Diabetic cardiomyopathy (DCM) is characterized by a cascade of myocardial changes that occurs in diabetes mellitus with fibrosis, hypertrophy and microcirculatory abnormalities. These cardiovascular complications compromise cardiac performance ultimately resulting in cardiac failure. A high prevalence of cardiac failure is seen in individuals with diabetic cardiovascular complications, with DCM as one of the key determinants 1. DCM is marked by diastolic dysfunction early in the disease progression 234, with its reported occurrence even in patients with well-controlled diabetes in the absence of clinically detectable cardiac disease 5. In addition reports also suggest subtle systolic dysfunction later during the course of diabetes that evades detection with echocardiography 2. Meanwhile it has been suggested that detection of systolic dysfunction might require highly sensitive techniques 3.

Magnetic resonance imaging (MRI) has proven to be a powerful and robust noninvasive imaging modality for structural and function evaluation of the rat heart 6. However, in vivo cardiac MRI studies using diabetic rat models are very limited. For example, Al-Shafei and colleagues 78 performed elaborate MRI studies with a 2 T magnet on streptozotocin- (STZ) diabetic Wistar rats to assess abnormalities of myocardial structure and cardiac cycle events in diabetes.

Understanding the course of pathological events in an appropriate model is the key for developing therapeutic strategies aimed at preventing the heart failure. In order to evaluate the cardiac performance in vivo we used MRI, a robust technique for resolving cardiac functional information and the reference standard for real time three dimensional visualization of myocardial structure 910. In a previous study, we demonstrated the merits of high resolution MRI in visualizing the diabetic heart and characterized the structural properties of non-beating myocardial tissue in the STZ-diabetic Sprague-Dawley rat 11. As an extension of our previous study, we have characterized the cardiac dysfunction associated with diabetes in this model. In particular, we report quantitative measurements on left ventricular end-diastolic and end-systolic volumes and demonstrate that these parameters are different for STZ-diabetic Sprague-Dawley rats compared to control rats.

The STZ induced diabetic rats used in our experiments are reminiscent of a model of uncontrolled hyperglycemia due to absolute insulin deficiency. The later feature closely captures the metabolic condition of type 1 diabetes. The STZ rat model has been used to study both tissue pathology 1314 and therapeutic interventions 1516 in type 1 diabetes. There has been a growing interest in the application of MRI to obtain structural and functional information from a variety of tissues including the eye 17, the kidney 18 and the heart 7819 that are targeted by diabetic complications.

Functional sensitivity of imaging modalities poses a major challenge for delineation of abnormalities of cardiac function in DCM 3. However, limitations on functional sensitivity might be lowered with the use of robust non-invasive techniques such as MRI. MRI has evolved as a powerful tool for the evaluation of cardiac function in both humans and experimental animal models of cardiovascular pathology 620. Hence MRI can be applied to the study of cardiac structure and function in DCM. In particular, the MRI study of cardiac abnormalities in DCM provides unique insights into cardiac dynamics that may remain undetected otherwise, with the use of other techniques. For example, echocardiography fails to capture the real state of the tissue due to intrinsic assumptions of tissue geometry 21. Our gravimetric finding of higher heart to body mass ratio in the diabetic group when compared to the control (Table 1) is suggestive of cardiac hypertrophy and altered ventricular geometry in this rat model at 8 weeks of diabetes. The LV wall volume, calculated from MR images of diabetic rats was not significantly different from that of controls. However with body mass normalization, the mean LV wall volume of the diabetic group became significantly higher than the control group. This supports our gravimetric results and indicates LV hypertrophy in this model of DCM, and is in agreement with our earlier findings reported on non-beating diabetic hearts 11.

In this study we utilized EKG gating to correlate the image acquisition with electromechanical end diastole to obtain functional information on the diabetic LV. The use of cine MRI to image the LV along the cardiac short axis provided excellent temporal resolution to delineate volume changes. The high contrast between the blood and endocardium allowed us to perform the planimetry on LV cavities from all images representing the ten phases of cardiac cycle. LV volume calculations showed a significant reduction of 28% in the mean end-diastolic volume of the diabetic group compared to controls. It has been suggested that the reduction of end-diastolic volume might be the undesirable consequence of an adaptive mechanism of stiff myocardium, in an effort to compensate for poor contractility by increased pressure during experimental cardiomyopathy 22. A stiff myocardium is characteristic of STZ induced diabetes of similar duration 11. However the difference in mean end-diastolic volume between groups disappeared when normalized for their body mass suggesting that the role of abovementioned early diastolic adaptive mechanism is plausible in DCM. Meanwhile the end-systolic volume of the diabetic rats increased 10% compared to controls. This difference between groups was also present after body mass normalization suggesting systolic volume dysfunction in this model. As a consequence of disparity between control and diabetic rats in phase volumes, the stroke volume and ejection fraction declined (48% and 28%, respectively) in the diabetic group compared to controls.

The LV end-diastolic volume, stroke volume and ejection fraction displayed significant changes with diabetes in this study, in accordance with a previous report 8. However, in contrast to our finding of an increase in LV end-systolic volume with diabetes (8 weeks diabetes duration), the previous study (9 weeks diabetes duration) observed no change in this parameter 8. In addition, the difference in body mass normalized end-diastolic volume between groups was insignificant in our study. Meanwhile recent MRI analyses of cardiac function in 8 weeks STZ-diabetic Wistar-Kyoto rats showed no significant difference in the LV end-diastolic volume, end-systolic volume, stroke volume and ejection fraction from age matched controls 23. These results may reflect the difference in the strain of rats, since this factor has been shown to clearly influence DCM in the STZ model of type 1 diabetes 24. Strain differences exist in their susceptibility to DCM with STZ induced diabetes in rodent models 2425 even though the diabetic cardiovascular complications closely imitate the human condition 26. In addition, echocardiographic differences in performance have been detected in the two widely used diabetic rat models, viz. STZ-diabetic Wistar 27 and STZ-diabetic Sprague-Dawley 28 rats. The differences in DCM susceptibility and cardiac performance may underlie the manifestation of cardiac functional abnormalities in these models of type 1 diabetes. Meanwhile stroke volume and ejection fraction were decreased in the STZ-diabetic Sprague-Dawley model used in our study, a finding in agreement with results from the STZ-diabetic Wistar model 8, suggesting that the overall cardiac performance is compromised in both models of type 1 diabetes. Interestingly, in a rat model of type 2 diabetes with DCM, the LV end-diastolic volume remained comparable to the age-matched controls while the end-systolic volume was increased due to poor longitudinal contractility of LV 19. In addition to compromised myocardial contractility in diabetes, the hemodynamic consequences of increased vascular resistance and compromised isoflurane induced vessel wall relaxation may also affect the cardiac cycle systole in these diabetic models 2329.

The derivatives of volume with respect to time (dV/dt) of the end-diastole to systolic transition and end-systole to diastolic transition were also significantly different between the diabetics and controls, substantiating the pathological changes involving both active (myocytes) and passive (matrix) components, respectively, in the dysfunction of diabetic LV 23. The complex pathology of DCM that limits normal ventricular function involves both cardiomyocyte loss 3031 and interstitial collagen accumulation 111632. The loss of force-producing myocytes may underlie contractile dysfunction whereas the accumulation of interstitial collagen might produce difficulties with passive stretch of myocardium during diastole thereby compromising ventricular relaxation in DCM 3133. Accordingly, we speculate that a compromised systolic function in this rat model, suggesting loss of contractility may have been the result of myocyte loss due to apoptosis/necrosis that characterize the middle stage of human DCM 3. Cardiomyocytes demonstrate both impaired contractility and delayed relaxation in mice models of DCM as well 3435. Interestingly, the flow velocity of the diabetic group (dV/dt) in this study was not different from the control group at the 6–7 phase transition. This indifference in dV/dt at the end-systolic phase transition suggests that the compromised compliance of the diabetic myocardium is not global, encompassing the entire cardiac cycle. This unexpected result also demonstrates the ability of MRI to provide unique insights that may fail detection otherwise by methods both invasive and non-invasive. However the insignificant difference in dV/dt between the diabetic and control end-systolic phase transition requires cautious interpretation since dV/dt measures are restricted as an indirect index of flow velocity only under assumptions of linear relationship between the variables concerned.

Although our choice of division of the cardiac cycle into equi-duration phases in this study was arbitrary, thus facilitating the evaluation of LV volume with respect to time as a perfectly smooth function between phases, our results nevertheless agree to a substantial degree, with previous reports on normal and diabetic left ventricular function utilizing a slightly different methodology 829.

In conclusion, the results from our investigations indicate that the functional manifestation of DCM in the STZ rat model of subchronic diabetes include early diastolic flow adaptation, systolic volume dysfunction and cardiac cycle phase dependent diminution of LV kinetics. Our study also demonstrates that in vivo MRI is capable of evaluating the cardiac dysfunction in this model of diabetes.

DCM – Diabetic cardiomyopathy

LV – Left ventricle

MRI – Magnetic resonance imaging

STZ- Streptozotocin

The author(s) declare that they have no competing interests.

RL carried out the diabetes induction procedures and rat maintenance, helped with data analysis, statistical procedures and drafting the manuscript. MB conceived of the MRI methods and helped with drafting the manuscript. BAH carried out the MRI procedures. MDA helped with data analysis and statistical procedures. IVS conceived of the study, participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.