Exercise capacity and cardiac hemodynamic response in female ApoE/LDLR−/− mice: a paradox of preserved V’O2max and exercise capacity despite coronary atherosclerosis

We assessed exercise performance, coronary blood flow and cardiac reserve of female ApoE/LDLR−/− mice with advanced atherosclerosis compared with age-matched, wild-type C57BL6/J mice. Exercise capacity was assessed as whole body maximal oxygen consumption (V’O2max), maximum running velocity (vmax) and maximum distance (DISTmax) during treadmill exercise. Cardiac systolic and diastolic function in basal conditions and in response to dobutamine (mimicking exercise-induced cardiac stress) were assessed by Magnetic Resonance Imaging (MRI) in vivo. Function of coronary circulation was assessed in isolated perfused hearts. In female ApoE/LDLR−/− mice V’O2max, vmax and DISTmax were not impaired as compared with C57BL6/J mice. Cardiac function at rest and systolic and diastolic cardiac reserve were also preserved in female ApoE/LDLR−/− mice as evidenced by preserved fractional area change and similar fall in systolic and end diastolic area after dobutamine. Moreover, endothelium-dependent responses of coronary circulation induced by bradykinin (Bk) and acetylcholine (ACh) were preserved, while endothelium-independent responses induced by NO-donors were augmented in female ApoE/LDLR−/− mice. Basal COX-2-dependent production of 6-keto-PGF1α was increased. Concluding, we suggest that robust compensatory mechanisms in coronary circulation involving PGI2- and NO-pathways may efficiently counterbalance coronary atherosclerosis-induced impairment in V’O2max and exercise capacity.


Results
Exercise capacity of female ApoE/LDLR −/− and C57BL6/J mice. To measure effects of atherosclerosis development on exercise capacity of female ApoE/LDLR −/− mice we measured whole body maximal oxygen consumption (V'O 2max ), maximal velocity (v max ) achieved and maximal distance (DIST max ) covered during incremental test for V'O 2max measurement by young 3-month old female ApoE/LDLR −/− mice with early atherosclerosis and older 6-8-month old female ApoE/LDLR −/− mice with advanced atherosclerosis 26 as compared with age-matched control C57BL6/J mice.
Body mass of young female ApoE/LDLR −/− mice was higher than control C57BL6/J mice while for older female ApoE/LDLR −/− and C57BL6/J mice we found the opposite (Fig. 1A) and it increased with age in both mouse strains. Therefore, the data regarding V'O 2max are presented both in absolute (Fig. 1B) and relative values (Fig. 1C). There was no difference in V'O 2max expressed in relative values between the mouse strains (Fig. 1C). However, when expressed in absolute values, changes in V'O 2max reflected changes in body weight (please compare Fig. 1A,B).
Measurement of blood count revealed that both younger and older female ApoE/LDLR −/− mice had higher number of circulating erythrocytes when compared with age-matched C57BL6/J mice (Table 1). However, number of erythrocytes declined with age in female ApoE/LDLR −/− mice what corresponded with age-dependent decline in v max and DIST max covered by female ApoE/LDLR −/− mice in the incremental test for V'O 2max measurement (Fig. 1D,E).
Since there was no statistically significant difference in Ach-induced changes in coronary flow, Achinduced changes in 6-keto-PGF 1α release and changes in 6-keto-PGF 1α release (basal and after rofecoxib treatment) between 2, 4 and 8-month old C57BL6/J mice, the results for C57BL6/J mice were averaged.

Discussion
In this study we characterized whole body maximal oxygen uptake (V'O 2max ), exercise capacity and NO-and PGI 2 -dependent coronary endothelial function of female ApoE/LDLR −/− mice with respect to age-related peripheral atherosclerosis progression described elsewhere 26 and age-related coronary atherosclerosis progression shown here. We also characterized cardiac haemodynamic response to dobutamine-induced stress in 6-8 month old female ApoE/LDLR −/− mice with advanced coronary atherosclerosis in comparison with age-matched C57BL6/J mice. Surprisingly, we found that V'O 2max considered a key index of cardio-respiratory capacity and running exercise capacity of female ApoE/LDLR −/− mice were preserved irrespective of atherosclerosis progression (Fig. 1). Moreover, both younger (3-month old) female ApoE/LDLR −/− mice and older (6-8-month old) female ApoE/LDLR −/− mice with early and advanced peripheral (as reported earlier 26 ) and early and advanced coronary (as shown here) atherosclerosis achieved higher velocities and covered longer distances than age-matched C57BL6/J mice (Fig. 1D,E). Excellent running exercise capacity of female ApoE/LDLR −/− mice was associated with preserved NO-dependent function of coronary circulation (Fig. 2) and increased COX-2-dependent PGI 2 production (Fig. 4C). Moreover, cardiac performance at rest as well as its inotropic and lusitropic reserve were well preserved in 6-8-month old female ApoE/LDLR −/− mice ( Fig. 6) with advanced coronary atherosclerosis (Fig. 5G-I) compared with age-matched C57BL6/J mice as evidenced by similar FAC, and slice-derived ESV and EDV in basal conditions as well as in response to dobutamine. Altogether, the present work demonstrated that development of atherosclerosis did not compromise V'O 2max of female ApoE/LDLR −/− mice and did not influence their cardiac function even at the advanced stage of atherosclerosis with advanced atherosclerotic plaques in the coronary circulation and even during stress conditions. Running exercise capacity of 6-8-month old female ApoE/LDLR −/− mice was only slightly lower than in young female ApoE/LDLR −/− mice (Fig. 1D,E). Nevertheless, it was still better than in age-matched C57BL6/J mice. We claim that robust compensatory mechanisms in coronary circulation, including increased vascular responsiveness to NO (Fig. 3) and increased generation of PGI 2 (Fig. 4B,C) that are triggered in the early phase of atherosclerosis development, could, at least partly, account for preserved V'O 2max and running exercise capacity of female ApoE/LDLR −/− mice as well as their full cardiac adaptation to exercise occurring even at the stage of advanced atherosclerosis at the age of 6-8 months.
Given that it is widely known that atherosclerosis progression severely impairs cardiac reserve and exercise capacity in humans 4-6 and in atherosclerotic ApoE −/− mice of both genders 11,13,14,17 , our results regarding V'O 2max , running exercise capacity and cardiac function of female ApoE/LDLR −/− mice are surprising. Indeed, exercise capacity of female ApoE/LDLR −/− mice was preserved despite peripheral (as reported previously 26 ) and coronary (Fig. 5) atherosclerosis progression what could be a factor that could limit their exercise capacity 30 . These results can be only explained by important compensatory mechanisms triggered by double knockout of ApoE and Ldlr genes that preserved cardiac haemodynamic response to exercise of female ApoE/LDLR −/− mice (Fig. 6) and assured sufficient adaptation of blood flow and oxygen delivery to their cardiac muscle during exercise which is the major determinant of V'O 2max (for review see 31 ). These compensatory mechanisms could be, at least partly, associated with estrogens known to exert cardioprotective and vasoprotective effects. In fact, it was shown that in female atherosclerotic LDLR −/− mice estrogens up-regulate cyclooxygenase 2 (COX-2) -derived production of PGI 2 via estrogen receptor subtype alpha 32 .
It is also worth noting that oxygen extraction ratio within coronary circulation is already nearly maximal at rest and, thus, the major mechanism involved in increasing oxygen delivery during exercise is based on coronary vasodilation (and increased coronary blood flow), while in the skeletal muscles an increase in oxygen delivery to working muscles is brought about by an increase in its extraction ratio and recruitment of capillaries, as well as increased blood flow 31 . Considering this, it is the coronary and but not the skeletal muscle circulation that represents a locus minoris resistantiae in the regulation of V'O 2max and exercise capacity. Accordingly, compensatory mechanisms aimed to preserve V'O 2max in female ApoE/LDLR −/− mice should be, first of all, present within the ACh-induced 6-keto-PGF 1α release in isolated heart from ApoE/LDLR −/− mice compared with age-matched C57BL6/J mice. Since there was no statistically significant difference in Ach-induced changes of 6-keto-PGF 1α release between 2, 4 and 8-month old C57BL6/J mice, the results for C57BL6/J mice were averaged; n = 15 for C57BL6/J mice, n = 7 for 2-month old ApoE/LDLR −/− mice, n = 9 for 4-month old ApoE/LDLR −/− mice, n = 11 for 8-month old ApoE/LDLR −/− mice. (C) Effects of COX-2 inhibitor (rofecoxib, 10 μM) on basal 6-keto-PGF 1α production in the isolated heart of ApoE/LDLR −/− mice. Since there was no statistically significant difference in changes of 6-keto-PGF 1α release between 2, 4 and 8-month old C57BL6/J mice (basal and after rofecoxib treatment), the results for C57BL6/J mice were averaged; n = 11 and 5 for C57BL6/J mice, n = 8 and 6 for 2-month old ApoE/LDLR −/− mice, n = 9 and 6 for 4-month old ApoE/LDLR −/− mice, n = 9 and 6 for 8-month old ApoE/LDLR −/− mice with respect to basal and after-rofecoxib 6-keto-PGF 1α release. Statistical analysis was performed using t-test or non-parametric Mann-Whitney test (if the assumptions were not completed). *indicates P < 0.05 for ApoE/LDLR −/− vs. control mice. # indicates P < 0.05 with rofecoxib vs. without. Data are presented as the mean ± SEM.
Scientific RepoRts | 6:24714 | DOI: 10.1038/srep24714 coronary circulation to preserve cardiac haemodynamic response to exercise. Indeed, our results fully support this assumption. In line with fully preserved basal cardiac function and stress response to dobutamine (Fig. 6), NO-and PGI 2 -dependent function of coronary circulation in female ApoE/LDLR −/− mice was not compromised (Figs 2,3 and 4). On the contrary, clear-cut compensatory mechanisms in these mice were detected such as increased COX-2-dependent basal PGI 2 production, increased COX-1-dependent PGI 2 release in response to ACh as well as increased coronary vessel responsiveness to exogenous NO most likely linked to upregulation of soluble guanylate cyclase (sGC) activity in coronary vessels. Interestingly, these mechanisms were already evident in hearts of female ApoE/LDLR −/− mice as early as at two months of age, suggesting that these compensatory mechanisms of coronary circulation are triggered before peripheral atherosclerotic plaques development 26 and are probably induced by early coronary atherosclerosis (Fig. 5A-C). Considering vasoprotective and anti-atherosclerotic activity of NO-sGC and COX-2-PGI 2 pathways, it is tempting to speculate that these compensatory changes within coronary circulation could mitigate the development of symptomatic ischaemic heart disease in female ApoE/LDLR −/− mice and explain why their cardiac function was not compromised (Fig. 6) despite extensive coronary atherosclerosis at the age of 6-8 months (Fig. 5G-I).
To counteract detrimental results of peripheral endothelial dysfunction in female ApoE/LDLR −/− mice, compensatory mechanisms induced by atherosclerosis also developed within peripheral circulation 26 . Namely, the cyclooxygenase-2 (COX-2)/PGI 2 pathway was upregulated in the aorta of female ApoE/LDLR −/− mice as early as at two months of age and remained elevated in older animals. Although we did not analyse  endothelium-dependent regulation of blood flow in skeletal muscles, it could well be that exercise-induced increase in blood flow in skeletal muscles was also preserved due to compensatory mechanisms triggered by NO-deficiency in female ApoE/LDLR −/− mice.
The mechanisms controlling coronary and muscle blood flow during exercise are not entirely understood and involve multiple endothelium-dependent vasodilatory pathways compensating one another. NO, PGI 2 and RBC-derived ATP were claimed to be the major mediators involved 31 . It is widely known that NO is an important coronary and peripheral vasodilator during exercise 33 . However, in view of the available data, NO action alone could not explain the huge exercise-induced increase in skeletal muscles 34 and coronary blood flow 35 . There is growing body of evidence that exercise-induced PGI 2 might play an important role in regulation of coronary and muscle blood flow. It was demonstrated that the magnitude of exercise-induced PGI 2 release correlated positively with V'O 2max in healthy humans 36 , while plasma nitrite concentration did not increase even upon vigorous exercise (for review see) 31 . Moreover, training-induced increase in V'O 2max was also associated with increased exercise-induced PGI 2 release 37 . Furthermore, administration of PGI 2 to patients with pulmonary hypertension increased their exercise capacity 38 . Extrapolating from human studies, increased PGI 2 release into coronary (Fig. 4C) and peripheral 26 circulation of female ApoE/LDLR −/− mice could account for their preserved V'O 2max (Fig. 1B,C) and, paradoxically better running exercise capacity (Fig. 1D,E) as well as their preserved basal and dobutamine-stimulated cardiac function (Fig. 6).
Interestingly, there was also a higher number of circulating erythrocytes in female ApoE/LDLR −/− mice compared with age-matched C57BL6/J control mice (Table 1). This could represent yet another mechanism that assured adequate oxygen supply to working tissues. Apart from oxygen transport, haemoglobin inside red blood cells reduces plasma nitrite (NO 2 − ) to NO in hypoxic conditions that is subsequently released and contributes to local vasodilation 31,[39][40][41] . When oxygen tension decreases within coronary microvasculature, red blood cells also release ATP which is a factor increasing coronary blood flow during exercise 42 . Increased number of circulating erythrocytes in female ApoE/LDLR −/− mice points to increased erythropoesis in these animals (probably to counteract hypoxia). Indeed, erythropoietin production was shown to be higher in mice with atherosclerosis 43 and it activated mitochondrial biogenesis to couple red cell mass to mitochondrial mass in the heart 44 . Therefore, even though we claim that upregulation of PGI 2 -dependent pathways within the coronary circulation (Fig. 4B,C) seems to represent the major compensatory mechanism in female ApoE/LDLR −/− mice that contribute to their cardiac haemodynamic adaptation to exercise, many other possible mechanisms (i.e. enhanced RBC-driven ATP release, erythropoietin-mediated mechanisms, mitochondrial biogenesis in the heart, and shift in substrate metabolism in heart and skeletal muscles) could also contribute to their extraordinary exercise capacity. Moreover, higher maximal running velocity found in female ApoE/LDLR −/− mice compared with age-matched C57BL6/J mice was accompanied by similar or even higher V'O 2max in female ApoE/LDLR −/− mice (compare Fig. 1D,B,C, respectively). This could be explained by higher muscle mechanical efficiency (for overview see 45 ) in female ApoE/LDLR −/− mice, similar to in the trained vs. untrained humans (see 46 ) and/or by higher capacity of anaerobic glycolysis and grater fatigue resistance in the locomotor muscles of the female ApoE/LDLR −/− mice allowing them to maintain the exercise long after reaching their V'O 2max . Indeed, compensatory mechanisms might operate also in the skeletal muscles of female ApoE/LDLR −/− mice enabling them to run on the treadmill much better than their age-matched C57BL6/J counterparts with similar V'O 2max .
To conclude, we claim that robust compensatory mechanisms in the hearts of female ApoE/LDLR −/− mice including increased vascular responsiveness to NO and increased generation of PGI 2 could, at least partly, explain preserved whole body V'O 2max as well as full cardiac adaptation to exercise of female ApoE/LDLR −/− mice despite peripheral and coronary atherosclerosis progression. Our results demonstrate surprisingly efficient adaptive mechanisms in the coronary circulation in female ApoE/LDLR −/− mice which when understood better, may be exploited therapeutically to limit impending impairment of cardiac output and the risk of perfusion deficit of the heart with coronary atherosclerosis. Incremental exercise for V'O 2max measurement. All animals were acclimatized to the closed treadmill for metabolic measurements (Columbus Instruments, Columbus, OH, USA) for three days before the experiment. Randomly selected mice from all groups (n > 9) underwent maximal incremental testing with measurement of respiratory metabolic performance according to the protocol described before 48 . Briefly, mice were run on a single-line metabolic treadmill equipped with a shock grid at 0°. The exercise protocol started at 5 m·min −1 and the speed was increased by 4 m·min −1 every 3 min until exhaustion. Data on whole body oxygen consumption (V'O 2 ) were collected.

Animals. ApoE/LDLR
Scientific RepoRts | 6:24714 | DOI: 10.1038/srep24714 Blood collection. Mice were anesthetized by sodium pentobarbital/pentobarbital (5:1) overdose (1 ml·kg −1 , intraperitoneally) and the rib cage was cut along the sides to expose the heart. Blood was collected from the right ventricle into a syringe with nadroparin (10 U · ml −1 ) and the blood count was performed at animal blood counter Vet abc (Horiba Medical, France).
Measurements in isolated perfused heart. The detailed technique of isolated heart perfusion to study alterations in coronary flow in mouse model according to Langendorff is described elsewhere 49 . The animals were anesthetized with mixture of ketamine (100 mg · kg −1 ) and xylazine (10 mg · kg −1 ). The hearts were precisely excised from the animal body and placed in ice-cold Krebs-Hanseleit bicarbonate buffer, where extraneous tissues were quickly removed. The aorta was attached to the prepared aortic cannula and isolated hearts were retrograde perfused with Krebs-Hanseleit buffer under 80 mmHg pressure and at 37 °C. It took approximately less than 30 seconds to mount the heart for Langendorff perfusion (hearts that required longer time for mounting were discarded from analysis). Coronary flow (CF) during the perfusion time was monitored by ultrasonic flow meter (Transit Time Flowmeter TTFM, HSE-Harvard and Transonic System Inc., USA) and flow measurements were displayed throughout the experiment at all time points. Two electrodes were placed on the right atrium of the heart model generating 400 impulses per minute. To characterize the responsiveness of coronary vessels to vasodilation several drugs were given at the volume of 10 μL each: bradykinin (Bk: NO -dependent, in 2 (n = 10), 4 (n = 10) and 8 (n = 13) months old female ApoE/LDLR −/− mice and 2-4 controls (n = 16)), Acetylcholine (ACh: PGI 2 -dependent, in 2 (n = 23), 4 (n = 16) and 8 (n = 22) months old female ApoE/LDLR −/− mice and 2-4 controls(n = 25)) as well as two exogenous, endothelium-independent NO -donors with different mechanism of NO release sodium nitroprusside (SNP in 2 (n = 7), 4 (n = 8) and 8 (n = 12) months old female ApoE/ LDLR −/− mice and 2-4 months old controls (n = 14) and S-Nitroso-N-acetylpenicillamine (SNAP in 2 (n = 14), 4 (n = 14) and 8 (n = 15) months old female ApoE/LDLR −/− mice and 2-4 months old controls (n = 16)). ACh (0.3 nmol) was injected as one bolus, whereas Bk, SNAP and SNP were injected as multiple doses (0.03; 0.1; 0.3 and 1.0 nmol). To study the contribution of NOS and cyclooxygenase (COX) to the coronary flow response separately, the nonselective NO -synthase inhibitors, NG -nitro-L-arginine methyl ester (L-NAME 500 μM in 2 (n = 6), 4 (n = 6) and 8 (n = 8) months old female ApoE/LDLR −/− mice and 2-4 months old controls (n = 16)) and COX-2 selective cyclooxygenase inhibitors -Rofecoxib, 10 μM (in 2 (n = 6), 4 (n = 6) and 8 (n = 6) months old female ApoE/LDLR −/− mice and 2-4 months old controls (n = 5) were used. After experiments all hearts were weighed and alterations in coronary flow were expressed in ml · min · g −1 of the wet ventricular weight. For determination of the concentration of the 6-keto-PGF 1α in effluent from the coronary vessels, samples of the effluent (500 μl) were collected and stored at − 70 °C until analysed by commercially available enzyme immunoassay kits (Cayman Chemical Co, Ann Arbor, MI). 6-keto-PGF 1 α concentration was expressed in mg · l −1 (or ng · ml −1 ) of coronary flow (ml · min −1 ). Assessment of coronary atherosclerosis in coronary arteries. The hearts were dissected, embedded in the OCT compound (CellPath, Oxford, UK), and snap-frozen or fixed in 4% paraformaldehyde (pH 7.4), paraffin embedded and serially sectioned. Then, 10 μm thick serial cross-section slides of the heart were cut in the apex region, at the middle (at papillary muscle level) and upper part of the heart (aortic roots), to evaluate the coronary atherosclerosis in distal, intramuscular and proximal parts of coronary arteries. The sections were stained with the oil red-O for lipids detection in the atherosclerotic plaque or with the OMSB staining (Orceine with Martius, Scarlet and Blue) for histological characterisation. Photographs were acquired using the Carl Zeiss HmR monochromatic digital camera.
Cardiac performance. Cardiac function was assessed using cardiac magnetic resonance imaging (MRI) with a 9.4T Bruker BioSpec system (Bruker, Ettlingen, Germany) equipped with BFG-113/60-S gradient system and 36 mm 1H quadrature volume resonator. Physiological parameters were monitored and gated using Model 1025 Monitoring and Gating System (SA Inc., Stony Brook, NY). Mice were measured under anesthesia maintained at 1.7% of isoflurane (Aerrane, Baxter) in a 1/2 oxygen/air mixture delivered via nose mask with monitoring of ECG, respiration and body temperature (37 °C). MRI cine scans were acquired with retrospectively gated IntraGate FLASH 2D sequences with the following parameters: echo time 1.49 ms, repetition time 4.3 ms, flip angle 18°, number of repetitions 250, field of view 30 × 30 mm 2 , matrix 256 × 256, slice thickness 1.0 mm and reconstructed with IntraGate 1.2.b.2 macro (Bruker) to 60 frames per cardiac cycle. LV images were analysed using the Segment package 17 (v.1.9 R2626, Medviso AB, Sweden) and papillary muscles were included in the cavity volume. Single-slice volume was plotted against the time and used for regional cardiac functional parameters: end-diastolic/end-systolic volume (EDV/ESV), fractional area change (FAC), stroke area change (SAC) and LV filling and ejection rates (FR, ER normalized to the individual SAC and R-R values).

Statistical analysis.
Statistical analyses were performed in GraphPad Prism Software (GraphPad Software, USA) and STATISTICA 10 software (Stat-Soft Inc., USA). Results from exercise performance test, body mass and blood count were analysed with two-sided t test or non-parametric Mann-Whitney test based on the results of D' Agostino and Pearson omnibus normality test. Results from in vivo basic cardiac performance as well as from isolated perfused hearts were analysed by the t-test for independent variables. If parametric tests assumptions were not fulfilled (normality of distributions and homogeneity of variance), non-parametric Mann-Whitney U test or Friedman test was used. For in vivo cardiac reserve assessment, a repeated-measures ANOVA followed by a Tukey post-hoc test was used. All data were presented as mean values ± SEM.