Maternal exercise during pregnancy promotes physical activity in adult offspring

Abstract

Previous rodent studies have shown that maternal voluntary exercise during pregnancy leads to metabolic changes in adult offspring. We set out to test whether maternal voluntary exercise during pregnancy also induces persistent changes in voluntary physical activity in the offspring. Adult C57BL/6J female mice were randomly assigned to be caged with an unlocked (U) or locked (L) running wheel before and during pregnancy. Maternal running behavior was monitored during pregnancy, and body weight, body composition, food intake, energy expenditure, total cage activity, and running wheel activity were measured in the offspring at various ages. U offspring were slightly heavier at birth, but no group differences in body weight or composition were observed at later ages (when mice were caged without access to running wheels). Consistent with our hypothesis, U offspring were more physically active as adults. This effect was observed earlier in female offspring (at sexual maturation). Remarkably, at 300 d of age, U females achieved greater fat loss in response to a 3-wk voluntary exercise program. Our findings show for the first time that maternal physical activity during pregnancy affects the offspring’s lifelong propensity for physical activity and may have important implications for combating the worldwide epidemic of physical inactivity and obesity.—Eclarinal, J. D., Zhu, S., Baker, M. S., Piyarathna, D. B., Coarfa, C., Fiorotto, M. L., Waterland, R. A. Maternal exercise during pregnancy promotes physical activity in adult offspring.

Environmental exposures during critical periods of embryonic, fetal, and early postnatal life affect the development of metabolic regulatory mechanisms, with lifelong consequences for susceptibility to disease (1, 2). Maternal overnutrition and obesity, for example, can promote obesity and related metabolic disorders in offspring (3). In fact, such developmental programming can have transgenerational effects that perpetuate overweight and obesity across successive generations (4). Targeting interventions to critical periods of development may therefore be an effective approach to curbing the global obesity epidemic (1, 5).

Most studies of developmental programming of obesity and energy balance have focused on food intake (6), largely neglecting energy expenditure and, in particular, physical activity (the component most amenable to change) (6, 7). Physical activity can be categorized as either voluntary exercise or spontaneous physical activity (SPA). Voluntary exercise describes physical activity that is not essential to survival and homeostasis, and SPA is nonvolitional activity that results from daily living (fidgeting and random muscle contractions) (8). In rodent studies, voluntary exercise is typically modeled by wheel running, whereas cage activity reflects SPA (8).

Rowland (9) proposed that one’s propensity for physical activity is set at a specific level known as the “activity-stat,” a component of the control center that regulates individual energy balance. Although there are clearly genetic influences on the activity-stat (10), its setting may also be affected by environmental influences during critical periods of development. For example, rodent studies have shown that maternal food restriction (11, 12), consumption of a low-protein diet (13, 14), and hyperleptinemia (15) during pregnancy affect physical activity in adult offspring. Our recent studies in 2 models of developmental programming of energy balance also indicate early influences on physical activity. In the small-litter (SL) mouse model of early postnatal overnutrition, adult female SL offspring exhibited decreased physical activity and energy expenditure (16). In the agouti viable yellow (A^vy) mouse model of maternal obesity, offspring of obese A^vy/a dams experienced fetal growth restriction but, when cross fostered to lean a/a dams, exhibited postnatal catch-up growth that was followed, in the females only, by persistently blunted physical activity and increased adiposity (17). We noted that obese adult A^vy/a females are extremely inactive (17), and therefore, in addition to growth restriction, their offspring are exposed to low levels of movement during fetal development. This finding suggests that perhaps maternal physical activity before and during pregnancy affects the establishment of the activity-stat in the offspring.

Previous rodent studies have tested whether maternal exercise before and during pregnancy affects metabolic outcomes in the offspring. In particular, 3 different groups recently reported that maternal physical activity before and during pregnancy improves overall glucose homeostasis, apparently by increasing insulin sensitivity (18–21). Some of these studies also reported increased lean mass and reduced adiposity in the offspring (18, 20). We postulated that these metabolic improvements result from an increased propensity for physical activity in the offspring of exercised dams. Indeed, there is precedent to suggest that maternal physical activity during pregnancy affects neurologic development in the offspring, leading to persistent improvements in short-term memory (22) and enhanced spatial learning (23).

We therefore set out to test the hypothesis that when female mice are allowed to engage in voluntary exercise (wheel running) before and during pregnancy, their offspring will be more physically active.

MATERIALS AND METHODS

Mouse husbandry and diet

This study was approved by the Baylor College of Medicine Animal Care and Use Committee, and animals were maintained in accordance with all relevant federal guidelines. Male and female C57BL/6J isogenic mice were obtained at 8 wk of age from The Jackson Laboratory (Bar Harbor, ME, USA). The study was conducted in 2 phases over the course of 1 yr (February 2014 through February 2015). Phase 1 yielded 4 L and 2 U litters; phase 2 yielded 4 litters per group. The mice were fed either a fixed-formula, soy protein–free maintenance diet (2020X; Harlan Teklad, Madison, WI, USA) or, during mating, pregnancy, and lactation, a fixed-formula, soy protein–free reproductive diet (2919X; Harlan Teklad).

Experimental design

The experimental design is illustrated in Fig. 1. In each phase, 12–16 virgin female mice, aged 8 wk (F0) were caged individually in rat cages equipped with a solid-surface running wheel (diameter 14 cm; Pets International, Ltd., Arlington Heights, IL, USA) and odometer (Cateye Velo 7; CatEye America, Boulder, CO, USA). A preliminary test was conducted to select dams to be included in the experiment. The females were allowed access to running wheels for 2 wk and were ranked based on average daily running distance during the last week (referred to as preliminary test d 1–7). In each phase, the most consistently active females were chosen for the study. Selected females were randomly assigned to either the U or L group and caged with their respective running wheels. In the L group, the wheels were locked to prevent running. After this 1-wk acclimation period, male C57BL/6J mice were introduced into the cages. The females were checked each morning for a vaginal plug. Once a plug was observed, the male was removed from the cage. Wheel activity (distance and average velocity) was recorded daily during the acclimation and mating period, as well as throughout pregnancy and lactation. Wheels of the U dams were locked on postnatal day (P)10 to prevent the pups from using them. Only offspring from a birth litter of 7–10 pups were included; a total of 6 U and 8 L litters were studied.

Figure 1.

The overall study design.

Metabolic cage studies

On P21, offspring were weaned and selected for metabolic cage studies. From each litter, 1 male and 1 female mouse were chosen based on body weight (closest to the average sex-specific body weight in the litter). These mice were then placed individually into Comprehensive Laboratory Animal Monitoring System (CLAMS) cages (Columbus Instruments, Columbus, OH, USA). For ∼2 wk, food intake, energy expenditure (by indirect calorimetry), and physical activity were monitored (17). Each CLAMS cage was equipped with optical beams to measure home cage activity (SPA), and an instrumented running wheel (diameter, 9.8 cm) to measure wheel activity (voluntary exercise). It should be noted that wheel running does not cause any beam breaks, so cage activity is independent of wheel running. This process was repeated at P60 and P160. During the periods between the metabolic cage studies, the mice did not have access to running wheels. Across the 2 phases of the study, 12 U offspring (6 male and 6 female) and 12 L offspring (6 male and 6 female) were studied in the CLAMS cages. The same mice were studied in the metabolic cages at each age; all remaining offspring were observed longitudinally for body weight and composition. During the CLAMS studies, the cages were calibrated every 5–6 d, in accordance with the manufacturer’s guidelines and protocols published elsewhere (17). At each study period, minimal energy expenditure was estimated for each mouse by averaging the lowest energy expenditure values from each of 3 d.

Body weight and body composition

Offspring body weight was recorded at P1, P21, P60, P120, P160, and P300. Body composition was measured on P21, P60, and P160, before the selected offspring were placed into the CLAMS cages. After the end of P160 measurements, body weight and composition were measured in all offspring, including those that were not studied in the CLAMS cages. Body weight was measured with a calibrated integrating scale and body composition by quantitative magnetic resonance (EchoMRI-100; Echo Medical Systems LLC, Houston, TX, USA), according to the manufacturer’s instructions.

P300 voluntary exercise study

At P300, female offspring (n = 12 U and 11 L) were individually housed in rat cages, each equipped with a running wheel (14 cm diameter) and an odometer (Cateye Velo 7) for a period of 3 wk. Body weight and composition were measured before and after the 3-wk period.

Statistical approaches

To eliminate data entry errors, all data in the study were entered independently into 2 Excel spreadsheets (Microsoft, Redmond WA, USA) and verified electronically. Study phase (phase 1 or 2) was included in all statistical models, but was not significant in any. Daily running distances over the 1 wk preliminary test (before group assignment) of L and U dams were compared by repeated-measures ANOVA (Proc Mixed, autoregressive covariance structure; SAS Institute, Cary, NC, USA), with “day” as the variable in the repeated effect. Group differences in body weight and body composition at single ages (e.g., at P1) were compared by repeated-measures ANOVA, with litter as the variable in the repeated effect (thus adjusting for the nonindependence of pups born to the same dam). Body weight comparisons across multiple ages were performed by repeated-measures ANOVA (SAS Proc Mixed). To adjust for potential effects of housing in the CLAMS cages, with access to running wheels, inclusion in the CLAMS studies was a categorical variable in the body weight and composition analyses. For the CLAMS studies, at each age, the first several days in the metabolic cages were considered an acclimation period, and the data were discarded. This method yielded 8 or more consecutive days of data for all 24 mice at each age. Data were sampled at 30 min intervals, so each day yielded 48 data points per mouse for each variable. These data (food intake, wheel running, cage activity, and energy expenditure) were analyzed by repeated-measures ANOVA, with age (d) and time (at 30 min intervals) as variables in the repeated effect. Models for intake (g/30 min), distance traveled on wheel (km/30 min), cage activity (counts/30 min), and expenditure (kcal/30 min) included main effects of group, time, and age (d), as well as time × time, time × time × time, time × group, and age × group interactions. Study phase (phase 1 or 2) was evaluated in all models but was omitted, as it was never significant. Models for Intake and Expenditure included lean mass (g) and fat mass (g) as independent variables to adjust for effects of mouse body weight and composition (24). Initially, the full CLAMS models included sex and sex × group interaction, and light (on or off), and light × group interaction. The sex × group and light × group interactions were generally significant, so models were run separately by sex and light. The P300 body composition data were analyzed by repeated-measures ANOVA (with “week” as the repeated effect).

RESULTS

Maternal voluntary exercise before randomization and before and during pregnancy

A preliminary test was conducted to ensure that all randomized dams engaged in similar levels of voluntary exercise. Dams with average daily running distance ∼10 km and the least day-to-day variation were selected (Fig. 2A) and randomly assigned to either the U or L group. The results of randomization (Fig. 2B) show that there were no overall differences between the 2 groups during the 7 d preliminary test (P = 0.58).

Figure 2.

Preliminary and experimental data on maternal exercise. A) A 7 d preliminary test was conducted to determine the activity level of the dams. Compared to those not selected, selected dams showed less variation in running distance by the end of the preliminary test. B) Of the dams studied, there was no significant difference in baseline exercise level among those randomized to the U or L groups. P = 0.58. C) Running distances of the U dams were recorded daily during acclimation, mating, pregnancy, birth, and early lactation. [Note that during the mating period (average. 2.0 d), we did not distinguish between male and female running.] Daily running distance decreased over pregnancy and was nearly zero during lactation. D) Average running velocity of U dams was fairly stable during early pregnancy but declined precipitously during the third trimester.

U and L dams were then individually caged with their respective running wheels for 1 wk before mating. For mating, males were in the cages for an average of 2 d (range, 1-4). Average daily running distance of U dams declined from 15.1 to 10.5 km during the acclimation period, and progressively diminished during pregnancy (Fig. 2C). Notably, however, even at the beginning of the 3rd trimester of pregnancy, U dams were still running ∼3 km each day. Aside from a small postpartum surge, running was minimal during lactation. Average running velocity also decreased during pregnancy (Fig. 2D). Hence, although the U dams still ran ∼1 km/day even in late pregnancy, average running velocity was about half that at the beginning of pregnancy.

Maternal exercise before and during pregnancy increases offspring birth weight, but does not affect adult body weight or body composition

On P1, there was no group difference in litter size [L = 8.5 ± 0.4 (means ± sem) and U = 8.0 ± 0.4 pups/litter, P = 0.4], but U pups were significantly heavier than L pups (P = 0.007) (Fig. 3A). By P21, there were no longer group differences in body weight, and no differences were observed thereafter (Fig. 3B). Likewise, we detected no differences in body composition (% fat) between U and L offspring at P160, in either males or females (Fig. 3C). These growth and body composition data represent both the offspring that underwent CLAMS studies and those that did not. In females, inclusion in the CLAMS studies had no effect on growth or P160 body composition (P = 0.44 and P = 0.51, respectively) but among males, those included in the CLAMS studies followed a slightly lower body weight trajectory (P = 0.001) and were slightly leaner at P160 (P = 0.007). There were no significant group × CLAMS interactions in either sex.

Figure 3.

Group differences in body weight were found only at P1. A) P1 body weight of U pups (n = 6 litters) was higher than that of L pups (n = 8 litters). P = 0.007. Box plots show median (thick line), 25–75th percentiles (box), and 5–95th percentiles (whiskers) of 48 U and 68 L pups; sex was not ascertained. B) There was no group difference in body weight at P21, P60, P120, or P160 (n = 23–41 mice/group per sex at each age). C) There was no group difference in percentage body fat at P160 (n = 23–41 mice/group per sex).

Maternal exercise before and during pregnancy increases offspring physical activity and energy expenditure

Our previous studies (16, 17), indicated that developmental programming of physical activity behavior is often sex specific. We therefore wished to determine whether potential sex-specific effects of maternal exercise were evident in the offspring directly after weaning or only after sexual maturation. Accordingly, physical activity (both wheel running and home cage activity), energy expenditure, and food intake data were collected from male and female U and L offspring at P21 (weaning), P60 (sexual maturation), and P160 (adult).

In a few cases, subtle group differences in food intake and wheel activity were evident during lights-on (resting) periods (Fig. 4). However, because food intake and physical activity are relatively low during the lights-on period, our description of the results focuses on the dark period. Relative to L males, U males exhibited no dark-period differences in food intake or energy expenditure at any age (Fig. 4A, 4D). U males showed no differences in physical activity at P21 or P60, but by P160 they ran more (Fig. 4B) and moved in their cages more (Fig. 4C) than L males. (P21 energy expenditure data were not collected because of an instrument malfunction.)

Figure 4.

Food intake, physical activity, and energy expenditure of L and U offspring. A) Active period (lights-off) food intake was slightly higher in U than in L females at P60. The opposite trend at P160 was not statistically significant (Table 1). B) Wheel running distance during the active period was higher in U than in L females at P60 and P160. This effect was observed in males at P160 only. C) U males showed higher cage activity only at P160. In females, overall cage activity was higher in U than L females at P60; at P160, a significant group × time interaction (Table 1) reflected higher activity during the early dark period. Increased cage activity was also observed in U vs. L males, but only at P160. D) Dark-period energy expenditure was slightly higher in U than in L females at P60. (In each panel are shown mean ± sem for 6 mice of each sex per group across 10, 9, and 8 d at P21, P60, and P160, respectively. The same 6 male and 6 female mice per group, each from a different litter, were studied at each age. Overall effect of Group during each 12-h period: *P < 0.05; **P < 0.01; ***P < 0.001. Significance of group × time interactions is listed in Table 1.

More consistent and substantial effects, particularly in physical activity and energy expenditure, were observed in female offspring (Fig. 4). At P60, dark-period food intake was slightly higher in U than in L females (Fig. 4A); however, the opposite trend was observed at P160. At P21, there were no group differences in wheel activity (Fig. 4B) or total cage activity (Fig. 4C) among the females. By P60, U females ran more (Fig. 3B) and moved more than L females (Fig. 3C). Consistent with this result, overall dark-period energy expenditure was slightly higher in U than in L females (Fig. 4D). The increase in wheel running observed in P60 females persisted to P160 (Fig. 4B). Moreover, although the overall trends for increased cage activity (Fig. 4C) and energy expenditure (Fig. 4D) in U females at P160 were not statistically significant, the significant group×time interactions (Table 1) reflect a sustained elevation in cage activity and energy expenditure of U females during the early part of the dark cycle.

TABLE 1.

Summary of CLAMS statistical analysis: dark period only

		Male						Female
		P21		P60		P160		P21		P60		P160
Variable	Factor	F	P	F	P	F	P	F	P	F	P	F	P
Intake	Group	0.0	0.96	1.6	0.21	0.0	0.92	3.5	0.06	8.0	0.005*	1.4	0.23
	Time × group	1.6	0.21	0.0	0.93	0.2	0.69	0.0	0.87	0.1	0.7	0.6	0.45
Running	Group	0.1	0.81	3.6	0.06	8.0	0.005*	2.9	0.09	11.6	0.0009*	5.4	0.02*
	Time × group	3.1	0.07	1.6	0.21	47.6	<0.0001*	1.2	0.28	9.6	0.002*	79.6	<0.0001*
Cage Activity	Group	3.7	0.06	2.2	0.14	7.8	0.005*	2.6	0.11	4.1	0.04*	3.1	0.08
	Time × group	1.7	0.20	0.0	0.88	10.7	0.001*	1.4	0.25	21.4	<0.0001*	23.1	0.003*
Expenditure	Group	No data		1.1	0.31	1.0	0.31	No data		5.0	0.03*	1.4	0.23
	Time × group			0.2	0.62	20.7	<0.0001*			2.2	0.14	29.2	<0.0001*

In addition to the group effect and time×group interaction, all models also included Time (30 min increments; linear, as well as second- and third-order polynomial), age (d), and age × group. The models for intake and expenditure included lean mass (g) and fat mass (g). Experiment phase (phase 1 or 2) was evaluated in all models, but was not significant and so was omitted from the final models. Variable units were intake (g/30 min), running (km/30 min), cage activity (counts/30 min), and expenditure (kcal/30 min).

^*Statistically significant (P < 0.05).

Maternal exercise promotes fat loss in adult female offspring that perform voluntary exercise

It may seem inconsistent that we observed increased physical activity (Fig. 4B) and a trend toward decreased food intake (Fig. 4A) in U female offspring but no differences in body weight or body composition (Fig. 3B, C). As one potential explanation, we considered whether resting metabolic rate might be lower in U offspring. Mice were not denied access to food during the CLAMS studies, so it was not possible to measure resting metabolic rate. However, estimates of minimal energy expenditure showed no group differences at any age (data not shown). Because only a subset of offspring had access to running wheels and only during the study periods (P21, P60, and P160), we hypothesized that, when provided continuous access to a running wheel, adult U females would run more and achieve greater fat loss than L females. To test this theory, we studied P300 females and measured changes in body composition during a 3-wk period of voluntary exercise training in their home cages (running wheel exposure). (Due to a limited number of running wheels, we studied only females in this P300 test.)

There were no differences in body weight or composition (% fat) between U and L females at baseline, and both groups achieved dramatic loss of body fat during the 3 wk exercise program (Fig. 5). U females, however, achieved a greater reduction in adiposity than L females (P = 0.03). Because of technical difficulties, we did not obtain data on running distance during this test. However, based on our observations at P60 and P160 (Fig. 4), it is likely that U females lost more fat because they ran more. These data indicate that the effects of maternal physical activity before and during pregnancy persist to influence energy balance in advanced adulthood.

Figure 5.

At P300, U females showed a greater reduction in adiposity in response to a voluntary exercise program. Before exercise, U and L females had comparable body composition (% fat). Over a 3 wk period of housing in a cage with a running wheel, the U females achieved a greater reduction in percentage of fat than did the L females (P = 0.03).

DISCUSSION

We found that maternal exercise directly before and during pregnancy promotes physical activity in the offspring. In female offspring this effect was first evident around sexual maturity, and persisted into later adulthood. Increased physical activity was also observed in male offspring, but not until later in adulthood.

Previous studies have shown that nutrition can affect developmental programming of physical activity (12, 16, 17, 25, 26), but we are aware of only one (27) that examined effects of maternal exercise on offspring physical activity. Consistent with our findings, that study reported a significant positive correlation between maternal running distance during pregnancy and offspring running distance in adulthood. Because outbred ICR mice were used, however, those authors concluded that this correlation was likely due to genetic inheritance (27). Our findings in an isogenic population of mice therefore provide the first clear indication that maternal voluntary exercise before and during pregnancy affects the establishment of lifelong propensity for voluntary exercise in her offspring. Since maternal running during lactation was nearly negligible (Fig. 2B), this effect almost certainly occurred either directly before or during pregnancy.

Rather than study a large number of mice, we employed several design refinements to achieve high power. The first was to minimize the influence of genetic variation by studying inbred mice. The second was to include a preliminary test to assess voluntary exercise in the dams. Even among inbred dams, it is possible that subtle de novo genetic variation could influence both their voluntary exercise level and that of their offspring. The preliminary test eliminated such potential confounding, while also ensuring that all U offspring were exposed to a consistent level of maternal exercise. Third, recognizing the nonindependence of offspring born to the same dam, our metabolic cage studies included only 1 male and 1 female offspring of each dam. Last, we performed our metabolic and physical activity measurements over an extended period at each age; even after excluding data during the acclimation periods, at least 8 d of data per mouse were analyzed at each age. The consistent findings of a significant increase in voluntary exercise in U adults, in both males and females (Fig. 4B), not only demonstrates a robust effect but also indicates that our study was appropriately powered.

We believe our results may help explain previous observations that maternal exercise during pregnancy improves offspring glucose tolerance and insulin sensitivity. Using designs similar to ours, 4 previous studies reported that maternal voluntary exercise before and during pregnancy improves glucose tolerance in adult offspring (18–20, 28). Three of these (18–20) also showed that the improved glucose tolerance was associated with enhanced insulin sensitivity. Because physical activity is well known to improve both glucose tolerance and insulin sensitivity (29, 30), some of the metabolic effects of maternal exercise during pregnancy on adult offspring (18–20, 28) may simply be a consequence of increased physical activity in the offspring. Consistent with previous studies (18–20, 28), we did not detect an overall difference in body weight or adiposity in U offspring in normal housing conditions. We found such an effect only by providing adult offspring sustained access to running wheels. It is important in future studies to determine whether more dramatic effects of maternal exercise are unmasked by providing the offspring with permanent access to running wheels, perhaps in combination with a high-fat diet challenge.

In addition to genetic factors (10), environmental influences during critical ontogenic periods, including fetal (12, 17), early postnatal (16), and even postweaning development (31) can influence the entrainment of the activity-stat. Our current findings confirm the important role of early environment in the establishment of the activity-stat and provide the novel insight that physical movement during fetal development may affect the development of central mechanisms regulating physical activity behavior. Indeed, gravity is necessary for appropriate development of the vestibular system in the mammalian CNS (32), underscoring the potential for acceleration (movement) to broadly affect fetal brain development. Our current understanding of central regulation of volitional physical activity remains rudimentary (8), so further extensive study is needed to elucidate the specific mechanisms underlying the programming effect reported here. Recent reports indicate that not just maternal but also paternal physical activity may have long-term consequences for offspring metabolism and body weight (33, 34) and brain function (35).

We found that offspring born to exercised dams had higher body weight at P1 (an approximation of birth weight), indicating that maternal exercise had a positive effect on fetal growth. Clinical studies have reported conflicting data on the effects of maternal exercise during pregnancy on birth weight. In both a small observational study (36) and a randomized controlled trial of home-based stationary cycling (37), maternal exercise during pregnancy predicted decreased birth weight. Other clinical studies, however, have reported contrary results. For example, a retrospective case–control study concluded that light physical activity during the second trimester of gestation protects against low birth weight (38). Overall, a recent meta-analysis (5322 pregnancies) found that maternal exercise during pregnancy does not change the risk of delivering a small-for-gestational-age infant, but reduces the risk of large-for-gestational-age birth (39). Hence, our findings are not inconsistent with the human data and suggest that maternal exercise, especially initiated during or before early pregnancy, may result in improved placental development or function.

Clinical studies have also attempted to assess long-term effects of maternal exercise on the offspring. The previously mentioned observational study (36) found lower body weight and adiposity at age 5 in the offspring of mothers who exercised before and during pregnancy. Two additional human observational studies examined associations between maternal exercise and the activity level of the offspring. In the first (40), the physical activity levels of more than 5000 preteen children were analyzed in comparison with their parents’ activity levels. Parental physical activity levels during pregnancy and in the children’s early life correlated positively with those in their children. In the second, the activity levels of 554 preschool children correlated positively with those in their mothers (41). In all of these observational studies, however, these positive correlations could merely reflect effects of parenting or genetic differences, rather than developmental programming. Prospective randomized controlled studies of the offspring of women assigned to exercise interventions are needed to determine whether maternal physical activity during pregnancy has a positive effect on spontaneous physical activity in humans.

In summary, among the many studies on developmental programming of physical activity, ours is the first to show that maternal exercise before and during pregnancy influences the propensity for physical activity in the offspring. If similar effects can be confirmed in humans, this would suggest a straightforward and effective intervention to combat the current worldwide epidemic of physical inactivity (42) and obesity.

Glossary

CLAMS

comprehensive laboratory animal monitoring system

L

locked

P

postnatal day

SL

small litter

SPA

spontaneous physical activity

U

unlocked