a digital electro-manometer (Type 330; Hugo Sachs Elektronik, March-Hugstetten, Germany). At the end of the study (after 10 wk), blood samples were collected for determination of plasma concentrations of ET-1. Then, all pigs were killed stress free by intravenous injection of an overdose of pentobarbital via the jugular vein catheter. Hearts were immediately excised and placed in cold, oxygenated Krebs bicarbonate buffer solution of the following composition (10−3 mol/l): 118 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 25 NaHCO3, and 8.3 glucose, pH 7.4. Coronary conduit and small arteries were isolated from the territory of the left anterior descending coronary artery for in vitro functional studies (5, 6). Furthermore, specimens of left anterior descending coronary artery and related coronary small arteries, liver, pancreas, and abdominal aorta were fixed for histology.
Segments of conduit arteries (∼4 mm length) were suspended in large organ baths containing Krebs bicarbonate solution aerated with 95% O2-5% CO2 (435, 6). Both vessel types were subjected to the same protocol with vascular responses reflected in changes in isometric force as recorded with a Harvard isometric transducer. Following a 30-min stabilization period, the internal diameter was set to a tension equivalent to 0.9 times the estimated diameter at 100 mmHg of effective transmural pressure (27). After this normalization procedure, segments were exposed to depolarization by 10−1 mol/l potassium chloride (KCl; Sigma-Aldrich, Zwijndrecht, The Netherlands) to determine maximal contractile responses. Upon washout, in a first set of segments, endothelium-dependent vasodilation to bradykinin (BK; 10−10 to 10−6 mol/l; Sigma-Aldrich) was recorded upon preconstriction with the thromboxane-A2 analog U46619 (10−7 mol/l). To quantify contributions of non-NO vasodilator substances to BK-induced dilation, the concentration-response curve (CRC) to BK (10−10 to 10−6 mol/l) was constructed upon 30 min of preincubation with 10−4 mol/l of the NO-synthase inhibitor N-nitro-l-arginine methyl ester HCl (l-NAME; Sigma-Aldrich) in a second set of experiments. CRCs to ET-1 (10−10 to 10−7 mol/l; Sigma-Aldrich) were constructed from a third set of segments. In this set, endothelium-independent but NO-mediated vasodilation to the exogenous NO-donor S-nitroso-N-acetylpenicillamine (SNAP; 10−7 and 10−6 mol/l; Sigma-Aldrich) was also examined. To investigate the specific contributions of ETA and ETB receptors to the ET-1 response, CRCs to ET-1 (10−10 to 10−7 mol/l) were constructed after 30 min of preincubation with 10−6 mol/l ETA receptor antagonist BQ123 (Sigma-Aldrich) or 10−8 mol/l ETB receptor antagonist BQ788 (Sigma-Aldrich), as described previously (59).
Baseline diameters of the coronary small arteries of the functional experiments could not be accurately assessed because segments were isolated under nonpressurized conditions. However, using the measured optimal tension and point of passive stretch according to the Laplace formula obtained during the normalization procedure, the passive baseline inner-diameters were calculated (2, 27). To examine whether the small arteries had undergone changes in their passive tension characteristics, the values of the relaxed steady-state diameters at each wall tension of the normalization procedure (n = 24–29 vessels/group) were fitted by an exponential equation according to Halpern et al. (27).
Protein expression of ETA and ETB receptors was examined by Western blotting. Lysates of segments of coronary small arteries were loaded on gel and blotted. ETA and ETB expression was determined using specific primary antibodies (1:1000; sc-21193 and sc-21199; Santa Cruz Biotechnology, Heidelberg, Germany) and corrected for β-actin expression (1:100,000; A5441 Sigma-Aldrich).
Fresh sections of coronary conduit and small arteries, pancreas, liver, and abdominal aorta were fixed in 4% buffered formaldehyde for further histological analysis. Sections of paraffin-embedded coronary arteries were stained with resorcin fuchsin stain and hematoxylin eosin stain to assess the presence of intimal hyperplasia and vascular cell nuclei, respectively. In sections of pancreatic tissue, insulin content was assessed using a primary antibody against insulin (Sigma-Aldrich). Staining was visualized using a horseradish peroxidase-labeled secondary antibody and diaminobenzidine as a chromogen (DakoPatts, Amsterdam, The Netherlands). Insulin stained areas were measured using Clemex Vision PE (Clemex Technologies, Longueuil, Quebec, Canada) and are presented as a percentage of total analyzed area. Sections of liver tissue were stained with hematoxylin eosin stain. Lobular steatosis was scored (0 = fat accumulation in 0–5% of the hepatocytes per lobule; 1 = 5–33%; 2 = 33–66%; and 3 = >66%; Ref. 14). In the formaldehyde-fixed abdominal aorta, the amount of fatty streaks as a marker of early atherosclerosis development (AHA type II lesion; Ref. 60) was quantified as percentage of total area after Sudan IV fat staining (Sigma-Aldrich; Ref. 17).
Comparison of normally distributed model characteristics was performed by one-way ANOVA. For skewed data, median values with interquartile ranges were calculated and the Kruskal Wallis test was used. Vasodilator responses were expressed as percentage of preconstriction to U46619 (BK) or ET-1 (SNAP). Vasoconstrictor responses to U46619, l-NAME and ET-1 were normalized to 10−1 mol/l KCl. Statistical analysis of CRCs was performed using two-way ANOVA for repeated measures, followed by Bonferroni's post hoc correction. For each CRC reaching a plateau, the concentration necessary to produce 50% of its maximal response was determined using logistic function (5, 6). The maximal response (Emax) of a vasoactive substance was assessed when appropriate. Associations between univariate data were evaluated by either Pearson or Spearman rank correlation. StatView 5.0 (SAS Institute) and SPSS 17.0 (IBM) were used for the analyses. Data are given as means ± SE or as median with interquartile ranges. Two-tailed P < 0.05 was considered statistically significant.
Gross energy intake was highest in SFC pigs, intermediate in control pigs, and lowest in SFC + DM pigs (54 ± 3.9 vs. 48 ± 1.7 vs. 44 ± 4.9 MJ/day; P < 0.05) without differences in hemodynamic parameters (Table 2). Although SFC + DM pigs had the lowest intake, they had a positive energy balance during the whole study period with no signs of urinary ketone body excretion. Therefore, the SFC + DM pigs did not show severe body wasting or signs of dehydration and did not require insulin therapy. In line with the lower levels of food intake, bodyweight at death was lowest in the SFC + DM group (P < 0.05). In contrast, relative liver weight of these pigs was significantly higher, with pronounced steatosis on histology (P < 0.01). Moreover, the abdominal aorta of SFC + DM pigs showed augmented fat staining (P < 0.02). The percentage of fatty streaks was increased >10-fold compared with SFC pigs, while control pigs demonstrated little fat staining. These fatty streaks consisted of several intimal cell layers with fat deposits (data not shown). In contrast, intimal hyperplasia was not observed in either coronary conduit or small arteries in any of the groups (Fig. 1). In summary, SFC + DM pigs showed a positive energy balance without insulin therapy, pronounced liver steatosis, and early atherosclerosis development only in the central aorta.
Plasma analysis showed significantly increased fasting glucose and fructosamine levels in SFC + DM pigs (P < 0.01), indicative of elevated glucose levels over the preceding week (Table 2). In contrast, fasting insulin concentrations were not altered. Therefore, a DM state was created with no absolute insulin deficiency. However, the ketone body β-hydroxybutyrate (which was below detection limit in control and SFC pigs) was significantly increased in SFC + DM pigs (P < 0.02), indicating a functional shortage of insulin in these animals. Indeed, immunohistochemistry confirmed a reduction in pancreatic insulin producing β-cells (P < 0.01). Also, nonesterified fatty acids levels were elevated in SFC + DM pigs (P < 0.04), again indicative of reduced insulin functionality. Although triglyceride levels tended to be highest in SFC + DM pigs, statistical significance was not reached. SFC + DM pigs also showed the most profound changes in lipid profile, characterized by significantly elevated total cholesterol and very low density lipoprotein levels and a reduced low-density lipoprotein-to-high-density lipoprotein ratio (all P < 0.02). Furthermore, these pigs showed a trend towards an increased inflammatory response with tumor necrosis factor-α (P = 0.07). In addition, the meal tolerance test showed sustained elevated postprandial plasma glucose levels in SFC + DM pigs together with a markedly reduced postprandial insulin response and a tendency toward an elevated triglyceride response as shown in Fig. 2. Taken together, SFC + DM pigs showed a markedly increased set point for plasma glucose, postprandial hypoinsulinemia with evidence of reduced insulin functionality, and dyslipidemia. No signs of liver or renal toxicity were observed between control, SFC and SFC + DM pigs (aspartate aminotransferase: 19 ± 1.6 vs. 19 ± 1.5 vs. 18 ± 2.0 U/l, P = NS; creatinine: 117 ± 5.0 vs. 124 ± 8.5 vs. 101 ± 6.4 10−6 mol/l, P = NS). Therefore, DM was safely induced, as previously described (28).
Coronary Vasomotor Function
SFC + DM and SFC groups showed similar baseline contractile responses to KCl and preconstriction to U46619 (both, P = NS). Segments from both groups dilated similarly to BK in a concentration-dependent manner (pEC50 SFC + DM vs. SFC: 7.9 ± 0.2 vs. 7.5 ± 0.3). Relaxations amounted up to 50% of the preconstriction level (Fig. 3A). Pretreatment with l-NAME almost completely abolished dilation to BK in both groups (P < 0.01, CRC l-NAME vs. no l-NAME), so that the effect of l-NAME was not different between both groups (Fig. 3B). ET-1 produced vasoconstriction in a concentration-dependent manner up to 125% of the response to KCl (pEC50 SFC + DM vs. SFC: 7.7 ± 0.2 vs. 7.5 ± 0.1) that was similar in both groups (Fig. 3C). Finally, the exogenous NO-donor SNAP induced identical vasodilator responses in both groups (Fig. 3D). These data are highly similar to those previously observed in control pigs (5).
The passive baseline inner-diameters of the coronary small arteries were similar (SFC + DM vs. SFC vs. control: 364 ± 19 vs. 333 ± 26 vs. 337 ± 19 μm; P = NS). Moreover, the small arteries showed similar baseline contractile responses to KCl and preconstriction to U46619, indicating equal vasoconstrictor potential between the groups (KCl SFC + DM vs. SFC vs. control: 3.6 ± 0.4 vs. 2.8 ± 0.5 vs. 2.8 ± 0.3 g; U46619: 2.5 ± 0.4 vs. 1.9 ± 0.3 vs. 2.4 ± 0.4 g; both, P = NS). Small arteries dilated to BK in a concentration-dependent manner in all groups. Relaxations amounted up to 90% of the preconstriction level but differed significantly between the groups (Fig. 4A). SFC + DM pigs showed a right-ward shift of the BK CRC (pEC50 SFC + DM vs. SFC vs. control: 8.0 ± 0.3 vs. 8.6 ± 0.2 vs. 8.5 ± 0.3; P < 0.01 CRC SFC + DM vs. SFC and control) without altering its maximal effect. After preincubation with l-NAME, the BK CRC of all groups was shifted to the right (P < 0.01, CRC l-NAME vs. no l-NAME), and the three curves were no longer significantly different. These findings indicate that the non-NO component of BK-induced vasodilation was identical in all three groups (Fig. 4B) and that the diminished response to BK in the absence of l-NAME was due to decreased NO production rather than a decrease in non-NO vasodilating substances, such as prostaglandins and endothelium-derived hyperpolarizing factors. The effect of l-NAME in the small arteries was more modest than in the conduit arteries, indicating a smaller NO component and a more pronounced non-NO component of the BK response, in agreement with previous results (5, 6). ET-1 induced vasoconstriction in a concentration-dependent manner (Fig. 4C). However, only SFC + DM pigs showed a clear downward shift of the ET-1 CRC (P < 0.01, SFC + DM vs. SFC vs. control), with a greatly diminished maximal response to the highest concentration of ET-1 (Emax: SFC + DM vs. SFC vs. control: 53 ± 16 vs. 136 ± 19 vs. 109 ± 8; P < 0.05). The exogenous NO-donor SNAP induced similar vasodilator responses in all groups, indicating intact vascular smooth muscle cell function (Fig. 4D).
ET receptor subtypes.
In both SFC and control groups, ETA receptor blockade, but not ETB receptor blockade, attenuated the vasoconstrictor responses to ET-1 (P < 0.05 CRC ETA vs. ET-1 and ETB), indicating a predominant role for ETA receptors in response to exogenous ET-1 (Fig. 5). However, for the SFC + DM pigs only a trend for ETA receptor contribution could be noted (P = 0.07 CRC ETA vs. ET-1), indicating a loss of ETA dominance. Systemic ET-1 plasma levels were not different between the groups [SFC + DM vs. SFC vs. control: 1.26 (1.08–1.35) vs. 1.10 (0.96–1.28) vs. 1.07 (0.93–1.29) pg/ml; P = NS].
ET receptor expression.
To examine whether altered expression of ETA and ETB receptors contributes to the reduced ET-1 responsiveness of small coronary arteries of the SFC + DM pigs, expression of ETA and ETB receptor protein in small arteries of the three groups were compared. No differences in ETA expression were noted between the groups (P = 0.26; Fig. 6A). Similarly, no alterations in ETB expression were observed (P = 0.62; Fig. 6B).
The relaxed wall tension-strain relations of small arterial segments revealed no significant differences in optimal circumference among the three groups. However, the stiffness coefficient β was significantly higher in SFC + DM pigs compared with control pigs (SFC + DM vs. SFC vs. control: 8.1 ± 0.4 vs. 7.1 ± 0.4 vs. 6.6 ± 0.4; P < 0.01 SFC + DM vs. control), suggesting an increase in stiffness in the SFC + DM pigs.
Correlation of vascular responses to metabolic alterations.
The small sample size in the present study does not allow firm conclusions regarding correlations between vascular responses in vitro and metabolic abnormalities in vivo and should therefore be considered to be hypothesis generating only. Nevertheless, it is of interest to note that the maximal ET-1 response of coronary small artery function (ET-1 Emax) correlated well with the overall BK responsiveness (BK pEC50; r = 0.53; P < 0.05), i.e., a reduced ET-1 response was associated with a reduced BK response. Moreover, ET-1 Emax correlated with the model characteristic of hyperglycemia: e.g., fructosamine (r = −0.49; P < 0.05), in contrast to BK pEC50 (r = −0.19; P = NS).
The principal findings of this study of coronary function early in the process of diabetic atherosclerosis development are as follows: 1) BK-induced endothelium-dependent vasodilatation was impaired in small coronary arteries of SFC + DM pigs, which was attributable to a loss of NO-mediated vasodilation. 2) Although plasma ET-1 levels were not significantly changed, coronary small vessels of these pigs showed reduced ET-1-induced vasoconstriction, due to a decreased ETA receptor dominance. 3) Moreover, these vessels showed signs of structural changes reflected in an increased stiffness coefficient. 4) In contrast to these observations in small arteries, the coronary conduit arteries of SFC + DM pigs did not show any alterations in vascular function at this stage of the disease.
Coronary endothelial dysfunction plays a key role in the pathogenesis of DM-associated cardiovascular events (46, 51). However, the affected vascular processes are not well understood and difficult to study in humans, hence requiring relevant animal models. Although small animal models are highly relevant to unravel disease mechanisms, their ability to fully mimic human disease is limited (64). Furthermore, while isolated vessels from other vascular beds, such as aorta, basilar and mesenteric arteries (12, 55, 56), or total heart perfusion (8, 61), have been studied in rodents, separate evaluation of coronary conduit vs. small artery function is difficult in small animals. Since human and porcine coronary anatomy and physiology are highly similar (19), a pig model to study CAD development with focus on vascular function appears to be highly relevant.
Coronary vascular alterations have been studied extensively in high-fat-fed prediabetic large animal models, including both dogs (33, 36, 58) and pigs (9, 10, 44), the results of which could be of value for patients with the metabolic syndrome (32). Also, atherosclerotic DM porcine models have been developed and successfully employed (16, 17, 23, 37, 41, 63). However, limited data on coronary endothelial function are available from these DM models and usually only in the presence of advanced CAD. Dixon et al. (17) studied DM minipigs, fed a high-fat, high-cholesterol diet, as early as 12 wk after DM induction and noted an increased percentage of contractile oscillations of coronary conduit arteries with unaffected endothelium-dependent vasodilation. Unfortunately, potential alterations in coronary microvascular function were not studied. Mokelke et al. (42) examined a specific aspect of coronary dysfunction at 20 wk of study duration in a diabetic, dyslipidemic swine model. Differential regulation of potassium currents in both coronary conduit and microvessels was studied, but the development of overall coronary endothelial dysfunction was not assessed. Therefore, information regarding the exact nature of endothelial alterations early in DM-associated coronary atherosclerosis is still incomplete and therefore new, well-characterized DM models continue to be of interest.
Using a recently developed DM type 2-like pig model (34, 35), we created conditions of hyperglycemia, hypoinsulinemia with reduced insulin functionality and dyslipidemia without requiring insulin replacement therapy and without signs of STZ-induced liver or kidney toxicity. These metabolic conditions mimic the human presentation of DM (11), in particular of advanced stage DM type 2 with hypoinsulinemia, dyslipidemia, and secondary pathology including liver steatosis and central atherosclerosis formation. We acknowledge that the model characteristics do not fully meet human DM type 2 or type 1 criteria. However, increased CAD risk caused by micro- and macrovascular complications is seen in both types of human DM. Therefore, classification may be less important than understanding the vascular alterations that mediate this complication (7). In the present DM model, we separately studied in vitro coronary function of both conduit and small sized arteries. The coronary arteries showed no signs of atherosclerosis yet. In addition, endothelium-dependent vasodilation of these arteries was unaltered and we observed no contractile oscillations during preconstruction, which contrasts with the results of Dixon et al. (17). However, we studied vascular function in normal sized pigs after 10 wk instead of 12 wk of study duration. The coronary small arteries showed changes towards increased vascular stiffness, comparable to human hypertensive DM type 2 pathology (54). Of notice, our results indicate that this process already starts before the onset of hypertension (25). Moreover, in these small arteries we confirmed the presence of endothelial dysfunction due to impaired NO-mediated vasodilation, in accordance with observations in the coronary microcirculation of DM patients (47), as well as in the coronary microcirculation (1) and the peripheral circulation (12, 49) of other large animal models of DM. This impaired endothelium-mediated, NO-dependent coronary microvascular dilation might be caused by reduced NO production, although we did not observe differences in total endothelial NO synthase expression in the small coronary arteries of the three groups (unpublished data). Alternatively, endothelial NO synthase uncoupling or NO scavenging, for example, by oxygen-derived free radicals, may have reduced NO bioavailability, as observed shortly after DM induction (1) or after long-term high-fat feeding (36) in dogs.
Interestingly, our results also demonstrate unaltered ET-1 plasma levels together with a reduction in coronary microvascular vasoconstriction to exogenous ET-1. ET-1 plasma levels have been reported to be either unchanged or elevated in DM patients, depending on disease duration and complications (30). The unaltered response to U46619 indicates that a general abnormality in vasoconstriction was not responsible for the reduced ET-1 response that we observed. In addition, evidence for a reduced responsiveness to ET-1 has been shown previously in the peripheral circulation of DM type 2 patients (15, 21, 38, 53, 54). However, such information specifically focusing on the coronary circulation of DM patients is not available. Only altered coronary responses to ET-1 have been reported after ischemia reperfusion in the specific setting of coronary intervention in DM type 2 patients with advanced CAD (22, 48, 62). In agreement with the present findings of control and SFC pigs, unaltered vasoconstriction to ET-1 was found between control and prediabetic dogs on a high-fat diet (33). In addition, unpublished data of our group show unaltered ET-1 responses in small coronary arteries of diabetic pigs fed an unsaturated fat rich diet for 10 wk. Therefore, specifically the combination of DM and an atherogenic diet appears to alter ET-1 responsiveness. In contrast, other DM animal models have reported enhanced ET-1 vascular responsiveness (31, 55, 56), also in the coronary circulation (8, 61). These differences between DM models are not readily explained but may involve differences in vascular beds (55, 56), genetically modified strains (8, 55, 56), and different temporal stages of DM-induction with subsequent other metabolic alterations (31, 61). In the present model, ET-1 reactivity strongly correlated with hyperglycemia in atherogenic conditions, suggesting a mutual interaction between metabolic and vascular aspects. However, future studies are required to investigate these correlations in more detail.
The hyperglycemic state is thought to affect endothelial function via several biochemical pathways (3), unified by Brownlee (13) and Giacco and Brownlee (24). According to this mechanism the production of reactive oxygen species generated by mitochondrial uncoupling causes increased polyol pathway flux, increased intracellular formation of advanced glycation end-products (AGEs), increased expression of the receptor for AGEs and its activating ligands, activation of protein kinase C, and overactivity of the hexosamine pathway (3). High glucose levels also cause the formation of extracellular AGEs, which may affect endothelial function directly or via receptor-mediated mechanisms. In general, AGEs result in altered cellular signaling, promotion of gene expression, release of proinflammatory molecules and again the generation of oxidative stress (4). The results of the present study show evidence of chronic hyperglycemia, inflammation, and coronary endothelium-dependent microvascular dysfunction making the contribution of both reactive oxygen species and AGEs likely. Indeed, AGE-associated oxidative stress has been shown to reduce the bioavailability of endothelium-derived NO in several ways (4), which could result in blunted NO-mediated dilation. In addition, both AGEs (52) and oxidative stress (18) result in increased ET-1 expression, which via increased interstitial ET-1 concentrations and subsequent receptor desensitization could explain the reduced ET-1-mediated microvascular constriction of the present study.
In healthy humans and pigs, ET-1 contributes to basal coronary vascular tone mainly via the ETA receptor (40, 50). In contrast, in our model of early DM-related atherosclerosis development, ETA-receptor-mediated vasoconstriction was markedly reduced, indicating a switch in ET-1 receptor subtype contribution. Since both ETA and ETB receptor expressions were not significantly different between the three groups of pigs, specific ETA receptor desensitization possibly played a role. However, the signaling cascade of ET-1 via the ETA receptor involves several steps, culminating in an increase in intracellular calcium concentrations (29). Hence specific postreceptor signaling mechanisms associated with altered vasoconstrictor responses may be involved. Clinical studies (15, 21, 39) have reported unaltered, enhanced, and decreased ETA receptor mediated responses in the peripheral circulation, which might be explained by differences in disease duration of the populations studied. Also, in animal models differences in receptor expression have been reported. High-fat-fed prediabetic dogs showed a decrease in ETA receptor transcription of both coronary conduit and small arteries, although ET-1 responsiveness was still unaltered (33). An advanced atherosclerotic DM pig model demonstrated a switch in ET-1 receptor contribution with reduced ETA dominance in vascular smooth muscle cells of coronary conduit arteries (37). Moreover, although a DM type 2 mouse model showed increased ET-1 induced coronary vasoconstriction, ET-1 receptor subtype contribution was altered with ETA reactivity depending on intravascular NO availability (8). Also in the present model, coronary small artery alterations in NO and ET-1 systems were associated. Therefore, altered ET-1 responsiveness is likely to be accompanied by altered ET-1 receptor subtype contributions, either by receptor desensitization or at postreceptor level, and reflects the balance between both NO and ET-1 systems within a specific vascular bed.
In conclusion, in the present model of early atherosclerosis and DM, the loss of NO-mediated coronary microvascular dilation may be counteracted by the blunted ET-1 constriction, which may facilitate hyperemia. Indeed, according to the hemodynamic hypothesis of DM microangiopathy (38), an initial increase in microvascular blood flow followed by microvascular stiffness and disturbed autoregulation could finally result in decreased microvascular perfusion and exaggerated DM associated CAD development. Future studies with this DM model are warranted to prove this concept in vitro and to assess its relevance to the coronary vascular bed in vivo.
No conflicts of interest, financial or otherwise, are declared by the author(s).
M.v.d.H., O.S., S.J.K., R.D. and R.d.V. analyzed data; M.v.d.H., O.S., S.J.K., H.M.M.v.B., E.C.E., D.J.D., A.H.J.D., and W.J.v.d.G. interpreted results of experiments; M.v.d.H. and O.S. prepared figures; M.v.d.H. drafted manuscript; M.v.d.H., O.S., S.J.K., E.C.E., D.J.D., A.H.J.D., and W.J.v.d.G. edited and revised manuscript; M.v.d.H., O.S., S.J.K., R.D. and R.d.V. performed experiments; O.S., S.J.K., A.H.J.D., W.J.v.d.G. conception and design of research; D.J.D., A.H.J.D., and W.J.v.d.G. approved final version of manuscript.
- Copyright © 2012 the American Physiological Society