Measurement of Endothelium-Dependent Vasodilation in Mice—Brief ReportSignificance
Objective—Endothelium-dependent, flow-mediated vasodilation after an increase in shear stress at the endothelial lining of conduit arteries during reactive hyperemia after ischemia is a fundamental principle of vascular physiology adapting blood flow to demand of supplied tissue. Flow-mediated vasodilation measurements have been performed in human studies and are of diagnostic and prognostic importance, but have been impossible because of technical limitations in transgenic mice to date, although these represent the most frequently used animal model in cardiovascular research.
Approach and Results—Using high-frequency ultrasound, we visualized, quantified, and characterized for the first time endothelium-dependent dilation of the femoral artery after temporal ischemia of the lower part of the hindlimb and demonstrated that the signaling was almost exclusively dependent on stimulation of endothelial nitric oxide synthase, similar to acetylcholine, completely abolished after pharmacological or genetic inhibition of endothelial nitric oxide synthase and endothelial denudation, substantially impaired in mice of increasing age and cholesterol-fed ApoE knock outs and increased by the dietary polyphenol (−)-epicatechin. Intra- and interindividual variability were similar to the human methodology.
Conclusions—The physiology of flow-mediated vasodilation in mice resembles that in humans underscoring the significance of this novel technology to noninvasively, serially, and reliably quantify flow-mediated vasodilation in transgenic mice.
- blood flow velocity
- Doppler duplex ultrasonography
- endothelial nitric oxide synthase
- femoral artery
The vascular endothelium is crucially involved in the fundamental regulation of blood flow matching demand and supply of tissue. After transient ischemia, arterial inflow increases. As a response to increased shear forces during reactive hyperemia, healthy arteries dilate via release of NO or other endothelium-derived vasoactive substances. This endothelium-dependent flow-mediated vasodilation (FMD) is impaired in atherosclerosis.1 As the presence of endothelial dysfunction is closely associated with cardiovascular risk and outcome, the measurement of FMD in the brachial artery has become a standard method for the assessment of endothelial function in patients and to evaluate therapeutic interventions targeting atherosclerosis. To date, this has been impossible in mice, although highly desirable to study mechanisms affecting endothelial function and in particular femoral artery dilation in transgenic mouse models.2 Therefore, we here characterize the physiology of FMD in living mice using a methodology analogous to humans and present important examples of vascular pathologies.
Materials and Methods
Materials and Methods are available in the online-only Data Supplement.
In healthy humans, the relative increase in brachial artery diameter at 45 to 60 sec after resolution of forearm ischemia is typically in the 5% to 10% range. To be able to detect such small diameter changes expected in mice (10–30 μm change at a baseline diameter of ≈200–300 μm) with a high heart rate, we used a high-resolution, high-frequency digital imaging platform and a 30 to 70 MHz linear array microscan transducer (Vevo 2100, VisualSonics) specifically designed for small animal vascular ultrasound. As FMD is temperature-dependent, the body core temperature in anesthetized mice was kept at 37±1°C by using a heated examination table that was also equipped with electrocardiogram electrodes (Figure 1A). The fur was removed from the hindlimbs and prewarmed ultrasound gel was applied in excess to the proximal inner thigh. The ultrasound probe was attached to a stereotactic holder and was manually aligned with the femoral vein visible at the upper inner thigh as the vein runs parallel with the artery, which in turn is not easy to visibly identify. As an initial proof of concept, we used a protocol with a 5 minutes period of total occlusion of the femoral artery to induce reactive hyperemia and monitor changes in vascular hemodynamics (Figure 1B), including flow velocity and diameter after distal hindlimb ischemia as compared with parameters before occlusion essentially as established in the human forearm. A vascular occluder (5 mm diameter, Harvard Apparatus) was placed around the lower limb to induce occlusion of the distal hindlimb as an ischemic trigger. Because of the small diameter of vessels and the anatomic proximity of femoral artery (Figure 1C) and vein, it was key to identify the artery for later semiautomated analysis off-line (Figure 1D). Using colored duplex ultrasound mode, the artery could be easily identified by the typical pulsatile blood flow pattern as opposed to the adjacent vein with continuous blood flow and compressibility. Arterial blood flow was confirmed by pulsed wave Doppler (Figure 1E and F). After 5 minutes of ischemia, deflation of the cuff led to an immediate increase in flow velocity, which in turn went along with an increase in wall shear stress of the upstream conduit artery. This was followed by a significant dilation with a maximum at 60 to 90 sec. Our first experiments showed a similar response pattern of FMD in mice (Figure 1G, femoral artery) as compared with humans (Figure 1H, brachial artery), yet with a greater effect size (human FMD 7% versus mouse FMD 20%).
As depicted in Figure 2A, 2D, 2G, and 2K (black symbols), we observed that increasing times of ischemia lead to a linear increase in the maximum of flow velocity at the onset of reperfusion, indicating dilation of downstream resistance arteries. This in turn led to parallel increases in wall shear stress and consecutively dilation of upstream femoral artery. Interestingly, 1 to 5 minutes ischemia was followed by a temporal increase in diameter with maximal values at 60 to 90 sec, whereas 10 minutes occlusion led to significantly greater dilatory response that was sustained for ≥180 sec. In humans, the rational for using a 5 minutes forearm occlusion protocol is that FMD is almost entirely mediated by endothelial nitric oxide synthase (eNOS),3 and FMD can be seen as an in vivo read-out of eNOS activity depending on the methodology used to measure it.4 As the physiology of FMD in mice was not established, we induced hindlimb ischemia by 1, 3, 5, and 10 minutes lower limb occlusion again after infusion of L-NAME, a competitive NOS inhibitor (Figure 2D–2F), in the same mice. Our results show that the flow velocity and immediate postocclusive wall shear stress increase was not affected by L-NAME. However, the vasodilator response of the upstream femoral artery of ≤5 minutes of lower limb occlusion led to an almost entirely NOS-dependent vasodilation of the femoral artery ≤60 sec after cuff release. The lack of femoral artery vasodilation was accompanied by significantly slower return of wall shear stress to baseline. Furthermore, we observed that with increasing ischemia times beyond 5 minutes and in the late reperfusion phase, beyond 60 sec after cuff release, NOS-independent vasodilating mechanisms were recruited again, resembling a flow response as observed in human beings (data not shown).
Using a 5 minutes occlusion protocol with 60 sec post occlusion measurement of FMD (Figure 2M and 2L), we further evaluated the signaling pathways involved in FMD. Apart from NO, several other endothelium-derived mediators, such as cytochrome P450 epoxygenase (CYP 2C9), generated endothelium-derived hyperpolarizing factor5 and prostacyclin,6 but also arginase may contribute to maintenance of vascular tone.3 To evaluate these potential other mechanisms in the context of our protocol, we measured FMD before and after injection of inhibitors with a washout phase of 7 days between measurements (Figure 2M). Furthermore, endothelium-independent vasodilation was also measured at 2 minutes after nitroglycerin (Figure 2N). The injection of neither the COX inhibitor indomethacin, the CYP 2C9/endothelium-derived hyperpolarizing factor inhibitor sulfaphenazole, nor the arginase inhibitor nor-NOHA (Nω-hydroxy-L-arginine) affected FMD as measured at 60 sec after ischemia in wild-type mice. FMD was practically abolished in eNOS-knockout (KO) mice, and this was not significantly further decreased by additional NOS inhibition with L-NAME. This implies that eNOS is the predominant NOS isoform responsible for FMD and that in eNOS KO animals compensatorily increased nNOS does not compensate for eNOS KO. Systemic injection of the endothelium-independent vasodilator nitroglycerin, on the contrary, led to a similar significant vasodilation of the femoral artery in both wild-type mice and eNOS-KO mice (Figure 2M), and this response was not altered by any of the inhibitors (data not shown). To further validate our approach, we also performed experiments comparing FMD with a classical endothelium-dependent vasodilator acetylcholine (Figure 2O). Our results showed that intra-aortic injection of acetylcholine led to a dose-dependent, almost instantaneous vasodilation of the femoral artery at a similar magnitude as observed during FMD. These data are in line with the proposition of eNOS being exclusively responsible for femoral artery FMD in the early phase of reactive hyperemia and can therefore be used as an in vivo read-out of eNOS activity in mice.
To provide biochemical insights that eNOS is indeed phosphorylated in this model, we harvested the thigh muscles at the time of peak FMD values (60 sec after 5 minutes of lower limb occlusion) and from the nonischemic contralateral leg as a control (Figure 3A and 3B). Western blot analysis followed by semiquantitative densitometric analysis of X-ray films demonstrate significantly increased phosphorylation on eNOS serine 1178 in ischemic as compared with nonischemic control leg (Figure 3A and 3B). As expected, eNOS protein levels were not altered in the 5 minutes protocol (Figure 3A). To further confirm that the observed FMD responses indeed depend on the arterial endothelium, we performed femoral artery denudation experiments using a wire (Figure 3C). This led to an almost complete abolition of FMD at 1 hour. Together with the fact that the endothelium-independent nitroglycerin mediated vasodilation remained unaffected confirms that the FMD response requires the presence of an arterial endothelium. Interestingly, we observed a partial recovery of FMD at 24 hours, suggesting that this model may also be useful to study endothelial functional recovery during vascular regeneration and reendothelialization.
It is well established that endothelial dysfunction progresses with age.7,8 We compared FMD in young, 3-month-old mice and old methuselah mice (>24 months of age). The aged mice exhibited a markedly decreased FMD by almost 50% as compared with young animals (Figure 4A). Furthermore, our data suggested that after injection of L-NAME, FMD decreased in both young and aged mice to similar low values, suggesting that the main reason for impaired FMD in old mice could be attributed to impaired NO-bioavailability.9
We moved on to study FMD longitudinally during diet-induced accelerated atherosclerosis. Chow-fed apolipoprotein E KO (ApoE KO) mice are known to develop foam cell lesions at 10 weeks of age. This process can be accelerated by Western diet10 and mimics the evolution of atherosclerotic lesions in humans that usually develop over several decades. We followed FMD in 10-weeks-old ApoE KO mice before and during accelerated development of atherosclerosis via Western diet. FMD was measured weekly for a time period of 6 weeks (Figure 4B). Our results showed that in ApoE KO mice, the initiation of Western diet leads to a progressive decrease in FMD. Initial FMD of 24% gradually decreased over time to 12% at week 6. It was previously shown in humans that dietary bioactives can positively affect endothelial function. Flavanols, a subgroup of dietary plant–derived bioactives, have gained increasing attention, as clinical studies have shown that a higher intake of flavanol-containing foods can increase arterial function in individuals at risk for cardiovascular disease and with established cardiovascular disease.11 Flavanols are one of a few bioactives known today, for which causality between the intake and an improvement in arterial function has been demonstrated.12 We demonstrate in our model that intragastric application of (−)-epicatechin (2 mg/kg body weight) but not vehicle acutely increased FMD at 1 hour. Endothelial function is a major determinant of arterial stiffness. Because of the recorded electrocardiogram and Doppler flow recording, it was possible to simultaneously calculate pulse wave velocity. Aged mice exhibited a marked increase in pulse wave velocity compared with young mice (Figure 4D). Furthermore, pulse wave velocity increased significantly in ApoE KO on initiation of the Western diet, indicating progressive arterial stiffness during accelerated atherogenesis, indicating that arteriosclerosis developed in parallel (Figure 4E). Epicatechin, however, led to a significant decrease in pulse wave velocity along with improved FMD values (Figure 4F).To our knowledge, this is the first time that a model is presented, which enables the longitudinal in vivo analysis of the atherosclerotic process over a relatively short time period, allowing to put changes, both positive and negative, in endothelial function and vascular physicomechanics in direct relation with morphological changes in conduit arteries.
To establish the reproducibility of our FMD model, intra- and interobserver variability was assessed by taking repeated measurements in mice.
Mice were either measured twice by the same sonographer or by 2 different sonographers. For intraindividual variability, the mean difference between measurements was 0.4% with a standard deviation of 0.8% and 95% limits of agreement of −1.1% to 2.0% (Figure 5A). The interobserver differences were 0.4% with a standard deviation of 1.3% and 95% limits of agreement of −2.2% to 3.0% (Figure 5B). Thus, this study shows that, presuming optimal sonographer training and constant environmental factors, FMD in mice can be assessed serially in a way that provides good intra- and inter-observer reliability with a low bias. Based on the standard deviation of differences between repeated measurements (intra-individual variation 0.8%) and standard deviation of FMD measurements in wild-type mice (1.4%), we performed a sample size analysis demonstrating the necessary n-value to detect a significant (power 0.8, α-error probability of 0.05) change in FMD in future cross-over and parallel group studies (Figure 5C and 5D).
Taken together, we present the fundamental physiology of FMD in the murine hindlimb using a new technique in mice that is analogous to human FMD measurements, the most widely used prognostically validated measure of endothelial function in humans mainly reflecting eNOS activity. The methodology is reproducible, enables to visualize changes in endothelial function, and to simultaneously investigate involved signaling cascades. Using young and aged wild-type, eNOS KO, and ApoE KO mice, we demonstrate that this approach can be used to assess in vivo vascular dysfunction in a range of genetically altered strains, allowing to noninvasively and longitudinally study pathophysiological processes and may be useful to investigate the potential and mechanisms of new therapeutic interventions.
Sources of Funding
C. Heiss, A. Rodriguez-Mateos, and M. Kelm are senior investigators in the FLAVIOLA research consortium of the European Union (FP7-KBBE-2008-2B). Additional funding was provided by the Deutsche Forschungsgemeinschaft (KE405/5-1, IRTG1902 TP9, and FOR809 TP7 to M. Kelm; IRTG1902 TP1 to J. Altschmied; and TP2 to J. Haendeler). A. Rodriguez-Mateos is funded by the Forschungskommission of the Medical Faculty, University Duesseldorf.
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.114.304699/-/DC1.
- Nonstandard Abbreviations and Acronyms
- endothelial nitric oxide synthase
- flow-mediated vasodilation
- Received June 18, 2014.
- Accepted October 8, 2014.
- © 2014 American Heart Association, Inc.
- Kelm M
- Sidney S,
- Rosamond WD,
- Howard VJ,
- Luepker RV
- Meir KS,
- Leitersdorf E
- Heiss C,
- Keen CL,
- Kelm M
- Schroeter H,
- Heiss C,
- Balzer J,
- Kleinbongard P,
- Keen CL,
- Hollenberg NK,
- Sies H,
- Kwik-Uribe C,
- Schmitz HH,
- Kelm M
Studying endothelium-dependent flow-mediated vasodilation has been impossible in transgenic mice to date. Using high frequency ultrasound, we reliably measured diameter changes of the femoral artery after ischemia in vivo. This was dependent on stimulation of endothelial nitric oxide synthase, completely abolished after pharmacological or genetic inhibition of endothelial nitric oxide synthase, similar to classical acetylcholine-mediated vasodilation, is endothelium-dependent, substantially impaired in aged mice and in ApoE knockout mice fed a high-fat Western-type diet, and improved by the polyphenol (−)-epicatechin. The flow-mediated vasodilation response in mice resembles that in humans. This novel technique allows for the first time to noninvasively, serially, and reliably quantify flow-mediated vasodilation in mice.