Electrocardiogram (ECG) Parameters and Measurements
Next we explored if a mismatched LD cycle altered cardiac function as determined by an electrocardiogram. To explore this possibility, we next examined changes to ECG parameters recorded at 3 timepoints: 1, 2 and 4 months of age. Given the difference in heart and cardiomyocyte size, BTBR and B6 were analyzed separately. ECG parameters were manually quantified and compared between LD cycles and sex. Cardiac parameters of interest are heart rate, RR interval, QTc interval, JT interval, and R amplitude.
Heart rate.
BTBR mice exhibited a slowing in heart rate as they aged (F(2,118) = 14.02, p < .001, ωp = .10, Fig. 1b), with the heart rate at 1 month being faster than at later time points (p < 0.05). Furthermore, females had slower heart rates than males (F(1,59) = 4.30, p = .042, ωp = .05). These main effects were modulated by a significant LD cycle by sex by age interaction (F(2,118) = 3.85, p = .024, ωp = .02) where heart rates tended to be lower in the mismatch LD cycle, particularly at later ages and more so in male mice (p < .01).
In B6 mice, while heart rates also tended to be slightly higher at earlier ages, this pattern was not significant (F(2,118) = 2.69, p = .074, Fig. 1a). Additionally, while heart rates tended to be lower in the mismatch LD cycle at later ages, this interaction was also not significant (F(2,82) = 2.97, p = .057).
RR Interval.
We next examined the RR interval. RR interval changed in BTBR mice over time (F(2,118) = 16.73, p < .001, ωp = .12, Fig. 1e) being shortest at one month of age and longest at two months of age. RR interval was also longer in female BTBR mice (F(1,59) = 6.23, p = .015, ωp = .08). RR intervals were slightly longer in the mismatched LD cycle, although this difference was not significant (F(1,59) = 2.91, p = .093). The lack of a main effect here might have been influence by a weak but non-significant 3-way interaction between sex, age, and LD cycle (F(2,118) = 3.00, p = .053, ωp = .02).
B6 mice displayed significant differences in the length of the RR interval between LD cycles overtime (F(2, 82) = 3.64, p = .031, ωp = .03), with B6 mice in the misaligned LD cycle showing significantly longer RR intervals at four months of age compared to those housed in a normal LD cycle (Fig. 1d; p = .004). None of the other main effects or interactions were significant.
QTc Interval.
As alterations in RR interval could indicate altered heart function, we next examined several more specific ECG measures. We first examined the heart-rate corrected QT interval (QTc). A prolonged QTc interval is associated with increased risk of ventricular arrhythmias and cardiac hypertrophy23,27,36 while a shortened QTc is associated with a variety of heart conditions, including sudden cardiac death36. BTBR mice house in the mismatched LD cycle had shorter QTc intervals (Fig. 1h; F(1,59) = 9.11, p = .004, ωp = .12). QTc was longest at 1 month of age, and shortest at 2 months of age (F(2,118) = 8.56, p < .001, ωp = .06). Males had longer QTc intervals than females (F(1,59) = 33.80, p = .004, ωp = .11). There were no significant interactions between sex, age and LD cycle.
With B6 mice, there was a significant shortening of QTc as they got older (F(2,82) = 3.68, p = .029, ωp = .03, Fig. 1g). Although female mice decreased consistently overtime, this interaction with age was not significant (F(2,82) = 3.01, p = .055). LD cycle did not significantly alter QTc interval in the B6 mice (F(1,41) = 0.42, p = .522, ωp = .00). There were no other main effects or interactions for QTc in B6 mice.
JT Interval.
We next examined if the treatments altered the JT interval. Because the J wave accounts for 90% of the repolarization of the murine heart, abnormalities in this segment have been associated with heart failure and aging25,26. The mismatched LD cycle led to a significant decrease in the JT interval in BTBR mice (F(1,59) = 10.96, p = .002, ωp = .14), although this difference was only observed at the 1 month of age timepoint (age x LD cycle interaction, F(2,118) = 13.63, p < .001, ωp = .10; Fig. 1k), after which the JT intervals normalized. Additionally, females had a shorter JT interval than males (F(1,59) = 33. 4.79, p = .033, ωp = .06). There were no other main effects or interactions for JT in the BTBR mice.
B6 mice housed in the mismatched LD cycle had a longer JT interval than those in the match 24h LD cycle (F(1,41) = 9.29, p = .004, ωp = .16, Fig. 1j) at one month old (p < .001). There was no main effect of sex or age, and no further interactions.
R Amplitude.
We then looked at R amplitude as it has been shown to increase during development and correlate with normalized heart weight27,37. Female BTBRs had the largest R Wave amplitudes (main effect of sex, F(1,59) = 7.80, p = .007, ωp = .10; Fig. 1n), although this was only observed when they were in the matched LD cycle (sex x LD cycle interaction, F(1,59) = 13.74, p < .001, ωp = .17). There was a main effect of session, with most groups showing an increase in R amplitude as they aged, apart from females in the short LD cycle (F(2, 118) = 4.78, p = .010, ωp = .04). The R amplitude of these females decreased overtime (LD cycle by sex by session interaction; F(2, 118) = 3.28, p = .041). However, these last two p values should be interpreted cautiously given violations to normality.
The R Wave amplitude in B6 mice significantly increased with age (F(2, 82) = 3.47, p = .036, ωp = .03, Fig. 1m). There was no significant effect of LD cycle (F(1,41) = 0.27, p = .605), nor any significant interaction with LD cycle and the other variables.
Glucose Tolerance Test (GTT)
Fasted mice had a baseline blood-glucose level measured prior to receiving a bolus of glucose (2g/kg). They then had their glucose levels assessed at 15-, 30-, 60-, and 120-minutes after the injection. Based on these measures, an Area Under the Curve (AUC) was calculated for blood-glucose levels.
Results from GTT tests revealed a significant interaction between strain and LD cycle (F(1, 51) = 28.89, p < .001, ωp = .32; Fig. 3a) and sex and LD cycle (F(1, 51) = 6.25, p = .016 ωp = .08). Male B6 mice show poorer glucose tolerance in a misaligned LD cycle compared to other B6 mice (p < .001), and misaligned female B6 mice displaying significantly larger AUCs than female B6 mice in raised in a normal LD cycle (p < .01). Male BTBR mice raised in a misaligned LD cycle had larger AUCs than females in a misaligned LD cycle (p < .01), but not males in a matched LD cycle (p = .212). There were no differences found between BTBR females (p = .289).
Insulin Tolerance Test (ITT).
Finally, we examined clearance of glucose from the blood following administration of insulin. Fasted mice had a baseline blood-glucose level measured prior to receiving a bolus of insulin (0.75IU/kg). They then had their glucose levels assessed at 15-, 30-, 60-, and 120-minutes after the injection.
Following an insulin bolus, glucose clearance was impaired in mice under mismatched LD cycles (Fig. 3b; main effect of LD cycle; F(1, 46) = 19.52, p < .001, ωp = .26). BTBR mice had a smaller area under the curve (main effect of strain; F(1, 46) = 4.31, p = .044, ωp = .06). Males had impaired responses to insulin compared to females (main effect of sex; F(1, 46) = 19.65, p < .001, ωp = .26). Female BTBR mice were more sensitive to insulin relative to BTBR males and all B6 mice (strain by sex interaction; F(1, 46) = 8.90, p = .004, ωp = .13)