Early life stress blunts time-of-day-specific autonomic function in adult male mice

Abstract

Early life stress (ELS) is an independent risk factor for cardiovascular disease, including hypertension. The current study was designed to test the hypothesis that exposure to ELS disrupts autonomic function in adult mice at baseline and in response to an acute behavioral stressor. Using the mouse model of ELS, maternal separation with early weaning (MSEW), we determined the effects of ELS on heart rate (HR), blood pressure (BP), locomotor activity (LMA), HR variability (HRV) and spontaneous baroreceptor sensitivity (sBRS) in adult male mice. Under basal conditions, normally reared (NR) and MSEW mice had similar HR, BP, LMA, and sBRS when characterized at different phases of the circadian cycle. Spectral analysis of HRV indicated that MSEW mice had significantly lower total power and low frequency (LF) power during the inactive or light period as well as a loss of diurnal rhythm of LF power compared to NR mice. When subjected to an acute behavioral stress, cage switch stress (CSS), NR and MSEW mice showed similar changes in HR, BP, and LMA. Frequency domain analysis of HRV indicated that total, LF, and high frequency power during CSS and recovery from CSS were all significantly decreased in MSEW mice compared to NR mice. Taken together, these findings indicate that exposure to ELS in mice disrupts autonomic function basally and in response to an acute behavioral stress.

NEW AND NOTEWORTHY

Early life stress is a known risk factor for cardiovascular disease in adulthood. The current study provides evidence for impaired regulation of the autonomic nervous system in adult mice after being exposed to maternal separation with early weaning, an established model of neglect.

INTRODUCTION

Early life stress (ELS), or childhood adversity, is an important determinant of future mental and physical health. ELS is a known independent risk factor for the development of a wide range of chronic diseases in adulthood, especially cardiovascular disease (1). This association was revealed with a large cohort study demonstrating a dose-response relationship of ELS experiences to ischemic heart disease in adulthood (2). In a longitudinal study, greater ELS exposure was found to be highly associated with increased blood pressure (BP) trajectories from childhood to adulthood compared to those not exposed to childhood adversity (3). Furthermore, young adults with two or more ELS experiences displayed increased plasma endothelin-1, diastolic BP, total peripheral resistance, and arterial pulse wave velocity (4). Recently, Deschenes and colleagues reported that depression, anxiety, and cardiometabolic dysregulation partially mediate the association between ELS and coronary heart disease (5). ELS also associated with abnormal cardiovascular and endocrine stress responses in adulthood (8, 9). Estimates of the general US population show that ~58% experienced at least one ELS experience (6) and ~15% report four or more types of ELS experiences (7), thus the effects of ELS over the lifespan constitute a major public health concern and economic burden.

Appropriate daily rhythms of cardiovascular parameters such as HR and BP are associated with maintaining health. HR and BP follow a diurnal rhythm (8, 9). During the night BP falls by 10 to 20% in healthy individuals and the loss of this dipping pattern is associated with increased morbidity and mortality (8). Nocturnal hypertension, where higher than normal BP occurs at night with normotensive BP during the day, is associated with increased all-cause mortality and risk of cardiovascular events (10). ELS is linked to poor sleep quality in both adolescence and adulthood (11, 12). Time of day dependent changes in cardiovascular function is important for minimizing disease risk, however, little is known how exposure to ELS may affect circadian-controlled physiology.

ELS ‘sets the stage’ for the development of cardiovascular disease yet mechanistic insight is lacking. We sought to determine the relationship between ELS and autonomic control of the cardiovascular system. Low HR variability (HRV) has been established as an indicator of cardiovascular disease risk while elevated HRV is associated with increased cardiovascular fitness and well-being (13, 14). Allowing for variable time between heart beats is thought to reflect flexibility in rapid adjustment to changing internal conditions or environmental contingencies (15, 16). HRV is a net consequence of the effects of many interdependent physiological control systems that normally interact on different time scales to maintain homeostasis. Many physiological systems influence HRV including BP, vascular tone, hormones, thermoregulatory systems, gastrointestinal activity, respiration, chemoreceptor and baroreceptor reflexes, biological clocks, and the presence or absence of physiological and psychosocial stressors (17). Two important frequency domain parameters used to assess the contributions of sympathetic and parasympathetic components controlling HRV are high frequency (HF) and low frequency (LF) power (18, 19). There is a consensus that HF power reflects vagal tone. In contrast, LF power is regarded as representing a collection of sympathetic, parasympathetic, and baroreceptor influences on the heart.

While rodent models have contributed significant insight into the mechanisms involved in cardiovascular risk of ELS, few studies have examined the connections between ELS and circadian cardiovascular parameters. Additionally, we are unaware of any studies showing the effects of ELS on measures of HRV in rodents. We hypothesized that ELS in mice disrupts the time-of-day dependent changes in HR, BP, locomotor activity (LMA), baroreceptor sensitivity and measures of HRV in adulthood. Furthermore, we also investigated these cardiovascular responses after an acute psychosocial stressor, cage switch stress, in adult male mice exposed to MSEW and age-matched controls.

METHODS

Maternal Separation and Early Weaning (MSEW) in C57BL/6J mice

All studies and procedures were conducted in accordance with the guidelines recommended by the National Institutes of Health Guide for the Care and Use of Animals and approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham (UAB). C57BL/6J mice were originally obtained from Jackson Labs (Bar Harbor, ME) and maintained in-house at UAB. Mice were provided ad libitum food (NIH 31 Irradiated Diet, Envigo) and water and maintained on a 12:12 hour light/dark cycle (lights on at 0700 h and lights off at 1900 h) with constant temperature and humidity. Normal rearing (NR) or MSEW groups were generated as previously described (20, 21). Briefly, breeding pairs were placed together and females monitored daily to determine the exact date of birth (postnatal day (PD) 0). Litters were randomly assigned to MSEW or NR groups and litter size varied from 4 to 9 pups. On PD2 to PD5, MSEW litters were separated from the dam for 4 hours each day (0900 h to 1300 h). During separation, litters were placed in an incubator set at 37.5±1°C with 60% humidity with bedding. During PD6 to PD16, MSEW litters were separated from dams for 8 hours each day (0900 h to 1700 h), and on PD17, MSEW litters were weaned to new home cages. NR litters remained undisturbed with dams until weaning at PD21. For both MSEW (n=16) and NR (n=12) groups, male littermates were weaned together in co-housed groups of 2 to 6 dependent on litter size. The animals in this study were taken from 6 MSEW litters and 6 NR litters, each born to a unique dam, across a full calendar year.

Telemetry

BP, HR, and LMA were recorded via DSI^® telemetry transmitters (PA-C10, Data Sciences International, St. Paul, MN) implanted in the carotid artery of 10-wk old male mice, as previously described (22, 23). Mice were moved to individual cages on the day prior to surgery and remained individually housed for the rest of the study. Animals were allowed to recover for at least 10 days after surgery prior to telemetry recording. Data were recorded for 1 minute every 10 minutes at 2000 samples/second for the 3-day baseline prior to the cage switch stress (CSS) protocol. Hourly averages were derived from the average of the 6 measurements each hour. Telemetry data from 1 NR animal presented without clear consistent blood pressure waveforms and was excluded from the study.

Cage switch stress protocol

At 13–14 weeks of age, individually housed mice were placed in clean cages the day before baseline recordings began. On the day of CSS testing, individual NR and MSEW mice were placed into a cage soiled by unfamiliar mice at 1300 h. Soiled cages were generated by placing a group of 3 male mice in a cage with 950 mL of bedding for 72 hours prior to the start of the CSS test. MSEW and NR mice were left undisturbed for 4 hours in the soiled cage and then returned to their home cages at 1700 h and allowed to recover for 1 h. Food and water were available for the entirety of the CSS protocol. Telemetry data was recorded continuously at 2000 samples/second on the day of the CSS protocol from 0700h to 1900h.

Analyses of blood pressure, heart rate, locomotor activity, heart rate variability, and spontaneous baroreflex sensitivity

BP, HR and LMA data were obtained from telemetry recordings using DSI^® Ponemah^® software. For baseline data, each hourly average was calculated from the 3 consecutive days of recording during that hour. During CSS, data was collected continuously and binned into 30-minute intervals. HRV analysis was conducted offline in Ponemah^®. Table 1 summarizes the time domain and frequency domain indices used in the current studies. BP waveforms were visualized manually to confirm markers, and any periods of noise or artifactual triggering were excluded from analysis. For time domain analysis, data were reported in 1-minute segments and outliers greater than 3 standard deviations were removed. For frequency domain analysis, data were analyzed using Hanning windows in 20 s segments with 50 Hz quadratic interpolation. Frequency cutoffs were set for low frequency (LF: 0.40 – 1.5 Hz) and high frequency (HF; 1.5 – 4.0 Hz) ranges based on prior published studies (19, 22). The maximum derivative of the blood pressure signal was used to obtain inter-beat intervals.

TABLE 1.

Definition of heart rate variability parameters

Selected Time Domain Measures
Variable	Units	Description
NN Interval	ms	All normal-to-normal (NN) intervals
SDNN	ms	Standard deviation of the averages of NN intervals in all 5-minute segments
RMSSD	ms	Square root of the mean of the sum of the squares of differences between adjacent NN interval
pNN6	%	Percentage difference between adjacent NN intervals that are greater than 6 ms
Selected Frequency Domain Measures
Total power	ms2	Sum of all power from 0.0–4.0 Hz
LF	ms2	Low frequency power 0.4–1.5 Hz
HF	ms2	High frequency power 1.5–4.0 Hz
LF/HF	Ratio	Ratio of low frequency power to high frequency power

Spontaneous baroreflex sensitivity was determined from pulse wave data with Hemolab software generously provided by Professor Harald Stauss (available at: http://www.haraldstauss.com/HaraldStaussScientific/hemolab) for 1-hour periods spanning the 1300 h and 0100 h time points on each of 3 consecutive baseline days. During the CSS, BP, HR, LMA, and HRV analyses were placed into 30-minute bins to assess the response to the acute psychosocial stressor.

Statistics

Data are presented as mean ± SEM with p<0.05 as the criterion for significance. Data were compared by repeated measures 2-way analysis of variance (ANOVA) with the factors time and ELS (MSEW or NR) in GraphPad Prism 9.1.2. (LaJolla, CA) using Sidak’s post hoc comparison.

RESULTS

Baseline cardiovascular parameters of MSEW and NR mice

During the 72-hr baseline period, both NR (n=12) and MSEW (n=16) male mice exhibited clear circadian rhythms for HR (Figure 1A), systolic and diastolic BP (Figure 1B), and LMA (Figure 1C) with the highest levels attained during the dark (active) period with no significant differences found between the NR and MSEW groups. During the 12-hr light (inactive) period (0700 h to 1900 h), HR, BP, and LMA measures were significantly lower in the NR and MSEW groups than during the 12-hr dark (active) period (1900 h to 0700).

Figure 1:

Hourly heart rate (1A), blood pressure (1B), and locomotor activity (1C) were similar between mice exposed to maternal separation with early weaning (MSEW) (n=16) and normally reared controls (NR, n=12) under baseline conditions in a 12:12 hr light/dark cycle. Gray background denotes night/lights off time period. Statistical analysis was performed by repeated measures 2-way ANOVA.

Heart rate variability analysis and spontaneous baroreflex sensitivity during baseline

Time domain analysis of HRV showed a clear circadian difference between the light and dark phases for both MSEW and NR mice as indicated by the differences in the N-N interval and the standard deviation of the N-N interval (SDNN; Figures 2A, 2B). The square root of the mean squared successive differences between adjacent N-N intervals (RMSSD) indicated a significant diurnal difference only in NR mice (Figure 2C). The percent of N-N intervals that differed by more than 6 ms (pNN₆) showed a clear circadian difference for both groups, but this time domain index, like the N-N interval and SDNN indices, was similar for NR and MSEW mice (Figure 2D).

Figure 2:

Time domain analysis of heart rate variability showed diurnal variation of both MSEW (n=16) and NR mice (n=12) in the N-N interval (3A) and standard deviation of the N-N interval (SDNN, 3B). Root mean square of successive differences between normal heartbeats (RMSSD, 3C) exhibited diurnal variation in control, but not MSEW mice. pNN₆, the percentage of consecutive N-N intervals which differ by more than 6 ms, was similar between the two groups (3D). Statistical analysis was performed by repeated measures 2-way ANOVA with Sidak’s post hoc comparisons of individual means. Data were calculated over the 12h inactive (light) and active (dark) periods.

Frequency domain measures of HRV during the 72-hr baseline period indicated significant differences in autonomic function between MSEW and NR mice. Total power decreased during the active phase of the circadian cycle in both MSEW and NR mice. However, compared to NR mice, total power was significantly lower in MSEW mice during the light period (Figure 3A) compared to NR mice. LF power was significantly higher during the inactive period compared to the active period for both NR and MSEW groups. MSEW mice had significantly lower LF power during the inactive period compared to NR animals. MSEW mice did not show evidence of a circadian rhythm in LF power whereas NR mice did have a circadian rhythm in LF power (Figure 3B). HF power and the LF/HF ratio were similar between the two groups across the circadian cycle (Figures 3C, 3D).

Figure 3:

Heart rate variability total power (2A) was lower during the 12 hr inactive period in MSEW exposed mice (n=16) compared to NR controls (n=12) under baseline conditions. Low frequency (LF) power (2B) was lower during the 12 hr inactive period in MSEW mice compared to control mice. High frequency (HF) power (2C) and the LF/HF ratio (2D) was similar between both groups. Statistical analysis was performed by repeated measures 2-way ANOVA with Sidak’s post hoc comparison. Data are presented as 12h averages of the inactive (light) and active (dark) periods.

Baroreceptor activity is crucial to the maintenance of healthy blood pressure as part of the feedback system integrating autonomic activity and heart rate. Since our results suggested autonomic impairment in MSEW mice during both light and dark periods, it was important to determine if spontaneous baroreceptor activation was affected in these animals. Spontaneous baroreflex sensitivity was similar between NR and MSEW mice during 1300 h to 1400 h (mid-inactive) and 0100 h to 0200 h (mid-active) time periods assessed during a 72-hour baseline period (Figure 4).

Figure 4:

Spontaneous baroreflex gain was similar in MSEW (n=16) and NR mice (n=12) between 0530 to 0630 hours and 1730 to 1820 hours under baseline conditions. Statistical analysis was performed by repeated measures 2-way ANOVA. Data were calculated over the 12h inactive (light) and active (dark) periods.

Effects of the acute stressor CSS on HR, BP, LMA, and HRV

HR, systolic and diastolic BP, and LMA all increased in response to CSS, but there were no significant differences between the NR and MSEW mice (Figure 5A, 5B, 5C). Following CSS, there were increases in these measures during the one-hour following CSS when they were returned to their home cages. Previous studies have shown that the overall response to being handled and placed in their home cage is much lower less than the test cage (24). Time domain parameters of HRV, including the N-N interval, SDNN, RMSSD, and pNN₆ were not significantly different between NR and MSEW mice during CSS and the period following (Figures 6A, 6B, 6C, 6D). However, frequency domain analysis showed that MSEW mice had lower total, LF, and HF power compared to NR mice (Figures 7A, 7B, 7C). While the overall HRV frequency response to CSS was attenuated in MSEW mice (middle panels), this can be attributed to a lower pre-stress level of total, LF, and HF power as can be seen when data are expressed as a change from baseline (right panels). The LF/HF ratio was similar between NR and MSEW mice during CSS (Figure 7D).

Figure 5:

Heart rate (5A), blood pressure (5B), and locomotor activity (5C) parameters were similar between MSEW (n=16) and NR mice (n=12) during cage switch stress (CSS), denoted by the yellow background. Statistical analysis was performed by repeated measures 2-way ANOVA.

Figure 6:

Time domain measures of heart rate variability were similar between MSEW (n=16) and NR control mice (n=12) during the CSS protocol. N-N interval (7A), SDNN (7B), RMSSD (7C), and pNN₆ (7D) parameters showed an effect of time but not stress during CSS. Statistical analysis was performed by repeated measures 2-way ANOVA.

Figure 7:

Heart rate variability was lower in MSEW mice (n=16) compared to NR mice (n=12) throughout the CSS protocol (left panels). Total (7A), LF (7B) and HF (7C) power were lower in the MSEW mice during the CSS protocol (left panels) as were the sums of total, LF, and HF for the 4-hour stress period (middle panels). Change from baseline (right panels) was similar for total, LF, and HF. LF/HF ratio (7D) was similar between both groups. Statistical analysis was performed by repeated measures 2-way ANOVA.

DISCUSSION

Previous investigations found that ELS impacts stress reactivity and increased risk of psychopathological conditions in adult humans(25–27). However, there has been a lack of mechanistic studies that are crucial to determining mechanisms contributing to pathophysiology in human patients. We report that time domain analysis of HRV indicated a diurnal pattern in N-N intervals and SDNN in both NR and MSEW mice. RMSSD displayed a diurnal difference in NR mice, but not in MSEW mice, suggesting a basal impairment in vagal control of the heart that is variable across the 24-hour day. Frequency domain analysis, which characterizes low and high frequency range of HRV, showed a significant diurnal difference in total power for both NR and MSEW mice. Notably, during the inactive (light) phase, total power was lower in MSEW mice compared to NR mice. Although no difference was found in HF power during the inactive phase between the two groups, LF power during the inactive phase was lower for MSEW mice as compared to NR. We believe our study is the first to report diurnal measures of HRV in MSEW-exposed mice. These findings suggest that exposure to ELS leads to autonomic dysfunction in adult male mice.

Previous human studies have reported attenuated heart rate reactivity in individuals who experienced adverse childhood experiences (28). These observations have been attributed to reduced heart rate variability consistent with our current findings (29). Using frequency domain analysis, adults exposed to abuse (emotional, physical, sexual) and neglect (emotional, physical) had reduced LF HRV suggesting a specific attenuation of the sympathetic system (30). Using time domain analysis, Bakema et al. found no association between HRV and ELS during resting conditions after accounting for co-variants including medication and sociodemographic factors (31). In our mouse model, we are able to analyze HRV using both time and frequency domain analysis that allows us to examine potential changes throughout the day and night. We only observed an attenuation in the frequency analysis in the MSEW mice that was not observed in the time-domain analysis. The reason for this inconsistency is not clear and so further, more direct measures of autonomic function are needed.

Adverse physiological and psychosocial stressors at an earlier time in life act as strong etiogenic factors setting the stage to prime an individual to be vulnerable to stressors encountered later in life (32). It is worth noting that many of the baseline measurements such as BP and HR in the MSEW model are similar to NR mice even with attenuated HRV. These findings are consistent with clinical observations that the response to secondary stressor or the “second hit” hypothesis are what leads to a more rapid rise in cardiovascular risk. Indeed we and others have observed that the response to a secondary stressor such as low dose angiotensin II can lead to an exaggerated response in the ELS models (33, 34).

The autonomic system provides a mechanism to “reset” a new homeostatic point in response to chronic stressors. To examine the effects of ELS on the autonomic system and response to stressful events in adulthood, we examined the relative contributions of the SNS and PNS regulation of HRV after challenging adult male mice with CSS. In both the NR and MSEW groups, CSS triggered comparable rapid increases in HR, BP, and LMA, which were then followed by gradual declines over the course of the remainder of the 4-hour test. Upon return to their home cage both NR and MSEW groups showed comparable increases in HR, BP, and LMA to levels comparable to those present immediately after CSS. Frequency domain analysis during CSS indicated significant lower total, LF and HF power with MSEW mice showing lower values compared to NR controls. This relative impairment in the capacity to increase total, HF and LF power is likely a result of the lower baseline HRV in MSEW mice suggesting an impaired parasympathetic drive to the heart, since HF power reflects parasympathetic tone and both LF and total power have a parasympathetic component. This overall ELS-induced deficit in parasympathetic control of the heart may constitute a reduced adaptive capacity to rapidly buffer against repeated or sustained exposure to stressors in adulthood. A limitation of our studies is that we are deriving HRV from the inter-beat intervals from the blood pressure signal rather than actual R-R intervals taken from electrocardiographic recordings.

Autonomic dysfunction measured as low HRV is a well-established characteristic seen in people with depression (35, 36) and is considered one of the major pathways linking major depressive disorder to adverse outcomes in cardiac diseases (37). Far less is known about associations between HRV and ELS and the few studies conducted in humans to-date have shown varying results at resting conditions (29–31). Our group and others have published that MSEW mouse models of ELS often present with similar phenotypes to NR animals until exposed to a second hit or stressor (25, 38). In the present study, the effects of ELS on HRV were more pronounced after experiencing an acute psychosocial stress in adulthood. This finding may underlie the heterogeneous response to ELS seen in human population studies. The inclusion of frequency domain analysis at baseline and after stressful events in human and animal model studies will help to tease out the effects of ELS on HRV.

The current study utilized CSS as an acute psychosocial stressor to display ELS-dependent effects on HRV. It is worth noting that the CSS protocol was performed during the light or inactive period of the day, during the same time period that distinct changes were seen in LF power in the ELS exposed mice at baseline. Acute psychosocial stress during one’s inactive period, when sleep typically occurs, may have unique and unknown effects on the stress response. It is tempting to speculate that time of day and the circadian changes in hormone status, feeding, or activity may contribute to the unique phenotype of MSEW mice. Future studies should examine CSS or MSEW protocols during the active or night period.

Although, we found no differences in sBRS between the NR and MSEW group during the inactive or active time period, it is worth noting that at this age, both groups had similar blood pressures and heart rates. It is possible that sBRS, a measure of how well the baroreceptor regulates beat-to-beat changes in BP, is not altered in MSEW animals on normal diets with physiologically normal BP during young adulthood. Alterations in sBRS may become apparent if MSEW mice are challenged secondarily with “second hits” such as high fat diets, as obese male MSEW mice have elevated systolic blood pressure and autonomic tone compared to age-matched controls (39). A significant limitation of our study is the use of only male mice. Our initial experiments led us to believe that the MSEW model was less severe in females, but this has not been followed up. In human studes, there have been reports that men and women have sexually dimorphic responses to ELS in adolescence and adulthood (40) and so further studies are needed to focus on the possible physiological interactions between sex and ELS. Additionally, researchers should be cognizant of the many variables, such as circadian rhythms, that can alter the effects of ELS in rodent models, which may contribute to some of the confounding results reported in the field (41).

Epidemiological evidence shows a robust association between ELS and cardiometabolic pathologies, including hypertension (3, 42–44). The pathophysiological pathway through which ELS leads to increased risk of CVD is likely to be multifactorial, with autonomic dysfunction as one of the potential underlying physiological mechanisms. Reductions in HRV predict poor cardiovascular health outcomes in populations without baseline CVD (45) and patients with CVD history (46, 47). HRV is a non-invasive index of autonomic regulation and spectral analysis of HRV offers a powerful tool to delineate the relative influence of sympathetic and parasympathetic determinates of HRV. It is crucial that future studies of HRV consider time of day as a biological variable to help overcome the gaps in the field. Identifying and defining functional mechanisms that link ELS to later adult cardiovascular impairments is likely to lead to discovering novel interventions or prevention strategies.