Magnetic resonance imaging of progressive cardiomyopathic changes in the db/db mouse

Patrick Yue, Takayasu Arai, Masahiro Terashima, Ahmad Y. Sheikh, Feng Cao, David Charo, Grant Hoyt, Robert C. Robbins, Euan A. Ashley, Joseph Wu, Phillip C. Yang, Philip S. Tsao


The db/db mouse is a well-established model of diabetes. Previous reports have documented contractile dysfunction (i.e., cardiomyopathy) in these animals, although the extant literature provides limited insights into cardiac structure and function as they change over time. To better elucidate the natural history of cardiomyopathy in db/db mice, we performed cardiac magnetic resonance (CMR) scans on these animals. CMR imaging was conducted with a 4.7-T magnet on female db/db mice and control db/+ littermates at 5, 9, 13, 17, and 22 wk of age. Gated gradient echo sequences were used to obtain cineographic short-axis slices from apex to base. From these images left ventricular (LV) mass (LVM), wall thickness, end-diastolic volume (LVEDV), and ejection fraction (LVEF) were determined. Additionally, cardiac [18F]fluorodeoxyglucose ([18F]FDG) PET scanning, pressure-volume loops, and real-time quantitative PCR on db/db myocardium were performed. Relative to control, db/db mice developed significant increases in LVM and wall thickness as early as 9 wk of age. LVEDV diverged slightly later, at 13 wk. Interestingly, compared with the baseline level, LVEF in the db/db group did not decrease significantly until 22 wk. Additionally, [18F]FDG metabolic imaging showed a 40% decrease in glucose uptake in db/db mice. Furthermore, contractile dysfunction was observed in 15-wk db/db mice undergoing pressure-volume loops. Finally, real-time quantitative PCR revealed an age-dependent recapitulation of the fetal gene program, consistent with a myopathic process. In summary, as assessed by CMR, db/db mice develop characteristic structural and functional changes consistent with cardiomyopathy.

  • diabetes mellitus
  • insulin resistance
  • heart failure
  • metabolism

congestive heart failure (CHF) is a significant yet often underappreciated complication of diabetes mellitus (25). While atherosclerotic coronary artery disease is highly prevalent and likely responsible for CHF in many diabetic patients, findings from several large-scale heart failure clinical trials reveal a 16–20% prevalence of diabetes in patients with nonischemic cardiomyopathy (9). Moreover, an analysis of hospital discharge data showed a 27% prevalence of diabetes in patients discharged with idiopathic cardiomyopathy, compared with 18% for control subjects (13). Collectively these data suggest an alternative mechanism for CHF in diabetics—one independent of the effects of epicardial coronary disease. However, studies investigating this unique form of “diabetic cardiomyopathy” (51) have failed to establish a unifying mechanistic basis for this phenomenon.

The C57BL/KLS-leprdb/leprdb (db/db) mouse, which has a mutation in the leptin receptor, is a well-established animal model of Type 2 diabetes mellitus (19). Leptin resistance results in hyperphagia and weight gain from birth. Homozygous db/db mice become noticeably obese by 3–4 wk of age and develop hyperglycemia at 4–8 wk. Serum insulin levels increase as early as 10–14 days, peak at 6–8 wk, then decrease precipitously afterward (although db/db mice continue to be hyperinsulinemic throughout life). This drop, which is believed to be secondary to pancreatic islet cell dysfunction, further exacerbates the hyperglycemia.

In addition to these characteristic phenotypic changes, db/db mice also develop cardiomyopathy. Metabolic experiments using cultured db/db cardiomyocytes have shown impaired glucose oxidation as early as 6 wk of age (1). Echocardiographic studies (7, 58), isolated working heart preparations (1, 11, 16, 31), and direct in situ ventricular pressure measurements (14, 49, 62) have also demonstrated structural and functional impairments consistent with cardiomyopathy. Although it is generally believed that cardiomyopathy exists in these animals, the onset and evolution of this phenomenon remain a relative unknown. Indeed, while many studies have documented myopathic changes in db/db hearts, comparatively few have investigated these changes longitudinally, especially over three or more time points (14).

Recently, cardiac magnetic resonance (CMR) imaging has emerged as a potentially valuable tool to accurately and reproducibly evaluate cardiac structure and function in small animals. As a three-dimensional imaging modality, CMR contains an inherent advantage over one- and two-dimensional techniques in that it permits measurement of ventricular volumes and masses without the need for assumptions regarding ventricular shape. Moreover, unlike invasive physiology-based protocols, it allows the animal to survive, thus facilitating serial evaluation of cardiac function over a prolonged period of time. Consequently, the availability of CMR potentially allows the investigator to monitor the progression of cardiomyopathy in a longitudinal fashion with significantly greater resolution and accuracy than before.

In this study, we used CMR imaging to comprehensively evaluate the course and natural history of the cardiomyopathy seen in db/db mice. Our results indicate that the hearts of these mice undergo a distinct period of early hypertrophic remodeling, followed by an attenuation of this remodeling and, ultimately, overt contractile dysfunction.



Female homozygous db/db mice (Jackson Laboratories, Bar Harbor, ME) were maintained on a normal chow diet and housed in a room with a 12:12-h light-dark cycle and an ambient temperature of 22°C. Unless otherwise stated, heterozygous db/+ littermates were used as control animals. All protocols were approved by the Administrative Panel on Laboratory Animal Care at Stanford University and were carried out in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care.

Animals were euthanized with a lethal dose of isoflurane. Immediately after death, wet heart weight (HW) and body weight (BW) were measured. Whole hearts were harvested and snap-frozen in liquid nitrogen for subsequent mRNA isolation, and a subset of hearts were fixed in 10% formalin for histological evaluation.

Insulin tolerance testing.

Insulin tolerance testing was performed on mice after a 6-h fast. At the time of testing, a bolus of human regular insulin (Eli Lilly, Indianapolis, IN; 0.75 IU/kg) was injected intraperitoneally. In blood derived from a tail nick, glucose levels were then determined with a FreeStyle blood glucose monitoring system (Abbott Laboratories, Abbott Park, IL) at baseline and 15, 30, 45, and 60 min after injection.

Insulin ELISA.

After a 6-h fast, roughly 200 μl of blood was obtained from each animal via retroorbital bleeding. Samples were placed in 500-μl tubes containing EDTA (Becton-Dickinson, Franklin Lakes, NJ) and centrifuged at 4°C and 13,200 rpm for 10 min. Approximately 50–75 μl of supernatant (i.e., serum) was then collected for further processing. Serum insulin concentrations were measured with a Mercodia mouse insulin enzyme immunoassay kit (Alpco Diagnostics, Salem, NH).

[18F]fluorodeoxyglucose positron emission tomography scanning.

To control for hyperglycemia, db/db mice were fasted for 12 h and control db/+ mice were fasted for 6 h before radioisotope injection. Animals were then injected with 218 ± 41.1 μCi [18F]fluorodeoxyglucose ([18F]FDG) via the tail vein. Sixty to seventy-five minutes after injection, animals were anesthetized with inhaled 2% isoflurane with an oxygen flow rate of 2 l/min and imaged with a P4 Concorde microPET system (Concorde Microsystems, Knoxville, TN). Images were reconstructed by filtered back projection, and three-dimensional regions of interest (ROIs) were drawn encompassing the heart. Counts per pixel per minute were converted to counts per milliliter per minute with a calibration constant derived from scanning a cylindrical phantom. For each ROI, counts per milliliter per minute were then converted to counts per gram per minute and divided by the injected dose to obtain the image ROI-derived [18F]FDG percent injected dose per gram of heart (% ID/g) as described previously (64).

Cardiac magnetic resonance imaging.

To prepare for scanning, induction of anesthesia was accomplished with 2% isoflurane and 1 l/min oxygen. Respiratory rate was monitored and used to manually calibrate the maintenance dose of isoflurane at 1.25–1.5%. Platinum needle ECG leads were inserted subcutaneously in the right and left anterior chest wall. Respiration was monitored with a pneumatic pillow sensor positioned along the abdomen. Body temperature was maintained at 36–37°C by a flow of heated air thermostatically controlled by a rectal temperature probe. Heart rate (HR), respiratory rate, and body temperature were recorded every 4 min during image acquisition.

Magnetic resonance images were acquired with a 4.7-T magnet (Bruker BioSpin, Fremont, CA) controlled by a Varian Inova Console (Varian, Palo Alto, CA) using a transmit-receive quadrature volume coil with an inner diameter of 3.5 cm. For particularly obese animals, a larger coil with an inner diameter of 6 cm was utilized. Image acquisition was gated to the ECG R wave (Small Animal Instruments, Stony Brook, NY). Coronal and axial scout images were used to position a two-dimensional imaging plane along the short axis of the left ventricular (LV) cavity. Gated gradient echo sequences were then used to acquire sequential short-axis slices spaced 1 mm apart from apex to base. For each sequence, 12 cine frames encompassing one cardiac cycle were obtained at each slice level with the following sequence parameters: acquisition time (TR) = 100–140 ms, echo time (TE) = 2.8–3.5 ms, number of repeats (NEX) = 8, field of view (FOV) = 30 × 30 mm, matrix = 128 × 128, flip angle = 60°.

For each short-axis slice, planimetry measurements of LV myocardial area (Fig. 1) were conducted off-line by tracing the epicardial and endocardial borders at end systole and end diastole with MRVision software (MRVision, Winchester, MA). For these purposes the papillary muscles were considered part of the LV cavity. Anteroseptal and posterior wall thickness measurements were performed on a midventricular slice at the level of the papillary muscles at end diastole. LV mass (LVM) was derived from the sum of the differences between the end-diastolic epicardial and endocardial areas from apex to base, adjusted for the specific gravity of myocardial tissue (1.055 g/ml). LV end-diastolic (LVEDV) and end-systolic (LVESV) volumes were calculated as the sum of the endocardial areas of each slice from the apex to the LV outflow tract at end diastole and end systole, respectively. LV ejection fraction (LVEF) was calculated as (LVEDV − LVESV)/LVEDV. Cardiac output (CO) was calculated as (LVEDV − LVESV)(HR).

Fig. 1.

Schematic representation of measured parameters with cardiac magnetic resonance (CMR) imaging. A: representative end-diastolic midventricular short-axis slice from a 13-wk db/+ mouse. Left ventricle (LV), right ventricle (RV), anteroseptal wall (AS), posterior wall (PW), and papillary muscles (PM) are indicated. B: luminal area (hatched) was determined by tracing the endocardial border of the LV. LV end-diastolic volume (LVEDV) was calculated by summing the luminal areas from apex to base. C: anteroseptal wall thickness (ASWT) and posterior wall thickness (PWT) were measured at the most anterior and most posterior aspects, respectively, of the LV. D: LV area (gray) was derived as the difference between the epicardial and endocardial borders of the LV. LV mass (LVM) was calculated by summing the LV areas from apex to base, corrected for the specific gravity of myocardial tissue (1.055 mg/μl). In this particular example, PWT and ASWT were 0.85 mm, luminal area was 9.01 mm2, and LV area was 11.4 mm2.

Pressure-volume loops.

Mice were induced with 2.5–3.0% isoflurane and 2 l/min oxygen, intubated, and ventilated with a rodent ventilator (Harvard Apparatus, Holliston, MA) at a maintenance dose of 2.5% isoflurane and 2 l/min oxygen. Body temperature was monitored with a rectal probe and maintained with a thermostatically controlled heating pad. Vascular access was obtained via the left jugular vein with PE-10 tubing. Albumin (12.5%) in normal saline was administered as a bolus of 35 μl/min for 5 min followed by a maintenance rate of 5 μl/min for the remainder of the procedure. A 1.4-F high-fidelity catheter tip containing both pressure and conductance transducers (Millar Instruments, Houston, TX) was inserted through the carotid, advanced in a retrograde fashion past the aortic valve into the LV cavity, and positioned in the LV outflow tract. To ensure proper placement, pressure-volume tracings were evaluated in real time to adjust the catheter as necessary. After baseline steady-state loops were obtained, the inferior vena caval segment immediately distal to the diaphragm was visualized and occluded with a vascular clamp for 5–10 s intervals while load-independent occlusion parameters were obtained. Loops were recorded with the Powerlab 5.4.2 system (ADInstruments, Colorado Springs, CO). Relative volume units (RVUs) were converted to true volumes with a standardized volumetric calibration cuvette (ADInstruments) containing six wells of known diameter, sized to calibrate for both large (50–60 μl) and small (5–10 μl) volumes. After each well was filled with heparinized blood from a freshly killed db/db mouse, RVUs were determined and, along with the volumes derived from the known dimensions of each well, used to generate a linear standard curve. This curve was used to convert RVUs to microliters.

Both steady-state and load-independent parameters were determined from the loops off-line with PVAN 3.4 software (Millar Instruments). The steady-state variables that were collected included LV end-systolic pressure (LVESP), LV end-diastolic pressure (LVEDP), end-systolic volume, end-diastolic volume, CO, stroke work (SW), peak instantaneous rate of LV pressure increase (dP/dtmax) and decrease (dP/dtmin), and the time constant of isovolumic LV relaxation (τ). Measured load-independent parameters included preload-recruitable stroke work (PRSW), the slope of the end-systolic (end-systolic elastance; Ees) and end-diastolic pressure-volume relationship (EDPVR), and the slope of the dP/dtmax-LV end-diastolic volume relationship (dP/dtmax-LVEDV).

Tail-cuff plethysmography.

Blood pressures were assessed with a BP-2000 (Visitech Systems, Apex, NC) tail-cuff plethysmography system. Conscious mice were prewarmed for 10 min in a thermostatically controlled restraint device. After a 2-wk run-in period, blood pressures were measured at 7, 9, and 11 wk of age. The mean of at least five separate measurements was taken to determine systolic (SBP) and diastolic (DBP) blood pressure. From these data, mean arterial pressure (MAP) and pulse pressure (PP) were calculated.

Real-time quantitative PCR.

After death, snap-frozen whole heart samples were homogenized and suspended in TRIzol (Invitrogen, Carlsbad, CA). RNA was extracted with an RNeasy Mini Kit (Qiagen, Valencia, CA), and cDNA conversion was performed with a SuperScript First-Strand cDNA synthesis kit (Invitrogen). The cDNA was then used as a template in a TaqMan real-time PCR assay with the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). All samples were run in triplicate. Specialized premade gene expression assay reagents for α-myosin heavy chain (α-MHC; catalog no. Mm00440354_m1), β-myosin heavy chain (β-MHC; Mm00600553_m1), sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2a; Mm00437634_m1), and atrial natriuretic peptide (ANP; Mm01255747_g1), purchased from Applied Biosystems, were used for these experiments.

Threshold cycles were placed in the logarithmic portion of the amplification curve, and each sample was referenced to 18S RNA amplification to control for the total amount of RNA. Fold difference between samples (relative quantification) was calculated with the delta-delta method [S1/S2 = 2Math ], where S1 and S2 represent samples 1 and 2 and T1 and T2 denote the threshold cycles for S1 and S2.

Assessment of cardiomyocyte hypertrophy.

At the time of death, a select group of hearts were fixed in 10% formalin, cut at the midventricular level, and embedded in paraffin blocks. The blocks were then sectioned into short-axis slices, which were subsequently stained with hematoxylin and eosin according to standard protocols. High-power fields (HPF) magnified to ×40 from the midportion of the LV free wall were photographed, taking care to avoid the epicardial and endocardial regions. Images were analyzed with Adobe Photoshop (Adobe Systems, San Jose, CA). Finally, the number of nuclei within each HPF was then counted.


Data were analyzed with GB-STAT 10.0 software (Dynamic Microsystems, Silver Spring, MD). Data are presented as means ± SE. For all tests, one-way ANOVA was performed. In addition, for pairwise comparisons between more than two groups, Scheffé's F post hoc analysis was performed. All P values <0.05 were considered significant.


Insulin resistance.

To confirm that the db/db strain was insulin resistant, we performed insulin tolerance tests on db/db, db/+, and C57BL/6J mice at 9 wk of age. In the db/db mice, glucose levels after insulin administration were strikingly elevated relative to both the db/+ and C57BL/6J groups, consistent with severe insulin resistance (Fig. 2). Additionally, fasting glucose and insulin levels were measured in db/db and db/+ mice at 5, 10, and 15 wk of age (Table 1). At 5 wk, glucose in the db/db group was not significantly different relative to control. However, by 10–15 wk, glucose levels were markedly increased. Fasting insulin levels were significantly elevated in db/db mice relative to db/+ mice at all ages.

Fig. 2.

Glucose levels 0–60 min after insulin administration for db/db, db/+, and C57BL/6J mice. Compared with both nondiabetic groups, db/db mice are severely insulin resistant. n = 4 animals for each group. Values are means ± SE. *P < 0.01 vs. db/+; #P < 0.01 vs. C57BL/6J.

View this table:
Table 1.

Fasting insulin and glucose measurements in db/+ and db/db mice

[18F]FDG positron emission tomography scanning.

Although decreased myocardial glucose uptake is known to occur in isolated cardiomyocytes from db/db mice (17), there remain concerns that the assessment of metabolic substrate utilization in the in vitro setting does not accurately reflect the in vivo condition (5). To address this issue, we measured myocardial [18F]FDG uptake in 9-wk db/db and db/+ mice. To control for the confounding effects of hyperglycemia in the diabetic group, the db/db mice were fasted for 12 h, while the controls were fasted for 6 h. Consequently, immediately before scanning, the glucose level in the db/db group (115 ± 20.1 mg/dl) was not significantly different from that in the control group (156 ± 8.1 mg/dl).

The % ID/g, a quantitative, normalized measure of myocardial [18F]FDG uptake, was significantly decreased by 40% in the db/db group (79.3 ± 16.9% ID/g) compared with the db/+ group (132.1 ± 4.1% ID/g) (Fig. 3). While the pretest glucose values for the db/db group were nominally decreased compared with control, any potential difference would have biased the results in the opposite direction. This finding is therefore consistent with impaired myocardial glucose uptake in db/db mice. Our data are complementary to those of Oakes et al. (43), who reported that db/db hearts were essentially unable to increase the rate of plasma glucose utilization in response to an acute glucose infusion. Together, these results provide a compelling case for the presence of primary myocardial insulin resistance in these mice.

Fig. 3.

Representative [18F]fluorodeoxyglucose (FDG) positron emission tomography scans of db/db (AC) and db/+ (DF) mice aged 9 wk. Individual frames are presented in the axial (A, D), coronal (B, E), and sagittal (C, F) orientations. Compared with control, myocardial enhancement in the db/db mouse is clearly reduced, indicating impaired glucose uptake and diminished insulin sensitivity. These findings are redemonstrated quantitatively in G. Here, the % FDG uptake normalized to heart weight (% ID/g) is decreased 40% in the db/db group. n = 4 for both groups. Values are means ± SE. *P < 0.05 vs. control.

Cardiac magnetic resonance imaging.

To evaluate for long-term changes in cardiac structure and function in our animal model, we performed CMR scans on db/db and db/+ mice. For both groups, mice were scanned at 5, 9, 13, 17, and 22 wk of age. To estimate the variance in HR over the course of each imaging session, HRs were recorded every 4 min during every scan. As the scans typically lasted between 12 and 30 min, four to eight HRs were obtained for each animal. For the 68 mice imaged for this study, the mean HR was 497 beats/min, while the SD and SE averaged 22.6 and 9.4, respectively. These data therefore indicate a relatively low amount of intrascan variance in HR. No differences in these parameters were observed between genotypes.

On initial review of the images, db/db mice were distinguished from their nondiabetic counterparts by a conspicuously larger body habitus (Fig. 4). Of particular note, there was an increased amount of pericardial fat in db/db mice relative to control. Otherwise, there were no apparent extracardiac differences between diabetic and nondiabetic mice.

Fig. 4.

Sample short-axis slices from CMR scans of db/db (A) and db/+ (B) mice at 9 wk of age. Relative to control, there is an increased amount of total body adipose tissue, as well as pericardial fat, in the db/db mouse.

From a structural standpoint, there were no differences in LVM between db/db and db/+ mice at 5 wk of age (Figs. 5 and 6A). However, at 9 wk, LVM was significantly increased in the db/db group relative to control. The disparity between groups persisted at 13, 17, and 22 wk. Similarly, wall thickness became augmented over time (Fig. 6, B and C). Compared with control, septal thickness was unchanged at 5 and 9 wk. However, by 13 wk there was a significant increase in the db/db group. This increase remained significant at 17 and 22 wk. Posterior wall thickness diverged similarly, with significant increases at 13, 17, and 22 wk.

Fig. 5.

Representative CMR images (top) along with low-power light microscopic slides (bottom) of db/db and db/+ mice at 5 and 17 wk of age. At 5 wk, there are no appreciable differences in cardiac size between db/db and db/+ mice, as visualized by CMR (A, B) and light microscopy (E, F). However, by 17 wk, db/db hearts are noticeably larger than their db/+ counterparts on CMR (C, D). This is confirmed by microscopy (G, H).

Fig. 6.

Structural and functional measurements on db/db and db/+ mice obtained via CMR imaging. Relative to control, early increases were noted in LVM (A), ASWT (B), PWT (C), and LVEDV (D). However, ejection fraction (LVEF; E) did not diverge significantly until later. Number of animals in each group is as depicted in Table 2. Values are means ± SE. *P < 0.05 vs. age-matched db/+; #P < 0.05 vs. 5-wk measurement for each respective genotype.

Regarding more functional parameters, there were no significant differences in LVEDV between db/db and db/+ mice at 5 and 9 wk (Fig. 6D). However, by 13 wk LVEDV was significantly greater, and it remained elevated at 17 and 22 wk (Fig. 7). An analogous pattern emerged with LVEF (Figs. 6E and 7). The LVEF of the db/db group did not differ significantly from that of the db/+ cohort at 5 and 9 wk. However, there was a small but significant decrease in LVEF relative to control by 13 wk. This difference persisted significantly at 17 and 22 wk. Finally, because HR was recorded at the time of scanning, we derived CO from these data (Table 2). There were no significant differences in HR between db/db and db/+ mice. Compared with the db/+ group, CO was slightly increased in db/db mice at all ages, although these differences were not significant.

Fig. 7.

Representative midventricular short-axis slices from cardiac magnetic resonance scans of db/db (top) and db/+ (bottom) mice at 22 wk of age. Frames are arranged from end diastole (left) to end systole (right). Both LV end-systolic and end-diastolic luminal areas are visibly larger in the db/db mouse, consistent with a decreased LVEF and an increased LVEDV. In this example, the db/db mouse had a LVEDV of 64 μl and a LVEF of 66.6%, while the db/+ mouse had a LVEDV of 50 μl and a LVEF of 70.7%.

View this table:
Table 2.

Heart rate and cardiac output as measured during cardiac magnetic resonance imaging

Taken together, these findings reveal a pattern consistent with ongoing cardiac remodeling in the db/db mouse. There are initial increases in LVM, wall thickness, and LV volume, followed by a period of continued LV dilatation and a progressive decrease in ejection fraction. Moreover, although CO was nominally similar in both groups, it can be argued that CO in the db/db mice was inappropriately low, given their markedly increased body habitus. These data therefore support the presence of diabetic cardiomyopathy in db/db mice.

Pressure-volume loops.

To provide adjunctive data for the CMR findings, we performed pressure-volume loops on mice at 15 wk of age. In the steady state, there were no significant differences in HR, CO, SW, LVESP, LVEDP, dP/dtmax, dP/dtmin, or the isovolumic relaxation time constant τ between db/db mice and their controls (Table 3). Overall these results match those of Buchanan et al. (14) and Reyes et al. (49), although a few discrepancies remain. Both studies report appreciably greater magnitudes for dP/dtmax and dP/dtmin, although the burden of anesthesia was larger in our study (2.5 l/min isoflurane) than that used by either Buchanan et al. (1–1.5 l/min isoflurane) or Reyes et al. (urethane/etomidate). Interestingly, for both db/db and db/+ mice, the CO in the steady state was roughly twofold lower than the CO as measured by CMR. While we cannot unequivocally reconcile this discrepancy, it should be noted that the two modalities measure CO under vastly different physiological states and testing conditions. Supporting examples of this include a difference in HR, an increased anesthetic dose, and the use of positive-pressure ventilation in the pressure-volume analysis.

View this table:
Table 3.

Hemodynamic steady-state parameters obtained during in situ left ventricular catheterization in 15-wk db/db and age-matched db/+ control mice

In addition, we evaluated several load-independent contractile parameters, as obtained during inferior vena cava occlusion (Table 4). Ees, PRSW, and dP/dtmax-LVEDV (which are indicators of systolic performance) were significantly decreased relative to control. Moreover, EDPVR was increased significantly in the db/db group, consistent with decreased ventricular compliance and diastolic dysfunction. These differences between groups are generally consistent with the results of Van den Bergh et al. (62). However, any direct comparison to our study should be undertaken with caution, as Van den Bergh used db/db and control mice with a different background strain (C57BL/6J) from ours (C57BL/KSJ).

View this table:
Table 4.

Hemodynamic load-independent occlusion parameters obtained during in situ left ventricular catheterization in 15-wk db/db and age-matched db/+ control mice

Blood pressure.

To evaluate for hypertension in db/db mice, blood pressures were measured with tail plethysmography. Blood pressures were recorded biweekly in db/db and db/+ mice (n = 6–8 for both groups) between 7 and 11 wk of age. Because of ongoing weight gain, the arterial waveforms of the db/db animals progressively dampened with age, thus precluding noninvasive blood pressure measurement after 11 wk. There were no significant differences in SBP, MAP, or DBP (Table 5). Of note, PP was significantly greater in db/db mice at 11 wk. Although this increase could represent a difference in peripheral arteriolar resistance, possible alternative explanations include a differential response of the diabetic mice to the stress of testing, inaccurate electronic signal transduction due to dampening of the arterial waveforms in older db/db mice, and statistical chance. Furthermore, in our pressure-volume loops, no significant differences between db/db and db/+ mice were observed in LVESV, LVEDV, or LV PP (Table 5). Overall, these blood pressure measurements suggest that hypertension is not a substantial contributor to the remodeling process.

View this table:
Table 5.

Blood pressure parameters measured over 2-wk intervals in db/+ and db/db mice

Morphometric analysis.

Immediately after death, hearts were excised and gently washed in PBS. Wet HW and premortem BW were measured and recorded (Table 6). As expected, the db/db group became markedly heavier over time. At 5 wk db/db mice were already significantly larger than their controls, and this difference was more pronounced at subsequent time points. Similarly, HW also increased as the animals grew older. While HW was not different between groups at 5 wk, there were significant increases in the db/db group at more advanced ages. At the microscopic level, specimens derived from 5-wk db/db mice were indistinguishable from their nondiabetic controls (Fig. 5, EH). However, by 17 wk, db/db mice displayed greater cardiac size and increased wall thickness compared with age-matched db/+ mice. Moreover, at high power, individual myocytes appeared larger in db/db mice than in their db/+ counterparts (Fig. 8). When specimens were compared quantitatively, 24% fewer myocyte nuclei per HPF were observed in the db/db group compared with the control group (67 ± 3.3 vs. 51 ± 3.4 nuclei for db/+ and db/db, respectively; P = 0.03). Overall, these data provide a corresponding anatomic substrate for the increased LVM and wall thickness observed on CMR imaging.

Fig. 8.

High-power light microscopy slides of myocardial tissue from db/db (A) and db/+ (B) mice aged 22 wk. Individual cardiomyocytes from the db/db image appear qualitatively thicker than those from the db/+ image, consistent with myocellular hypertrophy. Additionally, there is an increased number of cardiomyocyte nuclei, both qualitatively and quantitatively. n = 9 specimens for db/db and 15 for db/+. Values are means ± SE. *P < 0.05 vs. db/+ control.

View this table:
Table 6.

Heart weight and body weight in db/db and db/+ mice immediately after death

Real-time quantitative PCR.

To assess for molecular evidence of cardiac remodeling, real-time quantitative PCR was used to measure the mRNA expression of specific myocardial genes known to be affected by hypertrophy (Fig. 9). At 5 wk of age, the expression of α-MHC was not significantly different in db/db mice relative to control. However, α-MHC was decreased 2.8-fold at 10 wk (P < 0.01) (Fig. 9A). Although this reduction did not persist at 15 wk (1.3-fold downregulation; P = not significant), decreased α-MHC expression was again observed at 22 wk (1.6-fold downregulation; P < 0.01). With regard to the β-isoform (β-MHC), no significant changes in expression were seen (Fig. 9B). However, the ratio between β-MHC and α-MHC was significantly increased relative to control at 10 and 22 wk (Fig. 9C), indicating a shift from the adult α-MHC isoform toward the fetal β-isoform at these ages. These data are similar to those of Buchanan et al. (14), who also reported an age-dependent MHC isoform switch from α to β in db/db mice relative to control. Interestingly, the magnitude of change in the expression of both isoforms was uniformly greater in the study of Buchanan et al. than in our study. However, this difference could be explained by the fact that Buchanan et al. used wild-type C57BL/KSJ controls, instead of db/+ heterozygotes as we did.

Fig. 9.

Age-dependent trends in α-myosin heavy chain (α-MHC) mRNA expression (A), β-myosin heavy chain (β-MHC) expression (B), β-MHC-to-α-MHC ratio (C), sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2a; D), and atrial natriuretic peptide (ANP; E) in db/db mice. Expression levels are relative to age-matched db/+ control. These data suggest a progressive myopathic process. n = 6, 4, 7, 7 (db/db) and 6, 4, 7, 8 (db/+) animals for ages 5, 10, 15, and 22 wk, respectively. Values are means ± SE. *P < 0.05 vs. control; #P = 0.053 vs. control.

In addition to the MHC isoforms, we also evaluated the mRNA expression of SERCA2a. There were no differences in SERCA2a expression between groups at 5 and 10 wk (Fig. 9D). However, SERCA2a was downregulated in the db/db cohort at older ages, with 1.3-fold (P < 0.01) and 1.4-fold (P < 0.01) decreases at 15 and 22 wk, respectively. These findings are concordant with several papers reporting decreased SERCA2a expression in insulin-resistant animals (2, 18, 56), although a more recent report by Belke et al. (11) did not demonstrate a significant difference in SERCA protein expression between db/db and db/+ mice. It should be noted, however, that the experiments in the latter study were conducted on 12-wk mice, whereas our investigation used older mice (ages 15–22 wk). Moreover, our 10-wk mice, which were closer in age to those of Belke et al., showed no changes in SERCA2a expression—a finding not entirely inconsistent with the protein data.

Finally, we determined the myocardial mRNA expression of ANP. ANP expression remained unchanged at 5 and 10 wk but was increased 4.5-fold (P = 0.053) and 2.4-fold (P < 0.01) at 15 and 22 wk, respectively. Although to our knowledge no studies have evaluated ANP expression in db/db mice, ANP upregulation has been detected in other models of Type 2 diabetes (27, 30). Thus our data are in agreement with these reports. Interestingly, Christoffersen et al. (20) recently found no changes in ANP expression in ob/ob mice; it is possible that differences in strain, age, severity of diabetes, and/or cardiac function might have accounted for this discrepancy.


This study comprises the first use of CMR in a murine model of diabetes. In doing so, we have provided a description of the cardiac remodeling process with significantly more detail than has been reported previously. Initial manifestations, seen between 5 and 13 wk of age, include increases in LV mass, wall thickness, and end-diastolic volume. Later, starting at week 13, the mice enter a stage featuring a decreased ejection fraction, and, notably, no further changes in LV mass, wall thickness, or volume. These imaging findings are complemented by a progressive decrease in systemic and myocardial insulin sensitivity, as documented by insulin tolerance testing, serum glucose and insulin, and [18F]FDG scanning. The timing of the structural changes is therefore coincident with that of the insulin resistance and suggests that the ventricular remodeling seen in db/db mice may represent a true manifestation of diabetic cardiomyopathy.

The structural and functional changes in our db/db cohort closely mirror those of diabetic cardiomyopathy observed in the clinical setting. The natural history of this condition is marked initially by concentric hypertrophy and diastolic dysfunction. Echocardiographic studies of diabetic patients without cardiac complications have consistently reported increased LVM, increased wall thickness, and diastolic relaxation abnormalities (23, 38, 60), even without the potential confounding effects of systemic hypertension (55). Overt systolic dysfunction generally does not become manifest until more advanced stages of the disease (46), although subclinical contractile abnormalities are often detected in asymptomatic individuals with diabetes (26).

To provide a functional parallel to the imaging data, we performed pressure-volume loops on db/db and db/+ mice at 15 wk. Although we did not appreciate any significant differences in the steady-state measurements, the assessment of load-independent parameters, obtained during temporary caval occlusion, revealed several notable differences between groups. Ees, PRSW, and the slope of the dP/dtmax-end-diastolic volume relationship were all decreased in the db/db group, consistent with systolic dysfunction in the diabetic mice. Similarly, the slope of the EDPVR was increased in the db/db group, suggesting increased ventricular stiffness, decreased compliance, and diastolic dysfunction. While we cannot directly correlate these results with the imaging findings, the CMR data also provide evidence of systolic (i.e., decreased ejection fraction) and diastolic dysfunction (i.e., increased wall thickness). It is of interest that the load-independent parameters were significantly impaired in the db/db mice, while the steady-state parameters, which tend to be less sensitive indicators of contractile function, were not. Perhaps this should not be surprising, given that the contractile dysfunction seen on CMR is also somewhat subtle, as demonstrated by the relatively small magnitude of the ejection fraction differences between the db/db and db/+ groups.

Evaluation of the relative transcriptional activity of myocardial genes such as α-MHC, β-MHC, SERCA2a, and ANP provides further evidence of a myopathic process at the molecular level. Our data indicate that changes occur in the expression of these genes, and that these changes are coincident with the onset and progression of the structural and metabolic changes seen in the db/db group. The recapitulation of the fetal gene program (i.e., an MHC isoform switch from predominantly α to β, decreased SERCA2a, and increased ANP) classically occurs to varying extents in response to a number of mechanical insults, including thoracic aortic constriction (15, 34), pacing-induced heart failure (52), and aortocaval shunting (37). However, there is a growing body of evidence that certain metabolic disturbances, including hypothyroidism (24), dexamethasone treatment (21), gonadectomy (57), and a high-cholesterol diet (32), produce these changes as well. In light of our findings, and those of others (14, 39), it would appear that insulin-resistant diabetes is no different. Our results do not provide insights into the complex interplay among the molecular, metabolic, and structural changes seen in these mice; they only suggest an association. To further dissect the relationships between these observations, additional studies are clearly warranted.

Because of technical limitations, previous applications of CMR to diabetic models have been limited to animals no smaller than rats. Nevertheless, our findings are in agreement with the extant literature. Al-Shafei et al. reported a decrease in LVEF (4), along with increased LVM (3) and increased normalized LVEDV (4), in rats treated with streptozotocin. These findings were recently recapitulated by Loganathan et al. (41). Finally, in a study using Type 2 diabetic Goto-Kakizaki rats, Iltis et al. (33) also reported increased normalized LVM and decreased LVEF. Although it is problematic to compare imaging results across differing species, models, and types of diabetes, these studies, along with ours, nevertheless demonstrate that CMR is a feasible and effective method of assessing the structural and functional changes associated with diabetic cardiomyopathy.

Our results correlate with those of existing echocardiographic studies on db/db mice (7, 8, 16, 36, 45, 48, 58), although only two such investigations consisted of longitudinal evaluations with multiple time points. Semeniuk et al. (58) studied db/db mice at 6 and 12 wk of age. They demonstrated age-dependent increases in LV mass, wall thickness, and end-diastolic dimension but, contrary to our findings, did not see any differences in these parameters relative to control. Beyond the obvious differences in testing conditions, image acquisition, and data analysis between echocardiography and CMR, a more likely reason for this discrepancy may be the relatively short 6-wk interval between scans. Of note, fractional shortening (FS) in the 12 wk db/db group was 26% lower relative to control (0.60 vs. 0.44), indicating a considerably more impressive contractile deficit than in our mice. Although it is difficult to fully reconcile this finding with our data, Semeniuk et al. performed their studies on restrained, unanesthetized mice. It is therefore possible that the acute stress of test-related manipulation affected the results, a contention supported by the systematically increased FS across all study groups, diabetic and nondiabetic alike. The importance of this issue is further highlighted by the original report on the use of echocardiography on unanesthetized mice, in which the investigators described their careful efforts to train their animals before scanning (66). In this study, which, like that of Semeniuk et al., used 12 wk, 28-g controls, the FS in the control group was considerably lower (0.51). More recently, Barouch et al. (7) performed echocardiograms on db/db mice at 8 and 24 wk. Distinct from the findings of Semeniuk et al., but consistent with our findings, they demonstrated an age-dependent increase in LV mass and wall thickness relative to control, as well as a nonsignificant increase in end-diastolic dimension. Of note, LVM was greater in the study of Barouch et al. compared with our results, especially at 24 wk. However, as above, the inherent differences between CMR and echocardiography must not be discounted. It should also be pointed out that the mice in the study of Barouch et al. were heavier than those in our study.

Because of the potential confounding effects of systemic hypertension, we additionally measured blood pressures in db/db mice and their controls. While we did not discover any significant differences in systolic, diastolic, or mean arterial pressure, there exists some divergence within the literature with regard to the possibility of hypertension in db/db mice. For example, Bagi et al. (6) found a significant increase in noninvasively derived MAP, while more recently Oakes et al. (43) documented no differences in MAP with a pressure transducer inserted into the left carotid artery. Interestingly, both Bagi et al. and Oakes et al. reported blood pressures that were nearly 30 mmHg higher than ours. Possible explanations for this include differences in experimental protocol and sex (male in Bagi et al. and Oakes et al. studies, female in our study), although it is admittedly difficult to resolve such a large discrepancy. Importantly, both the db/db and db/+ groups were systematically hypotensive relative to those of Bagi et al. and Oakes et al., suggesting a difference in methodology rather than an intrinsic cardiac abnormality. In any event, our data nonetheless suggest that hypertension is unlikely to be a significant factor in the cardiac remodeling seen in db/db mice.

Separating the cardiotonic effects of obesity from those of diabetes in the db/db group is obviously a complex issue. However, it is important to mention that the increased LVM in our animals, as examined by CMR, is underlain not only by increased LVEDV but also by concentrically increased wall thickening. Given that 1) there is no significant hypertension in our animals and 2) in the clinical setting obesity-related cardiac hypertrophy in the absence of hypertension is typically eccentric (22, 42), it becomes more difficult to attribute the remodeling process solely to obesity.

While several mechanisms (e.g., defective calcium homeostasis, lipo- and glucotoxicity, decreased cardiac efficiency, mitochondrial dysfunction; for review see Ref. 5) have been put forward to explain the contractile dysfunction in db/db mice, the pathogenetic basis for the early structural remodeling and hypertrophy is less certain. The progression of changes observed on CMR imaging resembles the classic compensatory hypertrophy-decompensated failure sequence of overload-induced cardiomyopathy (35), although it is yet to be entirely determined what stressor (or stressors) this hypertrophic response is compensating for. While we have argued that hypertension itself is not a significant contributor, this does not rule out the presence of neurohormonal activators such as angiotensin II, norepinephrine, and endothelin in our model. These well-known modulators of hypertrophic signaling are increased in diabetes (12, 28, 61) and are therefore likely to be important to the remodeling process. An additional possible etiologic factor is hyperleptinemia. Indeed, leptin has been demonstrated to induce a hypertrophic response in cultured cardiomyocytes (47, 65), although leptin administration has also been shown to regress established hypertrophy in ob/ob mice (7).

Interestingly, several of the canonical metabolic disturbances in Type 2 diabetes do not cause cardiac hypertrophy and, in some cases, are paradoxically antihypertrophic. Insulin resistance itself has long been associated with hypertrophy (23, 29), although a direct causal link has never been established. Hypertrophy has also been associated with increased glucose uptake (59) and decreased fatty acid oxidation (53)—precisely the opposite of what is observed in db/db mice (1, 10, 14). Similarly, hyperglycemia, though believed to be a significant player in the loss of contractile function, is not known to induce hypertrophy, and in one study was even shown to inhibit angiotensin II- and phenylephrine-mediated hypertrophy (44). Finally, although hyperinsulinemia in isolation is prohypertrophic (54), in diabetes the associated insulin resistance implies defective insulin signaling by definition. Since the activity of downstream effectors such as phosphatidylinositol 3-kinase and Akt is therefore impaired, it is improbable that even supraphysiological levels of insulin could produce substantial hypertrophy in this setting.

It is admittedly uncertain whether cardiomyopathy contributes materially to the morbidity of db/db mice. As befits a systemic process, these animals suffer additional complications, such as nephropathy (40), delayed wound healing (63), and peripheral neuropathy (50). db/db mice typically survive only 10–14 mo; although the cause of death is believed to be cardiovascular in origin (8), the true mode of exodus has yet to be clearly identified. Nevertheless, nearly every study (including ours) that has serially evaluated ventricular function in these mice has documented decreased contractility over time, regardless of how the evaluation was performed. Further investigation using older animals will need to be conducted to provide more insight into this issue.

In summary, as assessed by CMR imaging, db/db mice undergo structural changes and ventricular remodeling consistent with diabetic cardiomyopathy. Similar to its human analog, this myopathy features an initial hypertrophic phase, characterized by increased wall thickness and LVM, followed by progressive dilatation and overt contractile dysfunction. The ability to assess the structural consequences of diabetes in this manner makes CMR a potentially valuable tool to study diabetic cardiomyopathy, as well as to test potential therapies designed to modify its natural history.


This study was supported by funds from a Stanford University Dean's Fellowship Award (P. Yue), an American Heart Association Beginning Grant-in-Aid (J. Wu), and National Institutes of Health Grants 5F32-HL-078108 (P. Yue), 5K23-EB-002063 (P. C. Yang), and 1R01-DZK-071333 (P. S. Tsao).


We thank Alexander J. Glassford for technical advice and editorial assistance.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract