INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

AN ESTIMATED 40 MILLION INDIVIDUALS in the United States are susceptible to chronobiological and sleep-related disease processes, which represents a serious public health threat.¹ The increased incidence of sudden cardiac death in the early morning hours, combined with disease processes such as sleep apnea and sudden infant death syndrome, exemplifies the need to better understand the modulation of cardiorespiratory control attributable to chronobiological mechanisms. The purpose of the present study is to characterize genetic factors that modulate variation in the circadian rhythm of behavior and physiological indicators of cardiorespiratory control.

The composite mechanisms underlying circadian rhythm generation have been recognized in a behavioral context as a complex trait (e.g., Ref. 17), implying that the genetic regulation is polygenic [i.e., determined by multiple quantitative trait loci (QTL)]. Variation between inbred mouse strains in behavioral circadian periodicity, phase, and amplitude have been established (e.g., among C57BL/6 and BALB/c mice), indicating a robust genetic component comprising a suite of QTL (1, 15, 17). With the use of recombinant inbred strains derived from C57BL and BALB/c progenitors, Mayeda et al. (4, 11, 12) determined seven potential QTL for circadian periodicity by using variation in physical activity patterns (Act), including candidate regions on mouse chromosomes 1, 2, 3, 5, and 16. From a recent genome-wide screen of F2 offspring derived from C57BL and BALB/c progenitors, Shimomura et al. (17) estimated that 14 QTL significantly affected the circadian period, phase, and amplitude in a complex epistatic interaction. As an alternative to QTL analysis, the classic report of Vitaterna et al. (28) used mutagenesis to identify the clock locus as a candidate gene that prolongs the free-running circadian period and abolished rhythmicity in mutants. This single locus is mapped to mouse chromosome 5, which is syntenic to human chromosome 4.² In summary, much of the evidence in mice concerning the genetic determinants of circadian pattern generation emphasizes variation in behavioral phenotypes by incorporating parallel strategies of QTL and mutagenesis techniques (21).

Our laboratory has focused on the genetic basis for variation in cardiorespiratory control mechanisms with the use of inbred mouse models. For example, among more than 10 inbred mouse strains, the C3H/HeJ (C3) and C57BL/6J (B6) strains occupy the two extremes of various strain distribution patterns for ventilatory responses, including breathing differences when exposed to room air and during acute hypercapnic and hypoxic challenges (26). By using quantitative genetic approaches, these studies have demonstrated that specific breathing traits are inherited by first- and second-generation offspring of C3 and B6 progenitors (25, 27). In addition, QTL have been identified that link specific breathing phenotypes to genomic regions on mouse chromosomes 3 and 9 (22-24). The motivation for the present study was based on recent evidence supporting a significant role of circadian rhythm in altering cardiorespiratory control mechanisms, including control of breathing during hypoxia and hypercapnia (16, 18-20). Hence, as a prelude to studying the impact of circadian rhythm on control of breathing in this model, differences between C3 and B6 strains were characterized to test the hypothesis that major genetic determinants modulate the circadian rhythm of Act, deep-body temperature (T_db), and heart rate (HR).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. Male C3 (n = 11), B6 (n = 11), and B6F1 (n = 8) mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and were housed four to six mice per cage in the animal facilities at the Johns Hopkins University. The reciprocal male C3F1 mice (n = 3) were propagated from a cross between C3 female and B6 male progenitors. The age (55-125 days of age) and presurgical body weight (23-34 g) distributions were similar among the four groups of mice. The environments before and during the experiments as well as animal handling were highly standardized. Water and mouse chow (Agway Pro-Lab RMH 1000) were provided ad libitum. All animal protocols were reviewed and approved by the Animal Care and Use Committee of the Johns Hopkins Bloomberg School of Public Health.

Surgical procedures. Act, HR [i.e., confirmed by electrocardiogram (ECG) recordings], and T_db were measured simultaneously with a transmitter implant and a radiotelemetry system (Data Sciences International, St. Paul, MN). The weight of the transmitter (model TA10ETA-F20) was ~3.5 g, and its dimensions were 2 cm long, 1 cm wide, and 0.7 cm deep. The implant surgery was initiated by obtaining the animal's presurgical weight and anesthetizing each animal with a mixture of acepromazine (0.5 ml at 10 mg/ml) and ketamine (5 ml at 100 mg/ml) at a dose of ~2 µl/g. The hair covering the abdomen and chest wall was clipped and further removed using a depilatory. Surgery was performed by placing the animal on a heating pad, applying betadine to the exposed region of skin, and establishing a sterile field surrounding the animal. An abdominal midline incision was made, and the transmitter was inserted and sutured to the abdominal muscle. The negative ECG lead was guided through the muscle and directed subcutaneously to the right shoulder. The positive ECG lead, also guided through the muscle, was directed laterally (left side) and positioned ~1 cm below the rib cage. Both leads were sutured to secure a lead placement resembling lead II in traditional human ECGs. Surgery was completed within 30 min, and recovery from anesthesia generally occurred within 60-90 min. Each animal was placed in a holding cage set on a heating pad for the first 24 h after surgery. Each animal was allowed to recover for at least 2 wk before the start of data collection. Individual cages were refreshed every Friday at 1400 h, and weekly body weights were recorded. Additional details concerning the telemetry system and output variables have been described elsewhere (7).

Data acquisition and analysis. Data were collected while each animal was individually housed in a facility maintained at 23-24°C with a light-dark cycle set at 12-h intervals (i.e., lights off occurring at 1800). For each individual animal, four 24-h samples were obtained for each variable at 30-min intervals. The measurement of Act was a cumulative determination obtained from translocation motion counts sampled during the last 10 min of each 30-min interval. Because the ACT measurements had much greater phase-dependent variances in the dark phase compared with light phase, a Box-Cox transformation was used, as shown by the following equation
This standard statistical procedure is intended to determine a power transformation, which confirms the data to be normally distributed with equal variances.

Statistical model. The initial analysis described above suggested to model Act, T_db, and HR measurements as the following
with the index i = 1, 2 representing strains C3 and B6, j = 1, ..., 11 representing the 11 mice in each strain (numbers assigned arbitrarily), k = 1, ..., 4 representing the 4 days of replicates, and t representing time in hours. Here, S_i represented the population daily average for mice of strain i and M_ij represented the deviation of mouse j from the population average. The periodic function, C_i(t) with a period of 24 h, represented the population circadian pattern for mice of strain i. The periodic function, D_ij(t), represented the deviation in circadian pattern of mouse j from the strain-specific circadian pattern. Last, _ijk(t) represented the environmental variation and the experimental measurement error. Previous studies (29) have used models in which C_i(t) was assumed to be a cosine wave with free parameters for the amplitude and phase. Estimates obtained from such models were unable to capture differences in shape of the circadian pattern between C3 and B6 strains. For this reason, the statistical model used in the current study assumed simply that the periodic function estimates, C_i(t), were smooth. The details regarding the fit of this statistical model have been reported elsewhere (5).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The average (±SE) 24-h daily mean, minimum, maximum, and span for C3, B6, and both F1 mice are reported in Table 2. Although the 24-h daily mean Act did not differ between C3 and B6 mice, the daily mean T_db was significantly (P < 0.01) lower in C3 compared with B6 mice. In contrast, the daily mean HR was significantly greater in C3 compared with B6 mice. These strain differences were also observed in both the average minimum and maximum T_db and HR. The 24-h daily average T_db and HR for F1 mice were significantly (P < 0.01) different from the C3 progenitor and did not differ from the B6 progenitor. Although the groups were not different in terms of the minimum activity, the minimum T_db of F1 mice was significantly (P < 0.01) greater than both progenitors. In contrast, the minimum HR was significantly (P < 0.01) lower in F1 mice relative to C3 mice but did not differ from B6 mice. In addition, the maximum Act and T_db of F1 mice was greater than both strains, whereas the maximum HR of the F1 was significantly (P < 0.01) greater than B6 mice but did not differ from C3 mice.

The periodic function estimates, C_i(t), for the C3 and B6 strains are shown in Fig. 1. Departures from the strain-specific 24-h daily average are illustrated for Act, T_db, and HR to highlight the shape differences between strains. The confidence bands did not intersect during various intervals for Act, T_db, and HR, which strongly refuted the null hypothesis, suggesting the C3 and B6 strains demonstrate similar shapes in circadian pattern. Strain-specific phenotypes between C3 and B6 mice were most prominent during intervals late in both the light and dark phases. The C_i(t) estimates for both F1 mice are also shown in Fig. 1. Although the shape of the circadian pattern did not differ between B6F1 and C3F1 mice for Act and T_db (i.e., the confidence bands intersect throughout the 24-h period; data not shown), the shape of the F1 responses differed from both C3 and B6 progenitor strains.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Strain differences in the shape of the circadian pattern for activity (Act; A), deep-body temperature (Tdb; B), and heart rate (HR; C) are depicted for C57BL/6J (B6; n = 11 mice) and C3H/HeJ (C3; n = 11 mice) progenitors and both F1 offspring [B6C3F1/J (n = 8 mice) and C3B6F1 (n = 3 mice)]. The light-dark cycle was set at 12-h intervals (i.e., bar along abscissa represents the dark phase). For each of the 3 measurements, deviations from the strain-specific 24-h daily average (denoted as 0 on each ordinate) were illustrated to highlight the shape differences between strains. The most prominent trait difference between C3 and B6 mice was associated with secondary peaks in Act, Tdb, and HR 1-2 h before the dark-to-light transition.

The strain differences during two periods of the dark phase are highlighted in Fig. 2. Although Act and T_db were similar between C3 and B6 strains early in the dark phase, these responses were significantly (P < 0.01) greater in both F1 offspring compared with the progenitors. In contrast, HR responses early in the dark phase were significantly (P < 0.01) lower in B6 relative to C3 and F1 mice. The most prominent strain difference between C3 and B6 mice occurred during a time interval late in the dark phase ~1 h before the dark-to-light transition. Here, secondary peaks in Act, T_db, and HR occurred in B6 and F1 mice that were not evident in C3 mice. In B6 mice, the secondary peaks in Act, T_db, and HR were similar in magnitude as the first peaks. In F1 mice, the secondary peaks were more modest relative to the first peaks, and the responses were intermediate in amplitude relative to both progenitors. Accordingly, during the dark phase, robust broad-sense heritability estimates suggested that the HR during the early period and Act and T_db during the late period may serve as phenotypes to pursue future QTL studies.

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Individual responses are depicted from 2 relatively stable 1-h periods early (hours 7.5-8.5 in Fig. 1) and late (hours 16-17 in Fig. 1) in the dark phase. For Act, Tdb, and HR, broad-sense heritability estimates (h2) are reported for each period. Results associated with the secondary peaks in Act and Tdb suggest that a robust genetic determinant modulates the strain variation in circadian rhythm late in the dark phase. Also, a significant genetic influence determines the strain difference in HR early in the dark phase. * P < 0.01 vs. C3 mice. § P < 0.01 vs. B6 mice.  P < 0.01 vs. both C3 and B6 mice.

As reported in Table 2, the consistently higher HR in C3 relative to B6 mice was a second prominent strain difference. In Fig. 3, strain differences in HR were accentuated during two periods of the light phase, periods where Act variation was modest or negligible between groups. During both the early and late periods of the light phase, HR was significantly (P < 0.01) lower in B6 and B6F1 mice compared with C3 mice. The C3 and B6 strain difference in HR was maximum late in the light phase, as suggested by a robust broad-sense heritability estimate.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Individual responses are depicted from 2 relatively stable 1-h periods early (hours 21-22 in Fig. 1) and late (hours 2-3 in Fig. 1) in the light phase. For Act, Tdb, and HR, broad-sense heritability estimates (h2) are reported for each period. During the light phase, where Act variations are modest or negligible between groups, a robust genetic determinant modulates strain variation in HR. * P < 0.01 vs. C3 mice. § P < 0.01 vs. B6 mice.  P < 0.01 vs. both C3 and B6 mice.  P < 0.01 vs. both B6 and B6F1 mice.

To compare phase interactions in the circadian pattern shapes of Act, T_db, and HR among C3, B6, and F1 mice, the C_i(t) functions for the three measurements were standardized (i.e., rescaled to a range between 0 and 1) and were plotted in Fig. 4 for each strain. A third prominent strain difference was related to the rise in Act, T_db, and HR that occurred in association with the light-to-dark transition. The rise in Act appeared to be similar between C3 and B6 mice but was earlier in F1 mice (Fig. 1). The rise in T_db occurred at a slower rate in C3 compared with B6 and F1 mice; however, F1 mice demonstrated a rise that was earlier than B6 mice. During the light-to-dark transition, the rise in HR was more gradual in C3 mice compared with B6 and F1 mice. The changes in HR occurred in closer proximity to the changes in Act and T_db in C3 mice. In contrast, the HR changes in B6 mice uniformly preceded the rise in Act and T_db. Although F1 and B6 mice demonstrated parallel HR responses to the light-to-dark transition, the magnitude of the HR response was greater in F1 mice.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Phase interactions between Act, Tdb, and HR are depicted for B6 (n = 11 mice; A) and C3 (n = 11 mice; B) progenitors and both F1 offspring [B6C3F1/J (n = 8 mice) and C3B6F1 (n = 3 mice)] (C). Scaling units between 0 and 1 were used to standardize Act, Tdb, and HR variables. Bar along abscissa represents the 12-h dark phase. Although the 3 variables change concurrently in C3 mice, HR precedes Act and Tdb in B6 and F1 mice.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of the present study was to characterize the circadian pattern in Act, T_db, and HR between C3 and B6 inbred mice. The results suggest that at least two major genetic determinants modulate the strain variation in circadian rhythm between C3 and B6 mice. One genetic determinant regulates the strain difference associated with secondary peaks in Act, T_db, and HR. At 1-2 h before the dark-to-light transition, secondary peaks are prominent in B6 mice and, to a modest extent, in F1 mice but not in C3 mice. A second genetic determinant regulates the strain differences associated with a consistently greater HR in C3 relative to B6 mice (e.g., ~70 beats/min at both the acrophase and bathyphase of the circadian period). Finally, both genetic determinants appear to be inherited independently in reciprocal first-generation offspring.

Because the C3 and B6 strain comparison has been used in many studies of lung physiology, the observations from the present study are especially important to the systems physiologist interested in the genetic modulation of cardiorespiratory control. For example, Watkinson et al. (29) studied the circadian rhythm of C3 and B6 mice to demonstrate short-term strain differences following ozone exposure, i.e., a potent oxidant air pollutant. The cosinor model used to characterize the ozone-induced strain differences in circadian rhythm did not emphasize the biphasic response in Act and T_db observed in B6 mice (see Fig. 3 in Ref. 29). In other genetic models of lung physiology, including studies of asthma, strain variation between C3 and B6 mice have been investigated to show that specific QTL influence differential lung inflammatory responses to ozone and other pollutant exposures (6, 13). In our laboratory, variation between C3 and B6 mice has been studied to explore the genetic determinants that regulate differences in the control of breathing (23, 24). Because circadian rhythm is known to influence control of breathing (16) and other lung physiological responses in health and disease (e.g., sleep apnea, nocturnal asthma), a long-term objective of this study considers the complex genetic interaction underlying the circadian regulation of the cardiorespiratory system.

The biphasic response pattern in Act and T_db of the B6 strain is indicative of a strong secondary arousal mechanism (i.e., the neurohumoral events associated with a secondary increase in Act and T_db) that is apparently absent in C3 mice. The secondary arousal in B6 mice was not associated with any obvious proximal stimulus from the environment. Indeed, immediately before the secondary arousal, a 2-h period of inactivity occurred in both C3 and B6 mice, and declines in T_db and HR were similar between strains (Fig. 1). In B6 mice, the magnitude of the secondary arousal appeared to resemble the anticipated primary arousal associated with the light-to-dark transition; that is, Act, T_db, and HR were similar between first and second peaks. Although a secondary arousal was evident in reciprocal F1 mice, the secondary peaks in Act and T_db were significantly attenuated compared with the primary arousal event. This observation suggested that first-generation offspring inherit an intermediate phenotype relative to the C3 and B6 progenitors. The intermediate phenotype of F1 mice is associated with only a modest change in HR late in the dark phase. Like the study of Watkinson et al. (29), this biphasic response pattern of B6 mice has been depicted for HR in data illustrations from other studies. For example, Li et al. (8) showed that a secondary rise in HR ~2 h before the dark-to-light transition was associated with an increase in mean arterial blood pressure (MAP) in B6 mice. Although the results of the current study suggest that this secondary arousal is a robust phenotypic difference between C3 and B6 mice, the specific genetic determinant is obviously unknown. One candidate hypothesis suggests that a spike in melatonin secretion from the pineal gland late in the dark phase may play a role in the secondary arousal mechanism (9).

The phenotypic differences in HR between C3 and B6 mice suggested that another major genetic determinant influenced the regulation of cardiac rhythm. The results from the present study (Table 2) are consistent with some but not all previous studies. For example, a recent study by Mattson (10) suggested that C3 mice demonstrated a significantly higher HR than B6 mice (i.e., average HR was 665 beats/min for C3 vs. 594 beats/min for B6), whereas MAP was similar between strains. To the contrary, Desai at al. (2) reported that baseline HR in C3 mice were significantly lower relative to B6 mice (i.e., average HR was 476 beats/min for C3 vs. 501 beats/min for B6 mice), whereas MAP was significantly depressed in C3 mice. The results from the present study appear consistent with the former study (10) and not with the latter study (2). The mechanism(s) by which HR differs between strains may be related to the balance in sympathetic and parasympathetic autonomic control of HR. For example, HR regulation may be more vagally mediated in B6 compared with C3 mice, resulting in a consistently lower HR. To the contrary, however, a recent study (3) in conscious B6 mice suggested that the intrinsic HR was likely determined by enhanced sympathetic activity because combined sympathetic and parasympathetic pharmacological blockade resulted in a reduced HR. Atropine alone did not alter HR in B6 mice (3). Given the results surrounding the B6 strain, one might postulate that the strain difference in HR regulation observed in the present study is attributable to an even greater sympathetic activity in C3 compared with B6 mice. Alternatively, other electrophysiological mechanisms may be contributing to the variation in cardiac rhythm between C3 and B6 mice.

The results from the present study also advance the role genetics play in regulating HR by demonstrating that parental phenotypes are inherited by first-generation offspring. Indeed, parental strain variation in HR occurred independent of activity (Fig. 1). However, HR in F1 mice was significantly greater than in B6 mice during the dark phase or active period (Fig. 2) and significantly lower than C3 mice during the light phase or inactive period (Fig. 3). Because Act during the dark phase was significantly higher in F1 mice, the unique HR phenotype of this offspring class may have been partially activity dependent. During the light phase, however, the strain differences in HR were less likely influenced by activity. Under these conditions, the HR phenotype of F1 mice appeared to resemble the B6 parental strain to a greater extent than the C3 strain (Fig. 3). In general, the results (Fig. 1 and Table 2) support the conclusion that F1 mice have a greater span in HR between the acrophase and bathyphase compared with both B6 and C3 parental strains.

The Act, T_db, and HR associated with the light-to-dark transition presented another potentially useful phenotypic difference to distinguish between C3 and B6 mice (Fig. 4). Although the HR was coincident with Act and T_db in C3 mice, the circadian regulation of HR varied more independently of Act and T_db in B6 and F1 mice. In B6 mice, for example, the separation between HR and Act was prominent at three different time periods: 1) at the light-to-dark transition, 2) between the two peaks in activity, and 3) at the dark-to-light transition (Fig. 4). Although the F1 demonstrated a similar separation between HR and Act, these responses were more coincident in C3 mice, especially after the peak in activity. Strain variation in the cardiorespiratory mechanisms associated with sleep-wake cycling may differ between C3 and B6 mice. For example, Schaub et al. (14) showed that HR in B6 mice fell during the transition from quiet wakefulness to non-rapid eye movement sleep (i.e., on average 610 to 566 beats/min) but did not change with the transition to rapid eye movement sleep (i.e., 564 beats/min). The results from the present study suggest that the average minimum HR in B6 mice is ~510 beats/min (Table 2). The bradycardia associated with the bathyphase of the circadian period may induce cardiovascular adjustments in advance of increases in activity associated with the light-to-dark transition. Alternatively, other anticipatory mechanisms (i.e., humoral) may modestly increase HR and T_db before increases in activity.

Although there was generally a tight coupling between Act and T_db among the groups during most periods of the circadian cycle, strain differences in T_db were not consistently attributable to Act (e.g., early light-phase, Fig. 3). The results, reported in Table 2, suggested that C3 mice demonstrated a modestly lower T_db (i.e., 0.3-0.4°C) compared with B6 and F1 mice. We considered this to be an important finding because our genetic studies of ventilation are technically and mechanistically dependent on strain variation in T_db. With respect to the technical importance, the measurement of tidal volume using whole-body plethysmography relies on a temperature gradient between animal and ambient conditions. Because our studies were conducted during the animal's natural light phase, an error of ~5% in computing tidal volume was estimated. Therefore, the strain differences in T_db cannot account for the C3 and B6 strain variation in tidal volume, which has been an informative phenotype in the genetic studies of differential breathing (24).

The current results are mechanistically important to the integration of metabolism, temperature regulation, and cardiorespiratory control; systems that are regulated by a circadian period. If a specific genetic variant determines the magnitude of a secondary arousal event, then the same gene likely modulates strain differences in specific metabolic, thermoregulatory, and cardiorespiratory outcomes. For example, the modest strain differences in T_db between C3 and B6 mice does not appear to account for the strain variation in breathing phenotypes observed in previous studies. On the other hand, if the same breathing studies were conducted during different periods of the dark phase rather than the light phase, the gene that determines the secondary arousal might likely modulate strain variation in chemical control of breathing. However, a cursory inspection of cosegregating QTL suggests that the genes determining differential breathing phenotypes are distinguishable from those associated with circadian pattern generation (4, 11, 12, 17, 22-24).

In summary, the results of the present study indicate that specific genes modulate differences in circadian rhythm between C3 and B6 mice. One gene determines the magnitude of a secondary arousal event, and another gene determines variation in the circadian regulation of HR. Heritability estimates and the response patterns of reciprocal F1 mice suggest that specific periods within the 24-h circadian cycle might be used in this genetic model to uncover novel genes that modify circadian periodicity. Finally, this genetic model may prove productive in demonstrating the significant genetic interaction between determinants of circadian rhythm and cardiorespiratory control.