The study of cardiomyopathies in small animals may contribute to our understanding of the cardiac pathophysiology and to evaluate experimental treatment strategies. The present study shows that myocardial perfusion imaging with contrast echocardiography can successfully be applied for the non-invasive evaluation of cardiac perfusion at rest and under hyperaemia in STZ-induced diabetic rats. Our results suggest that after six months, diabetes induces a functional alteration of the myocardial microcirculation that may explain left ventricular systolic dysfunction. Our data further indicate that microcirculation is already altered at rest and that the major determinant for decreasing myocardial perfusion reserve during hyperaemia, is the lower capillary recruitment in the diabetic group. Finally, histopathology findings demonstrate a reduced density of myocardial capillaries, in the diabetic group that can only be unmasked with hyperaemia using MCE.
Clinical parameters
The STZ induced diabetic rats used in our experiments are reminiscent of a model of uncontrolled hyperglycemia due to the direct pancreatic beta cell destruction and resulting insulin deficiency [
19]. Our glucometry and gravimetric values (Table
1) are in accord with other published values at similar time points [
20‐
22]. These STZ rats resembled more type 1 diabetes. The dose of STZ injected (45 mg/kg) was low in this study. However, it has been successfully applied in previous studies [
3,
14] and the impact on glucometry and gravimetric values was similar compared to the administration of a higher dose of STZ, leading to equivalent absolute insulin deficiency.
Myocardial perfusion
Microbubbles are excellent tracers of red blood cell kinetics. They are pure intravascular tracers. The method to quantify MBF is based on rapid destruction of these microbubbles by ultrasound, and subsequent assessment of the rate of replenishment into the myocardium within the ultrasound beam elevation. This method allows the assessment of the components of MBF: flow velocity and myocardial blood volume [
13]. In the presence of increased myocardial oxygen demand, there is arteriolar vasodilation, reducing the resistance at the arteriolar level, which enables a higher precapillary pressure, translating into increased red blood cell velocity across the capillary network, and opening dormant capillary networks in order to maintain mean trans-capillary pressure (recruitment), thus increasing the overall myocardial blood volume. Hence overall myocardial blood flow is increased.
The present study has shown that quantitative MCE is a robust method to assess the myocardial blood flow in small animals at rest and during hyperemia. The measurements are feasible and repeatable over time and between observers. As expected from physiology and previous in vitro and in vivo studies [
9,
10,
13,
23,
24], we were able to show the increase of MBF with hyperemia in normal rats. Moreover, our data were consistent with previous experiments in small animals, validating MCE measurements using microspheres [
15].
In our diabetic population, we could observe, at rest, a significant increase in β compared to controls. This suggests a decrease in arteriolar resistance and vasodilation of arterioles. Although this increase in red cell velocity at rest compared to controls persists under hyperaemia in the diabetics versus the controls, the change in β is not significantly different in diabetics compared to controls. This finding suggests the capacity of arterioles to further vasodilate, even in the diabetic group, in the absence of coronary artery lesions.
Structural alterations of the small vessels in diabetes have been incriminated in the development of diabetic cardiomyopathy, although this remains controversial. A reduction in capillary density and a significantly greater thickening of the capillary basement membrane has been shown compared to control subject in humans [
25] and more recently in rodents [
26]. Because rBV represents essentially capillaries, one would therefore expect to find a decrease in A in the diabetic rats, which is not the case at rest. However, during hyperaemia, there is less increase of rBV in diabetic rats compared to normal rats, indicating a less important capillary recruitment in diabetics. This may reflect the absence of compensation to an increased pre-capillary pressure mediated by arteriolar vasodilatation, as shown by the increased red cell velocity at rest and during hyperaemia compared to controls. In our study, histopathology did confirm a significant reduction in capillary density, in accordance with the previous studies [
25,
26]. The increase in pre-capillary pressure induced by the arteriolar vasodilation and the decrease of capillary density may be compensated by a capillary recruitment at rest in the diabetic group resulting in no significant change of rBV. We can hypothesize that hyperaemia is required to unmask the decrease of capillary density in diabetics, using MCE. As a net result, myocardial perfusion reserve is altered and may lead to relative ischemia. Similarly, a reduction of myocardial blood flow and significant increase in total coronary resistance during hyperaemia and consequent impairment of coronary flow reserve have been reported in type I young adult diabetic patients with no or minimal microvascular complications and without any evidence of coronary heart disease [
27]. Reduced myocardial flow reserve may lower the threshold for myocardial ischemia, particularly when coronary stenoses are present. It has been proposed that diabetic cardiomyopathy is a consequence of repeated episodes of myocardial ischemia resulting from these functional abnormalities in small vessels during increased myocardial demand. As shown by Litwin and colleagues, a real insulin-therapy that aims to normalize the glycemia, also corrects the cardiac abnormalities [
28]. An ongoing study investigates the effect of the correction of diabetes condition on the impairment of coronary circulation.
Left ventricular function
There remains controversy regarding diabetes-induced LV dysfunction, especially in type 1 diabetes, in the absence of documented coronary artery disease. Some authors have been able to detect early systolic LV dysfunction and dilatation of the left ventricle in STZ induced diabetic cardiomyopathy [
29,
30]. On the contrary, others were not able to demonstrate a remodelling and a significant alteration of the systolic function in a similar rat population [
31]. A recent study in patients with type 1 diabetes, even with the application of echocardiography, biochemical and morphologic techniques, failed to demonstrate that diabetes type 1 may actually precipitate myocardial dysfunction and no heart-specific, histological changes in the myocardium were found. However, as acknowledged by the authors of this latter study, all the patients were treated with intensive insulin therapy [
32].
Standard measurements of LV wall thickness and systolic and diastolic LV diameters by M-mode have extensively been described in normal and diseased rat models [
33‐
36]. The ratio data such as fractional shortening and ejection fraction have been demonstrated to be similar in rat and human echocardiography [
36]. Comparable results were obtained in the present study. In literature anatomical M-Mode and bi-dimensional echocardiography have been able to detect early systolic LV dysfunction and dilatation of the left ventricle in STZ induced diabetic cardiomyopathy [
22,
29,
37‐
39]. We observed an increase in mean normalized EDV of the diabetic group compared to controls. Meanwhile the normalized ESV of the diabetic rats also increased compared to controls but in lower proportion. The significant increase of normalized EDV with diabetes in this study is in accordance with previous reports [
22,
29,
40,
41]. However, in contrast to our findings, other authors have shown no significant change or even an decrease in this parameter [
42‐
44]. We also observed an increase in normalized ESV, suggesting a decrease in contractility, in accord with all previous studies [
22,
29,
40,
43,
44]. The differences in LV volumes may be due to the method used to normalize the data or to theabsence of normalization to body weight in other studies. Allometric relations exist between cardiac and body size measurements. However, the correct method to use in rat is unknown. We normalized LV mass to bodyweight as applied in the previous studies using the same animal model [
22,
41]. In addition, these differences may reflect the difference in the strain of rats (Wistar Unilever, Wistar Kyoto, Sprague-Dawley), since this factor has been shown to clearly influence cardiomyopathy in the STZ model of diabetes [
45].
Histopathology
In the present study, no significant pathological changes were observed in diabetic rats, regarding extra-cellular collagen deposits, endomyocardial necrosis in myocardium, and no microangiopathy involving arterioles, capillaries, venules, and hyaline arteriosclerosis was present, at the end of the follow up. Conversely, in our study, histopathology did confirm a significant reduction in capillary density. These results are in accordance with previously published works [
25,
26]. The muscular fibers were thinner in the diabetic rats: 14.6 ± 3.1 μm in controls and 11.4 ± 2.8 μm in diabetics (P = 0.043). This last finding was also consistent with previously published work (23).