Share this article on:

Exercise Attenuates Diabetes-Induced Ultrastructural Changes in Rat Cardiac Tissue


Medicine & Science in Sports & Exercise: November 2004 - Volume 36 - Issue 11 - p 1863-1870
doi: 10.1249/01.MSS.0000145461.38224.EC
Clinical Sciences: Clinically Relevant

Introduction/Purpose: Exercise is an effective nonpharmacological treatment in the prevention of mortality and morbidity due to cardiovascular disease in Type I diabetes. This study sought to explore the effects of endurance exercise on the ultrastructural changes seen in diabetic cardiomyopathy.

Methods: Seven-week-old rats were divided into three groups consisting of sedentary nondiabetic control, sedentary diabetic, and exercised diabetic animals. Diabetes was induced using streptozotocin injection, and the exercised animals were run daily on a treadmill for 9 wk. Changes in heart ultrastructure were analyzed using transmission electron microscopy.

Results: Ultrastructural changes in the left ventricle produced by diabetes included changes in myofibrillar arrangements, disrupted mitochondria, and increased cytoplasmic area with an increase in lipid amounts and an increase in individual collagen fiber cross-sectional surface area. Also, an increase in heterochromatin lining the nuclear envelope and an increase in invaginations of the nuclear membrane were observed in cardiomyocytes from diabetic rats when compared with the nuclei from nondiabetic cells. Exercise was found to significantly attenuate the diabetes-induced changes in collagen fibrils, cytoplasmic area, and level of mitochondrial disruption. In contrast, exercise did not appear to significantly influence myofibril volume density, lipid accumulation, or nuclear deformities.

Conclusion: These findings indicate that exercise restores specific ultrastructural characteristics of diabetic cardiomyopathy returning them toward nondiabetic phenotypes, particularly in the mitochondria and extracellular matrix proteins.

1Departments of Anatomy & Cell Biology, and 2Physical Therapy & Rehabilitation Sciences, University of Kansas Medical Center, Kansas City, KS

Address for correspondence: Lisa Stehno-Bittel, Ph.D., Department of Physical Therapy and Rehabilitation Sciences, University of Kansas Medical Center, MS 2002, 3901 Rainbow Blvd., Kansas City, KS 66160; E-mail:

Submitted for publication August 2003.

Accepted for publication July 2004.

We are indebted to Rajprasad Loganathan for assistance with data analysis. The suggestions pertaining to morphometric analysis from Dr. Kuen-Shan Hung, as well as his critical reading of the manuscript, are greatly appreciated. We thank Dr. Nataliya Kibiryeva, Seok Hun Kim, and Mike Wolfe for their assistance with exercise training the rats.

This work was supported by an NIH RO1 GMS grant to LSB, and a Lied Endowed Basic Science Pilot Research Grant and an American Heart Association Scientist Development Award to IVS.

The major cause of mortality in individuals with diabetes is heart disease. Although coronary artery disease may contribute to diabetic cardiomyopathy, myocardial dysfunction exists with diabetes that is independent of coronary artery disease (29). Diabetic cardiomyopathy is a complex result of insulin deficiency, thyroid hormone deficiency, and excessive sympathetic nervous system activation (7).

The clinical presentation of diabetic heart disease is a reduced rate of pressure development and pressure decline. Left ventricular volume is greater at the end of systole and diastole, and left ventricular fractional shortening and reduced ejection fraction are also seen (7,17,21). These features are indicative of a decrease in contractility. The midmyocardial region is most responsible for contractility. Further, ultrastructural changes have been noted in each myocardial layer in hypertension associated myocardial hypertrophy, a condition in which contractility is certainly impaired (19).

The benefits of exercise on the cardiovascular system have been well documented. An inverse dose–response correlation exists between the amount of physical activity and cardiovascular disease, coronary artery disease, and general mortality. With moderate aerobic exercise, an increase in cardiac mass is seen (15), which is associated with hypertrophy of the left ventricular wall and enlargement of the left ventricular cavity (8). Isolated cardiomyocytes from exercise-trained animals reveal an increase in myocyte length but no change in the sarcomere length (22). On an ultrastructural level, an increase in the width of myofibers and an increase in mitochondria to muscle volume density (10) with moderate aerobic exercise training have been reported.

Physiological improvements with endurance exercise training have been documented for patients with either Type I or II diabetes mellitus (3). Such findings are difficult to interpret, because most patients that regularly exercise also make a variety of other healthier lifestyle choices. Thus, the singular effects of exercise are challenging to identify. In diabetic animal models, endurance exercise diminishes the diabetes-associated bradycardia, leading to improved cardiac function in the diabetic animals (5,6,25). However, the cellular basis for these improvements has not been identified. This study sought to quantify the morphometric characteristics of diabetes-induced abnormalities of heart tissue, and to determine whether moderate levels of exercise could inhibit these abnormalities. The exercise-induced changes in diabetics were compared with the results obtained in hearts of nondiabetic animals to determine whether exercise could reverse the diabetic pathology.

Back to Top | Article Outline


Animal protocol.

All experiments were performed according to the American College of Sports Medicine animal care policy. Seven-week-old male Sprague Dawley rats (Harlan, Indianapolis, IN) weighing 250–270 g were used. Rats were divided into three groups: sedentary nondiabetic (control), sedentary diabetic, and exercise-trained diabetic. The exercised diabetic group began a progressive exercise program with treadmill running 5 d·wk−1. Each treadmill session began with a warm-up period. The endurance exercise program progressed as follows: week 1 began with 5 min·d−1 and ended the week at 25 min·d−1, 15 m·min−1; week 2 began at 30 min·d−1 and progressed to 50 min·d−1, 20 m·min−1. Not all animals could maintain this pace or time limitation; thus, exercise protocols were individualized. Running scores indicating the quality and the time run were obtained daily for each animal. The running goal for the remaining 7 wk of exercise remained steady at 60 min·d−1, at a pace of 20 m·min−1. The sedentary nondiabetic and diabetic rats were placed on a stationary treadmill.

Two weeks into the exercise protocol, the rats in the diabetic groups were injected intraperitoneally with 65 mg of streptozotocin (STZ, Sigma, St. Louis, MO) per kilogram of body weight. Nondiabetic controls were injected with vehicle. Diabetes was confirmed by measuring blood glucose levels using an Accu-Chek Advantage glucometer (Boehringer Mannheim Corporation, Indianapolis, IN). A nonfasting blood glucose level greater than 300 mg·dL−1 was considered diabetic. Blood glucose and body weight were monitored weekly. Animals were provided free access to food and water, and principles of institutional laboratory animal care were strictly followed. The experiments were terminated after 7 wk of diabetes (9 wk of exercise). The animals were killed with injection of pentobarbital.

Back to Top | Article Outline

Transmission electron microscopy.

For purposes of electron microscopy study, five to six animals per experimental group were used for a total of 16. Immersion fixation was completed on tissue from the midmyocardial region of the distal left ventricle by a method described previously (21). This region was chosen because it is the most representative of contractility. Samples were rinsed in cold phosphate-buffered saline (PBS) and placed in 2% glutaraldehyde at +4°C. This process took less than 10 s from biopsy of the beating heart to fixation. The tissue was rinsed in buffer and postfixed with 1% osmium tetroxide. Tissue was rinsed with distilled water before undergoing a graded ethanol dehydration series and was infiltrated using a mixture of one-half propylene oxide and one-half resin overnight. Twenty-four hours later, the tissue was embedded in Epon 812 resin (Electron Micro-scopy Sciences, Ft. Washington, PA). Eighty-nanometer sections were cut on an LKB Nova Ultratome and were placed on acid treated grids, which were stained using a double lead stain technique (4) with 0.5% lead citrate and 7% uranyl acetate. Images were captured from random sections using a JEM 100 CXII transmission electron microscope at 80 kV.

Back to Top | Article Outline

Morphometric analysis.

Determination of lipid incidence and mitochondrial quality, and cardiomyocyte nuclear analysis were made at final magnifications of 8700, 10,800, or 14,400×, respectively. The measurements of collagen diameter and fiber density were determined at a final magnification of 57,000×. All quantitative analysis was completed in a blinded fashion independently by two observers. The number of lipid droplets per area was counted manually. To determine the mitochondrial quality index, each mitochondrion was assigned one of the following grades: 1 = fully intact, 2 = less than 50% disruption of the organization of inner mitochondrial membranes, 3 = more than 50% disruption of organization of inner mitochondrial membranes, 4 = disruption of the outer mitochondrial membrane, and 5 = disruption of both inner and outer mitochondrial membranes. It should be noted that a grade of 2 or 3 could reflect either swelling that would displace the membrane or actual disruption of the membrane. Heterochromatin lining the inner nuclear envelope was calculated as the percentage of nuclei with dense heterochromatin > 0.25 μm.

Digital images were processed in Adobe Photoshop (Adobe Systems Inc.) for density measurements and analyzed with Scion Image beta 4.0.2 (Scion Corporation) when calculating cross-sectional surface area. Cytoplasm was defined as intracellular space not containing myofibrils, intercalated disc, mitochondria, or nuclei.

Back to Top | Article Outline

Data analysis.

Quantitative analysis of single cell morphology was completed on all animals within each group with cells taken from a minimum of five independent regions of each left ventricle to accurately represent the mean. Differences in percentages were analyzed using ANOVA. To account for within group variance, hierarchical ANOVA was used for all other tests. All bar graphs and Table 1 illustrate mean ± SE. P value < 0.05 was considered statistically significant.



Back to Top | Article Outline


Physiological effects of training.

Table 1 shows rat blood glucose levels, body and heart weights, and heart to body weight ratios measured at the termination of the experiment after 7 wk of diabetes (9 wk of exercise; 2 wk before diabetes induction, and 7 wk after STZ injection). All animals injected with STZ developed diabetes as judged by elevated blood glucose levels. The nonfasting blood glucose levels increased 471% with 7 wk diabetic rats. Importantly, exercised diabetic group showed a decrease of 14.7% in blood glucose level compared with sedentary diabetic group. The diabetic rats did not gain as much weight as control rats in the duration of the experiment. The heart to body weight ratios increased in both diabetic groups compared with control animals. This increase of approximately 25% in diabetic rats compared with control suggests that the animals developed diabetes-induced heart hypertrophy. There was no significant difference in heart to body weight ratio between sedentary and exercise-trained diabetic rats.

Thus, diabetic rats showed severe hyperglycemia that was slightly attenuated by physical exercise. Diabetic rats developed heart hypertrophy, which is a known feature of diabetic cardiomyopathy observed in experimental animal models of diabetes (24,30).

Citrate synthase levels were measured from the gastrocnemius muscle of the rats to verify the exercise training level. There was no difference in the citrate synthase levels between the nonexercised animals and the exercise-trained rats (I. V. Smirnova, N. Kibiryeva, E. Vidoni, R. Bunag, and L. Stehno-Bittel, unpublished data, 2004). This is likely due to the fact that animals were allowed to exercise at their own pace with a goal of 60 min·d−1. Higher intensity exercise protocols do result in changes in gastrocnemius muscle citrate synthase levels. Thus, we conclude that the exercise protocol used for this study resulted in moderate level endurance training.

Back to Top | Article Outline

Electron microscopy.

Comparisons of the size and circumference of the cardiac cells demonstrated no statistically significant difference between samples from the nondiabetic controls and the diabetic sedentary or exercise-trained rats. However, when compared with control heart tissue, the diabetic groups exhibited many differences in subcellular organelle morphology and distribution. Two of the most striking differences easily visualized were the condition of the myocardial fibrils and gross deformation of mitochondria with diabetes. Therefore, caution needs to be taken when interpreting changes in size and area of myofibrils and mitochondria due to the radically different quality of these organelles between the three groups.

Back to Top | Article Outline


Cardiac muscle fibers from nondiabetic samples were abundant, with regular arrays of myofibrils closely arranged within the sarcomere (Fig. 1). Diabetic muscle fibers often showed sparser fibrils within the sarcomere. Attempts to quantify the observed diabetes- and exercise-induced changes in the myofibrillar arrangement by measuring the total myofibril area (calculated as a percentage of the cytoplasmic area) found no statistically significant difference between the three groups. In general, the myofibril area comprised about 40% of the total intracellular area regardless of the group. Likewise, the sarcomere length remained relatively constant between all three groups (Fig. 1).

FIGURE 1—Cardiac myofibrillar arrangement: myofibril sarcomeres (s; from Z line to Z line) are indicated by white lines in the control, sedentary diabetic, and exercised diabetic samples. Rows of mitochondria (m) are interspersed between groups of fibers. Control tissue revealed compactly and neatly arranged myofibrils, whereas that of sedentary diabetic cardiac tissue showed disruption of myofibrils with thinning in some areas. Tissue from exercised diabetic rats did not appear to have as much myofibril disruption nor cytoplasmic space (CS) as tissue from hearts of sedentary diabetic rats. (L) designates lipid droplets (×8700).

FIGURE 1—Cardiac myofibrillar arrangement: myofibril sarcomeres (s; from Z line to Z line) are indicated by white lines in the control, sedentary diabetic, and exercised diabetic samples. Rows of mitochondria (m) are interspersed between groups of fibers. Control tissue revealed compactly and neatly arranged myofibrils, whereas that of sedentary diabetic cardiac tissue showed disruption of myofibrils with thinning in some areas. Tissue from exercised diabetic rats did not appear to have as much myofibril disruption nor cytoplasmic space (CS) as tissue from hearts of sedentary diabetic rats. (L) designates lipid droplets (×8700).

Previous studies have shown that the mitochondria to myofibrils ratio is an indicator of cell hypertrophy (11). In our work, there was no statistically significant difference in the ratio of the mitochondrial to myofibrillar area between the three groups of samples. However, the amount of swelling present in diabetic mitochondria suggest that this ratio is not likely an accurate indicator of cell hypertrophy, because the underlying assumption of the earlier work is that the cells have a otherwise normal morphology.

Back to Top | Article Outline


Overall, diabetic cardiac tissue revealed disrupted mitochondria of varying degrees when compared with the control, which was quantified with the mitochondrial quality index. Figure 2 illustrates the grading of mitochondria of cells from nondiabetic control, sedentary diabetic and exercised diabetic rats. In left ventricle from control rats the mitochondria were typically intact (Fig. 2A). The mitochondria from diabetic samples appeared swollen (grades 2–3) and the inner and/or outer membranes were disrupted (grades 4–5). Sedentary diabetic sections tended to have patchy areas of mitochondrial swelling and disruption. Sections from exercise-trained diabetic rats also showed patchy areas of mitochondrial swelling and disruption but to a lesser extent than the samples from sedentary diabetic rat hearts.

FIGURE 2—Illustration of the mitochondrial quality index using longitudinal sections of left ventricle. A) Control. Tissue from the nondiabetic control rats showed intact mitochondria. The

FIGURE 2—Illustration of the mitochondrial quality index using longitudinal sections of left ventricle. A) Control. Tissue from the nondiabetic control rats showed intact mitochondria. The

These qualitative changes could be compared by plotting the mitochondrial quality index score (Fig. 2B). The nondiabetic control group displayed 86% of its mitochondria intact whereas the sedentary diabetic group had 48%. Ninety-nine percent of cells from control animals had mitochondria with little (grade 2) to no (grade 1) alterations of the arrangement of inner mitochondrial membranes. Noteworthy is the fact that within the assignment of grade 2, mitochondria from nondiabetic samples typically showed less than 13% disarrangement, whereas the diabetic sedentary group displayed 41% disarrangement and disruption. One percent of the mitochondria from the control group displayed severe disruption of the mitochondrial membranes (grades 3–5), whereas the sedentary diabetic samples had 11% of the mitochondria with severe disruption.

The ability of exercise to alter the percentage of disrupted mitochondria was clear. Animals that were exercised had 59% of the cardiac mitochondria intact (compared with 48% in sedentary diabetic rats). Likewise, only 7% of the mitochondria from the cardiomyocytes of exercise-trained rats were severely disrupted, whereas sedentary diabetic samples had 11% disruption. This partial inhibition of the diabetes-induced mitochondrial disruption is obvious in Figure 2B. It is important to note that although the general trend of the mitochondrial quality index is that higher grades represent more severe mitochondrial disruption, there are exceptions. For example, in the images from the exercised diabetic animals in Figure 2A, a grade 3 mitochondria with severe disruption of the inner membranes is noted. Adjacent is a grade 4 mitochondria with a limited breach in the outer membrane, resulting in far less damage to the mitochondria, despite the higher grade. Thus, rather than discuss each grade separately, they were combined to represent fully intact to minimally damaged (grades 1–2) or moderately to severely disrupted mitochondria (grades 3–5).

Mitochondrial area (percent of total intracellular area) accounted for approximately 45% of the total intracellular area within each cardiomyocyte in normal cardiac tissue (Fig. 3A). The mitochondrial area demonstrated a significant reduction in the samples from the sedentary diabetic rats (36.0 ± 0.1%) compared with the nondiabetic control animals (44.5 ± 0.1%). However, the difference was not statistically significant between the diabetic groups or between nondiabetic control and diabetic exercised groups (41.0 ± 0.1%). Thus, exercise appeared to only partially restore the mitochondrial area.

FIGURE 3—Area of subcellular compartments. A) Mitochondrial area. The mitochondria from sedentary diabetic tissue occupied a smaller area within the cardiomyocyte (

FIGURE 3—Area of subcellular compartments. A) Mitochondrial area. The mitochondria from sedentary diabetic tissue occupied a smaller area within the cardiomyocyte (

Back to Top | Article Outline


The mitochondrial area reduction observed in the sedentary diabetic group was accompanied by a subsequent increase in cytoplasmic area that was also significantly different from the nondiabetic control cells (Fig. 3B). Normal tissue displayed cytoplasmic area that accounted for less than 13.72 ± 0.01% of the total intracellular area but was increased to 24.74 ± 0.02% in cells from sedentary diabetic rats. Again, exercise training inhibited the diabetes-induced change, resulting in significant decrease in cytoplasmic area in exercised animals compared with sedentary diabetic (18.23 ± 0.02%). There was no statistically significant difference in the total cytoplasmic area between cells from exercise-trained diabetic animals and nondiabetic controls.

Back to Top | Article Outline

Lipid droplets.

The incidence of lipids was significantly greater in both diabetic groups when compared with the nondiabetic rats (Fig. 4). However, the sedentary diabetic and exercised diabetic groups were not different from each other. Cells from the diabetic groups contained roughly five times the amount of lipid present in tissue in contrast to the control group. The proportion of lipid droplets associated with certain organelles was remarkably similar between all three groups. The majority of lipid droplets (70–72%) were associated with the mitochondria-myofibril interface. Eighteen to twenty percent of the lipid droplets were strictly found near mitochondria whereas, 7–9% of the lipid droplets were associated with myofibrils. The remaining 1–2% of the lipid droplets was associated with the extracellular matrix, vessels, or sarcolemma.

FIGURE 4—Lipid per intracellular area. Cells from sedentary diabetic and exercised diabetic left ventricle cardiomyocytes had significantly more lipid than cells from nondiabetic hearts (

FIGURE 4—Lipid per intracellular area. Cells from sedentary diabetic and exercised diabetic left ventricle cardiomyocytes had significantly more lipid than cells from nondiabetic hearts (

Back to Top | Article Outline


Differences in the morphology of the cardiomyocyte nuclei were illustrated in the appearance of invaginations and a thickening of the heterochromatin lining the nuclear envelope (Fig. 5A). Invaginations were defined as being at least 0.3 μm in depth and less than 1 μm in width. The percentage of nuclei demonstrating invaginations was significantly different between the control and the diabetic groups (P < 0.005) as seen in Figure 5B. The control group exhibited invaginations in 31 ± 13% of nuclei, whereas the sedentary diabetic and exercised diabetic groups exhibited invaginations in 90 ± 8 and 83 ± 11% of nuclei, respectively. There was an increased incidence (P = 0.035) of a thicker heterochromatin lining associated with the nuclear envelope (greater than 0.25 μm) in the cells from sedentary (25%) and exercised diabetic rats (26.1%) compared with the control rats (0%) as visualized in Figure 5C. The extent of nucleoli in the cardiomyocyte nuclei were essentially the same between all three groups. Thus, diabetes altered the morphology of the cardiomyocyte nuclei, and those changes were not altered by endurance exercise training of the diabetic animals.

FIGURE 5—Morphology of the cardiomyocyte nuclei: nuclei from cells of control rats (A) had few invaginations (

FIGURE 5—Morphology of the cardiomyocyte nuclei: nuclei from cells of control rats (A) had few invaginations (

Back to Top | Article Outline


Examination of individual collagen fibers within the interstitium revealed a significant difference in the cross-sectional surface area of the collagen fibers (Fig. 6). The cross-sectional surface area of individual fibers was significantly greater with sedentary diabetes (433,000 ± 13,000 nm2) as compared with nondiabetic controls (341,000 ± 10,000 nm2). This effect was completely inhibited by exercise training (339,000 ± 9,000 nm2). Interestingly, collagen fibers appeared to be denser within each individual fiber in samples from the sedentary diabetic rats when compared with the samples from control and exercise-trained diabetic rats (not shown). There was no difference in density measurements (number of collagen fibers per area) between the three groups. Thus, although the total number of collagen fibers/area did not change, the size of each fiber increased with diabetes, and exercise inhibited this effect.

FIGURE 6—Collagen fibers. Interstitial collagen fibers from the sedentary diabetic group (

FIGURE 6—Collagen fibers. Interstitial collagen fibers from the sedentary diabetic group (

Back to Top | Article Outline


This study demonstrates that 9 wk of moderate exercise is sufficient to reverse some of the phenotype of diabetic cardiomyopathy. Specifically, the mitochondrial quality, cytoplasmic area, and collagen cross-sectional area were returned toward nondiabetic values with exercise. Results from this study are summarized in Table 2. Diabetes induced several visible and calculable changes in the ultrastructure of the cardiomyocyte. Qualitatively the myofibrils were more disordered, the mitochondria were swollen and disrupted, and interstitial collagen was more abundant in samples from the left ventricle of diabetic rats. Quantitatively, the cytoplasmic area and density of lipid droplets were significantly greater in samples from sedentary diabetic animals, whereas mitochondrial disruption was more prevalent with diabetes. There were also alterations in the diabetic cardiomyocyte nuclear morphology, including an increased presence of invaginations and increased accumulation of heterochromatin. Finally, the size of individual collagen fibers was significantly increased in tissue from diabetic sedentary rats.



This study was the first to measure ultrastructural changes in diabetic heart in response to exercise. A progressive endurance exercise program appeared to partially inhibit the altered mitochondrial/myofibrillar area ratio induced by diabetes, and reversed the level of swelling and mitochondrial disruption induced by diabetes. Exercise also halted the diabetes-induced increase in cytoplasmic area and collagen fiber cross-sectional surface area in tissue from sedentary diabetic animals, both approaching the levels of the nondiabetic controls. Exercise, on the other hand, did not significantly affect the accumulation of lipid, as one might have speculated, nor did it appear to affect the changes in the diabetic cardiomyocyte nuclei.

In terms of the diabetes-induced ultrastructural changes, the results from this study are similar to findings from previous reports on diabetes-induced changes in the heart. One comparable study using the same animal model did not find any changes in the ultrastructure of cardiac tissue after diabetes induction at the 6-wk time point (16). However, whereas the previous study examined tissue from the right ventricle, this inquiry investigated changes in the left ventricle, which is more susceptible to diabetes-induced pathology.

The disorder of myofibrils that was noted with diabetes in this study is consistent with other reports that also showed no change in sarcomere length with diabetes (12). In addition, a loss of myofibrils in the diabetic cardiomyocyte is well noted (9). A more chronic model of diabetes demonstrated a concurrent reduction in the cardiomyocyte diameter and size (9,17), which was not observed in this study.

Diabetes-induced changes in the mitochondrial and cytoplasmic area are consistent with previous findings (9,17). Specifically, a reduction in the volume of cardiomyocyte mitochondria in STZ-induced diabetes mellitus (30) is consistent with our findings of reduced mitochondria density. The observation of patchy mitochondrial disruption is a key characteristic found in numerous previous studies (20,21,28).

Several studies reported increases in collagen bundles in interstitial spaces (1,17) and increased lipids (9,28) in the diabetic heart that progressed over time. Only one study measured individual collagen fibrils with diabetes and that was in tail tendon (23). These researchers found that the radius and gap depth of the collagen fibril was larger with diabetes, which is in partial agreement with an increase in the cross-sectional surface area found here. Finally, invaginations and thickened heterochromatin within the diabetic cardiomyocyte nuclei were noted, which has been reported previously (26).

The diabetes-induced ultrastructural changes described here are not unique to the STZ animal model. The spontaneous diabetic Bio Breeding (BB) rat model of Type I diabetes demonstrated ultrastructural deviations in the cardiac tissue similar to those described in this report. In the left ventricle of the BB rat (4–8 wk diabetic), the mitochondria displayed patchy swelling, with a loss of myofilaments and an increase in lipid droplets located near mitochondria (14). By 16 wk, membrane disruptions of the mitochondria and the sarcoplasmic reticulum were severe and myofilament loss was extensive. The consistency that mitochondria deformation in diabetic cardiac tissue occurs between studies and animal models illustrates the amplified fragility of the mitochondria membranes as a result of diabetes.

Biochemically, a host of findings support the observed changes in the subcellular ultrastructure in diabetes. Diabetes is associated with an abnormality in myocardial energy metabolism having reduced glucose transport, oxidation, reduced fatty acid oxidation and reduced ATP production. Variations in insulin levels can alter lipid composition in cardiomyocyte membranes (7). Metabolism shifts from using plasma glucose to the excess use of free fatty acids. The cell starving for glucose switches to a gluconeogenic pathway producing its own glucose stored as glycogen.

This study was the first morphometric analysis of ultrastructural changes in diabetic heart in response to exercise. The mechanisms responsible for the ability of endurance exercise to inhibit the morphological changes are a matter of speculation at this point, but several hypotheses can be put forward. Exercise appears to have an impact in decreasing the severity of the pathology in diabetic cardiac ultrastructure, most notably with a rescue of cytoplasm area and collagen fiber cross-sectional surface area. The explanation for improvement due to exercise is likely to be multifactorial. It has been proposed that hyperglycemia causes hyperosmolarity, shrinking ventricular cardiomyocytes and, in turn, reducing the degree of shortening during contraction, while the intracellular calcium transient is prolonged (13). Metabolic improvement of glucose transport and insulin intolerance in the exercised skeletal muscle leads to lower blood glucose levels that could then reduce the consequences of osmolarity. The exercise protocol used in this study resulted in the modest, however significant, reduction in plasma glucose levels obtained in the exercised diabetic group compared with the sedentary diabetic rats. However, the average glucose value for the exercised group was still 481 mg·dL−1, which is dramatically higher than the normal value of 120 mg·dL−1. Thus, it seems unlikely that a lower blood glucose level could account for the inhibition of diabetes cardiomyopathic changes.

It has been shown that in diabetic heart tissue there is a deficiency in antioxidant enzyme activity with an increase in lipid perioxidation (27). It is possible that exercise positively affects antioxidant enzyme activity, thus reducing cellular injury from oxidative stress and improving the fragility of the cell membranes. Perhaps the composition of lipids and/or proteins in the membrane has been altered making it more stable. Exercise decreased the levels of phospholipids: phosphatidylethanolamine, phosphatidylcholine, and cholesterol in cardiac mitochondria, resulting in a slight decrease in viscosity of the membrane (18).

Hyperglycemia associated with chronic diabetes produces increased levels of nonenzymatic glycation on numerous proteins, including collagen that ultimately causes functional impairment of many organs by altering their structural, biochemical, and metabolic properties (2). We found exercise attenuated the increase in collagen fiber cross-sectional surface area observed in diabetes. It seems unlikely that exercise would directly alter glycation levels that would result in changes in collagen size. Perhaps exercise altered the type of collagen expressed in the left ventricle.

The ability of exercise to decrease mitochondrial swelling and disruption, decrease the cytoplasmic area and alter collagen properties in diabetic animals is certain to have functional implications. The cell would be better equipped to respond to energy demands and less prone to damage caused by oxidative stress. Improvement of membrane stability would normalize the regulation of calcium and subsequently contractility, possibly leading to enhanced cardiac function. Testing these hypotheses will be the focus of future studies.

Electron microscopy is an invaluable tool for examining the ultrastructural pathology in the progression of diabetes. It can be used to tie the subcellular structures to some of the biochemical and functional changes observed in diabetic cardiomyopathy. Cardiac dysfunction in diabetes has been summarized as a combination of metabolic disturbances, abnormal contractile proteins, abnormal cell signaling, autonomic dysfunction, and interstitial fibrosis. Visualization of ultrastructural changes in the progression of diabetic cardiomyopathy can provide insight to the interplay between the above factors as well as contribute to assessment of how these factors might be positively impacted by exercise. It will be important in the future to correlate the development and progression of diabetic cardiomyopathic ultrastructure with the mechanical function of the heart and molecular properties like gene expression patterns and to identify the molecular pathways responsible for the improvement with exercise.

Back to Top | Article Outline


1. Balogh, I., M. Z. Koltai, and G. Pogatsa. Ultrastructural alterations in cardiac and skeletal muscles in experimental diabetes mellitus. Acta Physiol. Hung. 71:219–225, 1988.
2. Ceriello, A. Hyperglycaemia: the bridge between non-enzymatic glycation and oxidative stress in the pathogenesis of diabetic complications. Diabetes Nutr. Metab. 12:42–46, 1999.
3. Chipkin, S. R., S. A. Klugh, and L. Chasan-Taber. Exercise and diabetes. Cardiol. Clin. 19:489–505, 2001.
4. Daddow, L. Y. An abbreviated method of the double lead stain technique. J. Submicrosc. Cytol. 18:221–224, 1986.
5. De Angelis K. L., A. R. Oliveira, P. Dall’Ago, et al. Effects of exercise training on autonomic and myocardial dysfunction in streptozotocin-diabetic rats. Braz. J. Med. Biol. Res. 33:635–641, 2000.
6. DeBlieux, P. M., R. W. Barbee, K. H. McDonough, and R. E. Shepherd. Exercise training improves cardiac performance in diabetic rats. Proc. Soc. Exp. Biol. Med. 203:209–213, 1993.
7. Dhalla, N. S., X. Liu, V. Panagia, and N. Takeda. Subcellular remodeling and heart dysfunction in chronic diabetes. Cardiovasc. Res. 40:239–247, 1998.
8. Diaz-Herrera, P., A. Torres, J. A. Morcuende, J. M. Garcia-Castellano, J. A. Calbet, and R. Sarrat. Effect of endurance running on cardiac and skeletal muscle in rats. Histol. Histopathol. 16:29–35, 2001.
9. Fitzl, G., K. Welt, G. Wassilew, N. Clemens, K. Penka, and N. Mukke. The influence of hypoxia on the myocardium of experimentally diabetic rats with and without protection by Ginkgo biloba extract. III: ultrastructural investigations on mitochondria. Exp. Toxicol. Pathol. 52:557–568, 2001.
10. Frenzel, H., B. Schwartzkopff, W. Holtermann, H. G. Schnurch, A. Novi, and W. Hort. Regression of cardiac hypertrophy: morphometric and biochemical studies in rat heart after swimming training. J. Mol. Cell. Cardiol. 20:737–751, 1988.
11. Guski, H., G. Wassilev, R. Meyer, and H. David. Morphometric analysis of myocardial mitochondria in rats during adaptation to exercise. Cor Vasa 22:375–383, 1980.
12. Howarth, F. C., M. A. Qureshi, P. Lawrence, and E. Adeghate. Chronic effects of streptozotocin-induced diabetes on the ultrastructure of rat ventricular and papillary muscle. Acta Diabetol. 37:119–124, 2000.
13. Howarth, F. C., M. A. Qureshi, and E. White. Effects of hyperosmotic shrinking on ventricular myocyte shortening and intracellular Ca(2+) in streptozotocin-induced diabetic rats. Pflugers Arch. 444:446–451, 2002.
14. Hsiao, Y. C., K. Suzuki, H. Abe, and T. Toyota. Ultrastructural alterations in cardiac muscle of diabetic BB Wistar rats. Virchows Arch. A Pathol. Anat. Histopathol. 411:45–52, 1987.
15. Huonker, M., M. Halle, and J. Keul. Structural and functional adaptations of the cardiovascular system by training. Int. J. Sports Med. 17:S164–S172, 1996.
16. Jackson, C. V., G. M. McGrath, A. G. Tahiliani, R. V. Vadlamudi, and J. H. McNeill. A functional and ultrastructural analysis of experimental diabetic rat myocardium: manifestation of a cardiomyopathy. Diabetes 34:876–883, 1985.
17. Kawaguchi, M., M. Techigawara, T. Ishihata, et al. A comparison of ultrastructural changes on endomyocardial biopsy specimens obtained from patients with diabetes mellitus with and without hypertension. Heart Vessels 12:267–274, 1997.
18. Keatisuwan, W., M. Kinjo, and T. Koyama. Changes in phospholipid constituents in mitochondrial membranes after long lasting exercise in rat heart. Life Sci. 48:2173–2181, 1991.
19. Kimpara, T., M. Okabe, H. Nishimura, T. Hayashi, K. Imamura, and K. Kawamura. Ultrastructural changes during myocardial hypertrophy and its regression: long-term effects of nifedipine in adult spontaneously hypertensive rats. Heart Vessels 12:143–151, 1997.
20. Kumar, S. D., and S. S. Tay. Ultrastructural changes in the atrioventricular valves of streptozotocin-induced diabetic rats. J. Hirnforsch. 39:559–566, 1999.
21. Malone, J. I., D. D. Schocken, A. D. Morrison, and E. Gilbert-Barness. Diabetic cardiomyopathy and carnitine deficiency. J. Diabetes Complications 13:86–90, 1999.
22. Moore, R. L., T. I. Musch, R. V. Yelamarty, et al. Chronic exercise alters contractility and morphology of isolated rat cardiac myocytes. Am. J. Physiol. 264:C1180–C1189, 1993.
23. Odetti, P., I. Aragno, R. Rolandi, et al. Scanning force microscopy reveals structural alterations in diabetic rat collagen fibrils: role of protein glycation. Diabetes Metab. Res. Rev. 16:74–81, 2000.
24. Ren, J., and A. M. Bode. Altered cardiac excitation-contraction coupling in ventricular myocytes from spontaneously diabetic BB rats. Am. J. Physiol. Heart Circ. Physiol. 279:H238–H244, 2000.
25. Riggs, C. E., Jr., G. Collins, and M. Taylor. Streptozotocin-induced diabetes and the effects of endurance exercise training. Diabetes Res. 19:177–185, 1992.
26. Shimabukuro, M., S. Higa, T. Shinzato, F. Nagamine, I. Komiya, and N. Takasu. Cardioprotective effects of troglitazone in streptozotocin-induced diabetic rats. Metabolism 45:1168–1173, 1996.
27. Singal, P. K., A. Bello-Klein, F. Farahmand, and V. Sandhawalia. Oxidative stress and functional deficit in diabetic cardiomyopathy. Adv. Exp. Med. Biol. 498:213–220, 2001.
28. Tomita, M., S. Mukae, E. Geshi, K. Umetsu, M. Nakatani, and T. Katagiri. Mitochondrial respiratory impairment in streptozotocin-induced diabetic rat heart. Jpn. Circ. J. 60:673–682, 1996.
29. Trost, S. and M. LeWinter. Diabetic Cardiomyopathy 3:481–492, 2001.
30. Warley, A., J. M. Powell, and J. N. Skepper. Capillary surface area is reduced and tissue thickness from capillaries to myocytes is increased in the left ventricle of streptozotocin-diabetic rats. Diabetologia 38:413–421, 1995.


©2004The American College of Sports Medicine