Heart and lung weights in DCM mice
We first compared the heart to body weight (HW/BW) ratio with lung to body weight (LW/BW) ratio in WT and ΔK210 DCM mice at 1, 2, and 3 months of age () before starting the experiments with the running wheel. The average HW/BW ratio was approximately 1.8-fold higher in DCM mice than in WT mice at 1 and 2 months of age (, ), and even at birth, as reported previously 
. The average HW/BW ratio at 3 months of age in DCM mice (1.44±0.63%) was 2.9-fold higher than that in WT mice (0.50±0.05%). Comparison of the 10–90 percentile ranges between 2- and 3-month-old DCM hearts indicates that the HW/BW ratio in the upper (90th percentile) range increased from 1.08% to 2.48%, while that in the lower (10th percentile) range remained constant at 0.81% to 0.78% (, ). These results suggest that further cardiac enlargement proceeded during the period between 2 and 3 months of age in some DCM mice.
HW/BW and LW/BW ratios in individual WT and DCM mice at 1–3 months of age.
Averages and ranges of HW/BW and LW/BW ratios in individual WT and DCM mice at 1, 2, and 3 months of age.
The LW/BW ratio was measured to detect signs of CHF (, ). There were no significant differences in LW/BW ratios between DCM and WT mice at 1 and 2 months of age (approximately 0.8±0.1% and 0.6±0.1%, respectively). At 3 months of age, however, the average LW/BW ratio was significantly higher in DCM (1.21±0.73%) than in WT (0.60±0.08%) mice. Close examination revealed that one half of the DCM mice showed LW/BW ratios similar to those of the WT mice (0.5–0.7%), whereas the remaining half of the DCM mice showed clearly higher LW/BW ratios (1–3%). This suggests that at 3 months of age, some DCM mice developed pulmonary edema, probably secondary to CHF.
Voluntary wheel running activity in WT and DCM mice
We hypothesized that development of CHF may be reflected by a decline in voluntary exercise activity. To test this hypothesis, DCM and WT mice at 5 weeks of age or older were housed in a cage with a running wheel for 48 consecutive hours every 8 to 10 days, and voluntary running distances were recorded. This standard exercise protocol did not affect the survival curve of DCM mice. After learning how to run on the running wheel during 1 to 2 trials (2–4 days), mice at 2 months of age showed almost constant running activity. This wheel running behavior was absolutely voluntary, not forced or reward-associated. The voluntary running distances were recorded until sacrifice at approximately 3 months of age. Until 2 months of age, running distances per day were similar between WT and DCM mice (mean±SD, 7±2 km/day; 10–90 percentile, 4–11 km/day) (, left). Thereafter, at around 3 months of age, the running activity of WT mice remained unchanged (mean±SD, 7.2±2.6 km/day; 10–90 percentile, 4.2–11.4 km/day). On the other hand, the average running activity in DCM mice was significantly reduced; some DCM mice lost running activity, whereas others maintained a high activity ( right) (mean±SD, 4.9±4.2 km/day; 10–90 percentile, 0.02–10.3 km/day).
Relationship between running activity and LW/BW ratio and histological lung sections stained with Berlin blue.
Correlation between running activity and CHF
The relationship between running activity just before sacrifice and the LW/BW ratio was plotted ( left). Mice with lowered voluntary running activity showed larger LW/BW ratios; all mice with running activity of <1 km/day had considerably larger LW/BW ratios (mean±SD, 1.67±0.57%; range, 1.2–2.8%;), whereas all mice with running activity of >5 km/day had LW/BW ratios similar to those of the WT mice (mean±SD, 0.60±0.05; range, 0.5–0.7%) (, left and right). A group of mice with intermittent running activities between 3 and 5 km/day included both normal and high LW/BW ratios. These results suggest that low running activity (<1 km/day) is a good indicator of CHF. To confirm this prediction and compare running activity with other indicators of CHF, we obtained data related to CHF in DCM mice with high (HI: >5 km/day) and low (LO: <1 km/day) activities in subsequent experiments.
In the next experiment, we examined whether hemosiderin deposition was present in histological lung sections in the LO group of DCM mice. Hemosiderin-loaded macrophages are thought to appear in the lung as a result of phagocytosis of red blood cells that have leaked from alveolar capillaries secondary to pressure overload in left ventricular HF 
. Indeed, Berlin blue-positive alveolar macrophages were detected near capillaries in lungs from LO DCM mice, but never in lungs from HI DCM or WT mice ().
Correlation between running activity and other markers for HF
Several factors and markers, such as the size of the ventricular chamber, ECG signals, expression of the β-myosin heavy chain (β-MHC) isoform, and ANP and BNP levels, have been used to assess the severity of HF. We examined correlations of these factors with voluntary exercise activity.
First, echocardiography data were compared between the HI (n
5) and LO (n
4) groups (). Because DCM mice are highly sensitive to anesthesia including inhalation and intravenous anesthetics, these data were obtained from conscious mice. Four of eight LO mice died before measurements were obtained, which was probably related to restraint 
. A significantly higher left ventricular end-diastolic dimension (LVDd) and significantly lower ejection fraction (EF) were recorded in the LO mice that survived restraint than in the HI mice (). Thus, the LO group showed decompensated HF with ventricular dilation, consistent with one of the major phenotypes of DCM. However, these results indicate that echocardiography increases the risk of death in severe HF model mice.
Echocardiography and ECG data from WT and DCM mice.
ECG data of WT, HI, and LO mice were also compared (). No significant differences were observed in heart rates between WT and DCM mice (, left). The average QRS interval was significantly longer in the LO group than in the WT and HI groups (, middle), which is consistent with ventricular enlargement (). The average QT interval was significantly longer in HI and LO DCM mice than in WT mice (, right). Although the average QT interval was longer in LO than in HI mice, a significant difference was not detected (, right).
We next compared body and heart weights between the HI and LO groups to directly examine whether cardiac enlargement was associated with decompensated HF in these DCM model mice. Body weights were similar among the three groups, although those of the LO group were slightly but significantly smaller than those of the WT group (). The HW/BW ratio was significantly higher in the LO group than in the HI group; i.e., LO mice displayed more severe cardiac enlargement (). shows the relationship between HW/BW and LW/BW ratios. There was a positive correlation between HW/BW and LW/BW ratios (r2
0.33) in DCM mice, suggesting that CHF and secondary cardiac dilation proceeded almost concurrently.
Heart weights, myocardial fibrosis, and protein expression level of β-MHC in myocardium.
Cardiac fibrosis and the protein expression level of β-MHC were compared in the myocardium of the LO and HI groups because these factors were noticed in this DCM model as well as in other HF models 
. The LO group showed more advanced fibrosis than did the HI group (), consistent with prolonged a QRS on the ECG. The protein expression level of β-MHC, expressed as the relative expression of β-MHC to total (α-MHC plus β-MHC) MHC, was markedly higher in the LO than in the HI group (). These results confirm that cardiac fibrosis and up-regulation of β-MHC were much more severely pronounced in the deteriorated LO DCM mice.
Gene expression levels of ANP, BNP, and β-MHC, which have been related to HF, were quantitatively determined in the same set of left ventricular myocardial preparations from the HI and LO groups. The expression level of ANP was higher in most of the HI mice than in the WT mice, and was further elevated in the LO mice (, ). Although a significant difference was detected between the HI and LO groups (), there was considerable overlap in individual values between the two groups (). Similar results were obtained in terms of the expression levels of BNP and β-MHC (, middle and right, respectively).
mRNA expression levels of ANP, BNP, and β-MHC in left myocardium.
Expression of ANP, BNP, and β-MHC levels and HW/BW and LW/BW ratios in myocardium from WT and HI and LO groups of DCM mice.
To further evaluate quantitative correlations between expression levels and lung edema, the LW/BW ratio was plotted against the ANP level (). Linear regression analysis revealed no significant correlation. Similar conclusions were obtained with the expression of BNP and β-MHC (, middle and right). These results indicate that development of CHF/lung edema is not always reflected in the expression of ANP, BNP, and β-MHC.
On the contrary, there was a better quantitative correlation between the ANP level and cardiac enlargement than between the ANP level and lung edema; linear regression analysis revealed a significant correlation between the HW/BW ratio and ANP expression (, left). There was also a significantly positive correlation between the HW/BW ratio and BNP level (, middle) and between the HW/BW ratio and β-MHC level (right). These results indicate that the expression of ANP, BNP, and β-MHC better reflect cardiac enlargement than development of lung edema, which accompanies decompensated HF.
Causes of death in DCM mice
Finally, to determine whether ΔK210 DCM mice die with severe CHF or die suddenly without symptoms of CHF, we monitored the natural course of voluntary exercise activity in DCM mice until death. shows the representative data on running activity of DCM mice that died suddenly while maintaining high running activity. Among 29 mice examined, 15 maintained high running activity (>5 km/day) and died suddenly at 1.5 months or later. Their LW/BW ratios were similar to those of WT mice (see ; HD). In contrast, 12 DCM mice began to show decreased running activity at some point after 2 months, and then finally died within 10 to 20 days with running activity of <1 km/day (). All of these mice showed high LW/BW ratios (see ; LD). To our surprise, the number of mice in the middle group (MD) (see ) between the HD and LD groups (1–5 km/day) was smaller than expected (only 2 of 29). They decreased their running activity to 20% to 40% of the basal activity and died with a high LW/BW ratio, indicating lung edema (; MD). The fractions of SCD (HD) and CHF mice (LD+MD) estimated from running wheel activity were thus similar, approximately 50% each. The HW/BW ratio was smaller in the SCD group than in the CHF group () with a small overlap. Thus, low running activity not only reveals CHF, but also predicts the prognosis of DCM mice. The time course of running activity is also useful for the prediction.
Time courses of voluntary running activity and causes of death in DCM mice.
A previous report showed torsades de pointes at the time of death in DCM mice 
. In this study, similar ventricular arrhythmias were also captured at death in all DCM mice (n
3) with telemetry ECG recordings. Furthermore, we recorded ventricular arrhythmia in mice that ran >5 km/day (n
2). shows a representative ECG record obtained from a 2.5-month-old mouse with running activity of 7 km/day. This mouse survived ventricular arrhythmia that continued for 15 minutes and thereafter maintained high running activity for 1 month, occasionally showing less severe premature ventricular complexes and T wave alternans, and finally died of ventricular arrhythmia after development of CHF at 4 months. These results indicate that ventricular arrhythmia could occur in this DCM mouse model both before and after development of CHF, consistent with prolonged QT intervals before development of CHF. These results provide further evidence for lethal ventricular arrhythmia as a major cause of sudden death in inherited DCM before development of CHF.