University of Groningen the Rate of Living in Mice Vaanholt, Lobke Maria

University of Groningen

The rate of living in mice Vaanholt, Lobke Maria

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2007

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Vaanholt, L. M. (2007). The rate of living in mice: Impacts of activity and temperature on energy metabolism and longevity. s.n.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 06-10-2021 The Rate of Living in Mice: Impacts of activity and temperature on energy metabolism and longevity The research reported in this thesis was carried out at the Behavioural Biology Group at the University of Groningen, The Netherlands and at the Rowett Institute in Aberdeen, Scotland. All studies were approved by the Ethical Committee of the University of Groningen (DEC 2777(-1), DEC 3039(-1), DEC 3128 and DEC 4184A). Production of this thesis was partly funded by the University of Groningen and the research school of Behavioural and Cognitive Neurosciences (BCN). Additional financial support came from UNO Roestvaststaal BV in Zevenaar and Harlan Netherlands BV in Horst.

Cover: Lobke Vaanholt and Dick Visser Lay-out and figures: Dick Visser Photos: Lobke Vaanholt and Kristin Schubert Printed by: Ponsen en Looijen b.v., Wageningen

ISBN: 9789036729444 ISBN: 9789036729451 (electronic version) RIJKSUNIVERSITEIT GRONINGEN

The Rate of Living in Mice: Impacts of activity and temperature on energy metabolism and longevity

PROEFSCHRIFT

ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op vrijdag 23 maart 2007 om 16.15 uur

door

Lobke Maria Vaanholt

geboren op 20 februari 1979 te Enschede Promotores: Prof. dr. S. Daan Prof. dr. G.H. Visser

Beoordelingscommissie: Prof. dr. A.J.W. Scheurink Prof. dr. J.R. Speakman Prof. dr. R.J.G. Westendorp Contents

Chapter 1 General introduction 7

Part I – ACTIVITY & METABOLISM

Chapter 2 Wheel-running activity and energy metabolism in relation to ambient 19 temperature in mice selected for high wheel-running activity

Chapter 3 Behavioural and physiological responses to increased foraging effort 35 in male mice

Chapter 4 Plasma adiponectin is increased in mice selectively bred for 57 high wheel-running activity, but not by wheel running per se

Chapter 5 Responses in energy balance to high-fat feeding in mice selectively-bred 71 for high wheel-running activity

Part II – METABOLISM & AGEING

Chapter 6 Life span, body composition, and metabolism in mice selected for 91 high wheel-running activity and their random-bred controls

Chapter 7 Protein synthesis and antioxidant capacity in ageing mice: 111 effects of long-term voluntary exercise

Chapter 8 Ageing under cold conditions: effects on body composition, 127 metabolism and longevity

Chapter 9 Protein synthesis and antioxidant capacity in ageing mice: 147 effects of life-long cold exposure

Chapter 10 General Perspective 161

References 177

List of abbreviations 193

Nederlandse samenvatting – Dutch summary 195

Addresses of co-authors 203

Dankwoord – Acknowledgements 205

Chapter1

General introduction 8 Chapter 1

“Death is the only certainty you have in life”

In evolutionary biology, ageing is usually defined as a persistent decline in the age- specific fitness components of an organism due to internal physiological deteriora- tion. This definition integrates effects on reproduction and survival. Gerontologists simply define ageing as an increase in the likelihood that an individual will die in a certain time interval. As we age, intracellular processes degenerate and ultimately fail. This can lead to age-related diseases, such as cardiovascular disease, Parkin- son’s disease etc., and ultimately to death. There has been much speculation on the role of energy metabolism in the causation of these processes. This has led to the formation of several intriguing theories which attribute the causation of death ultimately to the very motor of life itself; the rate of living theory (Pearl, 1928; Rubner, 1908) and free radical theory of ageing (Harman, 1956). This idea was summarized by Murray (1926) in the statement:

‘If aliveness is measured by the velocity of chemical activity (heat production) an organism may in this sense be said to dig its own grave. The more abundant its manifestations of life, the greater will be its rate of senescence’.

The primary aim of this thesis was to investigate the relationship between ageing and metabolic rate. In two large-scale experiments I manipulated the energy expenditure of a group of animals by either increasing their physical activity or exposing them to cold. Survival curves were created for different experimental groups and I looked at changes that occurred in several physiological parameters that might be involved in ageing to explain differences that occurred in life span (Part II: Metabolism & Ageing). In addition, I explored the behavioural and physiological consequences of changes in energy balance in mice that had been selectively bred for high levels of spontaneous physical activity (Part I: Activity & Metabolism).

ACTIVITY & METABOLISM

Life history theory The evolution of life histories has been explained by the presence of limited resources that results in trade-offs between survival (maintenance of the body) and reproduction (Stearns, 2000). In times of plenty, resources can be allocated to growth and reproduction, but when resources are scarce, energy has to be allocated to enable survival of the individual and future success. In many species the reproductive season is tuned to coincide with the peak in food availability. When food is scarce, reproduction ceases and energy is allocated to increase the chance of survival into the next year. There is a large variation in the way animals deal with such trade-offs. When there is a genetic basis for these decisions, natural selection favours life-history traits that result in a higher fitness. The main environmental General introduction 9

factors influencing the available resources for endothermic animals are environmental temperature and food availability. In part I of this thesis we investigated the effects of low ambient temperatures or food availability on metabolism and the amount of voluntary activity mice were willing to engage. We used mice that had been selected by T. Garland Jr. for high wheel-running activity and their random-bred controls. Detailed description of the selection protocol and the main characteristics of these mice is provided in box 1.1. The amount of wheel-running activity was the selection criteria. After 31 generations of selection the mice ran approximately 2,7 times as much as control animals (Rhodes et al., 2005). With the selection for wheel-running activity other traits have been co-selected (i.e. small body size) and much research has been undertaken to uncover these co-selected traits. In chapter 2 we investigate mice exposed to various ambient temperatures. We measured wheel-running activity and metabolic rate simultaneously to determine whether high-activity mice have evolved to have a lower running economy and whether they would more likely use heat generated by activity to substitute for thermoregulatory heat at low ambient temperatures than control mice do. In chapter 3 we manipulated food availability using a system in which animals had to run in a running wheel for a set number of revolutions to obtain a food pellets. This approach was used to study effects of food availability on physiological and behavioural responses in control and selected mice. Previous studies in rats by T. Adage showed that rats with low spontaneous levels of wheel-running activity have more difficulties to cope with a workload schedule than rats with high spontaneous levels of wheel-running activity. Similar effects were expected between control and high- activity mice.

Exercise & obesity Obesity is becoming an increasingly prevalent health problem in affluent societies. It is often associated with metabolic derangements such as impaired glucose tolerance, insulin resistance, high blood pressure, dyslipidemia, and abdominal obesity. When these metabolic abnormalities are displayed in concert (often referred to as the “metabolic syndrome”), they entail a high risk of developing into life-threatening conditions such as cardiovascular disease and diabetes mellitus type 2 (for review see (Carroll and Dudfield, 2004; Moller and Kaufman, 2005)). Increased dietary fat intake in combination with a sedentary existence are factors precipitating the development of obesity and the associated metabolic syndrome. Adipose tissue produces several hormones, such as leptin and adiponectin, that are important for energy homeostasis. Levels of these hormones are associated with metabolic risk factors. Adiponectin levels are decreased and leptin levels increased in obese compared with lean subjects (Park et al., 2004). Mice that have been selected for high wheel-running activity (for detailed description see box 1.1.) have decreased levels of leptin even when correcting for fat mass (Girard et al., 2006). Leptin is produced by adipose tissue and informs the 10 Chapter 1

body about its available fat stores and is involved in regulating food intake. Selected mice have a high food intake to cover the increased costs of wheel-running activity (Swallow et al., 2001), and lowering levels of leptin may be an adaptation to increase food intake and maintain energy balance. High-activity mice have a lean phenotype (Dumke et al., 2001; Swallow et al., 1999) and adiponectin levels are thus expected to be increased in high-activity mice. This together with low levels of leptin might make high-activity mice less prone to develop metabolic derangements on a high-fat diet and would make these mice a suitable model to study the metabolic syndrome. In chapter 4 we measured hormone levels of leptin, adiponectin and corticosterone in aging male mice selected for high wheel-running activity and their random-bred controls. We studied correlations between the hormones and body composition. In chapter 5 we describe a study in which we exposed selected males and females to a high fat diet. Body composition, food efficiency, energy metabolism and glucose tolerance were tested to determine whether high-activity mice responded differently to a high-fat diet than controls.

METABOLISM & AGEING

“Rate of living” and “free radical” theory of ageing Instinctively we know that things (cars, machines) break down faster if you use them more often and more intensively. The same might be applicable to animal (and human) life. This notion that the rate of energy turnover determines the rate of breakdown is known as the “rate of living” theory (Pearl, 1928). In 1908, Rubner noted that food intake per gram decreased with increasing life span among five domestic animals (guinea pig, cat, dog, cow and horse). He calculated the energy intake per gram per life span (life-time energy potential, LEP) and found that the variation in LEP between species was small (1,5 fold), although the variation in body mass was very large. Including data for men the variation in life-time energy expenditure was slightly larger, but still only 5-fold. He concluded that mass-specific energy metabolism times the maximal lifespan was a constant (Rubner, 1908). Energy metabolism might thus be the factor that determines our life span. In 1928 Pearl postulated the “rate of living theory” that states that there is an inverse relationship between energy expenditure and life span (Pearl, 1928). An extensive body of evidence exists that is consistent with this theory. Comparative studies have shown that energy expenditure tends to show an inverse relationship with body size and longevity when compared across mammalian or bird species (Ku et al., 1993; Speakman, 2005a). Also evidence from intra-specific studies show evidence for the rate of living theory. Increasing ambient temperature (and thereby energy expenditure) decreased life span of nematodes proportionally (Van Voorhies and Ward, 1999). Honey bees that were forced to fly with extra loads had decreased life spans (Wolf and Schmid-Hempel, 1989), and flies prohibited to fly (thereby General introduction 11

decreasing metabolic rate) had increased life spans (Yan and Sohal, 2000). Brood size increases in kestrels resulted in increased energy turnover and a subsequent decrease in the survival of parents that had enlarged broods (Daan et al., 1996). In hibernating hamsters survival was higher than in hamsters that did not hibernate (Lyman et al., 1981). A moderate increase of the level of basal metabolism of young adult rats adapted to hypergravity compared to controls in normal gravity was accompanied by a roughly similar increase in the rate of organ aging and reduction of survival (Economos et al., 1982). In contrast there are also numerous studies that showed no relationship or a positive relationship between energy expenditure and life span (in mammals (Holloszy and Smith, 1987; Holloszy and Smith, 1986; Navarro et al., 2004; Speakman and Selman, 2003; Speakman et al., 2004)), and comparative studies show that for a certain body mass birds expend up to 4 times more energy than a mammal, and live longer (Speakman, 2005b). Another line of evidence comes from experiments on calorically restricted animals. Caloric restriction (CR; decreasing energy intake) is widely recognized as the only (non-genetic) manipulation that increases mean and maximum life span in mammals (first shown by (McCay et al., 1935)). In 1977 Sacher proposed that CR extended life span by decreasing metabolic rate. A study by Masoro et al. found that following the initiation of CR there was a brief period of reduced food intake per gram body mass, but this was followed by a lifetime where the intake per gram body mass was higher in CR rats than ad-libitum fed rats (Masoro et al., 1982). In a study where mass-specific 24-h metabolic rates were measured mass-specific (based on lean mass) metabolic rates were reduced upon the initiation of CR, but increased to levels higher than ad-libitum fed animals later on (McCarter and Palmer, 1992). Similar results were shown in rhesus monkeys (Ramsey et al., 1996). These studies disagree with a role for metabolic rate in the life extending effect of CR. Interpretation of the results is confounded because metabolic rate is usually normalized for body mass or lean mass, whereas the relative sizes of organs are not the same for animals that are CR or fed ad libitum (Greenberg, 1999b; Greenberg and Boozer, 2000). A related theory of ageing was suggested by Harman in 1956 known as the “free radical theory” (Harman, 1956). This theory specifies the reason why there should be a direct link between energy metabolism and the rate of ageing. Free radicals or radical oxygen species (ROS) are produced as by-products of normal oxidative phosphorylation, and can cause damage to macromolecules which may result in malfunction and eventually cell death (for review see (Beckman and Ames, 1998)). The body has evolved defense systems against these radicals in the form of antioxidant enzymes (e.g. superoxide dismutase, catalase and gluthione peroxidase) that scavenge ROS and transform them into less toxic products. A small amount of ROS escape conversion. If damage to macromolecules has occurred the processes of DNA repair and protein synthesis can repair most of this damage. Despite these defense systems a small amount of damage still occurs and this accumulates with age resulting in malfunction of cells and eventually death (see Figure 1.1. for a graphical representation of the process). When energy expenditure (and oxidative 12 Chapter 1

ambient METABOLISM activity temperature

uncoupling

ROS PRODUCTION

antioxidant enzymes

repair DNA, LIPID PROTEIN DAMAGE

AGEING

Figure 1.1. Schematic representation of the relationship between metabolism and ageing. When metabolism (=O2 consumption) increases reactive oxygen species (ROS) are produced that can cause damage to DNA, lipids and proteins which may result in ageing. Several defence mechanisms can slow down this process (in dark grey). First of all uncoupling of the electron transport chain from ATP production decreases the amount of ROS produced at a certain metabolic rate. Second, once ROS are produced they can be scavenged before they cause damage by antioxidant enzymes, such as superoxide dismutase, catalase and glutathione peroxidase. Third, oxidative damage to macromolecules can be repaired, i.e., DNA repair mechanisms, protein turnover. One can study these processes by experimentally increasing metabolic rate by decreasing ambient temperature or increasing activity.

phosphorylation) increases, the production of ROS will also increase. This would explain the relationship between ageing and metabolism proposed by the rate of living theory. The relationship between oxidative phosphorylation and ROS production is not linear. Oxidative phosphorylation takes place on the inner membrane of mitochondria as a result of the transport of electrons over the membrane (electron transport chain; ETC). The ETC consists of 4 complexes. NADH and FADH2 that have been formed in the tricarboxylic acid cycle (TCA) donate their electrons to subsequently complex I or II which are then passed on to ubiquinone (Q). Q moves across the membrane to complex III and the electrons are passed on to cytochrome C that moves on to complex IV where the electrons are accepted by molecular oxygen and combined with protons to form water (for a more detailed description see (Brand, 2000)). During this process protons are pumped across the membrane into the inner membrane space and a proton motive force builds up. When oxidative phosphorylation is coupled these protons are pumped back to the matrix via an ATP-ase pump resulting in the phosphorylation of ADP to ATP (ATP synthesis). General introduction 13

Free radicals are generated during oxidative phosphorylation when an oxygen molecule promiscuously reacts with one of the transported electrons before it reaches complex IV. This can for instance occur when the supply of ADP is limited thereby blocking up the system. Agents that increase respiration rate and thereby lower proton motive force (i.e., ATP synthesis) thus lower the rate of ROS production. ROS production is thus not linearly related to the rate of electron transport. The flow of electrons in the ETC is usually tightly coupled to the production of ATP, and it does not occur unless the phosphorylation of ADP can proceed. This prevents a waste of energy, because high-energy electrons do not flow unless ATP can be produced. If electron flow is uncoupled from the phosphorylation of ADP there would be no production of ATP, and the energy of the electrons would be wasted as heat. Uncoupling agents abolish the link between oxidation and phosphoryalation, allowing electron transport to proceed without coupled ATP synthesis, thereby increasing the respiration rate and lowering ROS production (Brand, 2000). Therefore, metabolic rate and free radical production are not necessarily linearly related. Many studies support the importance of antioxidants, oxidative stress and repair of oxidative damage for the ageing process. For instance, the importance of antioxidants enzymes is clear from studies with over-expression or knocking out of these enzymes. Overexpression of catalase and superoxide dismutase in Drosophila melanogaster increased median and maximum lifespan up to 30% (Orr and Sohal, 1994; Sohal et al., 1995), and mice lacking manganese superoxide dismutase died within 10 days (Li et al., 1995), whereas administration of superoxide dismutase- catalase mimetics increased lifespan up to three times in mice (Melov et al., 2001). In CR animals life span extending effects have also been attributed to differences in oxidative stress. Increased antioxidant enzyme activity, DNA repair and protein synthesis, and decreased numbers of oxidatively damaged molecules have been shown in CR animals (for reviews see (Gredilla and Barja, 2005; Tavernarakis and Driscoll, 2002; Yu, 1996)). Whereas the “free radical” theory has gained much support in recent years, the rate of living theory has been discarded as invalid by many researchers based on inter-specific comparisons and the lack of effects on (or increases in) energy metabolism in CR animals. This is remarkable since the free radical theory of ageing is itself the main theory postulating the mechanism connecting energy turnover and ageing. As argued by Speakman (2002; 2005a) the reasons to dispute the theory may not always be valid, because the arguments that are used to test the theory are fraught with problems. Firstly, maximum life span is not a good measure of ageing. Maximal life span is determined by a single point in every data base and is highly affected by the sample size used and also by the conditions in which animals are housed (i.e. laboratory or natural conditions). Secondly, basal metabolic rates have been used in most studies to estimate life-time energy potential. Basal metabolic rate is the metabolism of an animal when fasting and resting at thermo-neutral temperatures and contributes only 40% to the total daily energy expenditure. The 14 Chapter 1

latter is a better measure of metabolism. Using a single measure of metabolic rate in the life time of an animal might not be sufficient to make an accurate estimate of life-time energy potential. Thirdly, testing for consistency in life-time energy expenditure per gram of tissue by inter-specific comparisons between birds and mammals is not the best way to test the rate of living theory and inter-specific comparisons are complicated by the fact that animals from different species may reflect adaptive or genetic differences in free-radical production or differences in defence and repair mechanisms. Therefore, intra-specific comparisons are more convincing when looking at associations between energy expenditure and ageing. A fourth argument relates to the scaling of energy expenditure to body mass. Greenberg has shown that in cases where no relation was found between life-time energy expenditure per gram body mass and life span, a relationship does exist when one calculates the energy expenditure for certain metabolically active organs and relate this to life span (Greenberg, 1999). Life-time energy expenditure per gram dry lean body mass instead of total body mass might be a better measure to test the rate of living theory since this contains the tissue that is metabolically most active. A stronger correlation between energy metabolism and dry lean mass is usually found then between body mass and energy expenditure. In studies on energy expenditure and life span almost never the body composition and energy turnover are followed throughout life. In order to resolve some of the confusion in this area we carried out two large scale experiments. We manipulated energy metabolism by either increasing activity through selection (chapter 6) or by decreasing environmental temperature (chapter 8) and looked at the relationship between energy metabolism and survival in intra-specific comparison. Mice selectively bred for high wheel-running activity were used to investigate the effects of increased voluntarily exercise. In the cold experiment, c57Bl/6J mice were used that were subjected to 10°C compared to 22°C in control mice. An additional group that was exposed to cold early in life was added. This for the first time tests one basic implicit proposition in the rate of living and free radical theories: that the effects of energy turnover are cumulative. Energy turnover increase in youth should still have and effect in old age. In both experiments we paid specific attention to the effects of age and experimental manipulation (i.e. cold or activity) on two systems that are involved in defending the body against ROS, the antioxidant defence system and protein turnover (chapter 7 and chapter 9). General introduction 15

BOX 1.1: Mice selected for high wheel-running activity: selection procedure and main characteristics

Selection procedures were carried out by Theodore Garland Jr. et al. at the University of Wisconsin, Madison, USA and first described in 1998 by Swallow et al. (1998).

Selection procedure: Progenitors were 112 male and 112 female Hsd:ICR house mice (Mus domesticus) obtained from Harlan Sprague Dawley (USA) which are genetically heterogeneous (generation -2). These mice were randomly paired and their offspring (generation -1) was then randomly assigned to 1 of 8 lines so that every line contained 10 pairs of mice. These pairs were again randomly paired within each line. In the next generation (generation 0) selection for high wheel-running activity started. In 4 of the lines (control lines: 1, 2, 4 and 5) mice were paired randomly and at least 1 male and 1 female from each family were chosen as breeders (control lines). In the other lines (selected: 3, 6, 7 and 8) the highest running female and male from each family were chosen as breeders. Three extra female and male (second highest runners from highest running families; never 2 from the same family) were chosen to ensure 10 families per generation. Animals were paired at approximately 10 weeks of age and males were removed 15-18 days after pairing (gestation length: 19 days). All offspring of selection families were kept and 2 males and 2 females from each control family. Selection took place at 6-8 weeks of age, when mice were monitored for voluntary wheel running for 6 consecutive days (wheel circumference= 1.12 m). The average number of wheel revolutions on day 5 and 6 were used as selection trait. In April of 2002 80 breeding pairs (10 per line) of generation 31 of selection were shipped to the Zoological laboratory in Haren and a new breeding colony was started at these facilities without further selection preventing sibling mating. Offspring these mice were used in the experiments described in this thesis.

Main characteristics: The selection method has resulted in mice that run longer distances (more revolutions) per day, but not more time per day. High-activity mice thus run at higher velocities (Swallow et al., 1998). In addition, mice run more intermittently and in shorter bouts (Girard et al. 2001), probably to lower costs of running (Rezende et al., 2006). At generation 31 (the generation of the mice obtained by us) high-activity mice ran approximately 2,7 times more than control mice (Rhodes et al., 2005). On average female mice run more than males, but the selection procedure affected wheel-running activity in a similar way in both males and females (Morgan et al., 2003). High-activity mice are also hyperactive in their cages when deprived from wheels (Rhodes et al., 2001). Body mass at maturity is decreased in high-activity mice and food intake is increased compared to control mice (Dumke et al., 2001; Swallow et al., 2001). The reduction in body mass is mainly caused by a reduction in fat mass (Dumke et al., 2001; Swallow et al., 1999). Reproductive output is similar in control and high-activity mice (Girard et al., 2002). Control and high-activity mice do not differ with respect to their thermoregulatory mechanism (Rhodes et al., 2000). No differences in open field activity or defeca- tion have been found (Bronikowski et al., 2001). Leptin levels are decreased (even when correcting for fat content) in females (Girard et al., 2006), and corticosterone levels are elevated in high-activity mice relative to controls (specifically in females) (Girard and Garland, Jr., 2002; Malisch et al., 2006).

Part I

ACTIVITY & METABOLISM 18 Chapter2

Wheel-running activity and energy metabolism in relation to ambient temperature in mice selected for high wheel-running activity

Lobke M. Vaanholt, Theodore Garland Jr., Serge Daan, G. Henk Visser

Abstract Interrelationships between ambient temperature, activity, and energy metabolism were explored in mice that had been selectively bred for high spontaneous wheel-running activity and their random-bred controls. Animals were exposed to three different ambient temperatures (10, 20 and 30°C) and wheel-running activity and metabolic rate were measured simultaneously. Wheel-running activity was decreased at low ambient temperatures in all animals and was increased in selected animals compared to controls at 20 and 30°C. Resting metabolic rate (RMR) and daily energy expenditure (DEE) decreased with increasing ambient temperature. RMR did not differ between control and selected mice, but mass-specific DEE was increased in selected mice. The cost of activity (measured as the slope of the relationship between metabolic rate and running speed) was similar at all ambient temperatures and in control and selected mice. Heat generated by running apparently did not substitute for heat necessary for thermoregulation. The overall estimate of running costs was 1.2 kJ km-1 for control mice and selected mice.

Journal of Comparative Physiology B, 177(1); 109-118 20 Chapter 2

INTRODUCTION

Homeothermic animals maintain a rather constant body temperature over a wide ambient temperature range. At low ambient temperature resting homeotherms elevate metabolic levels to compensate for elevated heat loss, while at high ambient temperatures metabolic rates should be low to avoid hyperthermia (Mount, 1966; Tieleman et al., 2002). This temperature dependence of metabolic rate becomes more complicated when animals exhibit high locomotor activity, which is known to be energetically expensive (Taylor et al., 1970). In the cold, high levels of activity may be favourable if activity-related metabolic costs can be used for temperature regulation. The excess heat produced by activity might theoretically substitute for shivering thermogenesis during rest. In principle, if substitution takes place, then the cost of locomotion, formally measured as the energy turnover during activity minus the energy turnover during inactivity, will be reduced at low temperature. If no substitution takes place, then the costs for activity will be added to those for thermoregulation (addition). The empirical literature is ambiguous on this issue. Several studies demonstrate substitution (in White crowned sparrows, Zonotrichia leucophrys gambelii (Paladino and King, 1984), potoroo, Potorous tridactylus (Baudinette et al., 1993) deer mice, Peromyscus maniculatus (Chappell et al., 2004), rat, Rattus novegicus (Arnold et al., 1986; Makinen et al., 1996)), but there are also results consistent with addition (Kowari, Dasyroides byrnei (MacMillen and Dawson, 1986), Chipmunk, Eutamias merriami (Wunder, 1970) patas monkey, Erythrocebus patas (Mahoney, 1980)). This discrepancy among studies may well be related to different conditions. Activity may, for instance, simultaneously lead to reduced insulation and increased heat loss in situations where animals huddle or use bedding material while resting and could therefore mask substitutive effects of activity. If substitution occurs, this would lead to low net costs of activity at low temperatures and thereby should lead us to expect increased activity in the cold. We decided to exploit mice specifically selected for high activity to test the hypothesis of substitutive metabolic rate in this species. Swallow et al. (1998) have selected mice for high spontaneous wheel-running activity during many generations (for selection procedure see (Swallow et al., 1998)), which make these animals prof- itable to further explore interrelationships between ambient temperature, activity, and energy metabolism. Animals were bred under ambient temperatures of approximately 22°C. The intensity of spontaneous wheel-running activity has increased over generations and reached an apparent plateau around generation 16 (Bronikowski et al., 2001). In addition, selected animals have become smaller and leaner (Girard et al., 2006; Swallow et al., 2001), thereby diminishing whole-animal costs of running in these mice (Rezende et al., 2006). Smaller animals also have larger surface-to-volume ratios, which could make them more susceptible to heat loss at low ambient temperatures. During the selection process for high spontaneous wheel-running activity, animals seemed to exhibit annual cycles regarding Ambient temperature and wheel-running activity 21

their spontaneous wheel-running activity (Bronikowski et al., 2001) which might be attributed to variations in ambient temperature. In order to evaluate the effects of genetically increased activity in the selected mice, we studied animals from control and selected lines at various ambient temperatures and recorded their wheel-running activity, body temperature, resting metabolic rate and daily energy expenditure.

MATERIAL AND METHODS

Animals & housing House mice (Mus domesticus) that had been selected for high wheel-running activity and their random bred controls were obtained from the lab of Prof. Dr. T. Garland Jr, Riverside, CA, USA. Originally, eight lines of mice consisting of 10 pairs each had been created, four in which mice were randomly bred and four in which mice were selected for high wheel-running activity. Selection took place at 6-8 weeks of age during a 6 day trial on wheel running (1.12 m circumference). The most actively running female and male within each family were chosen as breeders for the next generation, without allowing sibling matings. Eighty breeding pairs (10 per line) from generation 31 of selection were sent to the Zoological Laboratory in Haren (NL) to start a breeding colony without further selection. In the present study, 16 male mice (8 control and 8 selected) at the age of 6-8 weeks were used from one of the control (lab designation is line 2) and one of the selection lines (line 7). The mice were individually housed in cages equipped with running wheels (Macrolon type II cages (15x30x15cm); UNO Roestvaststaal BV, Zevenaar, NL; adapted to fit in a wheel running with a diameter of 14 cm) and wood shavings as bedding two weeks prior to the experiments. The mice were on a 12:12 light-dark cycle (lights on at 8:00 CET) and food (Standard lab chow RMB-H (2181), HopeFarms B.V., Woerden, NL ) and water were provided ad libitum.

Experimental protocol At the start of the experiment animals were randomly divided into two groups (each consisting of 4 controls and 4 selected animals) and housed in two separate temperature-controlled rooms. The mice stayed in these rooms throughout the experiment. All animals were exposed to three ambient temperatures (10, 20 and 30°C) over a time course of three weeks. Each week ambient temperature was increased or decreased by 10 degrees, starting at 10°C in room 1 and at 30°C in room 2. Wheel-running activity was recorded on a PC-based event recording system (ERS) with 2-min resolution. Body weight was measured every day at 12 pm. At noon on day 6 of each stay at a set ambient temperature animals were put with their home cage in a respirometry chamber (25x35x25cm), in the same room . -1 . as they were housed. Oxygen consumption (V O2, l h ) and carbon dioxide (V CO2, l h-1) production was then recorded for each individual for 24 h by indirect calorimetry. Our 8-channel open circuit system has been described earlier by 22 Chapter 2

Oklejewicz et al. (1997). In brief, oxygen and carbon dioxide concentration of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck) from each chamber was measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and cardon dioxide by an infrared gas analyzer (Servomex 1440). The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic chambers. Oxygen and carbon dioxide analyzers were calibrated with two gas mixtures with known amount of O2 and CO2 prior to each measurement. Flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks) and set at 30 liter per hour. Of the respiration air a subsample was passed at a rate of 6 l h-1 through the drying system and subsequently through the gas analyzers. Ambient temperature in the chamber and cage were measured simultaneously. Data were collected every 10 minutes for each animal and automatically stored on a computer. Oxygen consumption was calculated according the equation 2 of Hill (1972) to correct for volume changes with respiratory quotient below 1 and expressed in standard temperature and pressure. The respirometric chambers fitted the complete home cage of the animals. Animals therefore did not need to be handled and had access to their own running wheel throughout the measurements. Water and food were provided ad libitum. Wheel-running activity was also measured throughout the respirometry measurement using the ERS. Body temperature was measured with a rectal probe (NTC type C, Ahlborn, Holzkirchen, Germany) immediately after the respirometry measurement. Body weight was also measured at this time. After these measurements the ambient temperature in the rooms was changed.

Data analysis Continuous recordings of wheel-running activity were available for day 3-5 in each condition, just prior to the respirometry. These data were used for further analysis, excluding days 1 and 2 after the temperature transition. Average wheel-running activity per day (distance run), time spent running and average running speed were calculated for each temperature. In addition, maximum wheel-running activity per temperature was calculated over the same days by determining the maximum amount run in a 2-min interval. The same variables of wheel-running activity were determined for the 24-h interval in the respirometer. Wheel-running recordings during this time were not available for all animals and sample size for controls and selected mice were, respectively, 5 and 3 at 10°C, 6 and 6 at 20°C and 5 and 6 at 30°C. Metabolic rate (MR, kJ h-1) was calculated using the equation MR= (16.18 x . . V O2) + (5.02 x V CO2 ) (Romijn and Lokhorst, 1961). Instead of using a fixed gas exchange conversion factor this versatile equation enabled the calculation of heat production of different nutritional states (see also (Gessaman and Nagy, 1988). Resting metabolic rate (RMR, kJ h-1) was defined as the lowest (running) mean metabolic rate recorded over half an hour anywhere during the 24 h measurement. The average metabolic rate over 24 h was used to calculate daily energy expenditure Ambient temperature and wheel-running activity 23

5 mouse 242 1600

4 ) -1 ) -1 1200 3

800 2

heat production (kJ h 400 1 running wheel activity (rev h

0 0 61812 external time (h)

Figure 2.1. Simultaneous measurements of RWA (white dots) and HP (grey dots) for a mouse representative of the group (10-min. averages) from 4 hours prior to the dark phase to 4 hours after the dark phase (black bar) at 30°C. At the flow rate employed a 30 min. time lag is detectable in our respirometry system and therefore data on HP were corrected with 30 min. to determine the relationship between heat production and running speed (black dots).

(DEE, kJ d-1). The body weight measured before and after the respirometry measurement was averaged and used to calculated mass-specific RMR and DEE (in kJ g-1 d-1). Independent t-tests were used to screen for differences between animals housed in the two separate rooms. No significant differences were found and data from both rooms were pooled for further analysis. For all traits, two-way repeated measures ANOVA were performed with a factor group (ctrl vs. selection), temperature (10 vs. 20 vs. 30°C) and group x temperature using SAS 9.1 (PROC MIXED). Body mass is known to have a strong influence on metabolic rate and analysis of resting metabolic rate (RMR) and daily energy expenditure (DEE) were done using models with or without body mass as a covariate. In addition, we were interested in the relationship between parameters of wheel-running activity (distance run, time run, average running speed and maximal running speed) and DEE, and these parameters were added as an additional covariate to body mass in the model one at a time to explore these relationships. Data was normally distributed and thus not transformed before analysis. When the ANOVA showed significant effects post hoc t- tests were performed. Significance was assumed at p≤ 0.05. All tests were two- tailed. To determine the relationship between running speed (V, km h-1) and heat production (HP, kJ h-1) at the different ambient temperatures, average wheel-running activity (while in the respirometer) and HP of most mice (for sample size see 24 Chapter 2

above) were calculated in 30-min. bins during the dark phase. We only used data from the dark phase (12 h) because mice are nocturnal and wheel-running activity is mainly limited to the dark phase (24 time points per mouse). We accounted for a 30 min. lag in the HP measurements caused by the low air flow rate through the respirometry system (see Figure 2.1). At each temperature and for both groups we calculated the average running speed and heat production per 30-min bin. The relationship between running speed and heat production for all groups was plotted in Figure 2.3. Using ANCOVA models we explored effects of group and temperature on the relationship between running speed and metabolic rate.

RESULTS

Body mass, food intake and wheel-running activity Table 2.1 shows the effects of ambient temperature on body mass, food intake and several measures of wheel-running activity in control and selected mice. We found no differences between control and selected mice in average body mass or food intake. Ambient temperature had no effect on body mass, but food intake was significantly higher at low ambient temperatures (10 and 20°C).

Table 2.1. Body mass, food intake and wheel-running activity of control and selected mice at various ambient temperatures.

Group Ambient temperatures P-values for repeated measures ANOVA 10°C 20°C 30°C Group Temp

Body mass (g) Control 28.7±1.5 28.2±1.6 28.0±2.2 0.151 0.085 Selected 27.2±2.2 26.8±1.9 26.6±2.3 Food intake (g d-1) Control 8.4±2.3 8.0±2.8 4.4±0.5 0.190 0.001 Selected 9.3±2.4 9.1±1.4 5.6±0.6 Distance run (km d-1) Control 7.1±4.1 10.7±3.7 8.8±3.2 0.024 0.003 Selected 9.6±3.0 14.1±5.1 13.1±3.4 Time spent running (h d-1) Control 6.1±2.4 7.7±2.1 7.8±1.8 0.048 0.001 Selected 7.5±1.1 9.1±1.0 9.5±2.3 Average speed (km h-1) Control 1.1±0.3 1.4±0.2 1.1±0.2 0.092 0.005 Selected 1.3±0.2 1.6±0.5 1.4±0.3 Maximum speed (km h-1) Control 2.3±0.4 2.5±0.4 2.3±0.3 0.048 0.092 Selected 2.5±0.4 2.7±0.6 2.9±0.2

Mean ± sd are given for several variables, for control and selected mice at 3 different ambient temperatures seper- ately. Two-way repeated measures ANOVA were performed on all variables with group as a between subjects factor and temperature (temp) and group x temp (GxT) as within subjects factors. P-values for effects of age and group are given in the table and are bold when the effect was statistically significant (p<0.05). No significant interaction effects between group and age were found (p>0.1), and p-values are therefore not shown in the table. Sample size was 8 in both groups, except for the measures of wheel-running activity where data of one mouse in the selected group were missing. Ambient temperature and wheel-running activity 25

As expected, selected mice had significantly higher wheel-running activity (expressed as time spent running or distance run) than control mice (see table 2.1). Ambient temperature significantly affected the distance run per day, running time per day and running speed. In both groups, wheel-running activity was significantly decreased at 10°C compared with 20°C. Maximum running speed was also significantly higher in selected mice than in control mice, but ambient temperature did not influence maximal running speed. Body mass was never a significant covariate in the models, indicating that body mass had no statistically detectable effects on food intake or any measure of wheel-running activity. This is most likely due to small variance in the body mass of the mice used for the experiments.

Metabolism and body temperature Animals were put in respirometry chambers for 24 h at different ambient temperatures to measure resting metabolic rate and daily energy expenditure in control and selected mice at these temperatures (see Figure 2.2). RMR was similar in control and selected mice and significantly decreased with increasing ambient temperature (see Table 2.2 for statistical analyses). Body mass was a significant predictor of RMR in the model , but did not influence the effects of group and temperature on RMR . DEE also did not significantly differ between control and selected mice and decreased with increasing ambient temperature. When body mass was included in the model as a covariate it significantly contributed to the explained variance in DEE and the group effect became significant, with a higher DEE in selected mice compared to controls. Post-hoc comparison showed that DEE was significantly different between lines at 30 and 20°C, but not 10°C. We were interested in how wheel-running activity as measured during the respirometry measurement contributes to the explained variance in DEE, and included activity variables into the

4 CTRL RMR

) CTRL DEE

-1 SEL RMR d SEL DEE -1 3 * * 2

1 heat production (kJ g 0 10 20 30 ambient temperature (°C)

Figure 2.2. Mass-specific resting metabolic rate (RMR) and daily energy expenditure (DEE) in mice selected for high wheel-running activity (SEL) and their random bred controls (CTRL) at various ambient temperatures. Values represent mean±SD. Asterisks (*) show at which temperatures DEE significantly differed between control and selected mice (p<0.05). 26 Chapter 2

Table 2.2. Results for repeated measures ANOVA on metabolic measurements and body temperature.

Variable n Group Temperature Covariate d.f. F p d.f. F p p

RMR 16 1,14 0.6 0.441 2,28 408.3 <0.001 none RMR 16 1,14 0.1 0.851 2,27 457.3 <0.001 Body mass 0.009

DEE 16 1,14 3.0 0.103 2,28 208.7 <0.001 none DEE 16 1,14 7.9 0.014 2,27 203.7 <0.001 Body mass 0.017 DEE 15 1,13 0.1 0.935 2,10 167.8 <0.001 Body mass 0.179 Distance 0.005 DEE 15 1,13 0.1 0.764 2,10 115.7 <0.001 Body mass 0.103 Time 0.042 DEE 15 1,13 0.2 0.679 2,10 107.5 <0.001 Body mass 0.556 Speed 0.059 DEE 15 1,13 0.1 0.971 2,10 117.2 <0.001 Body mass 0.968 Max speed 0.074

Body temp.16 1,14 0.9 0.356 2,28 11.9 <0.001 none

Repeated measures ANOVA were performed on all variables with group as a between subjects factor and temperature and group x temperature (GxT) as within subjects factors. In addition, where appropriate body mass and wheel-running activity variables were added into the model as covariates. Degrees of freedom (d.f.), F and p-values for each factor are given in the table. P-values are bold when the effect was statistically significant (p≤ 0.05). No significant interaction effects between group and age were shown and p-values are therefore not shown in the table. Sample size was 8 in both groups, except for the measures of wheel-running activity where data of one mouse in the selected group were missing.

model with body mass, one at a time. All variables were positively related to DEE. Only distance run and time spent running significantly contributed to the variance in DEE in these models They fully accounted for the group effect, but not for temperature. The effect of ambient temperature remained significant in these models. Body temperature at the different ambient temperatures was measured at the moment when animals came out of the respirometry chambers. In control mice body temperature was on average 36.7±0.5, 37.3±0.8 and 37.3±0.4 (mean±sd) at 10, 20 and 30°C, respectively, and in selected mice it was 36.7±0.4, 37.8±0.5 and 37.3±0.4. Body temperature did not differ significantly between control and selected mice and decreased with ambient temperature in both groups (see Table 2.2).

Cost of transport Estimates of the incremental cost of transport (COT, kJ km-1) are generally derived from the slope of the regression of heat production and running speed. The relationship between HP and running speed (V) for control and selected mice in the present study is shown in Figure 2.3 (see also Table 2.3). This figure plots the Ambient temperature and wheel-running activity 27

10°C 5

20°C

) 4 -1 30°C

heat production (kJ h 1

0 0.0 0.5 1.0 1.5 running speed (km h-1)

Figure 2.3. Heat production (HP, kJ h-1) of control and selected mice during voluntary running as a function of running speed (V, km h-1). Each symbol represents average running speeds and metabolic rates of mice at that temperature for each half hour of the dark phase. Control mice are in white and selected mice in black. Circles represent 10°C, triangles represent 20°C and squares represent 30°C. The solid lines are the regression lines for the different ambient temperatures (see text).

Table 2.3. Effect of temperature and group on linear regressions between running speed (km h-1) and metabolic rate (kJ h-1).

Temperature Slope Intercept R2

CONTROL 10°C 1.25±0.14 3.64±0.05 0.79 20°C 1.30±0.08 2.41±0.06 0.92 30°C 1.31±0.13 1.71±0.07 0.83 SELECTED 10°C 1.16±0.18 3.61±0.10 0.68 20°C 1.08±0.11 2.72±0.10 0.82 30°C 1.06±0.11 1.99±0.11 0.82

Using linear regression, slopes and intercepts of the relationship between running speed and metabolic rate (Figure 2.3) were determined (without body mass as a covariate). Slopes and intercepts for all separate groups are shown as mean±sem. All regressions were highly significantly different from zero (p<0.001).

interindividual average metabolic rate for each 30-min bin of running speed at each of the three temperatures. The figure clearly shows that at each temperature, the metabolic rates of both lines were distributed around the same positive regression with speed. The highest speeds were more often observed in the selected line. There was a thermal gradient, with higher metabolism at lower temperature, but at each temperature the slope appeared to be similar. We tested for effects of temperature and group in an ANCOVA model with HP as the dependent variable and running speed as a covariate, where we looked at effects of group (selection vs. con- 28 Chapter 2

trol), temperature, and their interactions with running speed. Temperature strongly affected the relationship between HP and running speed (F2,51=676.5, p<0.001). This supports the visual inspection of Figure 2.3, with obviously different intercepts (HP at zero running) at the ambient temperatures measured. There was no interaction effect between temperature and running speed, supporting similar slopes of all relationships (Slope=1.19, 95% CI: 1.09-1.29). Hence the incremental costs of running were equal at all temperatures measured. Group did not significantly affect the regression (F1,44=3.61, p=0.064). The slope of the regression between running speed and HP was slightly lower in selected mice (see Table 2.3), but not significantly so. Costs of running were thus similar in both groups. Even though body mass is known to affect COT, body mass did not contribute significantly to the explained variance in HP. Again this is probably caused by small variance in mass. The relationship between HP and body mass was positive in the models used, though. The only factor that significantly influenced the relationship between HP and running speed was thus ambient temperature. The solid lines in Figure 2.3 show the regressions for the three ambient temperatures measured with both groups combined and without taken body mass taken into account. The equations for these regression lines are: at 10°C: HP= 1.16 V + 3.63, at 20°C: HP=1.20 V + 2.55 and at 30°C: HP=1.20 V + 1.82 (p<0.001 for all regressions).

DISCUSSION

We explored effects of ambient temperature on wheel-running activity, body temperature and metabolic rate in mice that had been selected for wheel-running activity for 31 generations and their random bred controls. We expected that at low ambient temperatures the heat generated by activity might (partially) substitute thermostatic metabolic rate and therefore mice might run more in the cold. At high ambient temperatures animals were expected to reduce their activity to prevent hyperthermia, as has been shown in humans (Cheuvront and Haymes, 2001) and birds (Davies, 1982; Spinu et al., 2003). Ambient temperature did indeed significantly affect wheel-running activity, but opposite to the prediction on the basis of themogenetic substitution, wheel-running activity (distance run, time spent and average running speed) was decreased by approximately 60% in control as well as selected mice at low ambient temperature (10°C). As expected, selected mice ran a longer distance (+42%), more time (+22%) and at faster speeds (+19%) than control mice did. This difference between control and selected mice was no longer significant at low ambient temperature. The mice have been selected at ambient temperatures of approximately 22°C and at 10°C thermoregulatory costs might be too high for mice to maintain high levels of activity. Indeed, mice at 10°C had body temperatures decreased by approximately 0.6°C which could reflect difficulties to maintain constant body temperature. Lowering of body temperature may also be a strategy to lower costs for Ambient temperature and wheel-running activity 29

thermoregulation while resting. Body temperature was measured once in the middle of the light phase (rest phase) and the variation assessed between ambient temperatures may just reflect ambient temperatures at rest and may not have persisted while running. One could speculate that there is a restraint on running at low ambient temperatures (due to slower muscle contraction). Maximal running speeds, however, did not significantly vary between temperatures in this study. A study in deer mice like- wise provided no evidence for effects of temperature on wheel-running activity (Chappell et al., 2004). No differences in body temperatures were observed between control and selected mice at any of the ambient temperatures measured, which is in agreement with previous measurements of body temperature in mice of the same strain at an ambient temperature of 22°C (Rhodes et al., 2000). Regulation of body temperature at rest thus appears unchanged in mice selected for high wheel-running activity and there does not appear to be a difference in thermoregulatory capacity, at least at the temperatures studied here. Nonetheless, we can not exclude differences between the lines in shivering or non-shivering thermogenesis. Moreover, mice from selected lines show elevated heat shock protein 72 expression in the triceps surae muscle (Belter et al., 2004). As expected for a small endotherm, energy expenditure decreased with increasing ambient temperature. RMR increased 1.6-fold from 20 to 10°C and 1.8-fold from 30 to 20°C. DEE was also affected by ambient temperature with a 1.4-fold increase from 20 to 10°C and a 1.5-fold increase from 30 to 20°C. These results are similar to values found in a study in deer mice (Chappell et al., 2004) housed at 3, 10 and 25°C. Wheel-running activity (distance run and running time) was positively correlated with the simultaneously measured DEE. In concurrence with an increase in wheel-running activity, mass-specific DEE was significantly increased in selected mice compared with controls. RMR did not differ between control and selected mice, even though there are differences in body composition between them (Swallow et al., 2005; Swallow et al., 2001). Apparently, the costs for thermoregulation and maintenance of the body are similar in control and selected mice. Given that RMR did not differ between the groups and the group difference in DEE disappeared when correcting for variables of wheel-running activity, the difference in DEE between groups can be fully attributed to energy spent on activity. Total energy spent on activity was thus higher in the selected mice. This does not imply that there were differences in the costs per unit distance between the groups (COT). At all ambient temperatures COT were approximately 1.2 kJ km-1 (at an average body mass of 27.6 g), which is comparable to the COT 1.19 kJ km-1 obtained by forced locomotion on a treadmill by Taylor et al. (for a 21 g house mouse) (Taylor et al., 1970). COT is related to body mass, with higher costs of transport at higher body mass. In our study as well as previous work by Chappell et al. (2004) and Rezende et al. (2006), body mass was not a statistically significant predictor of COT. The incremental cost of terrestrial locomotion in relation to body 30 Chapter 2

mass can be estimated using the allometry given by Taylor et al. in 1982: COT (kJ km-1)= 10.7* mass (kg)0.684 (Taylor et al., 1982), and predicts a slope of 0.92 kJ km-1 for a 27.6 g animal. This is lower than the value we found for these mice. The mice measured by Taylor were forced to run on a treadmill. Animals on treadmills are forced to run at specific speeds, whereas voluntary running mice choose their preferred speed. This might render a different relationship between running speed and metabolic cost. A previous study on male selected mice at 22°C estimated a -1 -1 COT of 1.29 kJ km (when using a conversion factor of 20.1 J ml O2) (Rezende et al., 2006), which is very similar to the value of 1.2 kJ km-1 we obtained. The slight difference may easily be attributed to the different wheels used (plastic wheels with a 7 cm. radius in our study compared to metal wheels with a 18 cm. radius in the study by Rezende et al. (2006). The study by Rezende et al. demonstrated that whole-body COT during voluntary wheel running was significantly lower in the selected lines, when combining analysis of males and females (Rezende et al., 2006). When body mass and/or maximal speed were added as a covariate the difference disappeared. These factors apparently caused the line difference. Similar to our study, analyzing males alone did not render a significant effect of selection on COT. The novel result in our study is that COT was unaffected by ambient temperature. With decreasing ambient temperature the intercept of the relationship between metabolic rate of running speed did increase, indicating increased costs at rest at lower temperatures, as is also reflected in an increase in RMR. Heat generated by running apparently did not substitute for thermoregulation costs at low ambient temperature in our mice (Figure 2.3). At all ambient temperatures the slope of the relationship between metabolic rate and running speed was statistically indistinguishable. Contradictory evidence exists for other species of homeotherms, showing either addition or substitution of activity-generated heat for thermoregulatory heat at low ambient temperatures. Table 2.4 summarizes the results for two studies on birds and several on various mammals. We have listed whether heat generated by activity was additive or substitutive and at which temperature substitution first occurred. The two studies on birds indicate partial or complete substitution of exercise-generated heat production for thermoregulatory costs usually at low ambient temperatures and additive at moderately cold ambient temperature (Paladino and King, 1984; Pohl and West, 1973). In mammals the results are more scattered with cases of total, partial and no substitution (see Table 2.4). The ambient temperatures used vary widely amongst these studies. In our study the ambient temperatures applied might not have been extreme enough to show substitution of activity generated heat for thermoregulatory heat. However, there is no theoretical basis to assume that substitution should exclusively occur at very low ambient temperatures. At all ambient temperatures below the lower critical temperature substitution could occur to a certain degree. Also, when partial substitution occurs, these effects may be masked by differences in heat loss under resting or active conditions. For example, when an animal leaves a well-insulated resting place to become active, thermoregulatory costs may well simultaneously shoot up due to increased surface Ambient temperature and wheel-running activity 31 1993 2004 1996 1984 1986 1973 et al.

et al. et al. et al. et al. et al. treadmill treadmill treadmill treadmill treadmill treadmill treadmill hopping on rotating bars wheel running wheel running wheel running activity in cage (S) occurred (°C) Substitutive substitution 10-30 A - V this study 5 - 253 - 25 S S 5 10 F V Baudinette Chappell 0 - 55 A -10-304 - 24 A F S - Mahoney 1980 4 V F this study Arnold 5 – 40 S 5 F 1970 Wunder 10 – 35 A - F Yousef range (°C) or at which) (F) or Forced –25 - 15 A–20 - 22 - S V 0 Gates and Hudson 1979 F Makinen –50 – 30–10 – 25 S S -45 -10 F F 1973 and West Pohl Paladino omyscus maniculatus r canthis flammea attus norvegicus attus norvegicus otorous tridactylus otorous Mus domesticus leucurus Erythrocebus patas Erythrocebus P Cervus elaphus Pe Mus domesticus R R Eutamisa merriami Ammospermophilus A Zonotrichia leucophrys Overview of studies investigating whether cost running were substitutive or additive to heat necessary for thermoregulation. apiti andom-bred controls able 2.4. atas monkey otoroo ground squirrel MAMMALS: P T Species Latin nameP W Deer mouse TemperatureR (A) Additive Mice selected for high TemperatureRat (V) Voluntary Rat Chipmunk Reference Antelope BIRDS: Common redpoll White crowned sparrow wheel-running activity 32 Chapter 2

area and reduced insulation, and thus counteract substitutive effects of activity. In this case the net effect on costs of transport may not be different and partial substitution would not be noticed. In our animals housed in their home cage with bedding during the measurements, these effects may have been more pronounced than in other studies. At 10°C mice were less active and may have chosen to use shivering thermogenesis while well-insulated and curled up in their nest instead of using heat generated by wheel-running activity to offset increased heat loss (animals did not have nesting material, but did built small nests using wood shavings). Interestingly, in selected and control mice the cost of running was found to be similar and in both groups heat generated by activity could not substitute for heat necessary for thermoregulation at the lowest ambient temperature measured. In summary, mice that have been selected for high voluntary wheel-running activity had increased mass-specific daily energy expenditure, but did not differ from control mice with respect to resting metabolic rate. Wheel-running activity decreased at low ambient temperature (10°C) in both selected and non-selected mice and was unchanged at high ambient temperature (30°C) compared to control temperature (20°C). The cost of transport was similar between the lines. It was also indistinguishable between the ambient temperatures measured, indicating that the energy spent on activity was additive and did not substitute for heat necessary for thermoregulation.

Acknowledgements The authors thank Laura Ross, Edwin Alserda, Mark Doornbos and Els Van der Zee for help with the experimental procedures. All procedures concerning animal care and treatment were in accordance with the regulations of the ethical committee for use of experimental animals of the University of Groningen (DEC nr. 3039). TG was supported by US NSF grant IBN-0212567. Ambient temperature and wheel-running activity 33

Chapter3

Behavioural and physiological responses to increased foraging effort in male mice

Lobke M. Vaanholt, Berber De Jong, Theodore Garland Jr., Serge Daan, G. Henk Visser

Abstract Free-living animals must forage for food and hence may face energetic constraints imposed by their natural environmental conditions (e.g., ambient temperature, food availability). Simulating the variation in such constraints, we have experimentally manipulated the rate of work (wheel running) mice must do to obtain their food, and studied the ensuing behavioural and physiological responses. This was done in mice selectively-bred for high spontaneous wheel running and their randomly-bred controls to vary the amount of baseline wheel-running activity. We first determined the maximum workload for each individual. The maximum workload animals could engage in was ~23 km d-1 in both control and activity-selected mice, and was not associated with baseline wheel-running activity. We then kept mice at 90% of their individual maximum and measured several physiological and behavioural traits. At this high workload, mice increased wheel-running activity from an average of 10 to 20 km d-1, and decreased food intake and body mass by approximately 20%. Mass-specific resting metabolic rate strongly decreased from 1.43 to 0.98 kJ g-1 d-1, whereas daily energy expenditure slightly increased from 2.09 to 2.25 kJ g-1 d-1. Costs of running decreased from 2.3 to 1.6 kJ km-1 between baseline and workload conditions. At high workloads, animals were in a negative energy balance, resulting in a sharp reduction in fat mass as well as a slight decrease in dry lean mass. In addition, corticosterone levels increased, and body temperature was extremely low in some animals at high workloads. When challenged to work for food mice thus showed several physiological and behavioural adaptations.

Journal of Experimental Biology, In Revision 36 Chapter 3

INTRODUCTION

Free-living animals need to forage for food and they may face energetic constraints related to their natural environmental conditions (Speakman et al., 2003a), such as low ambient temperature and limited food availability. The main energetic costs for an endothermic and homeothermic animal with a large surface-to-volume ratio, such as a mouse, are of thermoregulatory nature (rather than those related to costs of locomotion (Carbone, 2005; Garland, 1983; Goszczynski, 1986)). Mice further need energy for maintenance of the body and for foraging activity. Excess energy can be used for non-essential physical activity (e.g., play behaviour), stored as fat or invested in growth and/or reproduction. When food is scarce, mice must invest more time (and energy) in foraging, and they may face constraints on the energy available for behaviour and maintenance functions other than foraging. They then need a physiological strategy to reallocate their limited energy. Fat reserves may provide energy for a short time (Bronson, 1987; Day and Bartness, 2001), but when food availability is low for extended periods animals must reallocate energy to systems that need it most from functions that are less crucial for survival. Reducing body mass and/or mass-specific resting metabolic rate is one strategy to reduce energetic demands (Deerenberg et al., 1998; Rezende et al., 2006; Speakman and Selman, 2003). Perrigo and colleagues have shown reduced investment in reproduction by female mice challenged to work for food (Perrigo, 1987; Perrigo and Bronson, 1985). Experiments by Adage et al. (2002) have shown that rats challenged to work for food undergo numerous physiological changes, including a reduction in body mass, blood glucose, and insulin levels, accompanied by increases in insulin sensitivity, ACTH, and corticosterone level. In these rats there was large inter-individual variation in the amount of wheel running rats could perform. The ability to maintain body mass during the working period could be predicted from the individual spontaneous wheel-running activity. This raises the intriguing question of whether spontaneous locomotor activity reflects the physiological capacity of individuals. To address this question, we have exploited the existence of replicate mouse lines that have been selectively-bred for high voluntary wheel-running activity (Swallow et al., 1998). We investigated the effects of an increase in foraging effort on behaviour, energy metabolism, body temperature, and body composition in both the selected lines and their random-bred control lines. Animals were housed in specialized cages with a running wheel and food dispenser. A steering computer controlled food rationing as determined by running-wheel activity. With this paradigm, as pioneered by Perrigo (1985; 1983), we could experimentally vary the wheel-running activity required to obtain a pellet of food. This is intended to mimic variations in the work animals would need to do to secure a living in nature under varying food availability. Unlike caloric restriction, this protocol more carefully mimics the natural conditions animals face. The present study had two aims: firstly, to investigate physiological and behavioural responses to high workloads and secondly, to investigate whether mice with a high Physiological adaptations to hard work 37

spontaneous level of wheel running would respond differently to the exposed challenge.

MATERIAL AND METHODS

Animals & housing Outbred Hsd:ICR mice (Mus domesticus) selected for high wheel-running activity over 31 generations and their random bred controls were obtained from Theodore Garland Jr. (for selection procedure see ((Swallow et al., 1998), see also (Garland, Jr., 2003)), and a breeding colony (without further selection) was started at the Zoological Laboratory in Haren, Netherlands. Sixteen male mice, 8 from one of the control lines (C; lab designation is line 2) and 8 from one of the selected lines (S; line 7) were used in the experiments. At 4–5 weeks of age, mice were housed individually in cages (30x30x40 cm) equipped with a plastic running wheel (Ø 14,5 cm, code 0131, Savic®, Belgium). They were maintained on a 12:12 light-dark cycle (lights on at 8:00 CET). Food (standard rodent chow RMB-H (2181), with a gross energy content of 16.2 kJ g-1, HopeFarms, Woerden, NL) and water were provided ad libitum. Spontaneous wheel-running activity was recorded automatically by a PC- based event recording system (ERS) and stored in 2-min bins. Body mass and food intake were determined throughout the whole experiment at 11:00 each day. When the animals worked for food, pellets (0.045 g per pellet) that were not eaten were removed, counted, and deducted from the total number of pellets the mice received. However, small, crumbled and wasted pieces of food (orts) were not removed, and hence represent an uncontrolled, but probably minor, source of error variance; see (Koteja et al., 2003). All procedures concerning animal care and treatment were in accordance with the regulations of the ethical committee for the use of experimental animals of the University of Groningen (License DEC 3039(-1)).

Experiment 1: Individual maximum workload All mice were kept for 30–40 days under ad libitum food conditions. At 8–9 weeks of age, food was removed and the running wheel was connected to a food dispenser (Med Associates Pellet dispensor ENV-203, Sandown Scientific, Hampton, United Kingdom) that released a food pellet (45 mg precision food pellets with a gross energy content of 13.4 kJ g-1, Sandown Chemicals, Surrey, UK) at a set number of revolutions (General Electric Series 3 Programmable Controller). The number of revolutions per pellet was established for each mouse by dividing its average spontaneous daily wheel-running activity over the previous week (= baseline wheel running) by 150. When running at baseline a mouse would thus receive 6.8 g of food (150 x 0.045), which is similar to the amount of food a mouse on ad libitum food would eat. On average mice had to run 218 (s.d. 54) revolutions per pellet at baseline level. All animals were kept at this level for two days, then the number of revolutions was increased by 15% of baseline every two days until the animal reached 38 Chapter 3

its maximum wheel-running activity. This maximum was defined as when a mouse started decreasing its wheel-running activity (running mean over three days) for three consecutive days. After the maximum was established, animals stayed in the same cages with a running wheel and received ad libitum food to allow recovery.

Experiment 2: Behavioural and physiological consequences of high workload Because we did not show any statistically significant differences in the response to workload between control (C) and activity-selected (S) mice in experiment 1 (see Results section), animals from both groups were pooled in experiment 2. These animals will be referred to as Workload mice (n=16). The workload mice were allowed to recover from experiment 1 for at least four weeks prior to the start of the experiment 2. Again food was taken away and the running wheels were connected to food dispensers via the computer system on day zero (t=0). Animals had to work at baseline level for two days and then over a period of 14 days the workload was increased by equal steps every two days until the workload had increased to 90% of the individual maximal wheel-running activity established in Experiment 1. Mice were kept at this level for 10 days and then ter- minated. To test whether the Workload mice had sufficiently recovered from experiment 1 and to enable comparisons of body composition an extra control group was used. Mice in this control group were housed in standard cages with a running wheel (15x30x15cm, Macrolon Type II long, UNO Roestvaststaal BV, Zevenaar, NL) when they were 4-5 weeks old, and received ad libitum food (standard rodent chow RMB- H (2181), HopeFarms, Woerden, NL) throughout the experiment. The group consisted of three mice from the C line and four from the S line. This group will be referred to as Ad-lib mice (n=7).

METABOLIC MEASUREMENTS In the Workload mice body temperature, daily energy expenditure (DEE, using the doubly-labeled water technique, DLW), and resting metabolic rate (RMR, indirect calorimetry) were determined twice, once during baseline (day -4 to 0) and once during workload (day 19 to 23, at 90% of maximal workload). In the Ad-lib group, DEE and RMR were determined once (at the same age as the working mice during the second measurements). The protocol for the measurements was as follows. First, mice were weighed on a balance to the nearest 0.1 g and body temperature was measured at 11:00 using a rectal probe inserted to a depth of approximately 10 mm (+ 0.1°C, NTC type C, Ahlborn, Holzkirchen, Germany). Thereafter we injected the animal with about 0.1 g doubly labeled water (2H and 18O concentrations of the mixture 37.6% and 58.7%, respectively) allowing an equilibration period of 1 hr. The precise dose was quantified by weighing the syringe before and after administration to the nearest 0.0001 g. After puncturing the end of the tail, an “initial” blood sample was collected and stored in three glass capillary tubes each filled with about 15 µl blood. These Physiological adaptations to hard work 39

capillaries were immediately flame-sealed with a propane torch for later analysis. Thereafter the mouse was returned to its cage. After 48 h a “final” blood sample was collected as described before, and the animal was weighed again. We collected blood samples of four sentinel mice from our breeding colony which had not been injected with DLW, to assess the natural abundances of 2H and 18O in the body water pools of the animals. Throughout these measurements the Workload mice were working for their food at 90% of their previously observed maximum (Experiment 1), and the Ad-lib mice had access to a running wheel. The next day at 12:00, animals were transferred to an 8-channel respirometry system to determine RMR. Mice were put in flow-through boxes (15x10x10 cm) . connected to an open-flow respirometry system where oxygen consumption (V O2, -1 . -1 l h ) and carbon dioxide production (V CO2, l h ) was measured simultaneously with ambient temperature and activity for 24 h, as described by Oklejewicz et al. (1997). In brief, oxygen and carbon dioxide concentration of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck, Damstadt, Germany) from each chamber was measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100, Crow- borough, United Kingdom) and carbon dioxide by an infrared gas analyzer (Servomex 1440), respectively. The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic cages. Flow rate of inlet air was set at 20 l h-1 and measured with a mass-flow controller (Type 5850 Brooks, Rijswijk, Netherlands). Data were collected every 10 minutes and automatically stored on a computer. Animals from the Workload groups received ~3 g of food (based on their food intake at that moment) and a piece of apple while in the respirometer. Animals from the other group (Ad-lib mice) had ad libitum food and a piece of apple. Metabolic rate (MR, kJ h-1) was calculated using the following equation: MR = . . (16.18 x V O2) + (5.02 x V CO2) (Romijn and Lokhorst, 1961). RMR was defined as the lowest value of metabolic rate in half-hour running means. RMR in this study thus represents the lowest metabolic rate of animal at room temperature (22°C).

MASS SPECTROMETRY The determinations of the 2H/1H and 18O/16O isotope ratios of the blood samples were performed at the Centre for Isotope Research employing the methods described in detail by Visser and Schekkerman (1999) using a SIRA 10 isotope ratio mass spectrometer. In brief, each capillary was microdistilled in a vacuum line. The 18 16 O/ O isotope ratios were measured in CO2 gas, which was allowed to equili- brate with the water sample for 48 h at 25°C. The 2H/1H ratios were assessed from H2 gas, which was produced after passing the water sample over a hot uranium oven. With each batch of samples, we analysed a sample of the diluted dose, and at least three internal laboratory water standards with different enrichments. These standards were also stored in flame-sealed capillaries and were calibrated against IAEA standards. All isotope analyses were run in triplicate. 40 Chapter 3

-1 The rate of CO2 production (rCO2, moles d ) for each animal was calculated with Speakman's (1997) equation:

rCO2 = N/2.078 * (ko – kd) – 0.0062 * N *kd

where N represents the size of the body water pool (moles), ko (1 d-1) and kd (1 d-1) represent the fractional turnover rates of 18O and 2H, respectively, which were calculated using the age-specific background concentrations, and the individual-specific initial and final 18O and 2H concentrations. The value for the amount of body water for each animal was obtained from the carcass analyses. The amounts of body water of the animals at baseline conditions were calculated from the body water vs body mass relationship of the 7 control animals. Finally, the rate of CO2 production was converted to energy expenditure assuming a molar volume of 22.4 l mol-1 and an energetic equivalent per l CO2 based on RQ measurements in our respirometry -1 setup (on average 22 kJ l CO2,(Gessaman and Nagy, 1988)).

BODY COMPOSITION After the respirometry measurement all animals were euthanized with CO2 followed by decapitation, and organs were dissected out and weighed to the nearest 0.0001 g. All tissues were stored at -20°C until further analysis. Dry and dry lean organ masses were determined by drying organs to constant mass at 103°C followed by fat extraction with petroleum ether (Boom BV, Meppel, NL) in a soxhlet apparatus.

HORMONES Blood samples were taken from the Workload mice from the tail tip during baseline (day –5) and workload (day 18) at 10:00 (one hour prior to daily weighing). Behaviour of the mice was noted prior to sampling, and all mice were at rest. Animals were not aneasthesized and samples were collected within 90 seconds of initial disturbance. Blood was collected in Eppendorf tubes with EDTA as anticoagulant and kept on ice until it was centrifuged at 2600 rpm at 4°C. The supernatant was collected and stored at –80°C. Corticosterone levels were determined using RIA (Linco Research, Nucli Lab, The Netherlands).

Data analysis Statistical analysis was performed using SPSS for windows (version 14.0). For experiment 1, we applied repeated measures ANOVA with line (C vs. S) as between-subjects factor and treatment (baseline vs. workload) as within-subjects factor. For experiment 2, paired t-tests were used to test for differences between baseline and workload conditions within the Workload animals and independent t- tests were used to test for differences between Ad-lib and Workload animals. All tests were two-tailed and significance was set at p≤ 0.05. Physiological adaptations to hard work 41

RESULTS

Experiment 1: Maximum workload Table 3.1 shows values of wheel-running activity, body mass, and absolute and mass-specific food intake in the workload mice during baseline and at maximum workload. We applied repeated-measures ANOVA to investigate differences between the workload and baseline conditions (within-subjects factor), and between the lines (C vs. S; between-subjects factor). Overall, wheel-running activity did not differ statistically between C and S mice (Table 3.1, no effect of line). However, as illustrated in Figure 3.1, post-hoc t-tests showed that spontaneous wheel-running activity under baseline conditions was significantly higher in S mice (14.7 km d-1, see Table 3.1) than in C mice (11.5 km d1; p=0.05). Body mass and food intake did not differ between C and S mice (Table 3.1). When challenged to work for food, all mice increased wheel-running activity (Figure 3.1). The maximum level of running did not differ statistically between C and S mice, and was on average 23.3 km d-1 in both groups (Table 3.1). This maximum level was independent of the spontaneous baseline wheel-running activity of the individual mice, as shown in Figure 3.1 (Pearson's r = 0.3, 2-tailed p = 0.26). At the maximal level of wheel running, body mass had decreased by approximately 16% and absolute food intake by 20% (significant effect of treatment; see Table 3.1). Mass-specific food intake did not differ between baseline and workload condition (no effect of treatment, Table 3.1). No significant interaction effects were seen between line and treatment. C and S mice thus responded similarly to the workload schedule, and both groups showed a similar increase in wheel-running activity and similar decreases in body mass.

Table 3.1. Experiment 1; Effects of maximal workload on main characteristics in control (C) and activity-selected mice (S).

Baseline Maximal workload P-values for repeated measures ANOVA C (n=8) S (n=8) C (n=8) S (n=8) d.f. Line Treat

Wheel-running activity 11.5±1.2 14.7±0.8 23.2±1.4 23.4±1.4 1,14 0.29 <0.001 (km d-1) Body mass (g) 30.9±0.5 30.6±0.5 26.0±0.3 25.5±0.5 1,14 0.82 <0.001 Food intake (g d-1) 5.7±0.1 6.0±0.2 4.6±0.3 4.6±0.2 1,14 0.44 <0.001 Food intake 0.46±0.01 0.43±0.01 0.40±0.02 0.41±0.02 1,14 0.56 0.02 (g g BM0.75- d-1) Mass-specific food intake 0.20±0.01 0.18±0.01 0.18±0.01 0.18±0.01 1,14 0.38 0.43 (g g-1 d-1)

Wheel-running activity, body mass, and food intake during baseline and at maximal workload in activity-selected mice and random-bred controls. Values given are mean±sem. Repeated-measures ANOVA with line (C vs. S) as between-subjects factor and treatment (baseline vs. workload) as within-subjects factor were performed. Inter- actions were never significant and are therefore not shown in the table. Significant results are shown in bold. BM=body mass, n=sample size, d.f. = degrees of freedom. 42 Chapter 3 ) -1 30

0 max. wheel-running activity (km d 0 5 1015 20 spontaneous wheel-running activity (km d-1)

Figure 3.1. Relationship between spontaneous wheel-running activity (RWA BL) and maximum wheel-running activity (RWA MX) in control mice (C, open circles) and mice selectively-bred for high wheel-running activity (S, closed circles). Linear regression gave the following equations: combining both groups: RWA MX = 0.35 RWA BL + 20.1 (R2=0.09, n.s.); for C mice: RWA MX = –0.22 RWA BL +27.6 (R2=0.02, n.s.), and for S mice: RWA MX = 0.84 RWA BL +15.2 (R2=0.47, n.s.).

2.5 ) -1 2.0

1.5

1.0

0.5 wheel-running activity (km h

0.0 8810 12 14 16 18 20 22 0246 clock time (h)

Figure 3.2. Circadian pattern of wheel-running activity in control (C; open symbols) and activity–selected mice (S; closed symbols) running spontaneously (circles) or running for food (triangles). Each symbol plots the mean distance ran in hourly bins (e.g., bin 12 = from 12:00 till 13:00). Vertical bars are inter-individual s.e.m. The black bar on top represents the dark phase.

Figure 3.2 shows the circadian pattern of wheel-running activity during baseline and workload. Under baseline conditions, mice mainly ran in the dark phase. A small peak in wheel-running activity after 11:00 (time of daily measurements) can be observed, probably due to disturbance. When challenged to work for food, the period of running was extended and mice started running more during the light phase. It appears that the mice shifted the onset of activity towards the time at which daily measurements took place. Physiological adaptations to hard work 43

Experiment 2: Behavioural and physiological consequences of high workload Experiment 1 showed no differences in wheel-running activity, body mass or food intake between C and S mice under the high workload conditions. In Experiment 2 we therefore pooled data from both groups (Workload mice, n=16) to study the effects of workload on behavioural and physiological traits. Effects of workload were investigated by comparing the baseline condition (ad libitum food) to the high workload condition (wheels attached to food dispensor) within these mice (using paired t-tests). For comparison of body composition, however, an additional control group of 7 age-matched animals housed with a wheel and ad libitum food was added (Ad-lib group). This extra control group also enabled us to determine whether the Workload mice had sufficiently recovered from Experiment 1 before the start of Experiment 2.

DEVELOPMENT OF BODY MASS, FOOD INTAKE, AND WHEEL-RUNNING ACTIVITY AT SUB-MAXIMAL WORKLOAD For daily measurements (body mass, food intake, and wheel-running activity) we calculated a baseline and workload value that was the average over one week (see Table 3.2). For the baseline condition, this was the week prior to the start of the training, and for the workload the week started when the animals were on a maximal workload for 2 days. To determine whether the animals had recovered sufficiently from Experiment 1, we first compared baseline data (Workload group) to data on animals in the Ad-lib

Table 3.2. Experiment 2; Main characteristics of Ad-lib animals, and Workload animals at baseline or workload conditions.

Ad-lib animals (n=7) Workload animals (n=16) Baseline Workload

Wheel-running activity (km d-1) 7.7±1.3 10.2±0.9b 20.2±1.5 Body mass (g) 34.2±0.8 34.6±0.5b 28.2±0.5 Food intake (g d-1) 4.3±0.4 6.4±0.2a,b 4.0±0.2 Mass-specific food intake (g g-1 d-1) 0.13±0.01 0.19±0.01a,b 0.14±0.01 RMR (kJ d-1) 49.5±1.9 49.3±1.2b 27.4±1.8 Mass-specific RMR (kJ g-1 d-1) 1.45±0.05 1.43±0.03b 0.98±0.05 DEE (kJ d-1) 62.6±2.9 72.3±1.7a,b 60.0±1.7 Mass-specific DEE (kJ g-1 d-1) 1.83±0.10 2.09±0.04a,b 2.25±0.07 Body temperature (°C) - 36.6±0.4b 35.4±0.8 Corticosterone (x103 ng ml-1) - 15±5b 222±47

Average(±sem) wheel-running activity, body mass, food intake, resting metabolic rate (RMR), daily energy expenditure (DEE), body temperature, and corticosterone level are shown for workload animals under baseline and workload conditions and for ad libitum fed mice. Values are mean±sem. One control animal died during the respirometry measurements and data on RMR thus were not available. a Indicates a significant difference between Ad-lib and Workload mice at baseline (independent t-test, p<0.05) and b indicates a significant difference within the Workload group between baseline and workload conditions (paired t-test, p<0.05). 44 Chapter 3

group of the same age using independent t-tests (see Table 3.2). Ad-lib and Workload mice under baseline conditions did not systematically differ in body mass or wheel-running activity (see Figure 3.3, triangles, and Table 3.2). Food intake was slightly lower in Ad-lib mice than in Workload mice (4.3 vs. 6.3 g d-1). These results indicated that mice had recovered sufficiently from the preliminary workload experiment and subsequently the new workload scheme was started. Figure 3.3 shows the changes in body mass and food intake that occurred in the Workload mice when put on a workload schedule. On day 0 wheels were attached to the food dispensers and the foraging effort was increased over 14 days up to 90% of the previously observed maximum for each mouse (training period). Mice were kept at this level for 10 days (workload period). Wheel-running activity showed a slight decrease just before the start of the training, which can probably be attributed to the manipulations done at this time (doubly-labeled water injections). Wheel-running activity increased steadily during the training period and reached a plateau of approximately 20 km d-1 at the highest workload (90% of maximum workload). In the Workload mice, body mass decreased significantly with approximately 20% from 34.6 to 28.4 g, and both absolute and mass-specific food intake decreased by approximately 30% at 90% workload compared to baseline. Wheel- running activity approximately doubled at high workload in the Workload mice (see Figure 3.3 and Table 3.2). We calculated the average time spent running by adding up all the 2-min intervals in which running occurred per day, and the maximum speed the mice ran (max distance covered per 2-min interval). This was done during baseline and workload to determine which strategy animals used to increase their wheel-running activity. During baseline, time spent running was 5.9 h (s.d.1.8), but this almost doubled to 11.5 h (s.d. 2.0) during workload. Max running speeds were 4.7 km h-1 (s.d. 0.8) and 6.3 km h-1 (s.d. 0.5) in baseline and workload phases, respectively (paired t- test; p<0.001 for both). Mice thus increased both time spent running (+94%) and maximum running speed (+34%). Multiple regression analysis showed that food intake was significantly, positively predicted by both body mass and wheel-running activity at baseline (Multiple regression: R2=0.49, p=0.012; body mass, p=0.018, wheel-running activity= 0.067), as well as during the high workload experiment (Multiple regression: R2=0.58, p=0.004; body mass, p=0.0012, wheel-running activity=0.002).

METABOLIC RATE Metabolic rate of the Workload animals was measured under baseline and workload conditions (Table 3.2). First, we compared resting metabolic rate (RMR) and daily energy expenditure (DEE) between Ad-lib animals and Workload animals at baseline (see Table 3.2). No significant differences were found for RMR, but DEE was significantly lower in the Ad-lib fed mice, which might be due to the slightly smaller cages they were housed in. Second, we compared RMR and DEE under baseline and workload conditions within the Workload group. At 90% of maximum workload, Physiological adaptations to hard work 45

baseline training workload ) -1 25

wheel-running activity (km d 0 38

26 body mass (g)

8 )

-1 6

2 food intake (g d

0 -10-50 5 10 15 20 time (d)

Figure 3.3. Development of wheel-running activity, body mass, and food intake during training and at a workload of 90% from the maximal capacity in Workload animals (C & S groups pooled). Spontaneous wheel-running activity is shown for the 2 weeks prior to the training period (day 0). Circles show the development of the different variables during the experiment in the Workload animals and triangles represent average values for mice in the Ad-lib group. mice decreased RMR by approximately 50%, from an average 49.3 kJ d-1 to 27.4 kJ d-1. The reduction in mass-specific RMR was about 1/3, from 1.43 to 0.98 kJ g-1 d-1. Both differences were statistically significant. Workload also influenced absolute and mass-specific DEE. Absolute DEE decreased on average from 72.3 to 60.0 kJ d-1 at high workload, but mass-specific DEE slightly increased from 2.09 to 2.25 kJ d-1. Both differences were statistically significant (Table 3.2). Looking at individual variation, all mice except one individual exhibited a decrease in DEE during workload (whole-animal values). 46 Chapter 3

100 A B

80 ) -1

metabolic rate (kJ d 20

0 22 26 3034 38 0 10 20 30 body mass (g) wheel-running activity (km d-1)

Figure 3.4. Relationship between body mass and metabolic rates (A) and between wheel-running activity and metabolic rates (B) at baseline (open symbols) and workload (closed symbols) conditions in Workload animals. Triangles represent the resting metabolic rate (RMR) and circles represent daily energy expenditure (DEE). Regression lines for all relationships are drawn. For equations of the regression lines, R-square, and p-values see Table 3.4. Results of multiple regressions are presented in the text.

We estimated the cost of activity (ACT, in kJ d-1) by deducting RMR from DEE (ACT was 23.0 and 32.6 kJ d-1 at baseline and workload respectively), and divided this by the amount of wheel running to estimate the costs per km. Costs of running were 2.3 kJ km-1 (s.d., 1.6) and 1.6 kJ km-1 (s.d., 0.3) at baseline and workload, respectively. This difference was significant (paired t-test, 2-tailed, p=0.026). It is well known that metabolic rates (RMR and DEE) are positively associated with body mass, and under baseline conditions this relationship was obvious in all mice, based on bivariate relationships (open symbols in Figure 3.4A, Table 3.3). However, when working for food there was no longer a statistically significant relationship between body mass and metabolic rates (closed symbols in Figure 3.4A, Table 3.3). We also performed multiple regression analyses with body mass and wheel-running activity as independent predictors of RMR or DEE. At baseline, the models including both body mass and wheel-running activity was significant (R2=0.48, p=0.015), but only body mass (p=0.007) and not wheel-running activity (p=0.148) significantly predicted RMR. The same was true for the relationship with DEE (R2=0.43, p=0.025; body mass; p=0.008, wheel-running activity, p=0.785). Body mass alone explained more of the variation in RMR and DEE than models that included wheel-running activity (see Table 3.3). At high workload, metabolic rates were better predicted by the amount of wheel-running activity than by body mass (see Figure 3.4B and Table 3.3). Multiple regressions for DEE or RMR with body mass and wheel-running activity were not significant (RMR: R2=0.25, p=0.182; body mass; p=0.709, wheel-running activity, Physiological adaptations to hard work 47

Table 3.3. Linear regressions of metabolic rates (RMR and DEE) on body mass or on wheel-running activity in workload mice.

Linear regression Slope Intercept R2 p Baseline Body mass vs. RMR 1.52 -3.3 0.38 0.011 Body mass vs. DEE 2.48 -13.5 0.50 0.002 Wheel running vs. RMR –0.35 52.9 0.06 0.370 Wheel running vs. DEE 0.61 66.2 0.09 0.260 Workload Body mass vs. RMR 0.13 23.8 0.01 0.390 Body mass vs. DEE 0.36 50.0 0.05 0.790 Wheel running vs. RMR –0.56 38.8 0.24 0.065 Wheel running vs. DEE 1.10 37.7 0.41 0.005

Slope, intercept, R2, and p-values are given in the table. Multiple regressions showed that at baseline RMR and DEE were better predicted by body mass alone and at workload by wheel-running activity alone. Results of multiple regressions with both body mass and wheel-running activity as independent predictors of RMR or DEE are described in the text. See also Figure 3.4.

p=0.071 and for DEE: R2=0.25, p=0.158; body mass; p=0.241, wheel-running activity, p=0.073). As shown in Table 3.3, wheel-running activity alone did significantly predict DEE (p=0.005), and approached significance for predicting RMR (p=0.065). RMR was negatively related to wheel-running activity, while DEE was positively related to wheel-running activity at workload. The animals that ran the most thus decreased their RMR the most, while increasing DEE. RMR and DEE at baseline did not relate to RMR and DEE at workload.

ENERGY BALANCE Figure 3.5 shows the energy budget of Workload mice at baseline and workload calculated over the days when DEE was measured in these mice. The figure shows the various components of the energy budget; gross energy intake (GEI), metabolisable energy intake (MEI), and daily energy expenditure (DEE) divided into resting metabolic rate (RMR) and energy spent on activity (ACT). GEI was calculated on the basis of the measured food intake and was 97.4 and 53.6 kJ d-1 in mice under baseline and workload conditions respectively (see M&M, for gross energy content of the food). Animals are not 100% efficient in metabolising their food and the actual amount of energy animals take out of their food can only be calculated when digestive efficiency and the amount of energy lost in the urine has been measured as well. Previous studies have shown a digestive efficiency of 79.1% in ad libitum fed mice, including loss of energy in urine (Hambly and Speakman, 2005). Under the assumption that workload did not alter digestive efficiency, MEI at baseline and workload was estimated using a digestive efficiency of 79.1%. Based on these values we can see whether animals were in a positive or negative energy balance. It is clear 48 Chapter 3

A absolute (kJ d-1) B mass-specific (kJ g-1 d-1)

GEI

GEI W 0.6 W DEE 20.4 DEE

+0.15 GEI DEE ACT ACT

+4.7 W 0.66 1.28 –0.74 ACT DEE MEI 0.4 23.0 GEI ACT 2.2 MEI W 32.6 –17.7 77.0 11.2 RMR MEI RMR 1.43 1.5 49.3 MEI RMR 42.4 RMR 0.98 27.4

baseline workload baseline workload

Figure 3.5. Energy budget of Workload mice during baseline and workload conditions. Panel A shows the absolute values and B the mass-specific values. To determine the energy balance we used measures of RMR and DEE. Energy for activity (ACT) was calculated by deducting RMR from DEE. In addition, the gross energy intake (GEI) was calculated on basis of the absolute food intake during the doubly labelled water measurements. Metabolisable energy intake (MEI) was then calculated from GEI assuming that digestive efficiency together with energy lost in the urine was 79.1 % (Hambly and Speakman, 2005). The white bars represent the energetic value of the food that is not metabolised (GEI - MEI = Waste, W). The numbers in the bars represent the amount of energy (either in kJ d-1 or in kJ g-1 d-1) spent on each part of the energy budget. The bracket shows the surplus energy available to the animals for growth.

from this picture that at high workload the proportion of energy used for RMR was strongly decreased and the energy available for activity had increased. At high workloads there was a negative energy budget of -17.7 kJ d-1 (or -0.74 kJ g-1 d-1) and the extra energy needed was obtained by reducing body mass by 0.8 grams on average. During baseline the energy budget was positive, +4.7 kJ d-1 (or +0.15 kJ g-1 d-1), and animals gained 1.0 gram body mass over the course of the measurements. Even after assuming an unlikely digestive efficiency of 100% in the workload animals, the energy budget would still be negative (-6.4 kJ).

BODY COMPOSITION We compared data from the animals in the Workload group with the animals in the Ad-lib group using independent t-tests to investigate the effects of workload on body composition (see Table 3.4). Body mass, total dry lean, and fat content were strongly decreased in animals in the Workload group. Fat content decreased the most, by 70%, from 3.1 to 0.9 g. Dry lean organ masses were significantly decreased in all organs of working animals compared to Ad-lib animals, except for Physiological adaptations to hard work 49

Table 3.4. Body composition of Ad-libitum fed mice and mice working for food.

Variable (g) Ad-lib animals Workload animals % difference Independent t-test (n=7) (n=15) tp

Body mass 31.4±0.9 25.9±0.4 –17 –6.8 <0.001 Dry lean mass 8.1±0.2 6.6±0.1 –19 –9.9 <0.001 Fat content 3.13±0.35 0.94±0.09 –70 –8.2 <0.001 Dl heart 0.04±0.001 0.03±0.001 –29 –4.8 <0.001 Dl liver 0.48±0.03 0.30±0.02 –37 –5.0 <0.001 Dl kidney 0.13±0.01 0.10±0.001 –29 –4.0 0.001 Dl brain 0.08±0.001 0.08±0.001 –3 –1.3 0.23 Dl stomach 0.04±0.001 0.04±0.001 +5 0.9 0.36 Dl intestines 0.29±0.01 0.34±0.01 +18 4.5 <0.001 Dl lung 0.04±0.001 0.04±0.001 +3 0.6 0.59 Dl skin 1.50±0.04 1.26±0.02 –16 –6.0 <0.001 Dl rest 5.48±0.12 4.35±0.05 –21 –10.6 <0.001 Fat heart 0.006±0.001 0.005±0.001 –18 –1.1 0.30 Fat liver 0.048±0.009 0.027±0.003 –43 –2.8 0.011 Fat kidney 0.026±0.003 0.006±0.001 –75 –7.7 <0.001 Fat brain 0.047±0.001 0.043±0.001 –10 –3.7 0.002 Fat stomach 0.008±0.001 0.005±0.001 –36 –4.2 <0.001 Fat intestines 0.070±0.010 0.039±0.002 –44 –4.2 <0.001 Fat lung 0.008±0.001 0.004±0.001 –51 –6.7 <0.001 Fat skin 0.85±0.120 0.16±0.028 –81 –7.6 <0.001 Fat rest 2.07±0.240 0.65±0.071 –69 –7.4 <0.001

Mean(±sem) total dry lean mass, fat content and the dry lean (dl) mass and fat mass of separate organs are shown for Workload and Ad-lib mice. One mouse died during the second respirometry measurement in the Workload group. % difference shows the change in mass between Ad-lib and Workload animals. Independent t-tests were performed to test for differences between groups and results are shown in the table.

the brain, stomach, and lungs that showed no difference, and the intestines that showed a significant increase in dry lean mass. Fat content also decreased significantly in most organs (except for the heart), with the largest decrease in skin (81 %) and the lowest in the brain (10%). We also calculated mass-specific organ masses to be able to correct for effects of body mass on body composition (data not shown). In these analyses, total fat content and fat content of all organs (except for heart) still showed a significant decrease. Total mass-specific dry lean mass did not differ between Ad-lib and Workload animals anymore; dry lean mass did significantly decrease in liver, kidney, skin and the remainder of the carcass, but it increased significantly in brain, stomach, intestine, and lung. The total fat content of the mice could be negatively predicted by the amount of wheel-running activity at workload (r=-0.67, p=0.006). 50 Chapter 3

BODY TEMPERATURE & PLASMA CORTICOSTERONE Body temperature of the Workload animals was measured in the light phase under baseline and workload conditions (see Table 3.2). Three out of 16 mice under workload conditions had extremely low body temperatures at the time of measurement (32.2, 32.5 and 26.8°C), but no significant differences were found within the workload mice between baseline or workload conditions. Plasma corticosterone levels were strongly affected by treatment. At high workload, corticosterone levels were approximately 15-times increased (see Table 3.2). Body temperature or plasma corticosterone did not correlate with wheel-running activity.

DISCUSSION

Challenging mice to work for food to mimic low food availability resulted in several physiological and behavioural changes that may be adaptative. All animals increased wheel-running activity by approximately 100%. This was mainly accomplished by spending more time running (including during the light phase), but running speed also increased. A shift in activity patterns towards the day in response to workload was shown before in Mus musculus (Perrigo, 1987). The increase in wheel-running activity was not sufficient to maintain adequate food intake, and body mass decreased (Figure 3.3). A detailed look at the body composition of the workload mice showed that the reduction in body mass was mainly caused by a reduction in fat mass. Total fat content was reduced by ~70% in Workload mice compared to mice in the Ad-lib group. Fat content of all organs (except for the heart) reduced significantly, with the most pronounced decreases in subcutaneous and intra-peritoneal fat and the smallest decrease in the brain. Similarly, dry lean mass of the brain was not significantly reduced; mass-specific dry lean mass of the brain even increased in the workload group. The brain is very important for the central regulation of bodily functions and is apparently protected in times of scarcity. A similar result was found in food- restricted rats, where brain mass was unaffected, but heart, kidney, and liver mass decreased (Greenberg and Boozer, 2000). Total mass-specific dry lean mass was similar in Ad-lib and Workload mice, but the distribution of dry lean mass over the body did change under high workload conditions. In liver, kidney, skin, and the remainder of the carcass, mass-specific dry lean mass was decreased, whereas it was increased in lung, stomach, and intestine. The increase in intestine mass and stomach mass could indicate that animals increased their digestive efficiency under workload conditions. This would enable them to get more energy from a gram of food. In addition, mice could have ingested their faeces (coprophagy) to increase their food efficiency even more. Further studies would be necessary to test these hypotheses. The strong reduction in fat content without a major change in dry lean mass is in agreement with observations by Perrigo and Bronson (1983) in pre- pubertal female mice. In their study, fat depots remained undiminished or above Physiological adaptations to hard work 51

control levels over a wide range of forced activity, even when accompanied by a moderate decrease in food intake, but at the maximum requirement of 225 revolutions per pellet (comparable to our conditions) females accumulated less body fat than ad libitum fed animals. Studies on food restriction in sedentary mice show con- trasting results on body composition changes with greater use of fat mass than dry lean mass (Greenberg and Boozer, 2000), defense of fat mass and reduction of dry lean mass (Hambly and Speakman, 2005), or no differential use of the different components (Selman et al., 2005). Corticosterone levels were increased at high workload and comparable to the values reported in response to restraint stress in male mice of this strain (Malisch et al., 2006). Baseline values were slightly lower than the ones reported in that study. We did not show a relationship between wheel-running activity (over 24h) and corticosterone or body temperature. Wheel-running activity in the 10–20 min. prior to measurements has been shown to correlate positively with both body temperature (Rhodes et al., 2000) and plasma corticosterone (Girard and Garland, Jr., 2002) in these lines of mice. Unexpectedly, selective breeding for high spontaneous wheel-running activity did not affect the response to a workload challenge, at least based on the two of eight total lines (see Swallow et al., 1998) that we studied here. Control (C) and activity-selected (S) mice did not differ with respect to their maximum wheel-running activity on a high workload (~23 km d-1; Table 3.1) and both groups showed similar decreases in food intake and body mass at the maximum workload. Also, spontaneous wheel-running activity at baseline did not predict wheel-running activity at workload (Figure 3.1). These results are in contrast to a similar study in rats (Adage et al., 2002). Based on measurements of spontaneous wheel-running activity they divided female Wistar rats from a single population into high or low spontaneous runners. They found that animals with high baseline running activity coped better on a workload schedule than rats with low spontaneous levels of wheel-running activity, and the former could also increase their wheel-running activity more. The rats with low spontaneous levels of activity markedly decreased in body mass, whereas rats that had high levels of spontaneous wheel running maintained body mass at the same workload level. The discrepancy between our study and the study of Adage et al. (2002) may represent differences between mice and rats in the regulation of wheel-running activity and body mass, and also may depend on differences in motivation to run. The rats were of similar age (3–4 months) to our mice, so age was probably not a factor. Resting metabolic rates and, to a lesser extent, daily energy expenditure showed a strong reduction under workload conditions (~50%), an effect that has been shown in several studies manipulating workload; in birds (Bautista et al., 1998; Deerenberg et al., 1998; Wiersma and Verhulst, 2005), hamsters (Day and Bartness, 2001), and mice (Perrigo, 1987), for summary see Table 4 in Wiersma and Verhulst (2005). In another study, an increase in daily energy expenditure has been shown (Wiersma et al., 2005), but that study used a variable rather, rather than the fixed 52 Chapter 3

reward ratio we used in this study. With increasing wheel-running activity, resting metabolic rate decreased and daily energy expenditure increased, but daily energy expenditure was lower under workload conditions than when animals were running spontaneously at a lower level. In principle, mice had unlimited access to food, but they stopped foraging at a point where their food intake was lower than the food intake of animals that had immediate access to food. Instead of increasing their food intake, animals compensated for the increased costs of activity by decreasing RMR. This may indicate the presence of constraints that prevent animals from increasing their activity further (see also Garland, 2003; Rhodes et al., 2005). First, the capacity for sustained, endurance-type activity can be a limiting factor. Second, time can be a limiting factor, and animals did extend their activity into the light phase on the workload (Figure 3.2), thus leaving less time to rest and sleep. All animals need to sleep to survive (Everson, 1995), and this may have limited the time mice had left to run. However, levels of running were much lower during the day than during the night, and animals only spent ~12 hours continuously running at high workloads, which would seem to leave enough time for rest. Third, digestive constraints could limit the intake of extra food. Total food intake was reduced at high workload compared to the baseline condition, and it is thus not likely that digestive constraints were at work in our mice. Moreover, when cold-exposed, these mice can increase their food intake by much greater amounts (Koteja et al., 2001) than were ever exhibited in the present study. Another possible constraint is metabolic. When we looked at mass-specific metabolic rates, RMR was reduced in mice at high workload, but DEE was slightly increased. Several lines of evidence indicate that maximum metabolic rates are limited by the intrinsic physiology of the animal (Drent and Daan, 1980; Speakman and Krol, 2005). When animals reach this maximum level they can no longer increase their activity (energy expenditure) to obtain more food. The maximum sustainable level of energy expenditure in laboratory mice subjected to forced exercise (Mus musculus) has been measured at 2.94 kJ g-1 d-1; see Table 2 in (Hammond and Diamond, 1997). With a daily energy expenditure of 2.25 kJ g-1 d-1, our mice were not yet at this highest sustainable level. Interestingly, measurements with the same lines of mice used here showed that animals exposed to cold with wheel access had an average daily energy expenditure of approximately 3.2 kJ g-1 d-1 at 10°C (Vaanholt et al., 2007) and animals exposed to severe cold without wheel access attained values as high as 5.7 kJ g-1 d-1 at -5°C (Koteja et al., 2001). Whatever the constraints on increasing activity further may have been, the mice in our study "chose" to compensate for the experimentally manipulated increase in energy expended on activity by reducing RMR. The mice that ran the most showed the greatest decrease in RMR. But how could they have accomplished this? First, reducing body mass reduces whole-animal RMR (Deerenberg et al., 1998; Speak- man and Selman, 2003). However, the reduction in RMR observed in the present study was much greater than expected based on changes in body mass alone, and at the high workload body mass did not significantly correlate with RMR. As proposed Physiological adaptations to hard work 53

by Rezende et al. (2006), in the lines of mice selectively-bred for high running, lowering of body mass may be a way to keep whole-animal energy costs of activity relatively low while selective breeding causes total running distance to increase. Similarly, in animals forced to work for food, lowering body mass may be a way to decrease costs of running and/or maintenance costs. Indeed, when we calculated the energy spent per km at baseline and workload condition, a reduction in whole- animal running costs of approximately 35% was found. The cost of transport (COT) estimated here (2.3 kJ km-1) is much higher than that reported previously ~1.2 kJ km-1 for these mice (Koteja et al., 1999; Rezende et al., 2006; Vaanholt et al., 2006). This discrepancy occurs because in this study we did not calculate COT based on the slope of the regression between running speed and energy expenditure, but instead made a crude estimate of COT by dividing ACT by the amount of wheel running. Animals also could have saved energy by reducing behaviours other than wheel- running activity, such as grooming or exploration, or they may have compensated by saving on maintenance processes. It has, for instance, been shown that zebra finches in energetically demanding situations refrain from mounting an immunological response to a novel challenge (Deerenberg et al., 1997) and that they invest less in regrowing feathers (Wiersma and Verhulst, 2005). Further research is necessary to determine whether similar effects may have occurred in our mice. Hypothermia, as we saw in several mice, and that has been reported in previous experiments manipulating foraging effort (Perrigo and Bronson, 1983) and in food- restricted animals (birds (Daan et al., 1989), and mice (Gelegen et al., 2006)), may also have contributed to the strong reduction in RMR. In the present study, mice were housed at 22°C, which is well below the lower critical temperature of mice; thermoregulatory costs could have been lowered even more by substituting thermoregulatory heat production for heat generated by activity. However, a previous study of these mice did not show substitution of thermoregulatory heat for heat generated by voluntary activity (Vaanholt et al., 2007). Lowering body temperature can be beneficial to save energy, but lowering body temperature may also impose a trade-off. When body temperature gets below the optimal temperature for enzymatic activity, protein turnover and/or cellular turnover in general decelerates, causing reduced repair of cellular damage or a reduction immunological defense (Deeren- berg et al., 1997). This will make animals more vulnerable and may reduce their life span. In addition, reduced body temperature may lower locomotor performance (Bennett, 1990) and impair various other physiological rate processes. Despite the strong reduction in RMR, the mice in our study were in a negative energy balance (Figure 3.5) and had to burn fat to meet their energetic requirements. Under baseline conditions mice were in a positive energy balance and gained body mass. Similarly, in humans prolonged imbalances in the energy budget (positive energy balance) have resulted in a strong increase of the prevalence of obesity. Dietary restriction alone generally only leads to weight loss during the period of restriction, and may mainly result in a reduction in lean mass, instead of fat 54 Chapter 3

mass; moreover, the lost weight is usually regained afterwards (in mice (Hambly and Speakman, 2005) and humans (Stiegler and Cunliffe, 2006)). The main reason for this is that animals have evolved ways to compensate during periods of food scarcity, including reducing resting metabolic rate and increasing digestive efficiency, or reducing activity. Dietary restriction together with exercise is the most advo- cated treatment for obesity. We showed in our working mice that combining both factors was indeed very effective in reducing fat mass. Because animals had not yet reached a new equilibrium in their energy balance, we do not know whether this loss of fat mass would be sustained over longer periods. Westerterp (2001) has shown that ‘novice’ trainees for the half-marathon lose body mass and concomitant- ly lower night-time metabolism (per gram fat-free mass), similar to what we saw in our mice. In summary, challenging mice to work for food resulted in several physiological changes. Mice readily increased wheel-running activity when they had to work for food, but they did not maintain food intake, and body mass subsequently decreased (mainly by a reduction in fat mass). Animals compensated for the increased energetic requirements by decreasing resting metabolic rate. The physiological responses were independent of inter-individual variation in spontaneous wheel-running activity, but wheel-running at the high workload was negatively related to RMR. The more they ran, the lower their RMR became. DEE showed an opposite relationship.

Acknowledgements The authors thank Gerard Overkamp for expert technical assistance with the workload equipment and Berthe Verstappen for performing the isotope analyses. Peter Meerlo enabled us to do the corticosterone measurements. We thank Kristin Schubert for comments on earlier versions of the manuscript. TG was supported by U.S. National Science Foundation grant IOB-0543429. S.D. was supported by the European 6th framework Integrated Program EUCLOCK. 55

Chapter4

Plasma adiponectin is increased in mice selectively bred for high wheel-running activity, but not by wheel running per se

Lobke M. Vaanholt, Peter Meerlo, Theodore Garland Jr., G. Henk Visser, Gertjan van Dijk

Abstract Mice selectively bred for high wheel-running activity (S) have decreased fat content compared to mice from randomly bred control (C) lines. We explored whether this difference was associated with alterations in levels of circulating hormones involved in regulation of food intake and energy balance, and whether alterations were caused by the presence of a running wheel. Plasma levels of leptin, adiponectin, and corticosterone as well as body composition were analyzed in male S mice housed with (+) and without (-) access to running wheels at ages of 10 and 18 months. These levels were compared to those found in C+ mice. Plasma corticosterone did not differ among groups. While plasma leptin levels tended to be lower in S+ mice as compared to S- or C+ mice, these differences were largely attributable to differences in fat content. Adiponectin levels were increased in S mice (+60%) compared to C mice, irrespective of wheel access. High levels of this hormone may be a trait co-segregated in mice bred for high wheel-running activity.

Hormone and Metabolic Research, In Press 58 Chapter 4

INTRODUCTION

It is generally accepted that moderate physical activity has a positive influence on health and life expectancy (Holloszy, 1988; Navarro et al., 2004). From a clinical standpoint, the most beneficial effect of regular exercise is that it prevents or antag- onizes increased adiposity and the adverse health risks (e.g. cardiovascular disease and metabolic syndrome, type-2 diabetes mellitus) that are associated with it (Holloszy, 1988; Novelli et al., 2004), for review see (Carroll and Dudfield, 2004). Indeed, physical activity promotes the breakdown of triglyceride stores inside adipose tissue and muscle, which in turn contributes to increased whole-body insulin action and improvement of tissue perfusion (Boden, 1997). Besides having a direct effect on metabolic and vascular processes, it has been pointed out that physical activity could also affect endocrine activity, which in turn could influence body adiposity (McMurray and Hackney, 2005). In this respect, adipose tissue has received considerable attention because it secretes the adipocyte hormones leptin and adiponectin. These hormones are extremely important in the long-term maintenance of energy balance and fuel homeostasis (Caro et al., 1996; Halaas et al., 1995; McMurray and Hackney, 2005; Ryan et al., 2003) and thus have a strong influence on sustainable health. Circulating levels of leptin have been shown to correlate highly with indices of fat content in many species (Park et al., 2004), and as such may signal the available amount of body fat to the brain (Halaas et al., 1995). In this way, leptin regulates appetite and metabolism in a coordinated fash- ion (Caro et al., 1996; Van Dijk et al., 1999). Adiponectin, on the other hand, does not correlate positively with indices of fat content, but was found to be negatively correlated with fat content in humans (Cnop et al., 2003; Park et al., 2004; Ryan et al., 2003) or be unrelated (Ferguson et al., 2004). Adiponectin has been implicated to stimulate fat oxidation in metabolically active tissue (Berg et al., 2002; Bruce et al., 2005; Fruebis et al., 2001; Yamauchi et al., 2002) and peripheral insulin sensitivity (Baratta et al., 2004; Schondorf et al., 2005; Yamauchi et al., 2001). Circulating adiponectin levels are reduced and leptin levels are enhanced in obese humans compared with lean individuals (Arita et al., 1999; Havel, 2001). This might contribute to the insulin resistance that is observed in obese subjects (Gil-Campos et al., 2004). Human studies on the interaction between physical activity and endocrine activity of adipose tissue are not conclusive, and outcomes may depend on the intensity of the exercise paradigm and the exercise capacity of subjects (Ferguson et al., 2004; Jurimae et al., 2005; Zoladz et al., 2005). In laboratory animals, however, leptin concentrations were found to be decreased in life-long voluntary exercising rats at 23 months of age (Novelli et al., 2004) and in hamsters that had been exercising voluntarily for 31 days (Coutinho et al., 2006), which may be consistent with a lowering of triglyceride stores in exercising animals. Mice selectively bred for high physical activity (Swallow et al., 1998) show a decreased body fat (Swallow et al., 2001) and a decreased leptin concentration in females at 3 month of age (Girard I., Rezende, E. L., and Garland, T., Jr., unpublished observation). The reduction in circulating Adiponectin in mice bred for high activity 59

leptin level, however, was greater than could be explained by the reduced fat mass alone. To further investigate the relation between physical activity, circulating adipocyte hormones, and body composition, the present study assessed these relationships in selectively bred high-activity male mice with chronic access to running wheels, and compared these effects to those found in their random-bred controls at different ages. Because relatively low plasma leptin levels in the high activity-selected females were associated with increased plasma corticosterone levels (Girard and Garland, Jr., 2002; Malisch et al., 2006) plasma levels of corticosterone were also assessed in the present study. To investigate the effects of wheel-running activity per sé on adipose and adrenal hormones and body composition, above-mentioned relations were also investigated in selectively-bred mice without the presence of a running wheel. These comparisons may shed light on the nature of changes that occur in the regulation of adipose and adrenal hormones; i.e., whether they are caused by high activity per se, or whether it is a trait that has genetically co-segregated with selection for high wheel-running activity.

MATERIAL AND METHODS

Animals & housing The progenitors for the original selection experiment (Swallow et al., 1998) were 112 pairs of outbred, genetically variable Hsd:ICR mice obtained from Harlan- Sprague-Dawley in Indianapolis, Indiana, USA. After initial generations of random mating, the selection procedure then employed 8 separate lines, 4 in which breeders were chosen randomly within each line (control or C lines) and 4 in which the highest-running males and females from each family were used as breeders (preventing sib-matings, selected or S lines). At generation 31, 80 breeding pairs, repre- senting all 8 lines, were shipped by air to the animal facility of the Biological Center in Haren, and a breeding colony was started. Male offspring from all 8 lines were used in the experiment described below. After weaning, mice were housed with their littermates until the age of 5 months when all animals were individually housed with or without a running wheel for the rest of their lives (Macrolon Type II, UNO Roestvaststaal BV, Zevenaar, NL; adapted to fit in a plastic running wheel with a 7 cm radius and 1 cm spacing between bars). Wood shavings were used as bedding material and all animals got a wooden stick. The animals had ad libitum food (Standard lab chow 2181, Hopefarms B.V., Woerden, NL) and water and were under a 12:12 light-dark cycle (lights on at 8:00). All procedures concerning animal care and treatment were in accordance with the regulations of the ethical committee for the use of experimental animals of the University of Groningen (DEC 2777(-1)).

Experimental procedures Three experimental groups were created; group 1) Control mice housed with a running wheel (C+), group 2) Selected mice with a running wheel (S+), and group 3) 60 Chapter 4

Selected mice without a running wheel (S-). Logistical constraints precluded inclu- sion of a fourth group, i.e., C- mice. For both C and S groups, mice from all four lines were represented, although not in equal proportions. Animals of two different ages were used, mice in the 10 month age group were 312±7 days old (mean±SD) and mice in the 18 month group were 559±10 days old. Every mouse was weighed once per month and left undisturbed except for the weekly cage cleaning. In the week prior to killing food intake was monitored over 3 consecutive days. In addition, wheel-running activity was assessed in the S+ and C+ groups over two weeks prior to sacrifice with a PC-based event recording system (ERS). Data on wheel- running activity was not available for all animals due to computer problems. At different ages (10 and 18 months) 7–8 mice per cohort were briefly anaesthesized with CO2 then killed by decapitation in the middle of the light phase. Mice were not fasted prior to blood sampling and all mice were asleep when they were taken for blood sampling. Trunk blood was collected in pre-chilled tubes with anticoagulant (EDTA) within 90 seconds from initial disturbance to the final drop of blood. Samples were spun down at 2600 g for 15 minutes at 4°C. Plasma was collected and stored at –80°C until later analysis for hormone levels. Corticosterone, leptin, and adiponectin levels were determined in duplicate with commercial radioimmunoassay kits (Linco Research, Nucli lab, The Netherlands). After blood collection animals were dissected and fresh mass of different organs (heart, liver, kidney, lung, stomach, intestines, skin) and the remainder of the carcass were weighed to the nearest 0.0001 g. The fat content of all animals was determined by drying (ISO 6496-1983(E)) all separate tissues at 103°C followed by fat extraction using petroleum ether (Boom BV, Meppel, Netherlands) and subsequent drying.

Data analysis To test for effects of treatment and/or age we applied ANCOVA models in the MIXED procedure in SAS for Windows (version 9.1). Group, age, and the group x age interaction were fixed effects. The main interest of this study was to determine effects of selective breeding and effects of the presence of a running wheel; therefore, a priori, we tested for differences between group 1 and group 2 (C+ versus S+), and between group 2 and group 3 (S+ versus S-), by adding contrasts to the model. Because we did not maintain a C- group and because sample sizes per group x age combination were relatively small (see Table 4.1), we did not include line as a random effect nested within linetype. Covariates were included as appropriate. For instance, leptin and adiponectin are produced by fat cells and have been shown to correlate with total fat content. Therefore, fat content was added into the model as a covariate for both hormone levels. Data were log10-transformed as necessary to improve normality and linearity of relations with covariates. Based on previous findings, one-tailed tests could be used for some traits; however, for simplicity two- tailed p-values are given for all variables. The significance level of p≤ 0.05 was used. Pearson correlations were used to explore relations between traits of interest, con- sidering each group separately. Adiponectin in mice bred for high activity 61

RESULTS

Body composition, food intake, and wheel-running activity Table 4.1 shows the main characteristics of the three experimental groups. As expected from previous studies, selected mice with a running wheel (S+) tended to have lower body mass (F1,40=17.4, p<0.001) and higher food intake than control mice with wheels (C+) (not significant: F1,40=3.72, 2-tailed p=0.061). Food intake also did not differ significantly between S+ and S- mice, but body mass of S+ mice was lower than S- mice (F1,40=5.4, p=0.025). Adding body mass to the model as a covariate did not alter these effects for food intake. Absolute fat content was approximately 50% lower in S+ mice compared to C+ and S- mice (F1,40=12.2, p=0.001 and F1,40=11.3, p=0.002, respectively) and absolute dry lean mass was higher in controls than both selected groups (F1,40=12.1, p<0.001). Both fat and dry lean mass strongly correlated with total wet body mass (see Figure 4.1) and therefore body mass was added to ANOVA models as a covariate. This analysis showed that fat mass was significantly lower in S+ mice compared to S- mice (F1,39=6.1, p=0.02) and that age did not affect fat mass. Dry lean mass did not differ significantly anymore between the groups or with age once wet body mass was included as a covariate.

Table 4.1. Main characteristics of mice from control lines housed with a running wheel (C+) and of mice from selected lines housed either with (S+) or without a running wheel (S-).

Age C+ S+ S-

Body mass (g) 10 42.5a (1.8) 36.7 (2.1) 39.6b (2.4) 18 44.4 (2.8) 33.4 (1.4) 39.5 (1.6)

Food intake (g d-1)104.3 (0.5) 6.8 (1.5) 5.3 (0.7) 18 6.2 (0.3) 6.4 (0.6) 5.9 (0.6)

Fat mass (g) 10 8.8a (1.2) 5.3 (1.9) 10.7b (1.7) 18 10.5 (2.3) 4.7 (0.7) 8.9 (1.7)

Dry lean mass (g) 10 9.4a (0.2) 8.3 (0.2) 8.3 (0.3) 18 9.7 (0.3) 8.1 (0.4) 8.8 (0.2)

Wheel-running activity (km d-1) 10c 14.0 (2.2) 18.2 (4.5) - - 18 7.4 (3.1) 9.4 (1.3) - -

Values are simple means (SEM). Two-way ANOVA was used to test for differences between groups and with age (months). Data were log-transformed as necessary to improve normality. a Indicates a significant difference between C+ and S+ mice (p<0.05), b indicates a significant difference between S+ and S- mice (p<0.05), c indicates a significant effect of age (p<0.05). Sample size was 8 per group, but in the S+ and S- groups it was 7 at 18 months. 62 Chapter 4

25 11

20 10

9 15

8 10 fat mass (g) drylean mass (g) 7 5

6 A B 0 30 40 50 60 30 40 50 60 body mass (g) body mass (g)

Figure 4.1. Relation between body mass, dry lean mass (A) and fat mass (B) in C+ (white circles), S+ (black circles) and S- (grey circles) mice.

Wheel-running activity assessed over the two weeks prior to sacrifice was calculated in km per day. As expected, wheel-running activity was higher in the selected mice and decreased with age. On average selected mice ran approximately 30% more than controls at both ages and wheel-running activity decreased approximately 50% in both groups between 10 and 18 months. The group effect on wheel-running activity was not significant, but the age effect was (F1,21=4.6, p=0.043).

Hormone concentrations Figure 4.2 shows leptin, adiponectin, and corticosterone levels in plasma for the different experimental groups and Table 4.2 gives an overview of the statistical analysis. Leptin levels were slightly lower in S+ mice compared with C+ and S- mice. This effect of group was statistically significant when comparing S+ with S- mice. No effect of age on leptin levels was shown. When total fat mass was added to the model as a covariate, no significant effects of age or group remained and total body fat was a significant predictors of leptin levels in the model. Adiponectin levels were increased by approximately 60% in both groups of selected mice compared with control mice at both ages (10 and 18 months) and decreased significantly with age. This group effect remained when fat content was added to the model as a covariate. Fat content was not a significant predictor of adiponectin levels in the model. Basal corticosterone levels did not significantly differ between control and selected mice with or without a wheel (See Figure 4.2 and Table 4.2) and did not change with age.

Correlations As shown in Table 4.3, body mass and fat content were negative predictors of wheel-running activity in control mice. In selected mice, a similar trend was visible Adiponectin in mice bred for high activity 63

) 8 -1 6

leptin (ng ml 2

0 24 C+ mice )

-1 S+ mice 20 S– mice 16 12 8 4 Figure 4.2. Leptin, adiponectin, and corti- adiponectin (mg ml costerone concentrations in C+ (white bars), 0 120 S+ (dark grey bars) and S- (light grey bars) ) mice. Values represent simple means ± SEM. -1 Sample size is 8 in C+ mice at all ages. In S+ 80 and S- mice the sample size was 8 and 7 at 10 and 18 months respectively. 40

corticosterone (ng ml 0 10 18 age (month)

Table 4.2. Results of two-way nested ANCOVA of leptin, adiponectin, and corticosterone plasma levels in control (C) and selected mice (S) housed with (+) or without (-) a running wheel.

ppp p Variable N Age C+ vs S+ S+ vs S- Covariate Covariate

Leptin (ng ml-1)450.346 0.091 0.029 none Leptin (ng ml-1)450.154 0.154 0.641 FAT 0.001

Adiponectin (ng ml-1)46 0.035 0.001 0.673 none Adiponectin (ng ml-1)46 0.037 0.003 0.939 FAT 0.534

Corticosterone (ng ml-1) 46 0.309 0.203 0.805 none

Tw o -way ANOVA were performed with age, group, and groupxage as fixed factors. A priori we tested for differences between C+ and S+ or S+ and S- mice. All data were log-transformed to improve normality. P-values of age effects and group (C+ vs. S+ and S+ vs. S-) are given in the table. No significant interaction effects were found and therefore p-values for interaction effects (agexgroup) are not shown. Fat mass was added into the model as a covariate for leptin and adiponectin. Total sample size (N) and degrees of freedom (d.f.) are given for all groups. Sample size was 8 per group, except for S+ and S- group at 18 months where n was 7. One leptin plasma sample gave a leptin concentration of 0 ng ml-1, so this sample was not used in the analyses (S+ group at 10 months). 64 Chapter 4

Table 4.3. Pearson correlations between body composition, wheel-running activity, food intake, and hormones

Body Mass Dry Lean Mass Fat Mass Running Adiponectin Leptin C+ Wheel-running activity –0.69** –0.68** –0.60* Adiponectin –0.07 –0.14 –0.04 0.28 Leptin 0.70** 0.56* 0.81** –0.28 0.04 Corticosterone 0.15 0.25 –0.09 –0.21 –0.52* –0.08 S+ Wheel–running activity –0.46 –0.06 –0.61# Adiponectin –0.18 –0.20 –0.48# 0.43 Leptin 0.53 0.01 0.79** –0.40 –0.20 Corticosterone 0.22 –0.03 0.17 –0.40 –0.15 0.01 S– Adiponectin 0.27 0.06 0.38 Leptin 0.83** 0.61* 0.85** 0.10 Corticosterone 0.10 –0.02 0.25 –0.07 0.03

* denotes p<0.05; ** denotes p<0.01, # denotes p>0.05 and p<0.1 (all unadjusted for multiple comparisons). Food intake did not correlate significantly with any of the other traits and results are not shown in the table.

but probably due to the small sample size (n=9), the correlations were not significant. Overall, lighter and leaner animals ran more than heavier and fatter animals. Food intake did not correlate with any of the measures of body composition. Plasma leptin levels positively correlated with fat in all groups and correlated with dry lean mass and body mass in C+ and S- mice, but not S- mice. Leptin did not correlate with food intake or wheel-running activity. Plasma adiponectin and corticosterone did not correlate with any of the measures of body composition nor with wheel- running activity and food intake in any of the groups. No significant correlations between the different hormones measured were found in selected mice, but in control mice corticosterone significantly correlated with adiponectin. Figure 4.3 shows the relationships between hormones and two measures of body composition, total fat, and dry lean mass.

DISCUSSION

Mice selectively bred for high wheel-running activity (S) generally have decreased body mass and fat content compared to control (C) mice (Dumke et al., 2001; Swallow et al., 2001). We explored whether these differences are associated with alterations in plasma levels of leptin, adiponectin, and corticosterone, and whether they depend on age, the actual levels of physical activity or other traits. Consistent with our expectations was the observation that wheel-running activity of mice was Adiponectin in mice bred for high activity 65

1.2 ) -1 0.8

0.4

0.0 log leptin (ng ml C+ mice S+ mice -0.4 S– mice

1.8 ) -1

1.4

1.0 log adiponectin (mg ml 0.6

2.6 ) -1

2.2

1.8

1.4

log corticosterone (ng ml 1.0 0.0 0.4 0.8 1.2 1.6 0.85 0.90 0.951.00 1.05 log fat content (g) log dry lean mass (g)

Figure 4.3. Correlations between hormones (leptin, adiponectin, and corticosterone) and body composition (fat mass and dry lean mass). Linear regressions were performed to examine significant relations. White dots represent C+ mice, black dots represent S+ mice and grey dots represent S- mice. Regression lines are shown in graphs when relations were significant (dotted line for C+, solid line for S+, dashed line for S-). a negative predictor of body fat content, and fat content in turn correlated positively with circulating levels of leptin. The levels of leptin were better predicted by fat content than by physical activity, which suggests that the leptin-reducing effects of physical activity may be mediated primarily through an effect on body fat content. In contrast to previous studies on these mice, at younger ages (Koteja et al., 1999; Swallow et al., 2001), we did not find a positive relationship between wheel-running activity and food intake in control or activity-selected mice. Food intake was 66 Chapter 4

increased in S+ mice compared to C+ mice in the present study, and sample sizes may have been too small to show effects of wheel running on food intake at the individual level. Leptin levels were found to be significantly decreased in S+ mice compared to S- mice (not corrected for fat mass), but not compared to C+ mice. The latter is at odds with an unpublished observations of decreased plasma leptin levels in young female S mice compared to C mice, even after correction for fat mass (Girard I., Rezende, E. L., and Garland, T., Jr., unpublished observation). The difference in outcome between this study and the one mentioned above may reflect a partial sex- dependent difference in the regulation of leptin levels. Indeed, women are known to have higher leptin levels than men with a similar fat content (Havel, 2001; Hickey et al., 1996). Leptin levels are known to decrease in voluntarily exercising rodents compared to sedentary controls (rats; (Novelli et al., 2004) and hamsters;(Coutinho et al., 2006)) and this was also shown in the present study when comparing S+ and S- mice. Obviously differences in activity were largest between the sedentary housed S- mice and the S+ mice than between the S+ and C+ mice and this might explain why leptin levels were not decreased in the S+ mice compared to C+ mice. The most important finding in the present study was the observation that plasma adiponectin levels were significantly increased in S mice compared to the levels found in their random-bred controls. This increase was found in S mice irrespective of the availability of running wheels and occurred in all of the separate selection lines (results not shown), which suggests that this increase is a genetically co-segregated trait with selection for increased wheel-running activity, instead of being mediated via increased physical activity per se. At present, we do not know whether this increased adiponectin is of high or low molecular weight (active or inactive form; (Fruebis et al., 2001)), or whether it is a cause of altered release or clearance. Future work is necessary to study this and to determine whether the increased adiponectin levels in activity-selected mice are associated with increased insulin sensitiviy. Also, if the increased levels of adiponectin are responsible for their low fat content, does this or the high adiponectin levels protect them against diet- induced obesity? There a number of reasons to suggest that increased circulating adiponectin levels might contribute to the different phenotypes seen in S mice compared to C mice. For example, Fruebis et al (Fruebis et al., 2001) found that chronic administration of gAcrp30 (i.e., adiponectin) caused weight loss in mice despite the fact that food consumption was unaffected. This is a phenotype which appears homologous to the one found in selected animals in this and previous studies (Swallow et al., 1999; Swallow et al., 2001). Fruebis et al. attributed the effect of adiponectin on body mass to increased fat oxidation, specifically in liver and muscle (Fruebis et al., 2001), and this was confirmed in subsequent studies (Berg et al., 2002; Bruce et al., 2005; Yamauchi et al., 2002). Although not directly assessed in the present study, we recently observed a decrease in the respiratory quotient (RQ) in S mice compared to C mice measured over a period of 24 hours in non-fasted animals (see Adiponectin in mice bred for high activity 67

Chapter 5), which indeed indicates higher levels of fat oxidation in selected mice. It might be speculated that an increased capacity to down-grade lipids in muscular tissue contributes the increased physical activity displayed by activity-selected mice. The effects of adiponectin on fat oxidation are believed to arise through stimulation of AMP-activated protein kinase (AMPK) (Berg et al., 2001; Yamauchi et al., 2002). Zhang et al. have shown increased levels of phosphorylated AMPK in aorta of male activity-selected mice compared with controls (Zhang et al., 2006), which is consistent with the observed elevated levels of adiponectin in this study. At present we do not know whether there is a positive relationship between adiponectin and AMPK at the individual level in the activity-selected mice. Previous studies have shown an increase in resting corticosterone levels in 2- month old male and female S mice (Girard and Garland, Jr., 2002; Malisch et al., 2006), without a change in the absolute corticosterone concentration in response to 40 min. restraint stress (Malisch et al., 2006). In the present study we investigated whether these differences were present in older animals. No statistically significant differences were found in corticosterone levels between C and S males at 10 or 18 months of age. Glucocorticoids increase in response to exercise (Droste et al., 2003) and can affect leptin and adiponectin concentrations (Droste et al., 2003; Fallo et al., 2004; Van Dijk et al., 1997). Nevertheless, we did not show any relations between wheel-running activity, plasma adiponectin or leptin and corticosterone levels in C+ and S- mice. In S+ mice, however, a negative relationship between corticosterone and adiponectin was found. In conclusion, mice from lines that have been selectively bred for high wheel- running activity over many generations show several endocrine alterations. Male activity-selected animals have higher adiponectin levels than their randomly-bred controls, independent of the presence of a running wheel. Leptin and corticosterone levels were unchanged in selected mice of 10 months and older, although a decrease in leptin and an increase in corticosterone were shown previously in younger females. These effects might be driven mainly by differences in wheel-running activity between the groups at young ages. Selected mice have been shown to be leaner than control mice (Dumke et al., 2001; Swallow et al., 2001) and together with the observed increase in adiponectin concentration this might render the mice less prone to develop insulin resistance and other aspects of the metabolic syndrome. This would make them a suitable model to study whether “physical activity genes” exist, and whether these could influence mechanisms underlying proneness for diet-induced obesity and related diseases.

Acknowledgements The authors thank Gerard Overkamp for expert technical assistance and Berber de Jong for help with the experimental procedures. Jan Bruggink is thanked for performing analytical procedures. This work was supported primarily by a Career Development Grant from the Dutch Diabetes Association (to GvDijk). Additional funding was provided by grants from the U.S. National Science Foundation to T.G., most recently IOB-0543429. 68 Chapter 4

BOX 4.1: Leptin, adiponectin and corticosterone in cold-exposed mice

Hormone levels of leptin, adiponectin and corticosterone were determined in plasma of male C57BL mice exposed to 10°C (CC) or 22°C throughout life (WW), and in mice exposed to 10°C till the age of 15 months and at 22°C afterwards (CW; for a detailed description of the experimental protocol see Chapter 8). Trunk blood was collected in tubes with anticoagulant (EDTA) at 3, 11, 19 and 27 months of age, centrifuged at 4°C for 15 min at 2600 g. Then plasma was collected and stored at –80°C for later hormone analyses (with RIA, Linco kits). Plasma samples at 27 months of age were not analyzed for leptin and adiponectin levels. Corticosterone measurements at 19 months were left out of statistical analysis: results were unreliable because the cold mice were probably disturbed prior to the measurements resulting in very high corticosterone values at this age in these mice. Figure 4.4 shows the results. Both, levels of leptin and adiponectin were decreased in mice exposed to the CC compared to WW mice (Two-way ANOVA with group and age as fixed factors (excluding CW mice); Effect of group: Leptin; F1,48=41.8, p<0.001, Adiponectin; F1,47=7.8, p<0.01). Age did not affect adiponectin levels significantly but leptin levels were significantly affected by age (Effect of age: Leptin; F2,48=21.4, p<0.001) and there was a significant interaction between group and age (GroupxAge interaction: Leptin; F2,48=10.1, p<0.001). Both hormones, leptin and adiponectin are produced by adipose tissue and specifically leptin is known to correlate strongly with fat content. Therefore, we also applied ANCOVA models with fat as a covariate. In these models there was still a significant effect of group on leptin levels (F1,47=14.9, p<0.001), but adiponectin was no longer significantly different between groups. In both cases fat content was significantly correlated to the hormone levels (p<0.001 and p=0.01 respectively for leptin and adiponectin). There was no longer a significant effect of age on leptin levels, but the group x age interaction remained significant (F2,47=5.5, p<0.01), indicating that leptin levels increased more with age in the WW mice than in the CC mice. We also tested for differences in leptin and adiponectin levels between CW, CC and WW mice using one-way ANOVA with a factor group. No significant effect of group on adiponectin level was found, but leptin did differ significantly between groups (p<0.05), and the CW and WW mice had significantly higher levels of leptin relative to CC mice (see figure 4.4). Corticosteron levels did not differ between groups and were not affected by age. These results agree with several other studies (Puerta et al., 2002; Imai et al., 2006), and may indicate a role for leptin and adiponectin in the extensive metabolic adaptations to cold. Adiponectin in mice bred for high activity 69

24 WW mice CC mice 20 CW mice ) -1 16

8 leptin (ng ml 4

0 16 ) -1 12

4 adiponectin (mg ml 0 200 ) -1 150

120

40 corticosterone (ng ml 0 3 11 19 27 age (month)

Figure 4.4. Leptin, adiponectin, and corticosterone concentrations in WW (white bars), CC (dark grey bars) and CW (light grey bars) mice. Values represent simple means ± SEM. Sample size is 5-8 in all groups.

Chapter5

Responses in energy balance to high-fat feeding in mice selectively-bred for high wheel-running activity

Lobke M. Vaanholt, Izabella Jonas, Mark Doornbos, Kristin A. Schubert, Csaba Nyakas, Theodore Garland Jr., G. Henk Visser, Gertjan van Dijk

Abstract Obesity is becoming an increasingly prevalent health problem among individuals in affluent societies, and can lead to life-threatening conditions like cardiovascular disease and type 2 diabetes mellitus. Increased dietary fat intake in combination with a sedentary existence are precipitating factors for the development of obesity and the associated metabolic syndrome. Whereas most animal studies have investigated effects of exercise using “forced” activity protocols, in the present study we investigated the role of “unforced” physical activity in the interaction between dietary fat on regulation of energy balance and fuel homeostasis. This was done using selectively-bred mice which display a high level of voluntary wheel-running activity relative to control lines. Control and selected males and females were fed with a standard lab chow or an iso-caloric 60% fat diet. All fat-fed mice rapidly increased in body mass and decreased food intake, with the exception of the selected females. Selected females did not develop diet-induced obesity and even increased their food intake on the fat diet. In addition, they had improved glucose tolerance on the standard diet compared to the other groups, but this difference was lost upon feeding the fat diet. Adiponectin was increased in selected mice (specifically in males) compared to control mice on the fat diet. Leptin and insulin levels increased on the fat diet in control females and both groups of males, but not in selected females. In conclusion, selected females fed a high-fat diet did not develop diet-induced obesity as the other groups did, but they did become slighty glucose intolerant relative to selected mice on standard chow. This effect could not be explained by differences in adiponectin levels and may be related to a failure to upregulate leptin. (Female) mice selected for high spontaneous wheel- running activity are an attractive model to further investigate mechanisms involved in the metabolic syndrome and associated type 2 diabetes mellitus. 72 Chapter 5

INTRODUCTION

Obesity is becoming an increasingly prevalent health problem among individuals in affluent societies, because it is often associated with metabolic derangements such as impaired glucose tolerance, insulin resistance, high blood pressure, dyslipidemia, and abdominal obesity. When these metabolic abnormalities are displayed in concert (often referred to as the “metabolic syndrome”), they have a high risk of developing into life-threatening conditions such as cardiovascular disease and diabetes mellitus type 2 (for review see (Carroll and Dudfield, 2004; Moller and Kaufman, 2005)). While the mechanisms underlying this epidemic are largely unknown, there is consensus that increased dietary fat intake in combination with a sedentary existence are precipitating factors (WHO, Carroll and Dudfield, 2004). A major part of the current knowledge on the aetiology of obesity and the metabolic syndrome has come from studies in which rodents are subjected to a high-fat diet (Storlien et al., 1986; Surwit et al., 1995; Ahren and Scheurink, 1998; Lin et al., 2000; Winzell and Ahren, 2004). In particular the use of selectively-bred and genetically-engineered rodents which display diet-induced obesity or obesity-resistance has been insightful to unravel the underlying mechanisms (for reviews see Tschop and Heiman, 2001; Carroll et al., 2004). Few animal studies, however, have address- ed whether and how physical activity could be involved in preventing dietary fat- induced obesity and associated metabolic derangements. In part, this is due to the methodological problems that arise when animals are “forced” to be active. For example, increased physical activity induced by forced treadmill running, water immersion etc. (Pellizzon et al., 2002; Jen et al., 2003) are perceived as psychologi- cal stress (Chennaoui et al., 2002). A way around these potential pitfalls is to investigate interactions between “unforced” physical activity, dietary fat intakes and energy balance in mice selectively bred for high voluntary wheel running relative to control lines (for selection procedure see (Swallow et al. 1998). Mice selected on increased wheel-running activity are leaner than their respective controls, and have relatively low circulating leptin levels (Girard et al. unpublished results). Even without the presence of a running wheel, these selected animals display increased locomotor activity (Vaanholt et al. 2006), and therefore may represent a relevant model for the study of increased spontaneous activity in the development of high-fat feeding induced obesity and metabolic derangements. In the present study female and male mice selected for high wheel-running activity and their random-bred controls were either exposed to a high-fat diet or remained feeding chow to test the hypothesis that selected mice are resistant to developing obesity and other indicators of the metabolic syndrome, such as impaired glucose tolerance and reduced lipid oxidation. Food intake and body weights were assessed, and total and resting metabolic rates were determined Glucose tolerance was assessed by repeated blood sampling after intra-peritoneal glucose injection. Finally, plasma levels of several metabolic hormones were determined and body composition analysis was performed. The results indeed point to a High-fat feeding in mice bred for high activity 73

major contribution of spontaneous activity in the prevention of high-fat induced obesity.

MATERIAL AND METHODS

Animals & housing Eighty breeding pairs of Hsd:ICR mice (Mus domesticus) selected on high wheel-running activity for 31 generations and their random bred controls were obtained from T. Garland Jr, Riverside, CA (For a detailed description of the selection procedure see (Swallow et al., 1998). From these founder mice, separate breeding lines were continued at our facilities in Haren without further selection for wheel-running activity. In the original selection protocol eight lines of mice were created (4 selected and 4 control). We used mice from one control line and one selection line in the present study (lab designated lines 2 and 7 respectively). Forty eight mice (25 male and 23 female) at the age of 5 months were used of the fourth generation of offspring (without selection), of which half belonged to the control line (C) and the other half tot the selected line (S). All mice were individually housed in standard cages (Macrolon Type II, UNO Roestvaststaal BV, Zevenaar, NL) in the same room with an ambient temperature of 22±1°C. At the start of the experiment all animals had ad libitum food (Standard lab chow RMH-B 2181, HopeFarms BV, Woerden, NL) and water; they were on a 12:12 light-dark cycle (lights on at 8:00). Wood shavings and EnviroDry® were used as bedding material. At the start of the experiments, they were 20 weeks of age. Food intake and body mass were measured daily (except for the weekends). After 4 weeks, half of the mice were put on a high-fat diet containing 60% of weight as fat (see Table 5.1 for the composition of the diets), which was equicaloric

Table 5.1. Composition of the diets.

Standard chow diet (RMH-B) - HC High fat diet HF Content (g kg-1) Energy (%) Content (g kg-1) Energy (%)

Protein 228 23 200 20.4 Crude (lab chow) 228 69 Added casein 131 Fat (saturated fat) 55 (0.7) 14 260 (94.3) 60.2 Lab chow fat 55 17 Added corn oil 74 Added beef tallow 169 Carbohydrates 625 63 190 19.4 Lab chow polysaccharides 600 182 Simple sugars 25 8

Energy 16.1 kJ g-1 16.3 kJ g-1 74 Chapter 5

to the standard lab chow. Food intake and body mass were measured daily until the mice had been on the fat diet for 9 weeks.

Glucose tolerance test In weeks 10 and 11 relative to the start of high fat feeding, all animals underwent a glucose tolerance test. Specifically, animals were food deprived for 6 hours (starting at lights on), and injected intra-peritoneally with 0.01 ml per gram body mass of a 10% glucose solution at noon (±1 hr). Drops of blood were collected via tail clip- ping just before glucose injection and at 20, 40 and 60 minutes following glucose injection. Blood glucose levels were determined immediately using a hand-held glucose analyzer (OneTouch Ultra, LifeScan). The glucose tolerance test was repeated 4 weeks later in the selected female animals. This time the mice received a dose that was comparable to the average dose that the control mice received before (0.35 ml of 20% glucose solution), and an additional blood sample was taken 120 min after the injection.

Respirometry In week 11-13 animals were moved into respirometric chambers to determine oxy- . -1 . -1 gen consumption (V O2, l h ) and carbon dioxide production (V CO2, l h ) by indirect calorimetry for 48 hours. Eight animals could be measured simultaneously. Oxygen and carbon dioxide concentration of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck) from each chamber was measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and cardon dioxide by an infrared gas analyzer (Servomex 1440). The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic chambers. Flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks). Ambient temperature in the chamber and cage, as well as activity (with passive infra-red detectors) were measured simultaneously. Samples were collected every 10 minutes for each animal and automatically stored on a computer. To reduce novel cage stress, the respirometric chambers (45x25x30 cm) were adapted to accommodate the home cage of the animal. Animals therefore did not need to be handled and stayed in their home cage during the entire measurements. Animals were measured at an ambient temperature of 22°C and food (standard chow or fat) and water were provided ad libitum. Heat production (HP, kJ h-1) was calculated using the following equation: HP= . . (16.18 x V O2) + (5.02 x V CO2 ) (Romijn and Lokhorst, 1961). Resting metabolic rate (RMR, kJ h-1) was defined as the lowest value of heat production calculated as the running mean over half an hour and was calculated for the first and second day in the respirometer separately. Maximal heat production was also calculated as the running mean over half an hour. In addition, the average heat production (daily -1 . . energy expenditure: DEE, kJ d ), respiratory quotient (RQ= V CO2/V O2) and PIR (passive infrared) activity were calculated for both consecutive days. High-fat feeding in mice bred for high activity 75

Wheel-running activity After the respirometry measurements, all mice were put in cages with a plastic running wheel with a 7 cm radius for a two-week period. Animals had never been exposed to running wheels prior to this time. The running wheels were attached to a PC-based event recording system, and total wheel-running activity was logged in 2 minute bins over a three week period.

Body composition & Metabolic hormones Following wheel-running activity measurements, animals were again housed sedentary for an additional month, and then animals were anaesthesised with CO2 followed by decapitation. Trunk blood was collected in tubes with EDTA as anti-coagulant, and organs and specific fat pads (i.e. retroperitoneal fat, gonadal fat, subcutaneous fat) were dissected out and weighed to the nearest 0.0001 g. All tissues were stored at –20°C until further analyses. Blood samples were centrifuged at 2600 g for 15 min at 4ºC, plasma was then collected and stored at –80°C for later analysis with RIA (Linco). Dry and dry lean organ masses were determined by drying organs to constant weight at 103°C (ISO 6496-1983(E)) followed by fat extraction with petroleum ether (Boom BV, Meppel, NL) in a soxhlet apparatus.

Data analyses Because males and females are known to differ with respect to their regulation of metabolic hormones, first we tested for effects of sex on all variables with an independent samples t-test and then tested the effects of diet and group were tested separately for males and females. To test for effects of diet and group we used GLM models in SPSS (version 12). Group, diet and group x diet were added as fixed factors. Where appropriate, covariates (i.e. body mass) were used in the models. Post- hoc independent t-tests were used to compare groups and diets. For several variables repeated measures ANOVA was applied . Group, time (repeated factor) and group x time were used as fixed factors in these models. The significance level was set at p≤ 0.05 and all tests were two-tailed.

RESULTS

Changes in food intake and body weight Body weights were significantly higher in males than females (t-test, p<0.001), but food intake was similar. Selected females had significantly lower body weight and increased food intake compared to controls when feeding chow. (ANOVA F3,22=14.3, p<0.001 and F3,22=7.7, p=0.01 respectively). In males of selected and control lines, there were no differences in body weight when feeding chow, but food intake was significantly increased in selected males compared to control males (F3,24=13.9, p<0.01). 76 Chapter 5

females males

body mass (g) 28 CTRL - normal diet CTRL - fat diet SEL - normal diet 24 SEL - fat diet

6 )

-1 5

2 food intake (g d 1

0 020405 1015 25 30 35 45 020405 1015 25 30 35 45 time after start diet (d) time after start diet (d)

Figure 5.1. Development of body mass (top graphs) and food intake (bottom graphs) in male and female control (circles) or selected mice (triangles). Running means over three consecutive days are shown. At day 0 animals were put on a fat diet or remained on standard lab chow.

Figure 5.1 shows the development of body weight and food intake for male and female control and selected mice after the mice had been put on a high-fat diet. In animals on the high-fat diet, we tested for differences in body weight or food intake between the weeks prior to and after the food manipulation using a repeated measures ANOVA. In female mice, the fat diet significantly influenced body weight (F1,10=11.36, p<0.01), and there was a significant interaction between group and diet (F1,10=7.5, p<0.05), indicating that control and selected mice responded differently to the fat diet. As seen in Figure 5.1, female control mice increased body weigh on the fat diet, whereas selected females did not change their body weight. Food intake was not significantly affected by diet in females overall. In males, body weight significantly increased (F1,11=10.2, p<0.01) and food intake significantly decreased (F1,11=17.1, p<0.01) on the fat diet, but there was no group x diet interaction. Control and selected males thus responded to the fat diet with similar changes in body weight and food intake. Closer inspection of the day-by-day changes in body weight and the associated food intakes revealed differences in food efficiency (FE= weight gained/gross energy consumed) between the control and selected females (Figure 5.2). Comparisons High-fat feeding in mice bred for high activity 77

females males

7 a CTRL - normal diet CTRL - fat diet a ac a 6 SEL - normal diet SEL - fat diet ab

) 5 ab a -1

food intake (g d 2

0 <–1 –1 to –0.25–0.25 to 0.25 to 1 <–1 <–1 –1 to –0.25–0.25 to 0.25 to 1 <–1 0.25 0.25 weight change (g d-1) weight change (g d-1)

Figure 5.2. Changes in daily food intake (g d-1) underlying body mass changes (g d-1) in control and selected mice fed standard or fat diet. Average food intake during the first month on the diet was calculated per animal for various categories of weight changes, by taken the average of the food intakes measured on days where body mass changes were smaller than –1.0, between –1.0 and –0.25, between –0.25 and 0.25 (constant body mass), between 0.25 and 1.0 or larger than 1.0 g (on two consecutive days). Two-way ANOVA was used to look at effects of group, diet and group x diet. a represents statistical differences between control and selected mice, b represents significant effects of diet, and c represents a significant interaction effect (group x diet) (p<0.05). between diets can easily be made without corrections since the high fat and the chow diet used in the present study are equicaloric. At several categories of body weight changes from one day to the next, control females had a lower food intake on the fat diet than on the chow diet. Selected females, however, increased food intake on the fat diet (interaction effect: F3,19=4.4, p<0.05). Both groups of males showed a decrease in food intake on the fat diet compared to standard chow (F3,18=6.6, p<0.05).

Glucose tolerance Results of the intraperitoneal glucose injection on blood glucose levels are shown in Figure 5.3. At all time points males had significantly higher glucose levels than females (t-test, p<0.01). In females, repeated measures ANOVA of the model in total revealed very strong trends of diet (F1,18=5.4, p=0.051), group (F1,18=3.6, p=0.075), and groupxdiet (F1,18=4.3, p=0.056) on glucose level. Post-hoc analysis revealed that at t40 selected females on a standard chow diet had lower blood glucose levels than control mice (p=0.073, p=0.005 and p=0.052 at T20, T40 and T60 respectively), and also compared to selected mice on a fat diet (p=0.071, p=0.008 and p=0.053 at T20, T40 and T60 respectively). No significant differences in glucose levels were found between control mice and selected mice on the high-fat diet (posthoc t-test: p>0.1). Glucose tolerance was analysed by comparing group and 78 Chapter 5

females males 16 ) -1 12

4 CTRL - normal diet CTRL - fat diet plasma glucose (mg ml SEL - normal diet SEL - fat diet 0 0204060 0204060 time after injection (min) time after injection (min)

Figure 5.3. IP glucose tolerance test results. Blood glucose levels measured 20, 40 and 60 minutes after an injection with a 20% glucose solution in control (circles) and selected (triangles) mice on a fat diet (black) or standard lab chow (white). Values given are mean±sem.

diet effects on Area-Under-the-Curve of the different glucose levels relative to those observed at baseline. Glucose tolerance decreased strongly in selected mice, (area- under-the-curve increased from 5.1 (s.d. 3.0) to 10.5 (s.d. 6.2) respectively on the fat diet), but this effect was not significant (post-hoc t-test; p=0.089). In controls there was no effect of the diet on glucose tolerance (area-under-the-curve; 8.4 (s.d. 3.1) and 8.4 (s.d. 4.1) respectively). In female selected mice the glucose tolerance test was repeated two weeks later, when the mice were injected with a higher dose of glucose (similar to the average dose control mice received previously). No significant differences were found between this test and the first one (data not shown). In males, no differences in blood glucose were found between the groups on either diet in response to the glucose injection (Two-way repeated measures ANOVA, p>0.05 for group, diet and dietxgroup interaction; see Figure 5.3, right frame). The area-under-the-curve was slightly increased in selected male mice on the fat diet compared to selected males on standard chow (=9.6 (s.d. 6.3) and 12.8 (s.d. 6.3) respectively), but this difference was not significant (p>0.1). Also glucose tolerance did not differ from that of control male mice (area-under-the-curve was 10.9 (s.d. 6.4) and 8.1 (s.d. 7.7) on standard chow and fat diet, respectively).

Energy expenditure All animals were put in a respirometric chamber for 48 hours to determine resting metabolic rate (RMR), daily energy expenditure (DEE), spontaneous activity and RQ. Values were highly repeatable measured on the two consecutive days. Data from day 2 were used in the analysis (results summarized in Table 5.2 and 5.3). In females, there were no effects of group or diet on RMR. However, when body mass was added to the model as a covariate selected females had a significantly higher High-fat feeding in mice bred for high activity 79

Table 5.2. Metabolic rates of mice selected for high wheel-running activity and their random- bred controls on a standard and fat diet.

Control Selection Standard chow Fat chow Standard chow Fat chow

Females: RMR (kJ d-1) 35.8 (4.1) 38.6 (2.2) 36.4 (1.7) 40.2 (1.4) DEE (kJ d-1) 47.8 (4.6) 50.9 (2.5) 59.0 (3.5) 70.5 (3.2) Maximum HP (kJ d-1) 74.3 (9.7) 72.8 (5.1) 84.1 (6.7) 113.8 (6.3) RQ 0.93 (0.01) 0.76 (0.03) 0.87 (0.01) 0.76 (0.02) Activity (number d-1) 1249 (254) 1047 (155) 5845 (1705) 7643 (1211)

Males: RMR (kJ d-1) 38.7 (1.7) 41.9 (1.7) 38.2 (2.0) 46.0 (0.6) DEE (kJ d-1) 51.1 (2.3) 54.1 (1.6) 59.9 (7.1) 67.6 (3.3) Maximum HP (kJ d-1) 72.5 (5.5) 73.5 (1.6) 86.1 (7.4) 103.9 (10.4) RQ 0.90 (0.01) 0.75 (0.01) 0.88 (0.01) 0.75 (0.01) Activity (number d-1) 784 (238) 1109 (147) 1115 (151) 3453 (1681)

Mean and (SEM) are given for metabolic rates, RQ and activity measured on day 2 of the respirometric measurements. N= 6 per group, with the exception of control females on a standard diet (n=4), control males on a standard diet (n=5) and control males on a fat diet (n=7).

RMR than control females. DEE was also significantly increased in selected females compared to controls (with or without body mass as a covariate). In selected females, the fat diet increased DEE; as shown by a significant interaction effect between group and diet. Results for maximal heat production were similar to those for DEE, with higher HP in selected females compared to controls and the highest levels in selected females on a fat diet. Selected females were far more active than controls in the respirometer, which is in agreement with their higher DEE. When feeding the high fat diet, control females slightly reduced spontaneous activity. In contrast, selected females further increased spontaneous activity when feeding the high fat diet to appr. 700% relative to controls. In males, there was no effect of group on RMR, but diet significantly increased RMR in both control and selected males. DEE was increased in selected mice compared to controls, and both control and selected males showed an increase in DEE on a fat diet compared with mice on a standard diet. As in females, maximal HP was higher in selected males compared to controls, but in males there was no interaction effect between group and diet. Also, activity did not differ significantly between control and selected male mice. A respiratory quotient (RQ) around 1 indicates that mice mainly utilize glucose, whereas at RQ values around 0.7 mice are mainly utilizing fat. As expected, RQ was lower in all mice on the high-fat diet. In addition to this, in females (not males) a significant effect of group and an interaction between diet and group was shown. 80 Chapter 5

Table 5.3. Respirometry, results for two-way ANOVA

Group Diet Groupxdiet Covariate Variable n d.f. F p F p F p p

Females RMR (kJ d-1)223,21 0.2 0.684 1.6 0.224 0.0 0.853 22 4,21 9.5 0.007 0.0 0.899 1.0 0.322 BW 0.001 DEE (kJ d-1)223,21 9.0 0.008 3.6 0.074 3.2 0.091 22 4,21 33.9 <0.001 0.8 0.372 10.3 0.005 BW 0.001 Maximum HP (kJ d-1)22 3,21 12.6 0.002 3.9 0.065 4.8 0.042 22 4,21 40.8 <0.001 1.0 0.330 14.1 0.002 BW 0.001 RQ 22 4,21 11.5 0.003 195.4 <0.001 7.3 0.015 Activity (number d-1)224,21 16.2 0.001 0.3 0.574 0.5 0.482

Males RMR (kJ d-1)243,23 1.3 0.373 11.4 0.003 1.0 0.173 24 4,23 2.1 0.161 7.1 0.015 2.3 0.143 BW 0.036 DEE (kJ d-1)243,23 10.4 0.004 4.9 0.039 2.3 0.144 24 4,23 9.8 0.006 6.4 0.020 2.4 0.137 BW 0.233 Maximum HP (kJ d-1)24 3,23 10.2 0.005 1.9 0.186 1.5 0.236 24 4,23 9.5 0.006 2.6 0.120 1.5 0.234 BW 0.340 RQ 24 3,23 0.4 0.553 180.7 <0.001 1.2 0.278 Activity (number d-1)243,23 2.4 0.136 2.4 0.137 1.4 0.255

Data recorded on day 2 of the respirometric measurement were analyzed with a two-way ANOVA with group, diet and groupxdiet (GxD) as fixed factors for females and males separately. Body weight (BW) is known to have a strong influence on metabolic rate and was added into the model as a covariate (Cov.) for these variables. Significant effects are shown in bold (p<0.05).

Post-hoc analyses revealed that selected females on standard chow had lower RQ values than controls on normal chow (t-test, p<0.05), whereas RQ values did not differ between control and selected females on a fat diet. RQ was similar between control and selected males on a fat diet.

Wheel-running activity To test whether animals differed with respect to their wheel-running activity as predicted by their selection history, animals were exposed to running wheels with a 7 cm radius after the respirometry measurements and wheel-running activity was measured for three weeks. Average wheel-running activity was 9.4±1.6 and 10.8 ±1.7 kilometers per day in control and selected mice, respectively. Overall, there were no significant effect of group and diet on wheel-running activity. Initially (during the first week of measurements) wheel-running activity was significantly increased in selected animals compared with controls (t-test, p<0.05), but this difference disappeared later. High-fat feeding in mice bred for high activity 81

Body composition Seperate fat pads were collected and weighed to determine the distribution of fat deposition throughout the body (see Table 5.4, and see Table 5.5 for statistical analysis). In control female mice, but not selected females, the amount of fat increased in all fat pads (except for the fat in the organs) when put on a fat diet. On the fat diet, control females increased their total fat content approximately by 40% whereas total fat content only increased by 5% in selected females. The distribution of fat (% of total fat) over the different fat pads was similar between the control and selected mice on standard chow and also between selected females on standard and fat chow. In control females on a fat diet, however, this distribution changed compared to control females on standard chow. The para-uterine fat pad represented approximately 2% of the total fat content on the standard diet and this increased to 14% on the fat diet. The intramuscular fat content decreased from 61 to 50 % between the standard and fat diet. The other fat pads remained similar. In both control and selected males, there was an increase of approximately 25% in the total amount of fat and in the separate fat pads (except for organ fat; see

Table 5.4. Fat content in mice selected for high wheel-running activity and their random-bred controls on a standard and fat diet.

Control Selected Mass (g) Standard chow Fat chow Standard chow Fat chow

Females Body mass 33.9 (2.0) 43.6 (2.8) 27.3 (2.8) 29.2 (1.1) Dry lean mass 7.2 (0.4) 7.4 (0.3) 6.3 (0.3) 6.2 (0.2) Total fat 21.3 (1.3) 29.4 (2.2) 16.6 (1.7) 17.5 (0.5) Para-uterine fat 0.42 (0.12) 4.03 (1.19) 0.13 (0.08) 0.21 (0.05) Retroperitoneal fat 0.10 (0.02) 0.48 (0.12) 0.04 (0.02) 0.09 (0.01) Subcutaneous fat 3.4 (0.2) 6.0 (0.7) 2.5 (0.6) 2.5 (0.2) Intramuscular fat 12.9 (0.8) 14.6 (0.6) 10.1 (0.9) 10.9 (0.3) Organ fat 4.5 (0.3) 4.3 (0.3) 3.9 (0.2) 3.7 (0.1)

Males Body mass 39.4 (1.8) 43.5 (1.5) 37.1 (1.9) 43.0 (3.2) Dry lean mass 8.2 (0.2) 8.3 (0.2) 7.7 (0.3) 7.9 (0.2) Total fat 28.0 (1.7) 33.9 (1.9) 25.7 (2.0) 33.5 (3.5) Epididimal fat 0.71 (0.11) 1.23 (0.21) 0.66 (0.19) 1.85 (0.40) Retroperitoneal fat 0.22 (0.06) 0.45 (0.14) 0.24 (0.08) 0.76 (0.20) Subcutaneous fat 4.6 (0.5) 6.4 (0.7) 3.5 (0.5) 6.2 (1.1) Intramuscular fat 17.7 (1.0) 21.0 (0.8) 16.5 (1.2) 20.0 (1.9) Organ fat 4.8 (0.2) 4.8 (0.2) 4.8 (0.1) 4.7 (0.2)

Mean and (SEM) are given in the table. Values given for the separate fat pads include the water content. N= 6 per group, with the exception of control and selected females on a standard diet (n=4 and n=5 respectively), control males on a standard diet (n=5) and control males on a fat diet (n=7). 82 Chapter 5

Table 5.5. Body composition, results for two-way ANOVA

Group Diet Group x Diet Covariate Variable N d.f. F p F p F p p

Females Body mass 21 3,20 20.2 <0.001 6.1 0.024 2.7 0.117 Dry lean mass 21 3,20 11.3 0.004 0.1 0.737 0.2 0.640 Total fat 21 4,20 8.4 0.010 8.3 0.011 4.9 0.042 DL 0.026 Para-uterine fat 21 4,20 6.7 0.020 6.8 0.019 6.3 0.023 DL 0.561 Retroperitoneal fat 21 4,20 3.8 0.069 7.4 0.015 4.5 0.048 DL 0.663 Subcutaneous fat 21 4,20 5.8 0.028 5.8 0.029 5.2 0.036 DL 0.189 Intramuscular fat 21 4,20 8.3 0.011 9.3 0.008 0.1 0.711 DL <0.001 Organ fat 21 4,20 0.1 0.749 2.4 0.134 0.4 0.528 DL <0.001

Males Body mass 24 3,23 0.4 0.535 5.1 0.036 0.2 0.662 Dry lean mass 24 3,23 3.8 0.064 0.5 0.469 0.0 0.851 Total fat 24 4,23 1.4 0.239 10.1 0.004 0.1 0.725 DL <0.001 Epididimal fat 24 4,23 6.3 0.021 11.5 0.003 2.1 0.165 DL 0.005 Retroperitoneal fat 24 4,23 4.3 0.050 6.6 0.019 1.0 0.312 DL 0.037 Subcutaneous fat 24 4,23 0.3 0.611 9.2 0.007 0.3 0.602 DL 0.002 Intramuscular fat 24 4,23 0.8 0.377 9.7 0.006 0.1 0.872 DL <0.001 Organ fat 24 4,23 0.7 0.416 0.6 0.453 0.1 0.881 DL 0.03

Data on body composition were analyzed with a two-way ANOVA with group, diet and groupxdiet as fixed factors for females and males separately. Dry lean mass was added as a covariate when testing for differences in fat mass and p-values are given in the table. Significant effects are shown in bold (p≤ 0.05).

Table 5.4 and 5.5) on the fat diet. The distribution of fat over the fat pads was similar between both groups when they were on standard chow, but on the fat diet the selected males stored more fat in the intraperitoneal and epididimal fat pads compared to controls (Table 5.4 and 5.5).

Metabolic hormones Plasma samples collected at decapitation were analysed for the concentration of several metabolic hormones, and the results are shown in Figure 5.4 (i.e., adiponectin, insulin, and leptin) and Figure 5.5 (i.e., T3 and T4). When feeding chow, plasma levels in females were generally higher of adiponectin, and lower of leptin and insulin than in males, but no differences were observed between control and selected mice. On the fat diet, plasma insulin levels increased in control females, an effect not observed in selected females (Two-way ANOVA: Group; F1,16=4.8, p<0.05), diet; F2,16=3.9, p=0.065, group x diet; F2,16=1.6, p>0.1). In males, insulin levels were similar in control and selected mice and slightly increased on the fat diet in both groups. This effect of diet was not significant. In both females and males a significant effect of diet on leptin levels was shown (females: F2,16=6.9, p<0.05; males: High-fat feeding in mice bred for high activity 83

25 chow

) fat -1 20

5 adiponectin (mg ml 0 4 )

-1 3

1 insulin (mg ml

8 )

-1 6

leptin (mg ml 2

0 control selected control selected females males

Figure 5.4. Plasma adiponectin, insulin and leptin levels of control and selected, male (right graph) and female (left graph) mice. Blood samples for determining hormone levels were taken when the animals had been on a fat diet (dark grey bars) or standard diet (light grey bars) for over eight months.

F2,21=8.0, p=0.01). Post-hoc tests showed that in control mice, but not in selected mice, leptin was significantly increased on the fat diet compared to the control diet. In addition, to the effect of diet, in females a significant effect of group (F1,16=6.7, p=0.02) and a significant interaction between group and diet was shown (F2,16=5.3, p<0.05). This indicates that the control mice increased their leptin levels more than the selected females did on the fat diet, which is also apparent in Figure 5.4. Leptin highly correlated with fat content, though, and when total fat content was added to the model as a covariate no significant effects of group, diet or group x diet remained. When feeding the high-fat diet, plasma adiponectin levels were significantly increased in selected males relative to those feeding chow, (Two-way ANOVA: group: F1,19=2.6, p=0.12 and diet; F1,19=5.3, p=0.032, groupxdiet; F2,19=9.8, p=0.006, and post-hoc t-test: p=0.021), and a strong trend 84 Chapter 5

2.0 chow fat 1.6 ) -1 1.2

0.8 T3 (mg ml 0.4

0.0 70 60

) 50 -1 40 30

T4 (mg ml 20 10 0 control selected control selected females males

Figure 5.5. Plasma T3 and T4 levels of control and selected, male (right graph) and female (left graph) mice. Blood samples for determining hormone levels were taken when the animals had been on a fat diet (dark grey bars) or standard diet (light grey bars) for over eight months.

towards increased levels was observed in females (post-hoc t-test: p=0.078). These effects of high fat feeding were absent in control males and females. When fat was added to the model as a covariate the differences between control and selected mice became more pronounced: controls showed a slight decrease in relative adiponectin levels (not significant), whereas selected mice still had increased adiponectin levels on the fat diet. Figure 5.5 shows plasma levels of T3 and T4. Plasma T3 levels were similar in control and selected females on both diets. In selected males on standard chow T3 levels were decreased compared to control males on standard chow. On the fat diet, however, plasma T3 levels increased significantly in selected males (post-hoc t-test, p<0.05), but not in control males (Two-way ANOVA: group; F1,19=6.8, p=0.017 and diet; F1,19=4.1, p=0.057, groupxdiet; F1,19=3.5, p=0.07). In selected females and males, plasma T4 levels were decreased compared to controls on standard chow, and T4 levels only tended to increased on the fat diet. In males a similar pattern emerged. Selected males had increased T4 levels on the fat diet compared to mice on standard chow, whereas control males had similar levels of T4 on both diets. These effects reached statistical significance (post-hoc t-tests: p=0.079 and p=0.052; Two-way ANOVA, group; F1,16=3.9, p=0.067 and diet; F1,16=0.3, p=0.58, groupxdiet; F1,16=2.8, p=0.112). When selected animals received fat food the T4 levels increased up to a similar level as that of controls. High-fat feeding in mice bred for high activity 85

DISCUSSION

It has been reported that feeding a high-fat diet increases the risk to attract obesity and symptoms of the “metabolic symptoms” in a number of species including man (Carroll and Dudfield, 2004; Stiegler and Cunliffe, 2006) and mice (Storlien et al., 1986; Surwit et al., 1995; Ahren and Scheurink, 1998; Lin et al., 2000; Winzell and Ahren, 2004). The present study investigated whether and how such an effect may be influenced by physical activity. For this purpose, we used a line of mice selectively bred for high-wheel runnning activity and randomly bred controls. Selected females in the present study did not develop increased adiposity on a high-fat diet, whereas control female and male mice markedly increased their body mass with 13% and 5% respectively. This increase in body mass was mainly due to an increase of adipose tissue mass which appeared to be equally distributed over all adipose tissue depots in the body. Interestingly, selected males did show some increased fat storage when feeding a high fat diet, but this was almost exclusively due to enlarged visceral depots to an extent that was even more pronounced than found in control males switched to a high fat diet. Important for consideration of these findings are the observations that spontaneous activity by infra-red detection was increased in selected – in particular female - mice relative to controls, and these differences were even more amplified in animals feeding the high fat diet. At the end of the study, the mice were subjected to running wheels for two weeks for charac- terization, but observed differences in spontaneous activity were not reflected by differences in wheel running activity among selected and control mice. This could be due to ageing, since previous studies indicated that differences between wheel running activity in selected and control mice were strongest at weaning and reduce over time (Morgan et al., 2003; Bronikowski et al., 2006). Since we wanted to avoid potential training-effects of wheel running at weaning, we do not know which animals in the present study would meet the standard breeding criteria of belonging to the highest running animal within each litter of his generation. In line with the generally higher level of spontaneous activity in selected mice was the finding that these animals had increased daily energy expenditure (DEE) compared to control mice when corrected for body weight. Again, this difference was most pronounced in females as they also showed the largest differences in overall spontaneous activity. Infra-red locomotion detection allowed us to dissociate resting metabolic rate (RMR) from DEE. These analyses revealed that selected females, but not selected males had a higher RMR than control animals. Thus, both RMR and spontaneous activity-related metabolic rate (by other defined as non-exercise activity thermogenesis; Levine, 2004) contributed to the overall increased DEE in selected females relative to controls. In addition to the effect of selection, in males the fat diet increased RMR and DEE further. Despite these increases in metabolism, both groups of males still increased body mass to a similar extent on the fat diet. In females, RMR and DEE were not increased on the fat diet, which might explain the extra weight gain in control females compared to males. 86 Chapter 5

Differences in activity (and metabolism) were more pronounced in females, and this is probably the mechanism that render selected females resistant to high-fat diet induced weight gain. In agreement with the higher metabolic rate in selected mice, mainly the males had increased levels of thyroid hormones (T3 and T4) when fed a high-fat diet. Despite the fact that selected females had lower body weight than controls, food intake of the latter was lower than that of selected females, particularly if this is calculated per gram body weight. No differences in food intake were observed in the males. Food intake usually decreases when animals are put on a fat diet, because food efficiency is elevated with elevated dietary fat content (Winzell and Ahren, 2004; Morens et al., 2005). Indeed, in control females and both groups of males we found an decrease in food intake and increase in food efficiency on the fat diet. When food efficiency is high, the amount of energy an animal obtains per gram of food is higher and therefore food intake should be reduced to prevent increases in body mass. In males and control females the decrease in food intake was not sufficient to prevent weight gain on the fat diet. In the selected females, these normal responses of high fat diet were totally absent, and food efficiency was even lower on the fat diet and a constant body mass was maintained. Thus, an increased metabolic rate in combination with a low food efficiency in selected females resulted in the maintenance of constant body mass on a fat diet in these mice. Apparently the selected females cope with the fat diet different than control mice and males do. Selected females did not store extra fat when given a high-fat diet. They could accomplish this by increasing mass-specific DEE. Fat was thus burned instead of stored. On standard chow, selected mice also burned more fat as was shown by a decreased RQ in selected females compared to control females. Female selected mice were resistant to developing obesity on a fat diet, but were they resistant to developing glucose intolerance on a fat diet as well? The active selected females did have a higher glucose tolerance on standard diet compared to controls. However, on the fat diet glucose tolerance was similar to control females. Compared to selected females on standard chow glucose tolerance was actually impaired in selected females on the fat diet, whereas in none of the other groups glucose tolerance was affected by the fat diet. Selected females responded differently to the fat diet compared to the other groups in our study and also compared to what has been shown from experiments on other mouse strains (Surwit et al., 1995; Ahren and Scheurink, 1998; Winzell and Ahren, 2004). Identifying the factors involved here, may shed light on the mechanisms involved in the regulation of energy homeostasis. We measured the plasma levels of several hormones involved in the regulation of nutrient metabolism. Adiponectin is a metabolic hormone secreted by adipose tissue and has an important role in improving insulin sensitivity (Yamauchi et al., 2001; Baratta et al., 2004; Schondorf et al., 2005) by stimulating oxidation of lipids in liver and muscle (Fruebis et al., 2001; Berg et al., 2002; Yamauchi et al., 2002; Bruce et al., 2005). Previously we found that male selected mice had higher basal plasma adiponectin High-fat feeding in mice bred for high activity 87

levels throughout life, even when they were feeding chow (Vaanholt et al., 2006). In this study, we did not show an increase in adiponectin levels in selected mice on a standard chow diet, although on the fat diet there appeared to be an increase in both selected males and females. As opposed to the animals in the present study that were 4th generation offspring of either control or selected lines (without further selection for wheel-running activity), the animals in the previous study were direct offspring from females and males that were choosen as breeders based on their high wheel-running activity (i.e., belonging to the highest runners per litter). Thus, it might be possible that hyperadiponectemia while feeding standard chow is only found in animals fit those criteria, while in others it may only become evident when feeding a high-fat diet. The improved glucose tolerance in selected female mice on the standard diet in the present study can thus not be explained directly by increased levels of adiponectin, but may be related to the decreased fat content in these animals. On the fat diet, selected females had increased plasma adiponectin levels and this may have suppressed plasma glucose levels through sensitization of the liver to circulating insulin. Basal plasma glucose levels did not differ between between control or selected animals (in males and females), though, and insulin, which promotes glucose disposal was increased on the fat diet in all groups, but not in the activity- selected females. Similarly to what we show here in all groups, except for the selected females, previous studies have also shown increases in insulin and leptin, involved in the regulation of food intake and fuel homeostasis, in response to high- fat feeding in rats (Iossa et al., 2003; Woods et al., 2004). Perhaps a failure to upregulate leptin in the selected females on the fat diet was crucial in the development of glucose intolerance relative to selected females on standard chow, since leptin has been shown in other studies to be important for insulin action (Shi et al., 1998). In conclusion, female mice that had been selected for high wheel-running activity over 31 generations responded differently to high-fat feeding than selected males and random-bred control mice. On standard chow selected females were capable of burning more fat, and they could clear glucose from the bloodstream faster than control mice on standard chow. Interestingly, when given a high-fat diet the selected females did not develop diet-induced obesity as was seen in the other groups, but they did become more glucose intolerant. The selected female mice might prove an important model to investigate resistance to high fat diet-induced obesity. Selected males on the other hand did develop in fact a higher level of viseral obesity than control mice. This unexpected finding might be an adaptive strategy since fat deposition in central place in the body (i.e., visceral fat) in stead of storing it in other parts might be an economical adjustment to the more active lifestyle of the selected mice.

PartII

METABOLISM & AGEING

Chapter6

Life span, body composition, and metabolism in mice selected for high wheel-running activity and their random-bred controls

Lobke M. Vaanholt, Serge Daan,Theodore Garland Jr., G. Henk Visser

Abstract The rate of living hypothesis proposes that an increased rate of energy turnover leads to shorter life, such that the product of mass-specific energy expenditure and life span remains unaffected. Increasing the daily rate of energy expenditure should decrease life span. We tested this hypothesis in males from lines of house mice that had been selectively bred for high voluntary wheel running. We compared three groups of mice: Selected lines housed with access to a running wheel (S+), Selected lines housed without access to a wheel (S-), and Control lines (not bred for high wheel-running activity) housed with wheel access (C+). Median life spans were similar in S+ and S- (735 and 725 d respectively), but both were significantly shorter than C+ (826 d). As expected S+ ran more than C+, although the difference diminished at later ages and was no longer statistically significant by 20 months of age. Subgroups were used for determination of energy turnover and of body composition at four different ages (2, 10, 18, 26 months). Resting energy metabolism was established by indirect calorimetry and overall daily energy expenditure by the Doubly Labelled Water method. Daily food consumption and energy expenditure on a mass-specific basis were greatest in S+ (+~30%), as would be expected, and were similar in C+ and S- mice. Results were similar for mass-specific values based on wet body mass and values based on fat-free dry body mass or on organ (heart, kidney, liver and brain combined) mass. Reduced longevity in the S+ group compared to the C+ mice is consistent with the rate of living hypothesis, but the similarity of life spans of S+ and S- mice is not. Since reduced longevity in the S strains does not require running activity, there may be other factors than energy expenditure that are instrumental in causing this reduction. 92 Chapter 6

INTRODUCTION

The “Rate of Living” theory proposed by Pearl in 1928 states that an increased rate of energy metabolism increases the rate of ageing and shortens life (Pearl, 1928). The idea was derived from a study by Rubner who showed in 1908 that the life- time energy potential (mass-specific energy metabolism times maximum life span) was remarkably size independent in animals over a wide range of body masses (Rubner, 1908). The energy that an animal spends per gram body mass in its life time was similar for a guinea pig and horse. In Rubner’s study humans were an exception. Humans spend more than predicted for their body mass, but even when including humans the variation in life-time energy potential was still much smaller than the observed variation in body mass. Thus, the original argument for the rate of living theory was based on interspecific allometric comparison. In comparing species we look at the product of evolution. Natural selection has acted on body mass, on energy metabolism and on life span in all species. It has been pointed out that in interspecific allometry rates of energy turnover tend tot increase with a mass exponent of ca. 0.7, while life span (and other time measures) increase with a mass exponent of 0.3 (Kozlowski and Weiner, 1996; Daan and Tinbergen, 1997). Their product (i.e., Life-time Energy Potential, LEP) is then proportional to mass to the exponent 0.3+0.7=1, and thus mass-specific LEP scales to the body mass to the power 0, i.e., is independent of body mass. This interspecific proportionality clearly does not prove that one causes the other. The rate of living theory obviously can not be tested by comparing species. It needs to be tested within species. Several experimental tests (Holloszy et al., 1985; Holloszy and Smith, 1986; Navarro et al., 2004; Speakman et al., 2004) have failed to support the hypothesis in rats and mice. Speakman (Speakman et al., 2002; Speakman, 2005a) has identified a number of problems associated with such tests. Firstly, maximum life span is not a valid measure of ageing, because it is measured in a single individual of the population and depends on the sample size used. Using the age reached by a group of animals (e.g. the oldest 5-10%) would provide a more reliable measure. Secondly, a single measurement of basal metabolic rate is usually used to calculate the total energy spent in a lifetime. Basal metabolic rate represents the lowest metabolic rate of an animal in rest at thermo-neutral temperatures and only comprises a small part (~40% depending on the environmental conditions) of the total energy budget. A better estimate of life-time energy expenditure would be total daily energy expenditure under standard housing conditions measured at different ages throughout life. A third problem arises when energy expenditure is expressed per gram body mass, as is done in the interspecific allometric scaling analysis. Not all components of body mass contribute equally to the metabolic rate. Greenberg (1999; 2000) has pointed out that in an ideal test of the rate of living theory one should express energy turnover rates not with respect to the whole body but with respect to the most relevant tissues, i.e., the most metabolically active organs (heart, liver, kidney, brain). Living fast, dying young? 93

We have heeded the warnings from these considerations in the present study. To address the question of the consequences of elevated energy turnover rates on life span we exploited strains that had been selectively bred for high wheel-running activity (S+ and S- mice, see (Swallow et al., 1998)). We studied the relationships between energy metabolism and life span in these mice housed with or without a running wheel, and compared them to randomly-bred mice housed with wheels (C+). Recently, another study has been published on the life spans of mice from these same strains (Bronikowski et al., 2006). This study demonstrated interesting consequences of selection of activity on patterns of body mass and food intake that are suggestive of differences in energy turnover. No direct metabolic rate measurements were included in this paper. Also, the significant differences in median life span between strains claimed in this paper are based on small sample sizes (n=20 per strain and sex combination) and on underestimation of the standard errors (see discussion). In our study we aimed to avoid this problem by using n=100 per group, of which a subgroup of 60 animals was left undisturbed throughout life to generate reliable survival curves. In a smaller subgroup of the population (n=40) body composition and metabolic rate were measured at various ages to obtain reliable estimates of the life-time energy potential (McCarter et al., 1985).

MATERIAL AND METHODS

Animals & housing HsD:ICR mice (Mus domesticus) selected for running wheel activity over 31 generations and their controls were used in these experiments. For a detailed description of the selection procedure we refer to Swallow et al, 1998 (Swallow et al., 1998). In short, 112 outbred Hsd:ICR house mice were obtained form Harlan Spraque Dawley (Indianaplois, IN, USA). These mice were randomly divided in 8 separate populations with 10 pairs per line, four lines were selected on increased running wheel activity and four were randomly bred as control lines. In each generation mice at the age of 6–8 weeks were tested in a running wheel (with a diameter of 30 cm) for 6 consecutive days after which pairs were formed on basis of their running wheel activity on day 5 and 6 of the test. The male and female of a family with the highest spontaneous wheel-running activity were chosen as breeders and the mice were paired randomly within each line preventing sibling matings. In the control lines the breeders were randomly selected. At generation 31 selected mice ran approximately 2,5 times more than control mice (Rhodes et al., 2005). Eighty breeding pairs (10 per line) from the lab of Theodore Garland Jr (University of California, Riverside, USA) were bred at the Zoological Laboratory in Haren. Animals were bred randomly within each line preventing sibling matings. Male mice, born between 31 July 2002 and 27 January 2003, from the first generation of offspring were used in the experiments described below. After weaning, mice were housed with their littermates till the age of 5 months when all animals were 94 Chapter 6

individually housed with or without a running wheel for the rest of their lives (Macrolon Type II, UNO Roestvaststaal BV, Zevenaar, NL; adapted to fit in a plastic running wheel with a diameter of 14 cm). Animals had ad libitum food (Standard Rodent Chow RMB-H (2181), HopeFarms, Woerden, NL) and water and were on a 12:12 light-dark cycle (lights on at 8:00).

Experimental protocol Three experimental groups were created: Group 1: control mice housed with a running wheel (C+); Group 2: selected mice with a running wheel (S+); and group 3: selected mice without a running wheel (S-). All groups consisted of 100 animals of which 40 mice were used to assess food intake, energy expenditure and body composition at different ages (test mice) and 60 animals were left undisturbed to determine the timing of spontaneous death (life span mice). Before the start of the experiments mice were randomly assigned to one of the two subroups. Every month body mass of all animals was measured on the same day. Due to differences in ages between the animals these data had to be sorted by age later on. The body mass data was categorized in 30 day blocks, starting at 15 days of age. Wheel-running activity was measured throughout life in the ‘life span mice’ with an event recording system (ERS). The system counts events (wheel revolutions) in 2-min bins and stores the counts in file. Every mouse was measured for a period of approximately 30 days at a time every other month due to limitations in the technical set up (max. 64 channels). This yielded data on the development of wheel-running activity with age from each individual mouse housed with a running wheel in the life span subgroups. In the test mice we measured food intake and metabolic rate at various ages (2, 10, 18 and 26 months). Daily energy expenditure (kJ d-1) was determined using the doubly labeled water technique. Before each trial, the mouse was weighed on a balance to the nearest 0.1 g. Thereafter it was injected with about 0.1 g doubly labeled water (2H and 18O concentrations of the mixture 37.6% and 58.7%, respectively) allowing an equilibration period of 1 hour. The dose was quantified by weighing the syringe before and after administration to the nearest 0.0001 g. After puncturing the end of the tail, an “initial” blood sample was collected and stored in 3 glass capillary tubes each filled with about 15 µl blood. These capillaries were immediately flame-sealed with a propane torch. Thereafter the mouse was put back in their home cage. After 48 hours the animal was weighed again and a “final” blood sample was collected as described before. Per sampling period, we collected blood samples of 4 mice which had not been injected with DLW, to assess the natural abundances of 2H and 18O in the body water pools of the animals. After the final blood sample was taken, animals were put in a clean cage with weighed food. Three days later the remaining food was weighed again. To be able to correct for the humidity of the air in the room a known amount of food was placed with the mice in the room that was also weighed three days later. A subsample of the food was taken and dried to constant weight in an oven at 103°C for 4 h (ISO Living fast, dying young? 95

6496-1983(E)) to be able to determine the dry mass of the food and make comparisons between different time points. After measuring food intake mice were moved to our respirometry room and put . -1 in flow-through cages where oxygen consumption (V O2, l h ) and carbon dioxide . -1 production (V CO2, l h ) was measured (described previously by Oklejewicz et al., 1997 (Oklejewicz et al., 1997)) simultaneously with ambient temperature and activity. Oxygen and carbon dioxide concentration of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck) from each chamber was measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and by an infrared carbon dioxide gas analyzer (Servomex 1440), respectively. The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic cages. Flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks). Data were collected every 10 minutes and automatically stored on a computer. All mice were measured for 24h at an ambient temperature of 22°C. Metabolic rate (MR, kJ h-1) was calculated using the following equation: MR = . . -1 16.18 x V O2 + 5.02 x V CO2 (Hill, 1972). Resting metabolic rate (RMR, kJ h ) was defined as the lowest value of metabolic rate calculated in half-hour running means. After the respirometry measurements, animals were weighed and sacrificed using CO2 gas followed by decapitation. Trunk blood was collected in tubes with anticoagulant (EDTA or heparin) for later hormone analyses (results are published elsewhere, Chapter 4). Heart, liver, kidneys, hind limb muscles, brown adipose tissue, white adipose tissue, intestines, stomach, lung, brain, testis and skin were dissected out and weighed to 0,1 mg. Subsamples of heart, liver, kidney, muscle, brown adipose tissue (BAT) and white adipose tissue (WAT) were immediately frozen at –80°C. The gut fill of stomach and intestines was removed and the samples were weighed again. Tissues were stored at –20°C until the water and fat content was determined. Water content was determined by drying for 4 hours to constant weight in an oven at 103°C following ISO protocol (ISO 6496-1983(E)). Fat was extracted using a soxhlet and petroleumether. Following fat extraction, samples were dried to constant mass at 103°C. Dry lean masses of the organs of which subsamples were taken, were calculated using the left over pieces with the assumption that the subsamples taken contained similar amounts of fat and water. In the estimation of the dry lean mass of the remainder of the carcass we assumed that the muscle taken out consisted of protein and that BAT and WAT was all fat.

Mass spectrometry The determinations of the 2H/1H and 18O/16O isotope ratios of the blood samples were performed at the Centre for Isotope Research employing the methods described in detail by Visser and Schekkerman (1999) using a SIRA 10 isotope ratio mass spectrometer. In brief, each capillary was microdistilled in a vacuum line. The 18 16 O/ O isotope ratios were measured in CO2 gas, which was allowed to equili- brate with the water sample for 48 h at 25ºC. The 2H/1H ratios were assessed from H2 gas, which was produced after passing the water sample over a hot uranium 96 Chapter 6

oven. With each batch of samples, we analysed a sample of the diluted dose, and at least three internal laboratory water standards with different enrichments. These standards were also stored in flame-sealed capillaries and were calibrated against IAEA standards. All isotope analyses were run in triplicate. The rate of CO2 pro- -1 duction (rCO2, moles d ) for each animal was calculated with Speakman's (1997) equation:

rCO2 = N/2.078 * (ko - kd) - 0.0062 * N *kd

Statistical analysis Results are reported as means ± SEM unless stated otherwise. To test for effects of group and/or age we applied ANCOVA models in the MIXED procedure in SAS for Windows (version 9.1). Group and age were added as fixed factors in the model. Because we used four replicated control and selected lines in our experiment, we applied nested ANCOVA models in the analyses of these animals, where replicate lines nested within group (group(line))was added as a random effect. In the test animals the different lines were not distributed evenly and random effects were not added to the models. Covariates were added to the models where appropriate (i.e. body mass for measures of body composition and metabolic rate). Data were log10- transformed or squared (see text) when necessary to attain a normal distribution of the data. The significance level was set at p≤ 0.05, and all tests were two-tailed.

RESULTS

Development of body mass and wheel-running activity Body mass was measured every month in all animals. Figure 6.1A shows the development of body mass in C and S mice housed with or without a running wheel. Already at weaning, body mass was reduced in S mice compared to C mice and this difference was maintained nearly throughout life in the two groups with running wheels (S+, C+) (2 months: ANOVA F1,8=9.4, p=0.016). When S mice were housed without a running wheel they did increase their body mass and from approximately 8 months of age S- mice had significantly higher body mass compared to S+ mice (F1,8=7.4, p=0.024). Living fast, dying young? 97

50 A

20 body mass (g) 10 C+ mice S+ mice S– mice 0 14 ) B -1 12 * * 10 * * 8 * 6

wheel-running activity (km d 0 0 200 400 600 800 1000 age (d)

Figure 6.1. Development of body mass (A) and wheel-running activity (B) in mice selectively bred for high-wheel running activity (S+, black circles) and randomly-bred controls (C+, white circles) housed with wheels. Grey circles represent activity-selected mice housed without a running wheel (S-). * in Figure 1B denotes a significant difference between the groups at this age (p<0.05).

Wheel-running activity was measured in each group for one month at a time, every two months (Figure 6.1B). For statistical analysis values were averaged over periods of two months. As expected, wheel-running activity was higher in S+ mice than in C+ mice (5–20 months, p<0.05). Wheel running decreased with age in both groups – a pattern that is common to most animals (Aschoff, 1962). S+ decreased their wheel-running activity slightly faster with age, and from approximately 600 days (20 months) of age no significant difference in wheel running was present anymore between C+ and S+ mice (see Figure 6.1B). To evaluate relationships between body mass and wheel-running activity we calculated for each mouse the average body mass and wheel-running activity at young (0–250 d), middle-aged (251–500 d) and old age (501–750 d). The relationship between body mass and wheel-running activity for middle-aged mice is shown in Figure 6.2. There appears to be no relationship in S+, and a negative association in C+ mice. Indeed at all ages there was a significant negative association between body mass and wheel-running activity in C+ mice (young: r=-0.49, p<0.05; middle-aged: r=–0.53, p<0.001; old: r=–0.50, p<0.001), but not in S+ mice. 98 Chapter 6

70 C+ mice S+ mice 60

40 body mass (g) 30

0 51510 wheel-running activity (km d-1)

Figure 6.2. Relationship between wheel-running activity and body mass in control (C+) and activity-selected mice (S+).

Survival Figure 6.3 shows the survival curves for the three experimental groups (A) and the finite mortality rates (B). Finite mortality rates (FMR) over intervals of 100 days were calculated using the following formula: FMR = 1- Ne/Nb (see (Krebs, 1994)), where Ne is the number of animals left over at the end of the interval, and Nb is the number of animals at the start of the interval. Table 6.1 presents different measures of longevity for different groups: the mean age at death, median (50%), 90% percentile and maximum age reached. Activity selected mice showed similar survival throughout life in the groups with and without running wheel (Figure 6.3A). Mortality early in life was slightly higher in the selected groups compared with the controls (Figure 6.3B), resulting in a lower median age at death in the selected mice. We tested for differences between groups using the Life Tables in the survival analysis of SPSS for windows (version 14.0) which uses the Wilcoxon (Gehan) test to compare survival distributions between groups. Overall no significant group effect was found (p=0.057), but pair- wise comparisons showed that mortality was higher in S+ and S- mice compared to C+ mice (p=0.046 and p=0.032 respectively).

Metabolism Average food intake over one week was measured once at the age of 668 (s.d., 70) days in all mice in the life span subgroup alive at that time. Average food intake was 4.7 g d-1 (s.d., 1.1; n=44) in C+, 5.3 g d-1 (s.d., 0.8; n=32) in S+ and 4.2 g d-1 (s.d., 1.0; n=30) in S- mice. Effects of group were tested using a one-way ANCOVA with group(line) as a random effect and age as a covariate testing a priori for differences between C+ vs. S+ and S+ vs. S-. Food intake was significantly increased in S+ mice compared to both C+ and S- mice (F1,9=8.7, p<0.05 and F1,9=7.8, Living fast, dying young? 99

100 A

% alive 40

1.0 C+ mice S+ mice S– mice 0.8

0.6

0.4

finite mortality rate 0.2 B 0.0 0 200 400 600 800 1000 age (d)

Figure 6.3. Effect of selection for activity on mice survival (A) and mortality rates (B; see text for formula). White circles represent randomly-bred mice housed with wheels (C+), black circles represent selectively-bred mice housed with wheels (S+), and grey circles are activity-selected mice housed without a running wheel(S-).

Table 6.1. Survival data.

Group n Mean SE Median SE 90% Maximum

C+ 60 787 23 826 24 965 1090 S+ 60 704 32 735 28 979 1099 S- 60 711 29 725 54 997 1098

Mean, median 90 percentile and maximum survival in control mice housed with wheels (C+), activity-selected mice with wheels (S+) and activity-selected mice without wheels (S-). All values are given in days. SE = standard error p<0.05 respectively). Food intake was not statistically associated with body mass or wheel-running activity. At 2, 10, 18 and 26 months of age food intake, metabolic rate and body composition were measured in a subgroup of S+, C+ and S- test mice. At 2 months groups present, S+ and C+, did not yet have their running wheels. Hence, S+ and S- were still in the same treatment. These results were tested separately from the rest with independent t-tests. 100 Chapter 6

8 ) 0.25 -1 d ) -1

-1 0.20 6 0.15 4 0.10

food intake (g d 2 0.05 metabolic rate (kJ g 0 0.00

) 2.5 -1 ) d -1

-1 2.0 40 1.5

1.0 20 0.5 metabolic rate (kJ d metabolic rate (kJ g 0 0.0

) 2.5 -1 ) d -1

-1 2.0 60 1.5 40 1.0

20 C+ mice 0.5

metabolic rate (kJ d S+ mice

S– mice metabolic rate (kJ g 0 0.0 0 200 400 600 800 0 200 400 600 800 age (d) age (d)

Figure 6.4. Absolute (left graphs) and mass-specific (left graphs) food intake, resting metabolic rate and daily energy expenditure in activity selected mice housed (S+, black circles) or randomly-bred controls (C+, white circles) housed with wheels at 4 different ages (2, 10, 18 and 26 months). Grey circles represent activity-selected mice at 10, 18 and 26 months of age that were housed without wheels (S-).

Figure 6.4 shows the food intake, resting metabolic rate (RMR) and daily energy expenditure (DEE) for S+, C+ and S- mice measured at different ages. At 2 months of age there was a significant increase in food intake in S+ mice compared to C+ mice (t-test, p<0.011). We tested for differences between the groups at 10, 18 and 26 months using two-way ANOVA with age, group and agexgroup as fixed factors. Here we did not add line effects nested within group to the model, because in some groups of the test animals not all lines were represented or the lines were not represented evenly. Body mass was entered to the models as a covariate where appro- Living fast, dying young? 101

priate (for measures of metabolic rate and body composition). At ages beyond 2 months no significant difference was found in food intake. RMR was similar in S+, C+ and S- mice around 47.5 kJ d-1 (s.d., 7.0). Indeed, there was no significant contribution to the explained variance from either age or group. Mass-specific RMR was maximal at the youngest age class (2 months). Absolute and mass-specific DEE was significantly increased in S+ mice compared to both S- and C+ mice throughout life (F2,56=7.6, p=0.001). On average DEE (kJ d-1) was 59.0 (s.d., 7.7), 69.5 (s.d., 11.9) and 61.2 (s.d., 7.6) in C+, S+ and S- mice respectively. With age, DEE significantly decreased (F2,56=4.7, p=0.013) and there was a significant interaction between group and age (F4,56=3.3, p=0.018).

Body composition Results on body composition are summarized in Table 6.2, and the results of statistical analysis on these data are shown in Table 6.3. Main differences between groups were found for dry lean mass, fat mass, heart, lung and skin mass. Dry lean mass was increased in S+ mice compared with S- mice, but did not differ from C+ mice. Fat mass was lowest in S+ mice, but differed significantly only from S- mice. Heart and lung mass were increased and skin mass was decreased in the running mice (S+, C+) compared to the sedentary mice (S-). Age significantly affected fat free mass, fat mass, heart, kidney, skin and remainder of the carcass. Fat free mass, heart, kidney mass significantly increased between 10 and 26 months, and skin and fat mass significantly decreased with age.

Life-time energy potential Traditionally, the life-time energy potential (LEP, kJ) has been estimated based on measurements of resting metabolic rate and maximum life-span (Rubner, 1908). Life is, however, not passed solely in the resting state, and the life-time energy potential should equal DEE times life span. RMR might be used as an estimator of DEE, but only if DEE and RMR have a fixed ratio. In most animals that is not the case for obvious reasons. Differences in activity will result in a higher DEE, but not RMR, as we showed in our mice (Figure 6.4). RMR was on average 83, 68 or 87 % of DEE in C+, S+ and S- mice respectively. Maximum life span represents only a single event in the colony and is therefore subject to a large variance and highly dependent on the sample size that is used. Using the 90% mortality yields a more reliable measure of life span (see also (Speakman et al., 2002)). Taking these considerations into account we estimated LEP based on measurements on DEE (measured with doubly-labelled water) and the age at which 90% of the animals had died. To incorporate changes that occur in DEE with age, we calculated the average DEE per group based on measurements at 4 ages throughout life (2, 10, 18 and 26 months, see Table 6.4). The resulting LEP values were 56908, 67075 and 60977 kJ in C+, S+ and S- mice respectively. On a whole-animal basis LEP is highest in the S+ group. We return to a possible approach to significance testing below. 102 Chapter 6 S+ and S- is assumed to be similar. BM= Body S+ and S- is assumed to be similar. 826 888 877 857 2101 Body composition in control (C) and selected mice (S) housed with (-) or without (+) a running wheel. est (g 20.4±1.7 18.6±2.4 26.9±4.1 22.9±4.0 25.4±4.7 28.2±5.4 21.4±2.9 25.7±3.1 23.8±2.7 22.3±3.0 23.3±3.3 able 6.2. alues given are mean ± two months of age all animals were still housed without a running wheel and body composition sem. At at (g) 3.4±0.4 2.4±0.2 8.8±1.2 5.3±1.9 10.7±1.7 10.5±2.3 4.7±0.7 8.9±1.7 4.9±1.2 3.4±1.2 8.0±1.7 GroupN88 BM (g) C+FFM (g) 32.9±1.1Dry lean (g) 29.5±0.8 8.0±0.2F S+ 30.5±1.2 28.1±1.1 7.6±0.3 42.5±1.8Heart (g) 33.6±1.0Liver (g) 36.7±2.1 0.18±0.01 C+ 9.4±0.2 31.4±0.7Kidney (g) 0.19±0.01 1.94±0.08 39.6±2.4 0.68±0.04Brain (g) 28.9±0.9 8.3±0.2 1.67±0.08 0.24±0.01 0.71±0.04Stomach (g) 44.4±2.8 S+ 0.51±0.01 0.24±0.03 0.23±0.01 8.3±0.3 2.16±0.04 33.9±1.2Intestines (g) 2.19±0.36 0.82±0.03 0.50±0.01 0.22±0.04 33.4±1.4 0.22±0.01 1.90±0.14Lung (g) 2.07±0.28 28.7±1.0 0.74±0.03 1.77±0.11 0.55±0.01 39.5±1.6 0.32±0.10 9.7±0.3 0.29±0.02Skin (g) 0.72±0.04 S- 0.50±0.17 30.6±0.7 2.59±0.27 0.55±0.01 0.29±0.07 0.28±0.03R 2.71±0.32 0.54±0.13 2.27±0.34 8.1±0.4 39.3±2.1 0.95±0.06 4.32±0.50 0.55±0.01 0.27±0.05 0.25±0.01 1.87±0.11 34.4±1.3 2.14±0.30 0.77±0.05 4.08±0.56 0.39±0.05 36.7±1.8 2.08±0.05 0.52±0.01 8.8±0.2 0.32±0.09 0.28±0.02 33.3±0.9 0.93±0.05 0.32±0.07 2.47±0.37 C+ 5.92±1.14 0.54±0.02 39.0±2.0 0.32±0.14 0.26±0.02 2.36±0.28 0.30±0.11 31.0±0.5 2.08±0.31 0.87±0.04 4.98±1.43 0.55±0.02 0.40±0.21 0.24±0.01 8.9±0.3 1.81±0.14 2.09±0.22 1.03±0.27 5.76±1.26 0.33±0.04 2.26±0.36 0.56±0.01 0.41±0.27 8.6±0.3 0.73±0.05 S+ 0.34±0.14 1.96±0.37 5.94±2.04 0.54±0.02 0.37±0.14 0.29±0.03 2.43±0.47 4.02±0.59 0.55±0.00 0.32±0.05 8.5±0.4 2.50±0.26 5.08±0.99 0.42±0.09 S- 0.30±0.10 4.73±1.52 0.35±0.09 4.15±0.75 5.28±1.33 C+ S+ S- T Age (months) V mass, FFM= Fat free mass. mass, FFM= Fat Living fast, dying young? 103

Table 6.3. Statistical analysis on body composition data.

Variable (g) Group Age Covariate Ndf F p F p p

Body mass 66 2,57 6.8 0.002 0.3 – none Fat free mass 66 2,56 13.8 0.001 7.4 0.001 BM 0.001 Dry lean mass 66 2,56 6.2 0.004 2.3 – BM 0.001 Fat662,56 13.8 0.001 7.4 0.001 BM 0.001

Heart 66 2,56 2.5 0.04 6.8 0.002 BM 0.06 Liver 66 2,56 1.4 – 2.7 – BM 0.001 Kidney 66 2,56 1.1 – 3.2 0.05 BM – Brain 66 2,56 0.4 – 1.2 – BM – Stomach 66 2,56 0.3 – 2.3 – BM 0.005 Intestines 66 2,56 0.8 – 1.0 – BM 0.001 Lung 66 2,56 3.4 0.04 0.9 – BM – Skin 66 2,56 3.5 0.04 7.5 0.001 BM 0.001 Rest 66 2,56 0.5 – 14.2 0.001 BM 0.001

Two–way ANCOVA were performed with group, age and groupxage as fixed factors. Body mass (BM) was entered to the models as a covariate to correct for effects of body mass on variables of body composition. Data for all groups at the ages of 10,18, and 26 months were analyzed. No significant interactions (groupxage) were found, except for intestine mass (F4,56=6.1, p=0.001), and data are therefore not shown in the table. – represents non–significant effects (p>0.05). Significant effects are shown in bold (p<0.05).

On the basis of interspecific comparisons Rubner concluded that LEP expressed per gram body mass is rather invariant across species (Rubner, 1908). This led to the premise of the rate of living theory that states that there is a negative relationship between mass-specific metabolism and life span (Pearl, 1928). LEP per gram body -1 mass (LEPBW) was 1431, 1956 and 1642 kJ g in C+, S+ and S- mice respectively. When expressed per gram body mass S+ mice thus still spend most energy in their life-time. This is due to the fact that S+ mice do have shorter life spans (Table 6.1) as predicted from the measurement of daily energy expenditure (Table 6.4). Total body mass contains water and fat that are not metabolically active. We therefore also calculated LEP per gram dry lean mass; LEPDL was 6335, 8248, 7349 kJ g-1 respectively. Dry lean body mass still includes matter such as skeleton and skin that is metabolically rather inactive. Greenberg has therefore proposed that we should go one step further and express LEP relative to the metabolically most active organs: the heart, liver, kidney and brain (Greenberg, 1999). Therefore, we calculated the sum of the dry lean weight of the heart, liver, kidney and brain (Organ mass, OM) and then calculated the LEP per gram of organ mass (LEPOM): 64255, 87463 -1 and 78228 kJ g in C+, S+ and S- mice, respectively. LEPOM is still highest in the S+ mice and the coefficient of variance is 15%. Table 6.4 provides a summary of these results. 104 Chapter 6

Table 6.4. Life-time energy potential.

C+ S+ S- CV (%) Sign.

Body mass (g) 39.8 34.3 37.1 Total dry lean mass (g) 9.0 8.1 8.3 Organ mass (dry lean, g) 0.89 0.77 0.78 DEE (kJ d-1) 59.0 68.5 61.2 0.001 -1 -1 DEEBM (kJ g d ) 1.5 2.0 1.6 0.001 -1 -1 DEEDL (kJ g d ) 6.6 8.4 7.4 0.001 -1 -1 DEEOM (kJ g d )678978 0.001 Max. Life span (90%, days) 965 979 997 LEP (kJ) 56908 67075 60977 8 0.002 -1 LEPBM (kJ g ) 1431 1956 1642 16 0.001 -1 LEPDL (kJ g ) 6335 8248 7349 13 0.001 -1 LEPOM (kJ g ) 64255 87463 78228 15 0.001

Life-time energy potential (LEP; kJ) is the product of energy expenditure and life span and was calculated using average daily energy expenditure (DEE, kJ d-1) measured at 4 ages (2, 10, 18 and 26 months) throughout life in the test group, and maximum life span (90 percentile) measured in the life span animals. In addition, LEP (kJ g-1) was corrected for various measures of body composition measured at the same ages: BM; body mass, DL; dry lean mass, OM; organ mass (sum of dry lean heart, liver, kidney and brain mass). CV represents the coefficient of variation calculated over the three groups (s.d. divided by mean x100%). Sign. shows the p-values for the two-way ANOVA performed to look at differences between the groups (see text for detailed description). n.s. is not significant (p>0.05).

The problem with these comparisons is obviously that we have only a single measure for life span in each group. We can thus not readily test for differences between groups in average LEP. We do however have individual values for total and mass-specific (per dry lean and wet mass) DEE in each group. We can test the group averages against each other, both before and after multiplying all individual data with the groups life span. The basic data as well as the results of testing are supplied in Table 6.4. We applied two-way ANOVA with a factor group, age and groupxage. Afterwards we estimated LEP for each individual, by multiplying the mass-specific value of DEE with the life span (90%) of its group; this was 965, 979 and 997 for C+, S+ and S- mice respectively. S+ mice had significantly increased DEE and LEP when calculated per gram body mass, dry lean mass or organ mass, and these effects were necessarily equal when testing DEE or LEP (p<0.001). Compared to S- mice, S+ mice thus spend more energy per gram body mass, dry lean mass or organ mass, but had similar life span.

DISCUSSION

We compared metabolic rates and life span of mice selectively-bred for high wheel- running activity and their randomly-bred controls. As expected, spontaneous wheel-running activity was increased in activity-selected mice and exercise declined Living fast, dying young? 105

with age. Between 5 and 20 months of age wheel-running activity was significantly higher in activity-selected mice, but afterwards levels of running dropped to similar levels in selected and control mice. These results are similar to those found in previous studies in these mice (Morgan et al., 2003; Bronikowski et al., 2006). They correspond to the general pattern of decline in spontaneous activity with age which has been known in rodents for nearly a century (Richter, 1922; Aschoff, 1962). Despite the fact that wheel-running activity was only increased up to 20 months of age, daily energy expenditure was increased throughout life in selected mice housed with running wheels. On average daily energy expenditure was increased with 14% (or 29% mass-specific DEE) compared to sedentary selected mice and controls with a running wheel. How did this effect life span? Activity-selected mice had a median life span approximately 730 days, apparently independent of the presence of a running wheel. This is approximately 100 days shorter than controls (with a running wheel). These differences are attributable primarily to increased early mortality in the activity-selected mice. Neither median nor maximum life span differed significantly between the groups. The results disagree with a previous experiment on longevity in activity-selected mice from generation 16 of the same strains of mice (we used mice from generation 31) (Bronikowski et al., 2006). This study yielded a different conclusion on life span. It states that ‘median life expectancy differed significantly between selection and control mice within both females and males (standard errors varied between one and two days) based on the nonoverlap of the 95% confidence intervals.’ (Bronikowski et al, loc cit p.1497). The standard errors of the median life span in the paper vary from 1.3 to 1.7 days (Bronikowski et al. loc cit. table 1), which is an order of magnitude smaller than the values we find (24 to 54 days, see table 1). This difference is in spite of the facts that survival curves are visually nearly identical, and that our data are based on n=60 per group, the Bronikowski data essentially on n=20. We have recalculated the median life span from the control active female life span data (Bronikowski et al. loc.cit. Fig 2A) using SPSS (version 14, Kaplan-Meijer survival analysis). This yields 799 days, s.e.m. 54.8 days, compared to the Bronikowski figure of 801 days, s.e.m. 1.7 days. We thus remain unconvinced that the difference between conclusions in the two studies is due to a change between generations 16 and 31. Our results also contrast with some other studies in rats and mice that showed beneficial effects of exercise (absence or presence of a running wheel) on life span (Holloszy and Smith, 1987; Navarro et al., 2004). Exercise increased median, but not maximal survival in these studies. Holloszy has suggested that exercise may bring about this rectangularization of the survival curves by counteracting deleterious effects of a sedentary life combined with overeating, making it possible for more of the animals to attain old age without slowing primary ageing (Holloszy, 1988). In rats under caloric restriction, not only median but also maximal life span is increased (McCay et al., 1935). If animals exercise during caloric restriction, this does not influence the survival curve and exercise thus does not have an extra beneficial effect (Holloszy and Schechtman, 1991). In our activity-selected mice 106 Chapter 6

food intake has increased during the selection protocol, and sedentary mice (S-) have greater body weight than mice housed with wheels. S- however did not experi- ence deleterious effects of their sedentary life style, and had similar life spans to S+ mice. This may be related to the fact that selected mice are more active per se. In a cage without the presence of a running wheel, we observed that activity of selected mice is also increased compared to control mice (infra red measurements in respirometer, data not shown). It is also possible that traits that have been co- selected during the selection process may play a role in this. Negative effects of increased work on life span, as expected on the basis of the rate of living theory, have been shown in several species (house flies (Sohal and Buchan, 1981), honey bees (Wolf and Schmid-Hempel, 1989), kestrels (Daan et al., 1996)). In agreement with this, comparing control and selected mice housed with a running wheel, the selected mice spent more energy and had decreased life spans. However, when comparing selected mice with or without a running wheel, a difference was found in energy metabolism but not in life span. These results indicate that the differences in life span between the groups may not be directly related to their metabolic rates, and thus contradict the rate of living theory. The differences we showed in life span between control and selected mice may be caused by any behavioural or physiological traits that differ between these mice. For instance, young selected females have increased corticosterone levels (Malisch et al., 2006) (not old males; see Chapter 4), and it has been shown in rats that animals with high corticosterone levels have significantly shorter life spans (Cavigelli and McClintock, 2003). There has been much debate over the optimal way to test the rate of living theory (Lynn and Wallwork, 1992; Greenberg, 1999; Speakman et al., 2002). On the basis of results from experiments on caloric restriction, some authors have discarded the rate of living theory (McCarter et al., 1985; Masoro, 1996). Calorically restricted rodents have greater median and maximum life span, but have similar mass-specific food intakes (Masoro et al., 1982) and mass-specific 24h metabolic rates are similar to ad libitum fed animals (McCarter et al., 1985; McCarter and McGee, 1989). Mass specific energy turnover disregards however that the metabolic rate is far from uniform in the body. Greenberg and Boozer (2000) have argued that one should express energy metabolism per unit of mass of the metabolically most active organs (heart, liver, kidney and brain), and that this organ-specific metabolic rate does decrease in calorically restricted animals (Greenberg, 1999), such that the results are consistent with the rate of living theory. In the present study we eliminated these problems involved. For the first time we accurately measured body composition as well as metabolic rates throughout life to obtain good estimates of LEP. Mass-specific life-time energy expenditure (LEPBW) is traditionally calculated to make inter-specific comparisons -1 -1 of LEP (Rubner, 1908). LEPBW was 1431, 1956 and 1642 kJ g Live in C+, S+ and S- mice respectively. Based on these values we would discard the rate of living theory because LEP is not a constant as proposed by Rubner (1908). Also when cal- Living fast, dying young? 107

culating LEP relative to dry lean mass (LEPDL) or the mass of the metabolically activity organs (LEPOM), S+ mice still have a higher LEP than C+ and S+ mice. In the calculations we do not differentiate between metabolic organs (heart, liver, kidney and brain) in energy expenditure, whereas Greenberg based his estimates on the different metabolic rates per organ (Greenberg, 1999). Complex mechanisms exist that protect the body from reactive oxygen species (ROS) that are inevitably produced during normal oxidative metabolism. These ROS have been shown to cause damage to macromolecules and could eventually lead to cell death (Beckman and Ames, 1998). The body is protected against these ROS among others by antioxidant enzymes. Antioxidants enzymes scavenge ROS before they can cause damage. Up-regulating antioxidant production protects the body against oxidative stress and potentially slows ageing. The antioxidants enzymes superoxide dismutase (SOD) and glutathione peroxidase (GPx) have been measured at various ages in the test mice (Chapter 7), but as in the study of Bronikowski et al. (2002),no differences in antioxidant levels were found between activity-selected and control mice. A positive relationship between SOD activity and metabolic rate was found, but the increase in SOD activity with metabolic rate was not high enough to increase SOD activity of selected mice above the levels measured in controls. Given the increased metabolic rate, and presumably ROS production, in selected mice, this would leave the selected mice more susceptible to oxidative damage. When the protection against ROS (enzyme activity per kJ) was calculated no differences were found between control and selected mice. Another important mechanism protecting animals against the accumulation of damaged proteins is protein turnover. Protein turnover involves the removal (breakdown) and replacement (synthesis) of inactive or oxidized proteins in the cell. Protein turnover therefore plays a potentially vital role in ageing (Sohal, 2002; Ryazanov and Nefsky, 2002; Yarasheski, 2003). In young selected mice an increase in protein synthesis in muscle, but not liver was found relative to control mice (both housed with a running wheel, Chapter 7), but at later ages there were no differences in protein synthesis rates. At old age, when oxidative stress starts causing problems, selected mice were thus not better protected against the accumulation of damaged proteins. In summary, mice bred selectively for high wheel-running activity have increased levels of wheel-running up to 20 months of age, which results in an increase in absolute metabolic rate of 14%. Life span was subsequently reduced by approximately 100 days compared to randomly-bred mice housed with a running wheel, but they did not differ from activity-selected mice housed without a running wheel. The resultant life-time energy expenditure (LEPBM) in activity-selected mice housed with wheels was increased with 27% compared to the two other groups. These results are not consistent with the rate of living theory and suggest that physiological and/or behavioural adaptations rather than differences in metabolic rate underlie the differences seen in life span between control and selected mice. 108 Chapter 6

Acknowledgements We thank Saskia Helder for taking excellent care of the animals, and Gerard Overkamp for technical assistance. Berthe Verstappen performed the isotope analyses. We also thank Peter Meerlo, Kristin Schubert, Alinde Wallinga, Mark Doornbos and Berber De Jong for their help at various stages in the project. S.D. is supported by EUCLOCK (EC 6th framework). 109

Chapter7

Protein synthesis and antioxidant capacity in ageing mice: effects of long-term voluntary exercise

Lobke M. Vaanholt, Gerald E. Lobley, Theodore Garland Jr, John R. Speakman, G. Henk Visser

Abstract Reactive oxygen species (ROS) are produced as by-products of aerobic metabolism and can cause damage to macromolecules (DNA, lipids and proteins) thereby contributing to ageing. Antioxidants scavenge ROS and lower their concentrations. In addition, oxidative damage caused to macromolecules can be repaired or replaced via protein turnover. Exercise increases metabolism and ROS production but also elevates protein turnover. The balance between these two responses may underlie the effect of physical activity on longevity. Effects of life-long exercise on antioxidant enzyme activities and fractional synthesis rates (FSR) were examined in heart, liver and muscle of mice selectively bred for high wheel-running activity at various ages (2-26 months). FSR decreased with age and were increased in muscle of in young, not old, activity-selected mice. FSR did not differ between control and activity-selected mice in liver. Enzyme activity of superoxide dismutase and glutathione perioxidase also decreased with age, and showed a peak at 10 months of age in liver. Selection for wheel-running activity did not affect antioxidant enzyme activity. Daily energy expenditure correlated positively with antioxidant levels in liver of control and activity-selected mice. This could indicate that increases in ROS production with raised metabolic rate result in up-regulation of antioxidant enzymes.

Physiological and Biochemical Zoology, In review 112 Chapter 7

INTRODUCTION

Reactive oxygen species (ROS), such as the superoxide anion (O2•-), hydrogen peroxide (H2O2) and the hydroxyl radical (•OH), are produced as by-products of aerobic metabolism in mitochondria and can cause damage to DNA, lipids and proteins (Beckman and Ames, 1998; Davies et al., 1982; Mecocci et al., 1999; Tyler, 1975). This damage to macromolecules can accumulate with age (Barja, 2004b) and may contribute to senescence and degenerative diseases associated with ageing (e.g. cardiovascular disorders, Parkinson’s disease) (McEwen et al., 2005; Melov et al., 1999; Wallace, 2005). An elaborate defence system consisting of endogenous antioxidant enzymes such as catalase (CAT), superoxide dismutase (SOD), glutathione peroxidase (GPx), and numerous non-enzymatic antioxidants, including vitamins A, E and C, glutathione (GSH), ubiquinone, melatonin and flavonoids, exist to scavenge ROS and thereby prevent deleterious effects (Beckman and Ames, 1998). A small amount of the ROS produced escape conversion and can still damage macromolecules. As a second line of defence, damaged macromolecules can be replaced, as occurs by protein turnover that involves the removal (breakdown) and replacement (synthesis) of inactive or oxidized proteins in the cell. Protein turnover therefore plays a potentially vital role in ageing (Ryazanov and Nefsky, 2002; Sohal, 2002; Yarasheski, 2003). Relationships between antioxidant enzyme activity, protein turnover and metabolism have been studied by experimentally increasing metabolic rate, e.g. by increasing physical activity. Although some results are contradictory, it is widely accepted that regular physical activity leads to an increase in both ROS production and activities of antioxidant enzymes, especially in muscle (reviewed in (Ji, 1999)). If the increase in antioxidant defences in response to exercise is greater than the increase in ROS production, this would lead to a better protection against oxidative damage. Exercise generally also has a stimulatory effect on protein synthesis rate, specifically in skeletal muscle, in rats (Hayase and Yokogoshi, 1992; Hernandez et al., 2000; Katzeff et al., 1994; Mosoni et al., 1995), and humans (Biolo et al., 1995; Chesley et al., 1992; Phillips et al., 1997; Rennie et al., 1981; Sheffield-Moore et al., 2004; Short et al., 2004). Maintaining high levels of antioxidant enzyme activity together with high rates of protein synthesis in old age would diminish the accumulation of damaged proteins and increase cell survival. For example, protein turnover and antioxidant enzyme activity have been shown to be increased in calorically-restricted animals, a nutritional condition that increases life span in several species, including mice (Weindruch et al., 1986) and rats (McCay et al., 1935); for recent review see (Masoro, 2005). Therefore, ageing may be ameliorated by mechanisms, such as exercise, that elevate protein turnover and/or antioxidant enzyme activity. Mice from lines selectively-bred for high wheel-running activity for 31 generations and their random-bred control lines (Swallow et al., 1998) were used to study effects of life-long voluntary exercise on antioxidant enzyme activities and protein turnover. Although the effects of exercise on antioxidant systems and protein syn- Antioxidants and protein synthesis in high-activity mice 113

thesis are often compared between young and old subjects the influence of exercise throughout life on oxidant systems and the relationship between metabolic rate and antioxidant activity on an individual level are not well studied. The main aim of the current study was to test whether long-term exercise, that increases metabolic rate, induces compensatory changes in antioxidant enzyme activities and protein synthesis rates. In addition, we studied combined effects of age and exercise on both systems and we explored the relationship between energy metabolism and antioxidant enzyme activity.

MATERIAL AND METHODS

Animals & housing Male mice, originally of the Hsd:ICR strain (Mus domesticus), from lines that had been selected for high wheel-running activity for 31 generations and their random bred controls were used (Swallow et al., 1998). Eighty breeding pairs of mice from the lab of Prof. T. Garland Jr. were used as the basis for a colony at the University of Groningen. In the original selection protocol (Swallow et al., 1998) eight separate lines were created (4 control and 4 selected) by breeding randomly (control) or selecting the most actively running (in revolutions per day) male and female of each family for breeding (selected lines). Mice from each of the eight lines were used in the experiments. Mice were housed with three litter mates from weaning until they were five months old, after which they were housed individually (Macrolon type II long cages, UNO Roestvaststaal BV, Zevenaar, NL) with wood shavings as bedding material and a running wheel (cages were adapted to fit a plastic running wheel with a 7 cm radius: (Vaanholt et al., 2006)). Food (standard rodent chow RMH-B (2181), HopeFarms, Woerden, NL) and water were provided ad libitum, and animals were on a 12:12 light:dark cycle. Two experimental groups were created: Control mice (C+) and Selected mice (S+) both housed with a running wheel from the age of five months. In the 2-month old group, animals were thus not yet housed with a running wheel. Metabolic rates of all animals were measured prior to sacrifice at four ages: 2, 10, 18, and 26 months. All procedures concerning animal care and treatment were in accordance with the regulations of the ethical committee for the use of experimental animals of the University of Groningen (License DEC 2777(-1) and 4184A).

Tissue collection At each age, five to eight mice per group were briefly anaesthesized with CO2 and then killed by decapitation. Animals were dissected and biopsies of hind limb muscle (only at 2 and 26 months), liver and heart were immediately frozen in liquid nitrogen and stored at –80°C for antioxidant enzyme measurements. Protein synthesis was assessed only in 2 and 26 month old animals. The 26 month old individuals were the same animals used for the antioxidant measure- 114 Chapter 7

ments, but for the 2-month old group different animals were used (n=8 per group). For logistic reasons, this group only contained one control (lab designation is line 2) and one selected line (line 7). Food intake and body mass were measured for two consecutive days prior to the harvesting of tissues for protein synthesis measurements. Protein synthesis was measured using the large-dose method as described by Garlick et al. (1980). Mice were given an intra-peritoneal injection of 150 mM 2 H5-phenylalanine (1.5 ml per 100 g animal). After 15 minutes the mice were euthanized using CO2, followed by decapitation. Trunk blood was collected in pre- chilled tubes with heparin as anti-coagulant. Blood samples were centrifuged at 2600 g at 4°C for 15 min, and the plasma was collected and stored at –80°C until analysis. Liver and hind-leg muscle were rapidly removed, weighed to 4 decimal places, rinsed in ice-cold saline, frozen in liquid nitrogen, and stored at –80°C until analysis. Exact times (nearest second) of injection and freezing of tissues were recorded.

Protein synthesis Free and protein-bound enrichments of phenylalanine in liver and muscle tissues were quantified as described by Wester et al. (2004). Approximately 300 mg of frozen tissue was homogenised on ice in 3 ml 7% (w/v) sulphosalicylic acid (SSA). Free phenylalanine was separated from protein-bound phenylalanine by centrifugation at 1000 g at 4°C for 15 minutes and the supernatant retained. The pellet was then washed three times with 3 ml 7% sulphosalicylic acid to remove free phenylalanine. The initial supernatant fraction (free pool) was passed through a 0.4 ml column of Dowex AG 50W-X8 (100-200 mesh) and the resin rinsed with 2x3.5 ml water before the phenylalanine was eluted with 2 ml 2M NH4OH and 1 ml water. The eluate was freeze-dried and stored at –20°C for later analysis. Half of the washed pellet (protein-bound pool) was transferred to a 8 ml screw-topped Pyrex hydrolysis tube and solubilised in 1 ml 0.5 ml 0.3 M NaOH for 30 minutes. A few phenol crystals were added and the sample was hydrolysed by adding 7 ml 4M HCl and heating on a dri-block at 110°C for 18 hours. Hydrolysates were dried under vacuum, resuspended in 1 ml 0,5 M sodium citrate (pH 6.2), and stored at -80°C until later analysis. For the plasma samples 150 µl was treated with 150 µl 15% SSA, centrifuged at 1000 g at 4°C for 10 minutes and 150 ml of the supernatant was passed through a 0.2 ml column of Dowex AG 50W-X8. Elution conditions and subsequent treat- ments were similar to the tissue free pool samples. Stable isotope enrichments of the tissue and plasma free pool were measured by gas chromatography mass spectrometry (GC/MS) after conversion to the tertiary- butyldimethylsilyl (TBDMS) derivatives (Calder and Smith, 1988). In the hydrolysed samples (protein-bound pool, low enrichment) phenylalanine was converted to phenylethylamine by enzymatic decarboxylation prior to forming the TBDMS derivative. This was separated by capillary column gas chromatography, and enrichments obtained from electron impact ionization selective ion monitoring (EI-SIM) Antioxidants and protein synthesis in high-activity mice 115

mass spectrometry was used to detect the TBDMS derivate of phenylethylamine (see Calder et al., 1992 and Slater et al., 1995). The fractional synthesis rate (FSR, % d-1) was calculated using the following equation: FSR = 100*(BP/FP)*1440/t, where BP is the bound pool of phenylalanine in mole percent excess (MPE), FP is the MPE of the free pool of phenylalanine measured in either plasma or tissue, and t is the time (min) between injection of phenylalanine and freezing of the tissue in minutes. The ratio between FP measured in plasma and liver or muscle was calculated. The ratio was 1.02±0.03 and 1.00±0.03 (mean±sd) in muscle and liver respectively. Neither ratio was significantly different from unity. From these data we concluded that the injected phenylalanine effectively equilibrated between tissue and plasma and remained so until time of death. In consequence, the FSR reported represent the FSR calculated based on plasma free phenylalanine as representative of the precursor pool.

Antioxidant enzyme activities & protein content Prior to enzyme activity determinations, tissue samples were homogenized by soni- cation in 20 volumes of ice cold 50 mM phosphate buffer. Following centrifugation (25 min at 3000 g), the supernatant fraction was collected, divided over several tubes and stored at –80°C for enzyme activity and protein measurements. Total superoxide dismutase (SOD) activity was determined at 25°C by the inhibition of the auto-oxidation of pyrogallol by SOD in the supernatant, following the method of Marklund and Marklund (1974). The reaction was followed spectrophotometrically at 420 nm in the following reaction mixture: 50 mM Tris-DTPA buffer, 15 µl supernatant and 15 µl pyrogallol in a total volume of 800 µl. Each triplicate measurement was preceded by a blank, containing only pyrogallol in Tris-DTPA buffer. One unit of SOD was defined as the amount of enzyme causing 50% inhibition of pyrogallol auto-oxidation. Glutathione peroxidase (GPx) activity was determined at 25°C via the oxidation of NADPH in the presence of reduced glutathione (GSH) and H2O2 (combining the assays of Paglia and Valentine (1967), and Lawrence and Burk (1976). The following reaction mixture was used: 4.28 mM sodium azide (to block catalase activity), 1.07 mM EDTA, 4.286 mM GSH, 0.214 mM NADPH, 1 U ml-1 GR in ice cold 50 mM phosphate buffer. 25 µl H2O2 and 25 µl sample were added to the reaction mixture. Reactions were followed spectrophotometrically at 340 nm in a total volume of 700 µl. To correct for spontaneous oxidation reactions independent of GPx, blanks without H2O2 were measured and subtracted from the assay values. One unit of GPx was defined as the amount of enzyme that oxidized 1 µmol of NADPH per minute in the presence of reduced glutathione. Protein content of the supernatant fraction was determined using a Bradford assay (Quick start Bradford protein essay kit 2; Biorad Laboratories B.V., Veenendaal, NL). 116 Chapter 7

Indirect calorimetry Prior to killing, resting metabolic rates (RMR) and daily energy expenditure (DEE) were measured for each animal (the same individuals for which antioxidant enzyme activity was measured) using an eight-channel indirect calorimetry system (described previously in (Oklejewicz et al., 1997; Vaanholt et al., 2006)). The mice . were put in flowthrough chambers where oxygen consumption (V O2, l h-1) and . carbon dioxide production (V CO2, l h-1) was measured simultaneously with ambient temperature and activity (passive infrared detectors). Oxygen and carbon dioxide concentrations of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck) from each chamber were measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and carbon dioxide by an infrared gas analyzer (Servomex 1440). The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried outlet air from the metabolic cages. The flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks). Computerised data were collected every 10 minutes. All mice were measured for 24h at an ambient temperature of 22°C. Oxygen consumption was calculated according the equation 2 of Hill (1972) to correct for volume changes with respiratory quotient below 1 and expressed in standard temperature and pressure. Metabolic rate (MR, kJ h-1) was . . estimated using the following equation: MR= 16.18 x V O2 + 5.02 x V CO2 (Romijn and Lokhorst, 1961). Resting metabolic rate (RMR, kJ d-1) was defined as the lowest value of metabolic rate calculated from cumulative means every 30 min (Vaanholt et al., 2006). Daily energy expenditure (DEE, kJ d-1) was calculated as the average metabolic rate during the entire 24-h measurement period.

Statistical analysis Results are reported as means ± SEM. To test for effects of treatment and/or age we applied ANOVA models in the MIXED procedure in SAS for Windows (version 9.1). Group, age, and group x age were added as fixed factors. Because we used four replicated control and selected lines in our experiment, we applied nested ANOVA models in the analyses of these animals, where replicate lines nested within group and line(nested within group) x age (i.e., an interaction term) were added as random effects. These random effects of line were not added when testing for differences in protein synthesis, because not all lines were represented in the 2-month old animals used for these measurements. Where appropriate, covariates were added into the models (e.g., metabolic rates, food intake). Adjusted means were calculated using the least squares means command in SAS MIXED. The outlier test described by Cook and Weisberg (1999) was used to test for outlying data points. One significant outlier was identified when testing protein synthesis in muscle (see Table 7.2). This data point was subsequently removed before final statistical analysis. Data on antioxidant enzymes were log10-transformed as necessary to improve normality. The significance level was set at p≤ 0.05, and all tests were two-tailed. Antioxidants and protein synthesis in high-activity mice 117

RESULTS

Body mass, food intake and fractional synthesis rates At both ages control mice had greater body mass than activity-selected mice and body mass increased with age in both groups (see Table 7.1). Interestingly, food intake did not differ between groups or with age. FSR decreased by approximately 15% with age in liver and muscle of both groups. This effect was significant in liver but not muscle (Table 7.1). In liver, no significant effects of group (selection vs. control) nor a group x age interaction was found. In muscle, the group effect reached significance (p=0.053) and post-hoc tests (Tukey) showed that at 2 months of age FSR was significantly higher in muscle of activity- selected mice relative to controls (p<0.05), but no difference between groups was found at 26 months of age. Also, no interaction between group and age was found. Food intake can influence protein synthesis and therefore we added food intake to the models as a covariate. In none of the models was food intake a significant covariate.

Antioxidant enzyme activity & metabolism The age-related development of antioxidant enzyme activities in the liver (SOD and GPx) and heart (SOD) are shown in Figures 7.1 and 7.2, respectively. In the liver, antioxidant enzyme activities (GPx and SOD) varied considerably with age (see Table 7.2) and were highest at 10 months of age (post-hoc Tukey; at 10 months SOD activity was significantly different compared to 2 and 26 months; no differ-

Table 7.1. Effects of selection for high locomotor activity activity on body mass, food intake, and fractional synthesis rates (FSR) in liver and muscle.

2 months 26 months p-values Variable name Control Selected Control Selected Group Age n8885 Age (d) 73±2 72±5 781±11 781±11 Body mass (g) 33.9±1.5 30.0±1.5 41.4±1.5 37.9±1.9 0.026 <0.001 Food intake (g d-1) 4.1±0.4 4.0±0.4 3.8±0.4 3.0±0.5 0.21 0.57

FSR Liver (% d-1) 69.3±2.6 69.6±2.6 56.7±2.6 63.4±3.3 0.23 0.003

FSR Muscle (% d-1) 5.1±0.5 7.1±0.5 4.9±0.5 5.2±0.6 0.053 0.065

Results for two-way ANOVA’s with a factor group, age and group x age are given in addition to least square (adjusted) means±SE for all groups. For analysis of food intake body mass was added to the model as a covariate. See text regarding analyses of FSR with food intake as a covariate. No significant interactions (groupxage) were found (p>0.1), and p-values are therefore not shown in the table. Bold values represent significant results (p≤ 0.05). Food intake data for two mice were missing. n=sample size. 118 Chapter 7

400 control selected

300

protein) 200 -1 SOD activity

(U mg 100

0 1.6

1.2 protein) -1

0.8 GPx activity 0.4 mol NADPH mg µ ( 0.0 2 10 18 26 age (month)

Figure 7.1. Superoxide dismutase (SOD; top graph) and glutathione peroxidase (GPx; bottom graph) activity in liver of mice selected for high wheel-running activity and their random-bred controls at different ages. One unit of SOD was defined as the amount of enzyme that causes 50% inhibition of pyrogallol auto-oxidation. One unit of GPx is defined as the amount of enzyme that oxidizes 1 µmol of NADPH per minute in the presence of reduced glutathione.

control 120 selected protein)

-1 80

SOD activity (U mg 0 2 10 18 26 age (month)

Figure 7.2. SOD enzyme activity in heart of mice selected for high wheel-running activity and their random-bred controls at different ages. One unit of SOD was defined as the amount of enzyme that causes 50% inhibition of pyrogallol auto-oxidation. Antioxidants and protein synthesis in high-activity mice 119

Table 7.2. Nested ANOVA on effects of group and age on antioxidant enzyme activities.

Group Age Group x Age Trait N d.f. F p d.f. F p F p

SOD Liver 57 1,6 0.02 0.89 3,11 5.9 0.012 1.4 0.29 GPX Liver 55 1,6 0.26 0.63 3,10 5.6 0.016 0.4 0.78 SOD Heart 57 1,6 0.11 0.75 3,11 3.7 0.047 1.2 0.35 SOD Muscle 29 1,2 0.05 0.85 1,2 0.5 0.56 1.1 0.42

Nested ANOVA models were performed in the MIXED procedure of SAS for Windows (version 9.1). Group, age and group x age were added as fixed factors and replicate lines nested within group and line(nested within group) x age were added as random effects to correct for line effects. To test for relationships between resting metabolic rate (RMR, kJ d-1) or daily energy expenditure (DEE, kJ d-1) and enzyme activity both were added to the model as covariates separately (data not shown in Table, see text). Sample sizes were: 2 months, n=8; 10 months, n=8 and n=5 in control and selected mice respectively; 18 months; n=8 and n=7, 26 months; n=8 and n=5. In 2 mice sample volume of liver was too small to measure both SOD and GPx enzyme activity, reducing the sample size for these measurements. N represents the total sample size and d.f. the degrees of freedom. Significant effects are in bold format (p≤ 0.05).

control 50 selected

40 protein) -1 30

SOD activity (U mg 0 2 26 age (month)

Figure 7.3. SOD enzyme activity in muscle of mice selected for high wheel-running activity and their random-bred controls at different ages. One unit of SOD was defined as the amount of enzyme that causes 50% inhibition of pyrogallol auto-oxidation. ences between the other ages, p<0.05). An effect of age on SOD activity in heart tissue was also found (Table 7.2). A similar pattern as seen in liver with a peak at 10 months can be seen in control mice (see Figure 7.2), however, post-hoc tests only showed a significant difference between SOD activity at 2 and 26 months (Tukey; p<0.05), indicating a decrease in SOD activity with age in heart. Selection for high wheel-running activity did not affect antioxidant activities in liver or heart (Table 7.2). SOD activity in muscle from control and selected animals was measured only at 2 and 26 months. There was no effect of either age or selection for activity (Figure 7.3, Table 7.2). Resting metabolic rate (RMR, kJ g-1 d-1) and daily energy expenditure (DEE, kJ g-1 d-1) were measured in all animals at all ages. Overall, mass-specific RMR was 120 Chapter 7

500 CONTROL SELECTED 2 month 10 month 400 18 month 26 month protein) -1 300

200

100 SOD activity (U mg

0 30 40 5060 70 80 30 40 5060 70 80 dailt energy expenditure (kJ d-1)

Figure 7.4. Relationship between daily energy expenditure and SOD activity in liver of control (left panel) and selected mice (right panel) at various ages.

1.20±0.26 (mean±sd) and 1.31±0.25 in control and selected mice respectively. Selected mice thus had a slightly higher RMR (+9%), but this difference was not significant (Two-way nested ANCOVA with body mass as covariate). Mass-specific DEE was on average 1.53±0.31 and 1.82±0.35 in control and selected mice respectively (+19%). Differences in daily energy expenditure were slightly larger at young ages, but overall no significant effect of group on DEE was shown (two-way nested ANCOVA). RMR and DEE were added into the models as covariate to test whether they affected antioxidant enzyme activity. In liver, but not heart, DEE significantly predicted both SOD (p=0.008) and GPx (p=0.037) enzyme activity. RMR also significantly predicted SOD (p=0.044), but not GPx (p=0.056) activity). Including RMR or DEE into the models did not alter the outcome. Figure 7.4 shows the relationship between DEE and hepatic SOD activity of control (left panel) and activity- selected (right panel) mice of all ages. In both groups a positive relationship between daily energy expenditure and antioxidant enzyme activity was found at all ages. Similar results were found for the relationship between resting metabolic rate and both antioxidant enzymes, and between DEE and GPx activity in liver (data not shown). In heart and muscle metabolic rate did not predict SOD activity. We calculated the protection against ROS for each individual mouse by dividing hepatic SOD enzyme activity by daily energy expenditure (protection in U mg-1 protein kJ-1 d-1) and plotted it against daily energy expenditure to see whether mice with a high metabolic rate were better protected against ROS (See Figure 7.5). No significant relationship was shown between protection and metabolic rate. Average protection of control and selected mice was 4.6±1.4 and 4.5±1.2 U mg protein-1 kJ-1 respectively and these values did not differ significantly. Antioxidants and protein synthesis in high-activity mice 121

8 9.4 ) -1

6 protein kJ -1 4

control protection (U mg selected 0 30 40 5060 70 80 dailt energy expenditure (kJ d-1)

Figure 7.5. Protection against ROS (SOD enzyme activity in liver per kJ energy expenditure) as a function of daily energy expenditure in control (white circles) and selected (black circles) mice. Linear regression showed no significant relationship between protection and energy expenditure. The arrow indicates an outlier that was left out of analysis.

DISCUSSION

In liver, age had a strong effect on protein FSR in both control and selected mice. On average FSR decreased with 16%. In muscle, selected animals had a steep decrease in FSR with age (36%), but control mice only showed a small decline (4%). The literature on this subject is ambiguous. Most studies have shown an age- related decrease in muscle FSR (Dorrens and Rennie, 2003; Lewis et al., 1985; Rattan, 1996) but other reports claim no difference (Sheffield-Moore et al., 2005; Volpi et al., 2001). Discrepancies between studies can have many possible explanations, such as tissue (muscle type) used, sex, diet and activity level of subjects. A common finding is an increase in muscle protein synthesis in response to exercise (Chesley et al., 1992; Hayase and Yokogoshi, 1992; Hernandez et al., 2000; Short et al., 2004). As expected, an increase in muscle FSR was found in 2-month old activity-selected mice. At this age the activity of the mice was not controlled as both groups without a running wheel and no record of activity was attempted. In a similar study, however, where animals were housed in a cage with a locked running wheel, overall activity recorded by passive infrared sensors was found to be increased by 130% in selected mice at 2 months of age (unpublished data). Therefore, the activity levels were probably also increased in the current 2-month old mice. This is supported by the observed increase seen in muscle FSR. Several studies have shown that exercising at old ages still induces an increase in FSR (Sheffield-Moore et al., 2004; Short et al., 2004; Yarasheski et al., 1993). At 26 months of age we did not find a difference in muscle FSR between control and activity-selected mice. Mice selected for wheel-running activity have been shown to 122 Chapter 7

run approximately 2.5 times more per day than control mice at 6-8 weeks of age (Morgan et al., 2003), but these differences disappear over time. Indeed, after approximately 60 weeks of age no differences exist in wheel-running activity (Bronikowski et al., 2006; Morgan et al., 2003). Thus, in old animals (26 months) wheel-running activity probably no longer differed between selected and control mice and this would explain why protein synthesis was not different between the groups at an old age. These findings would also suggest that prior exercise history has no long-term effect on muscle FSR. A more intensive workout at old age is probably necessary to evoke increases in protein synthesis rates. In liver, selection for activity did not affect FSR, but FSR did change with age. Previous studies have reported both increased (Mosoni et al., 1995) or decreased (Hayase and Yokogoshi, 1992) hepatic protein synthesis in response to exercise and the effects remain unclear. Another important factor in determining the extent of the effect of exercise on protein synthesis might be age. Lewis et al. (1985) studied the effects of caloric restriction on whole-body protein turnover in rats at different ages and found the most pronounced effect of caloric restriction on FSR at 12 months of age. At this age caloric restriction increased FSR by 45% whereas at 24 months the increase was only 6% and at 2 months there was a 2% decrease. Measurements at intermediate ages are thus important to give insights into the effects of exercise on FSR throughout life. In addition to effects FSR, age affected antioxidant enzyme activity of SOD and GPx in mouse liver and heart. In liver, peak enzyme activity occurred at 10 months of age with a subsequent decline, while heart SOD decline in enzyme activity between 2 and 26 months. There was no effect of age on SOD activity in muscle, although with the sampling protocol used (only measuring at 2 and 26 months of age) intermediate changes may have been missed. Conflicting data exist on the effects of age on antioxidant enzyme activities and comparisons across studies are complicated by the use of different species, ages selected and organs studied. Most studies have measured antioxidant enzyme activity at only two ages and found increased, decreased or no differences in antioxidant enzyme activities in various tissues (Gunduz et al., 2004; Kakarla et al., 2005; Rao et al., 1990; Sohal et al., 1990). Antioxidant enzyme activity has also been measured in C57BL/6J mice at 3, 11, 19 and 27 months of age and a similar peak in SOD activity in heart and liver was found at 11 months of age (see Chapter 9). In rats, SOD and CAT enzyme activity in the brain at five ages also followed a similar pattern in antioxidant enzyme activity as shown here, with a peak in activity at 12 months of age (Tsay et al., 2000). These results highlight the importance of measuring antioxidant enzyme activities at various ages. By doing this, discrepancies between existing data might be resolved. There were no significant effects of exercise (selection for activity) on antioxidant enzyme activities in liver, heart or muscle. This is in contrast to a study by Gunduz et al. (2004) comparing antioxidant enzyme activity in rats that underwent one-year forced swimming exercise and sedentary controls. In this study, exercising Antioxidants and protein synthesis in high-activity mice 123

rats had increased antioxidant enzyme activity in heart and liver relative to sedentary controls (Gunduz et al., 2004). Animals in our study exercised voluntarily and the difference in activity between the groups may not have been large enough to observe effects on antioxidant enzyme activities. In agreement with this, Selman et al. (2002) compared antioxidant enzyme activity in heart, liver and muscle of sedentary and voluntarily exercising voles and also observed no differences. Another study on voluntary exercising male rats has shown that GPx and CAT activities was not altered, while SOD activity increased in response to exercise (Yamamoto et al., 2002). In voluntary exercising female rats, increases in both SOD and CAT in liver of young (three-month) and old (twelve-month) were found (Kakarla et al., 2005). Female rats voluntarily undergo up to 10-fold more wheel- running activity than males (Yamamoto et al., 2002), which might explain the discrepancy between studies. In agreement with this, in the activity-selected strain used here, females run approximately 25% more than males (Morgan et al., 2003), and hepatic antioxidant mRNA expression (SOD2 and CAT) at 20-months of age was different for the females selected for activity compared with controls, but not for males (Bronikowski et al., 2002). Metabolic rate positively correlated with liver antioxidant activity in both control and selected mice at all ages (Figure 7.4). This indicates that individual mice with a higher metabolic rate protected themselves against increased ROS production by increasing antioxidant enzyme activity. Without measures of ROS this hypothesis is unproven and the relationship between antioxidant enzyme activity and metabolic rate can also be explained by another, equally tenable, hypothesis; the disposable soma theory (Kirkwood and Shanley, 2005), This theory suggests that protecting the soma is costly hence it is traded off against other effects. If indeed the costs of protection are high, increasing protection levels could cause a direct metabolic cost. The difference in metabolic rate between control and selected mice was small, and no significant difference in antioxidant enzyme activity was found between the groups. The small difference in metabolic rate may relate to the observation that mice selected for high wheel-running activity have increased running economy on a whole-animal basis, i.e. they spent less energy per distance trav- elled (Rezende et al., 2006). However, this effect was mainly seen in females and another study in male mice housed with wheels at 10, 20 or 30°C showed no differences in costs of transport between control and selected male mice (Vaanholt et al., 2007). In this study there was a significant increase in daily energy expenditure in activity-selected mice (Vaanholt et al., 2007). The measurements on daily energy expenditure in the current study were done in the respirometer (without wheels) and not under normal housing conditions (with wheels), which may have resulted in an underestimation of the daily energy expenditure in the home cage. If so, this may indicate that selected mice did not compensate for their higher metabolism with increased antioxidant enzyme activity. At present, we do not know whether the antioxidant system was affected in activity-selected mice in other tissues than the ones measured. 124 Chapter 7

Several studies have reported increased mortality or reduced life span following increased workload (energy turnover), e.g. hamsters Mesocricetus brandti (Lyman et al., 1981), kestrels Falco tinnunculus (Daan et al., 1996), honey bees Apis mellifera (Wolf and Schmid-Hempel, 1989), house flies Musca domestica (Yan and Sohal, 2000) and nematodes Caenorhabditis elegans (Van Voorhies and Ward, 1999). This is consistent with the “rate of living theory” (Pearl, 1928; Rubner, 1908) derived from inter-species comparisons, that postulates total energy turnover of homeotherms determines life span. Comparisons between species are not always straightforward (Perez-Campo et al., 1998; Tolmasoff et al., 1980), however, because several taxa, such as birds (Barja, 1998) and bats (Brunet-Rossinni, 2004), combine high metabolic rates with long life spans. Also, these comparative studies are complicated by the co-variation of traits with body mass and the lack of independence of the data due to a shared phylogenetic history (Speakman, 2005a). Several studies have shown that exercise has a beneficial effect on life span (Bronikowski et al., 2006; Holloszy, 1988; Navarro et al., 2004) and the beneficial effects of exercise on protein turnover rates and/or antioxidant enzyme activities might underlie this effect. Up- regulation of these systems would diminish the damage caused by ROS (that are produced in parallel with metabolic rate under certain conditions) and could thereby increase life span. In practice, we did not show any major effects of long-term voluntary exercise on either parameter. A correlation between metabolic rate and antioxidant enzyme activities was found, indicating that mice with a higher metabolic rate had more protection against ROS. Assuming that the amount of ROS produced increased linearly with increasing metabolism (with a gradient of 1 and intercept of 0), the net result would be a similar protection against oxidative stress in animals with a high or low metabolic rate. Indeed, as shown in Figure 7.5, the protection (enzyme activity per kJ) was similar at all metabolic rates. Selected mice had slightly higher metabolism without a change in antioxidant enzyme activity, which may have left them more vulnerable for damage caused by ROS. We calculated the protection by antioxidants per kJ energy expenditure and no differences in protection between the groups were shown. In summary, age strongly affected antioxidant enzyme activity, which showed a peak at 10 months of age in liver and a decline with age in heart. Protein synthesis rates also decreased with age in liver, and to a lesser extent in muscle. DEE and RMR correlated with antioxidant activity in the liver, showing that individual mice with high metabolic rates protect themselves against the increased ROS production by increased scavenging. Despite a small difference in metabolic rate, there was no difference in antioxidant enzyme activity between control and activity-selected lines of mice. Long-term voluntary exercise thus did not result in compensatory changes in antioxidant enzyme activities or protein synthesis rates. Antioxidants and protein synthesis in high-activity mice 125

Acknowledgements We thank Suzan Anderson and David Bremner for performing the analysis of fractional synthesis rates and Annemieke Meijer for performing the antioxidant enzyme activity measurements. Serge Daan is thanked for commenting on earlier manuscripts. L.M. Vaanholt was a recipient of an EU Marie Curie Training Site award (Mass School) to the Rowett Research Institute, Aberdeen, Scotland. TG was supported by US NSF grant IBN-0212567.

Chapter8

Ageing under cold conditions: effects on body composition, metabolism and longevity

Lobke M. Vaanholt, Serge Daan, G. Henk. Visser

Abstract The “Rate of Living” hypothesis proposes that variations in the rate of mass- specific energy expenditure are causally involved in senescence related mortality. Thus, increasing the daily rate of energy expenditure should decrease life span. We tested a prediction from the hypothesis by exposing mice to low ambient temperature, thereby aiming to increase energy expenditure. We compared groups of 60 mice each, housed at 22°C (warm, WW) and at 10°C (cold, CC) throughout their life. Precipitated death in the cold group would be consistent with the rate of living hypothesis but also with the hypothesis that life in the cold increases instantaneous mortality without an accumulating role for energy expenditure. We therefore included a third group of 60 mice exposed to 10°C early in life (age 2-15 months) and to 22°C afterwards (cold-warm, CW). Cold exposure increased overall daily energy expenditure, assessed by Doubly Labeled Water (DLW) by 46%. No differences in life span were shown between WW and CC mice (median life span was circa 830 days in both groups). Also exposure to cold only early in life did not affect life span compared to mice housed under warm temperatures throughout life (median life span was 751 days). In the cold, mice had lower body mass and fat content compared to mice at 22°C. No differences were found in UCP levels in BAT, WAT or muscle. Lifetime energy potential (calculated over the 90 percentile life span) was 65717 kJ in CC mice, 47941kJ in WW and 62550 kJ in CW, when expressed over the whole body metabolism. When expressed per gram body mass, lean dry weight or per gram metabolically highly active organ tissue (heart, liver, kidney, brain) these differences became more extreme. The study leads to a refutation of the rate of living theory.

Journal of Gerontology: Biological Sciences, In review. 128 Chapter 8

INTRODUCTION

The “Rate of Living” theory, proposed by Pearl in 1928, states that an increased rate of energy metabolism increases the rate of ageing and shortens life (Pearl, 1928). The theory was derived from a study by Rubner in 1908 who showed that the life- time energy potential (mass-specific energy metabolism times maximum life span) was size independent in animals over a wide range of body masses (Rubner, 1908). The energy per gram body mass that an animal spends in its life time was similar for a guinea pig and horse. In Rubner’s study humans were an exception. Humans spent more than predicted for their body mass, but even when including humans the variation in life-time energy potential was still much smaller than the observed variation in body mass. Rubner himself speculated that food intake rates per gram mass would eventually explain longevity. Thus, the original argument for the rate of living theory was based on interspecific allometric comparison. In comparing species we look at the product of evolution. Natural selection has acted on body mass, on energy metabolism and on life span in all species. It has been pointed out that in interspecific allometry whole-body rates of energy turnover tend to increase with a mass exponent of about 0.7 (and thus mass-specific exponent -0.3), while life span (and other time measures) increases with a mass exponent of about 0.3 (Kozlowski and Weiner, 1996; Daan and Tinbergen, 1997). Their product (i.e., Life- time Energy Potential, LEP) is then proportional to mass to the power 1, and thus mass-specific LEP scales to the body mass to the power 0. Effects of energy expenditure on life span are thus mass-independent. This interspecific proportionality clearly does not prove that one causes the other. The rate of living theory needs to be tested within species. Although there are studies supporting the theory within species (mice; (Johnson et al., 1963), Drosophila; (Loeb and Northrop, 1917), Kestrels; (Daan et al., 1996), honey bees; (Wolf and Schmid-Hempel, 1989)), several studies in rodents have yielded either no relationship or even a positive relationship between energy turnover and life span (Holloszy et al., 1985; McCarter et al., 1985; Holloszy and Smith, 1986; Navarro et al., 2004; Speakman et al., 2004). Speakman (2002; 2005a) noted a number of problems associated with such tests in the literature. Firstly, maximum life span is often used, but is not a valid measure of ageing, because it is measured in a single individual of the population and depends strongly on the sample size used. Using the maximum survival of a group of animals (top 5-10%) would be a more reliable measure. Secondly, a single measurement of basal metabolic rate is usually used to calculate the total energy spent in a lifetime. Basal metabolic rate represents the lowest metabolic rate of an animal in rest at thermo-neutral temperatures and only comprises a small part (40%) of the total energy budget, depending on environmental factors. A better estimate of lifetime energy expenditure would be total daily energy expenditure (DEE) under the normal housing conditions measured at different ages throughout life. A third problem arises when energy expenditure is expressed per gram body mass. Not all components of body mass contribute Ageing in the cold 129

equally to metabolic rate. Greenberg (1999; 2000) has pointed out that in an ideal test of the rate of living idea one should therefore express energy turnover rates not with respect to the whole body but with respect to the most relevant tissue, i.e., the metabolically most active organs (heart, liver, kidney, brain). A fourth concern is that experimental manipulation of energy turnover rate always involves the manipulation of something else (temperature, workload, nutrition etc.). This may make it hard to conclusively distinguish between proper attribution of the modified life span to energy turnover per se rather than to the conditions imposed. The solution to this problem may be found in letting the effects accumulate early in life, and study the natural mortality under standard conditions late in life. Since the rate of living theory implies a cumulative effect of energy turnover, we should expect the effects on precipitated death also if the manipulation is restricted to the early part of life, leaving conditions late in life unchanged. In the present study, we investigated the effects of elevated energy turnover rates on life span, and we have heeded the warnings from these considerations. We manipulated energy metabolism by exposing animals to cold (10°C) throughout life or only early in life, and studied the relationship between energy metabolism and life span. A large subgroup of animals (60 in each of the two experimental groups and in the warm control group) was left undisturbed throughout life to generate reliable survival curves. In a smaller subgroup of the population body composition and metabolic rate (resting metabolic rate and daily energy expenditure) were measured at four different ages to get good estimates of the life-time energy potential (LEP). The mechanism linking metabolic rate and ageing may lie in the inevitable production of free radicals during oxygen consumption, the so-called free radical theory of ageing, postulated half a century ago by Harman (Harman, 1956; Beckman and Ames, 1998). Free radicals (or reactive oxygen species, ROS) can cause damage to macromolecules, that could eventually result in cell death. The amount of radicals that are produced during oxidative phosphorylation highly depends on the speed of the process affected, i.e., by the supply of substrates, and the amount of uncoupling that occurs. Uncoupling proteins (UCP) uncouple oxidative phosphorylation from ATP production and energy is subsequently lost as heat (Brand, 2000). Specifically in animals exposed to cold that need to produce extra heat to maintain body temperature, uncoupling might be beneficial because it would enable the animal to spend more energy with minimal production of ROS. We therefore included in the analysis an assessment of UCP expression in several tissues of mice at different ages.

MATERIAL AND METHODS

Animals & housing Male C57BL/6JOlaHsd mice, 4 weeks of age, were obtained from Harlan Nether- lands BV, Horst. Animals were housed in groups of three at 22°C until the age of 2 130 Chapter 8

months when they were housed individually and at different ambient temperatures. At this time all mice were divided randomly over three experimental conditions. The first (control) group of mice was housed at 22°C throughout their lives (warm; WW), the second group, was housed at 10°C from the age of 2 months onwards until their spontaneous death (cold; CC) and the third group was housed at 10°C from 2 till 15 months of age and at 22°C from age 15 months onwards (cold-warm; CW). Animals were housed in Macrolon® type II cages (UNO Roestvaststaal, Zevenaar, NL) with Hemparade® (HempFlax B.V., Oude Pekela, Netherlands) as bedding material and EnviroDry® (BMI, Helmond, Netherlands) as nesting material. Animals had ad libitum food (Standard rodent chow, RMH-B, Hopefarms BV, Woerden, NL) and water and were on a 12:12 light-dark cycle throughout their lives. Body mass of all animals was measured once a month. Each experimental group consisted of 100 mice which was randomly split into two subgroups of 60 and 40 mice. The 60 were left undisturbed throughout their lives (except for cage cleaning and monthly measurements of body mass) and the time of spontaneous death was noted to construct survival curves (life span subgroup). Of the other 40 animals we used 8 mice at each of four ages (3, 11, 19 and 27 months) to measure food intake and metabolic rate and to collect tissues and determine body composition (test subgroup, see below). At the later ages when mortality had occurred, sample sizes were smaller than 8 in some of the groups (see samples sizes given below tables 8.2 and 8.3).

Test subgroup Food intake and metabolic rate were measured in each age sample of the test subgroup at the ages 3, 11, 19 and 27 months. Food intake (g d-1) was measured over a period of 3 days. To express food intake in metabolizable energy intake we corrected for changes that occur in wet food mass due to evaporation of water or uptake of water from the air. A tray with food was put in the room of which the weight change was measured over the same three day period. In addition, a sample was taken from the food that was dried to constant weight over 4 hour in an oven at 103°C (ISO 6496-1983(E)), to enable comparison between food intakes at different ages. Daily energy expenditure (kJ d-1) was measured for each mouse with the doubly labeled water technique. Before each trial, the mouse was weighed on a balance to the nearest 0.1 g. Thereafter it was injected with about 0.1 g doubly labeled water (2H and 18O concentrations of the mixture 37.6% and 58.7%, respectively) allowing an equilibration period of 1 hour. The dose was quantified by weighing the syringe before and after administration to the nearest 0.0001 g. After puncturing the end of the tail, an “initial” blood sample was collected and stored in 3 glass capillary tubes each filled with about 15 µl blood. These capillaries were immediately flame-sealed with a propane torch. Thereafter the mouse was put back in its home cage. After 48 hours the animal was weighed again and a “final” blood sample was collected as described before. Per sampling period, we collected blood samples of 4 Ageing in the cold 131

mice which had not been injected with DLW, to assess the natural abundances of 2H and 18O in the body water pools of the animals. The determinations of the 2H/1H and 18O/16O isotope ratios of the blood samples were performed at the Centre for Isotope Research employing the methods described in detail by Visser and Schekkerman (1999) using a SIRA 10 isotope ratio mass spectrometer. In brief, each capillary was microdistilled in a vacuum line. The 18 16 O/ O isotope ratios were measured in CO2 gas, which was allowed to equili- brate with the water sample for 48 h at 25ºC. The 2H/1H ratios were assessed from H2 gas, which was produced after passing the water sample over a hot uranium oven. With each batch of samples, we analysed a sample of the diluted dose, and at least three internal laboratory water standards with different enrichments. These standards were also stored in flame-sealed capillaries and were calibrated against IAEA standards. All isotope analyses were run in triplicate. The rate of CO2 produc- -1 tion (rCO2, moles d ) for each animal was calculated with Speakman's (1997) equation: rCO2 = N/2.078 * (ko - kd) - 0.0062 * N *kd

-1 - where N represents the size of the body water pool (moles), ko (1 d ) and kd (1 d 1) represent the fractional turnover rates of 18O and 2H, respectively, which were calculated using the age-specific background concentrations, and the individual- specific initial and final 18O and 2H concentrations. The value for the amount of body water for each animal was obtained from the carcass analyses. Finally, the rate of CO2 production was converted to energy expenditure assuming a molar volume -1 of 22.4 l mol and an energetic equivalent per l CO2 based on RQ measurements in -1 our respirometry setup (on average 22 kJ l CO2, (Gessaman and Nagy, 1988)). Three days after the DLW measurements, the same mice were moved to our 8- . -1 channel open flow respirometry system and CO2 production (V CO2, l h ) and O2 . -1 (V O2, l h ) consumption were measured in concurrence with ambient temperature and activity (passive infra-red, PIR) (described earlier by Oklejewizc et al. (Okleje- wicz et al., 1997)). Oxygen and carbon dioxide concentration of dried inlet and outlet air (drier: molecular sieve 3 Å, Merck) from each cage was measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and carbon dioxide by an infrared gas analyzer (Servomex 1440). The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic cages. Flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks). Data were collected every 10 minutes and automatically stored on a computer. All mice were measured for 24h at an ambient temperature of 22°C. Mice that were housed in the cold were then also measured for 24h at 10°C. This enabled us to determine whether changes had occurred in metabolic rate between the groups (when measured at a similar temperature) and to determine the resting metabolic rate that the mice experienced under the experimental conditions. Metabolic rate (MR, kJ h-1) was calculated using the following equation: MR= . . 16.18 x V O2 + 5.02 x V CO2 (Romijn and Lokhorst, 1961; Gessaman and Nagy, 132 Chapter 8

1988). Using this versatile equation we could accurately estimate the heat production of mice under various nutritional states. Resting metabolic rate (RMR, kJ d-1) was defined as the lowest value of heat production calculated as the running mean over half an hour. After the respirometry measurements, animals were weighed and sacrificed using CO2 gas followed by decapitation. Trunk blood was collected in tubes with anticoagulant (EDTA), centrifuged at 4°C for 15 min at 2600 g, plasma collected and stored at -80°C for later hormone (corticosterone, leptin and adiponectin) analyses (analysed elsewhere; see Box 4.1). Heart, liver, kidneys, intestines, stomach, lung, brain, testes, hind limb muscles, brown adipose tissue, white adipose tissue, and skin were dissected out and each weighed to 0,0001 g. Subsamples of heart, liver, kidney, hind limb muscle, brown adipose tissue (BAT) and white adipose tissue (WAT) were immediately frozen at –80°C. The gut fill of stomach and intestines was removed and the samples were weighed again. Tissues were stored at –20°C until the water and fat content was determined. Water content was determined by drying for 4 hours to constant weight in an oven at 103°C (ISO protocol 6496-1983(E). Fat was extracted using a soxhlet and petroleumether and samples were dried again for 2 hours at 103°C. Dry lean masses of the organs of which sub samples had been taken, were calculated using the left over pieces with the assumption that sub samples taken contained the same proportions of fat and water. In the estimation of the dry lean mass of the remainder of the carcass we assumed that the leg muscle taken out consisted of protein only and that BAT and WAT consisted of fat only.

mRNA isolation and UCP expression mRNA extraction, hybridisation and detection was performed as described by Trayhurn et al. (Trayhurn et al., 1994). In short, mRNA was extracted from BAT, WAT and hind limb muscle and subjected to agarose gel electrophoresis. The RNA was transferred to a charged nylon membrane by Northern blotting and fixed with UV light. Prehybridisation was performed at 42°C in prehybridisation buffer (for details see (Trayhurn et al., 1994)). Hybridisation was at 42°C overnight in the buffer together with an oligonucleotide (25–50 ng ml-1). Membranes were then washed, incubated with blocking buffer containing a poly-clonal antibody against DIG, washed again and subsequently incubated with the chemiluminescence substrate CDP-star (Roche, 25 mM stock). Membranes were exposed to film for 2–5 min at 37°C immediately after incubation with the chemiluminescence substrate. Probes, synthesised with DIG ligand, were: 5’ CGG ACT TTG GCG GTG TCC AGC GGG AAG GTG AT for UCP-1 in BAT, 5’ GTG GCA AGG GAG GTC ATC TGT CAT GAG GTT GG for UCP-2 in WAT, and 5’CCC TGA CTC CTT CCT CCC TGG CGA TGG TTC TG for UCP-3 in muscle. Blots were reprobed for 18S RNA (5’ CGC CTG CTG CCT TCC TTG GAT GTG GTA GCC G) to adjust for variations in RNA load- ing. Quantification of UCP mRNA levels was determined by creating density his- tograms in ImageJ and calculating the area below the curve. These values were Ageing in the cold 133

expressed as ratio’s relative to 18S data per gel. To enable comparisons between gels we blotted a reference sample on each gel and normalized data to that.

Data analysis General linear models were applied in SPSS for Windows (version 14.0). We tested for effects of group, age and interactions between group and age in cold and warm mice at the all ages measured using a two-way AN(C)OVA with group, age and groupxage as fixed factors. Subsequently differences between the three experimental groups at 19 and 27 months of age were analysed using one-way AN(C)OVA models with group as fixed factor. Survival curves were analyzed using the LifeTables option in the survival analysis in SPSS for Window was used to create life tables. Significance level was set at p≤ 0.05 and all tests were two-tailed.

RESULTS

Survival The survival curves of all groups are shown in Figure 8.1A, and a summary of the main descriptors of age at death is shown in Table 8.1. We calculated finite mortality rates (FMR) over intervals of 100 days using the following formula: FMR = 1- Ne/Nb (see (Krebs, 1994)), where Ne is the number of animals left over at the end of the interval out of Nb, the number at the start of the interval. The values are shown in Figure 8.1B. Both mean and median age at death were virtually identical for mice exposed to 10°C and 22°C throughout their life (CC and WW in Table 8.1). In mice that were exposed to cold only early in life, median age at death was 82 days less than the median in cold and warm mice (751 vs. circa 833 days; Table 8.1). This appears to be due to increased mortality in the CW group between 600 and 800 days of age (i.e., 150-250 days after the switch in ambient temperature, see Figure 8.1B). Maximal survival was similar in all groups. We tested statistically for differences between groups using the Wilcoxon (Gehan) test in life tables of the survival analysis of SPSS for windows (version 14.0). The Wilcoxon (Gehan) test compares survival distributions between groups. Overall and pair-wise comparisons showed no significant differences between any of the groups.

Development of body mass Figure 8.2 shows the development of body mass for the three groups. Initially CC mice had a similar growth rate as WW mice. After approximately 300 days CC mice stopped growing, whereas WW mice kept increasing in body mass up till about 500 days of age. When the temperature was switched from 10 to 22°C, CW mice immediately increased their body mass and reached a plateau at a level intermediate between cold and warm mice. For statistical analyses average body mass during three intervals was calculated (0–250 days, 251–500 days, 501–750 days) and we 134 Chapter 8

100 A

% alive 40

1.0

0.8

0.6

0.4

finite mortality rate 0.2 B 0.0 0 200 400 600 800 1000 1200 age (d)

Figure 8.1. Effect of cold-exposure on mice survival (a) and finite mortality rates (b; see text for formula). White circles represent mice housed at 22°C (WW), black circles represent mice housed at 10°C (CC), and grey circles are mice housed at 10°C early in life (up to 15 months) and at 22°C later on (CW).

Table 8.1. Survival data.

Group n Mean (d) sem Median (d) sem 90% (d) Max (d)

WW 57 801 25 832 21 1002 1071 CC 57 798 21 834 19 939 1053 CW 58 768 27 751 19 1035 1121

Mean age at death (±sem), median life span (50%), 90 percentile and maximum age at death in mice exposed to various ambient temperatures is shown in days.

tested for differences between groups at different ages using one-way ANOVA. At all ages the warm animals had significantly higher body mass compared to the CW and cold mice. In addition, in the oldest age group (501-750 days) the CW mice had significantly higher body mass compared to the cold mice (p<0.05). Ageing in the cold 135

20 body mass (g) 10

0 0 200 400 600 800 1000 age (d)

Figure 8.2. Development of body mass in mice exposed to warm (white circles) or cold (black circles) ambient temperature throughout life, and in mice exposed to cold only early in life and at warm temperatures later on (grey circles).

Body composition At four ages a subgroup of mice from each of the three treatment groups was sacrificed and body composition determined (see Table 8.2). It turned out that the significant reduction in body mass of mice exposed to cold compared to warm mice was due to reduced fat mass. Fat free mass and dry lean mass were even significantly increased in cold mice compared to warm mice (see Table 8.3 for statistical analysis). Several changes in organ mass occurred in mice exposed to cold relative to WW mice. In tests of significance between cold and warm mice we always included body mass as a covariate. Thereby significance only refers to changes in the proportion of organs relative to body mass. Cold exposed mice had significantly increased relative heart and kidney weights compared to warm mice, while relative skin mass was significantly decreased in cold mice. Relative liver, stomach, intestines and lung masses did not differ significantly between cold and warm mice. In the models applied we also evaluated effects of age on relative organ masses. All organ masses, except for the brain showed a significant increase with age. For heart and kidney mass a significant interaction between group and age was also found, indicating that heart and kidney mass increased more rapidly with age in cold mice than in warm mice (see table 8.3). To see whether body composition of mice that were exposed to cold early in life and housed at normal conditions later was more similar to cold or warm mice one- way ANOVA with a factor group was applied on data on body composition from 19 and 27 month old animals. In most cases CW mice showed an intermediate phenotype to warm and cold mice. Significant effects of group were found for relative dry lean, fat, heart, kidney and skin mass (p<0.01). In these cases organ masses in the CW mice were similar to those of warm mice and differed from those in cold mice. 136 Chapter 8 °C until the age of 15 months and 3111927 Body composition of mice housed at different ambient temperatures various ages. est (g) 17.0±0.3 17.0±0.1 24.4±1.6 20.7±0.4 22.9±0.3 20.1±0.6 24.3±1.7 21.1±2.1 18.0±0.5 18.4±1.1 able 8.2. at free mass (g)at mass (g) 22.6±0.3 23.6±0.2 24.2±0.5 4.8±0.5 24.6±0.3 4.5±0.2 25.5±0.6 13.2±2.0 27.0±0.6 8.0±0.6 26.7±0.3 10.3±0.9 27.1±0.9 26.5±1.1 5.5±0.6 25.8±0.6 10.5±2.1 7.6±2.8 3.2±0.6 4.4±1.2 T Age (months) Groupn816816776667 Body mass (g)F Dry lean mass (g)F 27.4±0.4 WWHeart (g) 28.1±0.3 6.1±0.1Liver (g) CC 6.4±0.1 37.4±1.9Kidney (g)Brain (g) 32.5±0.5 7.0±0.1Stomach (g) 35.8±0.5 WWIntestines (g) 0.16±0.01 6.8±0.1Lung (g) 0.18±0.01 32.5±0.8 1.46±0.20 0.42±0.01Skin (g) 1.41±0.03 37.3±2.3 0.18±0.01 7.3±0.1 CC 0.48±0.01R 0.47±0.00 0.16±0.01 0.22±0.01 1.77±0.11 1.51±0.06 0.47±0.02 0.48±0.01 34.7±3.2 7.2±0.1 0.16±0.01 1.63±0.03 0.19±0.01 1.62±0.05Mean (SEM) weights of various components the body are shown for mice housed at 22°C (WW), 10°C (CC) and 10 0.55±0.01 29.7±0.9 0.48±0.01 0.19±0.01 0.28±0.03 7.5±0.3 WW 1.73±0.08at 22°C thereafter (CW). Measurements were done four ages: 3, 11, 19 and 27 months. Sample size (n) for each group is shown. 0.18±0.01 1.89±0.16 0.52±0.01 0.46±0.01 0.18±0.01 30.2±1.3 0.21±0.01 1.74±0.07 0.20±0.01 1.77±0.04 0.59±0.01 3.7±0.2 6.9±0.3 1.78±0.13 0.48±0.01 0.18±0.01 0.24±0.01 0.57±0.01 0.21±0.01 CC 1.84±0.09 0.46±0.01 0.19±0.01 0.28±0.01 1.84±0.08 3.7±0.1 6.6±0.2 0.20±0.01 1.87±0.11 0.55±0.03 0.49±0.01 0.20±0.01 0.23±0.01 1.74±0.05 1.99±0.10 0.66±0.03 0.21±0.02 6.4±0.2 1.62±0.13 0.48±0.01 0.24±0.02 CW 5.5±0.5 0.58±0.01 0.23±0.01 2.16±0.07 0.48±0.00 0.23±0.02 0.23±0.02 2.39±0.34 4.1±0.1 0.48±0.01 0.20±0.01 2.16±0.07 0.25±0.02 WW 5.2±0.2 0.25±0.01 0.22±0.01 3.9±0.1 CC 5.2±0.4 CW 4.9±0.9 3.1±0.2 3.6±0.4 Ageing in the cold 137

Table 8.3. Statistics on body composition: testing for differences between cold and warm mice.

Group Age Group x Age Covariate Trait d.f. F p d.f. F p F p p

Body mass (g) 1,66 15.4 <0.001 3,66 21.1 <0.001 3.4 0.024 none Fat free mass (g) 1,65 7.2 0.009 3,65 16.9 <0.001 1.0 - BM 0.006 Dry lean mass (g) 1,65 12.1 0.001 3,65 8.7 <0.001 0.2 - BM <0.001 Fat mass (g) 1,65 7.0 0.010 3,65 17.6 <0.001 1.0 - BM <0.001

Heart (g) 1,65 46.9 <0.001 3,65 21.2 <0.001 3.7 0.016 BM - Liver (g) 1,65 0.2 - 3,65 2.2 - 0.5 - BM 0.001 Kidney (g) 1,65 104.9 <0.001 3,65 36.9 <0.001 4.6 0.005 BM <0.001 Brain (g) 1,65 1.6 - 3,65 0.9 - 1.5 - BM - Stomach (g) 1,65 0.3 - 3,65 20.8 <0.001 0.6 - BM - Intestine (g) 1,65 0.4 - 3,65 10.0 <0.001 0.8 - BM - Lung (g) 1,65 0.8 - 3,65 9.6 <0.001 1.2 - BM - Skin (g) 1,65 12.2 0.001 3,66 11.1 <0.001 1.3 - BM <0.001 Rest (g) 1,65 1.8 - 3,65 6.3 <0.001 1.0 - BM <0.001

Tw o -way ANCOVA was applied to data on body composition to test for differences between cold (CC) and warm (WW) mice and look for effects of age. Group, age and group x age were added to the models as fixed factors and body mass (BM) as a covariate (except when testing for differences in body mass). F and p-values are shown for all fixed factors and p-values for the covariate are shown. Absolute values were used in the analysis. Total sample size was 74, with 8, 8, 7 and 6 mice at 3, 11, 19 and 27 months respectively in warm mice, 16, 16, 7 and 6 in cold mice. Significant p-values (p≤ 0.05) are shown in bold. d.f.=degrees of freedom. See text for analysis of differences between the three groups (WW, CC and CW) at 18 and 26 months of age (One-way ANOVA).

Food intake and energy expenditure Figure 8.3 shows food intake, resting metabolic rate and daily energy expenditure for the different experimental groups and Table 8.4 gives a summary of the statistical analyses done comparing warm and cold mice. Food intake was significantly increased in mice exposed to cold compared to warm mice and increased with age. At 3 months of age no differences in food intake were shown, but food intake was strongly increased in cold mice at 11, 19 and 26 months of age compared to warm mice. Resting metabolic rate (RMR) of cold exposed mice was measured at two ambient temperatures; at the experimental temperature (10°C, RMREXP) and at the temperature the warm group was housed in (22°C, RMR22°C), to enable comparison of RMR at a similar temperature. In the warm group resting metabolic rate was only measured at the control temperature (22°C). Comparing measurements at 22°C (RMR22°C) in both groups, we found no significant differences in RMR between CC and WW mice (Table 8.4, data not shown). Age did affect RMR22°C and was slightly increased at 11 months of age compared to the other ages. Under experimental conditions, RMREXP (10°C for CC mice and 22°C for WW mice) was increased by approximately 60% in CC mice compared to WW mice (see Figure 8.3). Age affect- 138 Chapter 8

8 ) 0.25 -1 d ) -1

-1 0.20 6 0.15 4 0.10

food intake (g d 2 0.05 metabolic rate (kJ g 0 0.00 80

) 2.5 -1 ) d -1

60 -1 2.0

1.5 40 1.0 20 0.5 metabolic rate (kJ d metabolic rate (kJ g 0 0.0

) 2.5 -1 ) d -1

-1 2.0 60 1.5 40 1.0 20 0.5 metabolic rate (kJ d metabolic rate (kJ g 0 0.0 0 200 400 600 800 0 200 400 600 800 age (d) age (d)

Figure 8.3. Absolute (left graphs) and mass-specific (left graphs) food intake, resting metabolic rate and daily energy expenditure in mice exposed to cold (10°C, black circles) or warm (22°C, white circles) environments at different ages (3, 11, 19, 27 months). Grey circles represent mice at 19 and 27 months of age that were housed at cold temperatures up to 15 months and at warm temperatures afterwards.

ed RMREXP in both groups, but RMR decreased slightly faster with age in the cold mice (as shown by a significant interaction effect; see Table 8.4). Daily energy expenditure (DEE) was measured in the home cage of the animals (CC mice at 10°C and WW mice at 22°C) and was sharply increased in CC mice compared to WW mice, and decreased significantly with age in both groups. With one-way ANOVA we tested for differences between all three groups at ages 19 and 27. There was a significant effect of group on food intake, DEE and RMREXP (See Figure 8.3, p<0.001). Animals in the CW group differed significantly from CC mice and had similar levels of energy expenditure as mice in the WW group. Ageing in the cold 139

Table 8.4. Tw o -way ANCOVA on food intake and metabolic rate: testing for differences between cold and warm mice.

Group Age Group x Age Covariate Trait d.f. F p d.f. F p F p p

Food intake (g d-1) 1,64 16.9 <0.001 3,64 8.0 <0.001 4.0 0.011 BM - -1 RMR22°C (kJ d ) 1,65 0.03 - 3,65 11.9 <0.001 1.5 - BM 0.002 -1 RMREXP (kJ d ) 1,65 132 <0.001 3,65 11.3 <0.001 3.8 0.015 BM 0.027 DEE (kJ d-1) 1,66 237 <0.001 3,66 14.6 <0.001 1.9 - BM -

Tw o -way ANCOVA was applied to data on metabolic rates to test for differences between cold (CC) and warm (WW) mice and look for effects of age. Group, age and group x age were added to the models as fixed factors and body mass (BM) as a covariate. F and p-values are shown for all fixed factors and p-values for the covariate are shown. Absolute values were used in the analysis. Total sample size was 75, with 8, 8, 7 and 7 mice in the warm group and 16, 16, 7 and 6 in cold group at 3, 11, 19 and 27 months respectively. One mice (WW, age 27) died during the respirometry measurements and total samples size is 74 here. For food intake sample size is 73 because data for 2 mice was missing. RMR22°C compares RMR measured at 22°C in both the WW and CC group. RMREXP compares values measured at experimental conditions; 22°C for WW mice and 10°C for CC mice. Significant p-values (p≤ 0.05) are shown in bold. d.f.=degrees of freedom. See text for analysis of differences between the three groups (WW, CC and CW) at 18 and 26 months of age (One-way ANOVA).

Uncoupling proteins (UCP) mRNA expression of UCP1, UCP2 and UCP3 was measured in BAT, WAT and hind- leg muscle respectively in 3 and 27 months old mice (see Figure 8.4). In all three tissues and at both ages, UCP expression was on average higher CC mice compared to WW and CW mice, but not significantly so (Figure 8.4).

Life-time energy potential Traditionally, the life-time energy potential (LEP, kJ) has been estimated on the basis of measurements of resting metabolic rate and maximum life-span. Life is, however, not passed solely in the resting state, and the life-time energy potential should equal DEE times life span. RMR might be used as an estimator of DEE, but only if DEE and RMR have a fixed ratio. In most animals that is obviously not the case. In the mice in this study, RMR was on average 92% of DEE in WW, and 86 % in CC mice. These differences, although small, call for the use of DEE to calculate LEP as a more accurate estimate of the life-time energy potential. Maximum life span represents only a single event in the colony and is therefore subject to large variance and highly dependent of the sample size that is used. Using the 90% mortality yields a more reliable measure of life span (see also (Speakman et al., 2002)). Taking these considerations into account we estimated LEP based on measurements on DEE (with DLW) and the age at which 90% of the animals had died. To incorporate changes that occur in DEE with age, we calculated the average DEE per group based on measurements at 4 ages throughout life (3, 11, 19 and 27 months, see Table 8.5). On the basis of the rate of living theory we expect CC mice to have the shortest life span, followed by CW mice and WW mice. We focus here on the com- 140 Chapter 8

0.5 brown adipose tissue - UCP1

0.4

0.3

0.2

0.1

0.0 1.0 white adipose tissue - UCP2

0.8

0.6

0.4

RNA expression 0.2

0.0 0.8 muscle - UCP3

0.6 Figure 8.4. mRNA expression of various uncoupling proteins (UCP1, 2 and 3) in brown 0.4 adipose tissue (BAT), white adipose tissue (WAT) and muscle repectively. White bars are mice housed at warm temperature, black bars 0.2 are mice housed at cold temperatures and grey 0.0 bars are mice that were switched from cold to 3 27 warm temperatures at 15 months of age. age (month)

parison between CC and WW mice. CC mice had a slightly shorter life span than WW mice. LEP on a whole-animal basis was 47941 in WW and 65717 kJ in CC mice respectively. LEP on a whole-animal basis was thus considerably higher in CC than WW mice. Body mass is known to affect metabolic rate and to be able to compare LEP between animals, LEP should be expressed per gram body mass. LEP per gram body -1 mass (LEPBM) was 1416 and 2138 kJ g in WW and CC mice respectively. When corrected for body mass CC mice thus still spent about 50% more energy in their life-time. Total body mass contains water and fat that are not metabolically active, and dry lean mass may be a better representative of the metabolically active tissue in the -1 body. We calculated LEP per gram dry lean mass; LEPDL was 7029 and 9748 kJ g respectively. Dry lean mass still includes matter such as skeleton and skin that is metabolically rather inactive. Greenberg (1999) and Lynn (1992) have therefore proposed that we should even go one step further and express LEP relative to the Ageing in the cold 141

Table 8.5. Life-time energy potential.

WW CC CV (%) Sign. CW

Body mass (g) 33.8 30.7 32.0 Total dry lean mass (g) 6.8 6.7 6.8 Organ mass (dry lean, g) 0.61 0.65 0.63 DEE (kJ d-1) 47.8 70.0 0.001 60.4 -1 -1 DEEBM (kJ g d ) 1.4 2.3 0.001 1.9 -1 -1 DEEDL (kJ g d ) 7.0 10.4 0.001 8.9 -1 -1 DEEOM (kJ g d )78108 0.001 97 Max. Life span (90%, days) 1002 939 1035 LEP (kJ) 47941 65717 22 0.001 62550 -1 LEPBM (kJ g ) 1416 2138 29 0.001 1953 -1 LEPDL (kJ g ) 7029 9748 23 0.001 9224 -1 LEPOM (kJ g ) 78467 101119 21 0.001 99928

Life-time energy potential (LEP; kJ Live-1) is the product of energy expenditure and life span and was calculated using average daily energy expenditure (DEE, kJ d-1) measured at 4 ages (2, 10, 18 and 26 months) throughout life in the test group, and maximum life span (90 percentile) measured in the life span animals. In addition, LEP (kJ g-1 Live-1) was corrected for various measures of body composition measured at the same ages: BM; body mass, DL; dry lean mass, OM; organ mass (sum of dry lean heart, liver, kidney and brain mass). CV represents the coefficient of variation calculated over the two groups (s.d. divided by mean x100%). Sign. shows the p-values for the two-way ANOVA performed to look at differences between the CC and WW mice (see text for detailed description). For com- pleteness values for CW mice are also shown in the last column.

mass of the most metabolically active organs: the heart, liver, kidney and brain. We calculated the sum of the dry lean weight of the heart, liver, kidney and brain (Organ mass, OM; see Table 8.5), and calculated the LEP per gram of organ mass -1 (LEPOW): 78467 and 101119 kJ g in WW and CC, respectively. Table 8.5 provides a summary of these results. The problem with these comparisons is obviously that we have only a single measure for life span in each group. We can thus not readily test for differences between groups in average individual LEP. We do however have individual values for total and mass-specific (per dry lean and wet mass) DEE in each group. We can test the group averages against each other, both before and after multiplying all individual data with the group’s life span. The basic data as well as the results of testing are supplied in Table 8.5. We applied two-way ANOVA with a factor group, age and the interaction group x age on whole-body and mass-specific DEE. After- wards we estimated LEP for each individual, by multiplying the mass-specific value of DEE with the maximum life span (90%) of its group and ran the same statistical test. CC mice had significantly (ANOVA, F1,67=192.8, p<0.001) increased DEE and LEP and this increase in energy expenditure remained apparent when values were expressed relative to body mass, dry lean mass or organ mass. 142 Chapter 8

DISCUSSION

Despite a 50% increase in daily energy expenditure throughout life, no difference in survival was found between cold and warm mice. These results cast doubt on the rate of living theory that states that increased mass-specific metabolic rate reduces life span (Rubner, 1908; Pearl, 1928). However, a more precise formulation of the theory would have to take changes in body mass and composition into account. In Table 8.5 we have calculated the life-time energy potential (LEP) for all groups, based on measurements of daily energy expenditure and the 90 percentile life span. Rubner (1908) found that LEPBW (calculated using measures of food intake and maximum LSP, 100%) was size invariant for a variety of species. In our inter-specific comparison, LEPBW clearly increased due to life in the cold without shortening life. Expressing metabolic rate relative to dry lean mass, or metabolic (organ) mass did not change this conclusion. Previous studies investigating the effects of cold-exposure on life span in rats have found conflicting results. Kibler et al. (Kibler and Johnson, 1961; Johnson et al., 1963; Kibler et al., 1963) have shown that rats kept at 9°C continuously had shorter life span than control rats (28°C), and these studies have been cited as providing experimental support for the rate of living theory. Holloszy and Smith (1986) argued that the continuous cold exposure in Johnson study was a chronic stressor and thereby will have deleterious effects on health and longevity (Paré, 1965), mediated by chronic elevation of stress hormones (e.g., corticosteroids), regardless of whether or not there was an increase in energy expenditure. To minimize the nonspecific effects of chronic stress and maximize the effects of increased energy expenditure they therefore immersed rats (to the upper border of their scapulae) in water kept at 23°C for 4 h d-1 for 5 days per week (Holloszy and Smith, 1986). In our study we applied a similar protocol to that of Kibler and found no differences in corticosterone level (unpublished results, see Chapter 10) or life span between cold-exposed or control mice. These results indicate that cold-exposure was not a chronic stressor; at least it did not result in increased levels of corticostrone, and the effects of cold exposure on life span were similar to those found in the cold- immersed rats of Holloszy. They found cold-immersed rats had increased food intake and metabolic rate compared to controls housed at 20°C, but did not differ in life span. These findings join a growing body of evidence suggesting that (in mammals) life-time energy expenditure per se does not underlie senescence (Holloszy and Smith, 1986; Holloszy and Smith, 1987; Speakman et al., 2003b; Navarro et al., 2004; Speakman et al., 2004) and refute the rate of living theory. Even thought there is no direct relationship between metabolic rate and life span, metabolic rate is known to play a important role in ageing via the production of reactive oxygen species (ROS; (Harman, 1956; Beckman and Ames, 1998)). ROS are produced in mitochondria during oxidative phosphorylation and can cause damage to lipids, DNA and proteins, which may eventually result in cell death. If ROS are increased in the cold, how did the mice in the cold protect themselves against Ageing in the cold 143

the increased oxidative stress they faced due to their high metabolic rates? Several defense systems can reduce the amount of ROS or reduce the damage they cause: e.g., uncoupling oxidative phosphorylation from ATP production, enhancing antioxidant enzyme activity, and increasing protein turnover to remove damaged proteins from the circulations. Uncoupling oxidative phosphorylation from ATP production by enabling protons to leave the intracellular space via uncoupling proteins reduces the production of ROS (Brand, 2000). This process might in principle enable mice to combine high metabolic rates with long life spans, as has indeed been shown in mice (Speakman et al., 2004). Specifically in cold-exposed animals it would be beneficial to uncouple oxidative phosphorylation, because when uncoupling occurs energy is dissipated as heat to become available for thermoregulatory purposes (Brand, 2000). We measured UCP activity in BAT, WAT and muscle at two ages and found a slight, but not significant increase in UCP expression in cold mice. We can thus not rule out the possibility that differences in UCP expression have been present at intermediate ages, but we have no proof to support the idea that animals in the cold increased uncoupling. Antioxidant enzymes such as superoxide dismutase, catalase and glutathion peroxidase can scavenge ROS before they cause damage to macromolecules. We measured activity of several antioxidant enzymes in heart and liver of these mice at various ages (Chapter 9). No differences in SOD or GPx activity in liver and heart were found between warm and cold mice, except for SOD activity in liver of 19 months old animals, which was lower in cold-exposed mice. These findings do not support the idea that animals in the cold up-regulate their antioxidant enzyme activity to reduce ROS. With a similar antioxidant capacity the protection per kJ was actually lower in the cold mice. We also found no evidence for an increase in protein turnover in these animals (Chapter 9). At present we thus do not know how the mice in the cold dealt with the increased oxidative stress to enhance survival. Some studies (Holloszy and Smith, 1986) have shown a reduction in the occurrence of tumors in mice exposed to cold, which might extend their life span. Cold-exposure may thus have positive effects on other aspects that influence survival and thereby counteract the negative effects of ROS on survival. If that is true, the metabolic rate itself is apparently not a good predictor of life span. In cold environments homeothermic animals face high metabolic demands to maintain body temperature (+50% in mice exposed to 10°C relative to mice at 22°C). Based on simple physical principles we might expect an adaptive increase in body mass in cold environments. In contrast to this expectation, the mice exposed to cold (10°C) had reduced growth and adult body mass relative to control mice (22°C). Body mass was strongly correlated with resting metabolic rate and reducing body mass in that way helped to conserve energy (see (Deerenberg et al., 1998; Speakman, 2005a) and Chapter 4). Fat mass and skin mass were also reduced in cold-exposed mice, which may have compromised insulation and led to further heat loss. Barnett et al (1965) showed decreased body mass, fat mass and skin mass in mice bred at –3°C, but this did not result in reduced insulation, due to thicker fur. We did not assess insulation but we did observe larger nests and less activity in the 144 Chapter 8

cold, which may have reduced heat loss (see also (Barnett, 1965; Rowland, 1984)). In wild rodents a similar response to cold environments is known (Klaus et al., 1988), but an increase in body (and fat) mass in winter was shown in field voles (Krol et al., 2005). The higher metabolic rate and food intake at low temperatures imposed an extra burden on some organs, while others were unaffected. Heart and kidney mass were increased relative to body mass in the cold, but no changes in liver, stomach or intestine mass occurred despite the strong increase in food intake in cold mice. Mice bred in the cold are also smaller, have less fat, lower skin mass and increased heart and kidney mass compared to mice bred at high temperature (25). The differences in body composition we found between cold and warm mice were maintained with age. In conclusion, mice exposed to cold ambient temperatures responded with a number of physiological adaptations; a decrease in body mass, steep increase in food intake and metabolic rate and changes in body composition (reduction in fat mass). Daily energy expenditure was increased by 46% (mass-specific; 64%) in cold-exposed mice, but this did not result in an altered life span compared to mice housed at normal ambient temperatures. Also, mice that were housed in the cold only early in life and that had an intermediate energy expenditure had a similar life span to the other groups. Lifetime Energy Potential (LEP) was calculated relative to body mass, dry lean mass and metabolic organ mass (heart, liver, kidney and liver) and in all ways expressed LEP was about 40% greater in mice housed under cold conditions relative to mice housed under warm conditions. This is in disagreement with predictions from the rate of living theory. This study thus joins a growing body of evidence suggesting that (within species) energy expenditure per se is not a good predictor of the ageing processes that ultimately determine life span.

Acknowledgements We thank Saskia Helder for taking excellent care of the animals, and Gerard Overkamp and Jacqueline Duncan for technical assistance and help with experimental procedures. Berthe Verstappen performed the isotope analyses. We thank John Speakman for enabling us to do the UCP measurements. We also thank Peter Meerlo, Kristin Schubert and Jan-Albert Manenschijn for their practical help at various stages of the project. L.M. V. was recipient of a Marie Curie Training Site Award to the Rowett institute, Aberdeen. S.D. is supported by EUCLOCK (EC 6th framework). 145

Chapter9

Protein synthesis and antioxidant capacity in ageing mice: effects of life-long cold exposure

Lobke M. Vaanholt, John R. Speakman, Gerald Lobley , G. Henk Visser

Abstract Substantial evidence supports a key role for reactive oxygen species (ROS) in causing cumulative damage to cellular macromolecules, thereby contributing to senescence. Antioxidants can scavenge ROS while protein turnover removes and replaces oxidized proteins. How these defence systems vary with age and with metabolic demand is not well known. In the present study 2H5-phenylalanine was injected into young (3 months) and old (27 months) mice chronically exposed to cold to explore effects of cold exposure on age-related changes in liver and muscle protein synthesis. In addition, effects of cold exposure (10∞C) on antioxidant enzyme activities were investigated in two metabolically active tissues in mice at various ages (3-27 months). Cold exposure did not affect fractional synthesis rates (FSR) in liver or muscle. FSR rates did decrease with age in both tissues, and in liver this occurred more rapidly in cold-exposed animals than in controls. Antioxidant enzyme activity (SOD and GPx) was also affected by age. SOD activity peaked in 11 month old mice followed by a decline, while GPx activity slowly declined with age. SOD activity in heart was unaffected by cold. In liver, SOD activity was decreased in cold-exposed animals, but GPx activity was not. No relationship between energy expenditure and enzyme activity was found. Cold exposure increased metabolic rate by approximately 40% with no concurrent increase in antioxidant enzyme activity. Cold-exposed animals did not up-regulate either protein synthesis or antioxidant enzyme activities to provide protection against extra production of ROS. 148 Chapter 9

INTRODUCTION

Reactive oxygen species (ROS) are produced as by-products of aerobic metabolism in mitochondria. They can cause damage to macromolecules (lipids, DNA and proteins) (Beckman and Ames, 1998; Davies et al., 1982; Mecocci et al., 1999; Tyler, 1975), and thereby contribute to senescence and several degenerative diseases associated with ageing (e.g. cardiovascular disorders, Parkinson disease) (McEwen et al., 2005; Melov et al., 1999; Wallace, 2005). An elaborate defence system consisting of endogenous antioxidant enzymes, such as catalase (CAT), superoxide dismutase (SOD), glutathione peroxidase (GPx), and numerous non-enzymatic antioxidant (vitamins, flavenoids), exist that scavenge ROS to prevent deleterious effects (Beckman and Ames, 1998). Antioxidant enzymes cannot scavenge all the ROS produced; a small part escapes conversion and can damage proteins. This will impact on essential functions within the cell, such as maintenance of the structural archi- tecture and enzyme activity. Indeed, the accumulation of oxidized proteins with age has been reported for many experimental ageing models (Stadtman, 2004). Such accumulation could be due either to increased generation of reactive oxygen species or to reduced elimination of oxidized proteins (Shringarpure and Davies, 2002; Stadtman, 2004). Protein turnover, the composite of protein degradation (removal of inactive or damaged proteins) and protein synthesis (replacement with new, undamaged proteins) plays a potentially vital role in the ageing process (Rattan, 1996; Ryazanov and Nefsky, 2002; Sohal, 2002). Maintaining high rates of protein synthesis in old age could diminish the accumulation of damaged proteins and increase cell survival. Indeed, protein turnover is increased in calorically-restricted animals (Lambert and Merry, 2000; Lewis et al., 1985), a nutritional condition that increases life span in several species, including mice (Weindruch et al., 1986), and rats (McCay et al., 1935), for recent reviews see (Masoro, 2005; Merry, 2005). Other factors can also influence protein synthesis and protein degradation, including cold exposure (rats: (McAllister et al., 2000; Samuels et al., 1996); calves: (Scott et al., 1993), chickens: (Yunianto et al., 1997)) and age (mice: (Blazejowski and Webster, 1983); rats: (Lewis et al., 1985) and humans: (Short et al., 2004; Young et al., 1975); for reviews see: (Dorrens and Rennie, 2003; Rattan, 1996; Van Remmen et al., 1995; Ward, 2000)). Protein turnover accounts for 20-35% of the resting metabolic rate at ther- moneutrality (Lobley, 1990; Newsholme, 1987). It has been hypothesised that protein turnover would increase in cold-exposed animals to cope with the increased demands on heat production (McAllister et al., 2000). Indeed, in hearts of rats exposed to cold for 21d an increase in protein synthesis rates was shown (McAllister et al., 2000), although contradictory results have also been reported (Lindsay et al., 1988; Scott et al., 1993; Yunianto et al., 1997). Cold-exposure increases metabolic rate and ROS production and this may counteract any beneficial effect of increased protein breakdown. Alternatively, animals may adapt to the extra Oxidative stress in cold-exposed mice 149

ROS production by increased activity of antioxidant enzymes and, in concert with elevated protein breakdown, protect cells from problems arising from the presence of oxidized proteins. Such protection may vary with age, however, due to altered responsiveness in protein turnover and antioxidant systems. The current study addresses some of these questions. First, the impact of cold- exposure on age-related changes in protein synthesis and the antioxidant enzyme system was monitored in C57BL mice exposed to 10°C and 22°C throughout life. This included comparisons between liver and muscle (protein synthesis) and liver and heart (anti-oxidant activity). Second, the relationship between metabolism and antioxidant enzyme activity between individual animals was examined.

MATERIAL AND METHODS

Animals & housing Male C57bl6J mice were obtained from Harlan Nederland B.V. at the age of four weeks and individually housed in standard cages (15x30x15 cm, Macrolon type II, UNO Roestvaststaal BV, Zevernaar, NL) with standard bedding (Hemparade®, HempFlax, Oude Pekela, NL; and Envirodry®, BMI, Helmond, NL). Food (Standard rodent chow RMB-H (2181), HopeFarms, Woerden, NL) and water were provided ad libitum. Animals were divided over two rooms and housed at an ambient temperature of 22°C (Control group; WARM) or 10°C (Cold group; COLD). Animals were tested and sacrificed at four ages: 3, 11, 19, and 27 months.

Tissue collection At each sacrifice age, five to eight mice per group were lightly anaesthesized with CO2 and then killed by decapitation. Samples of liver and heart were quickly removed and immediately frozen in liquid nitrogen and stored at –80°C for antioxidant measurements. Protein synthesis was assessed only in 3 and 27 month old animals. The latter were the same animals used for the antioxidant measurements, but for the 3 month old group different animals were used (n=8 per group). Food intake (g d-1) and body mass was measured for two consecutive days prior to the sampling of tissues for protein synthesis measurements. Protein synthesis was measured using the large-dose method as described by Garlick et al. (1980). Mice were given an intra- 2 peritoneal injection of 150 mM H5-phenylalanine (1,5 ml per 100 g animal). After 15 minutes the mice were euthanized using CO2, followed by decapitation. Trunk blood was collected in pre-chilled tubes with heparin as anti-coagulant. Blood samples were centrifuged at 2600 g at 4°C for 15 min, and the plasma was collected and stored at –80°C until analysis. Liver and hind-leg muscle were rapidly removed, weighed to 4 decimal places, rinsed in ice-cold saline, frozen in liquid nitrogen, and stored at –80°C until analysis. Exact times (nearest second) of injection and freezing of tissues were recorded. 150 Chapter 9

Protein synthesis Free and protein-bound enrichments of phenylalanine in liver and muscle tissues were quantified as described by Wester et al. (2004) (Wester et al., 2004). Approximately 300 mg of frozen tissue was homogenised on ice in 3 ml 7% (w/v) sulphosalicylic acid (SSA). Free phenylalanine was separated from protein-bound phenylalanine by centrifugation at 1000 g at 4°C for 15 minutes and the supernatant retained. The pellet was then washed three times with 3 ml 7% sulphosalicylic acid to remove free phenylalanine. The initial supernatant fraction (free pool) was passed through a 0.4 ml column of Dowex AG 50W-X8 (100-200 mesh) and the resin rinsed with 2x3.5 ml water before the phenylalanine was eluted with 2 ml 2M NH4OH and 1 ml water. The eluate was freeze-dried and stored at –20°C for later analysis. Half of the washed pellet (protein-bound pool) was transferred to a 8 ml screw-topped Pyrex hydrolysis tube and solubilised in 1 ml 0,5 ml 0,3 M NaOH for 30 minutes. A few phenol crystals were added and the sample was hydrolysed by adding 7 ml 4M HCl and heating on a dri-block at 110°C for 18 hours. Hydrolysates were dried under vacuum, resuspended in 1 ml 0,5 M sodium citrate (pH 6.2), and stored at -80°C until later analysis. For the plasma samples, 150 ml was treated with 150 ml 15% SSA, centrifuged at 1000 g at 4°C for 10 minutes and 150 ml of the supernatant was passed through a 0.2 ml column of Dowex AG 50W-X8. Elution conditions and subsequent treat- ments were similar to the tissue free pool samples. STable isotope enrichments of the tissue and plasma free pools were measured by gas chromatography mass spectrometry (GC/MS) after conversion to the tertiary-butyldimethylsilyl (TBDMS) derivatives (Calder and Smith, 1988). In the hydrolysed samples (protein-bound pool, low enrichment) phenylalanine was converted to phenylethylamine by enzymatic decarboxylation prior to forming the TBDMS derivative. This was separated by capillary column gas chromatography and enrichments obtained from electron impact ionization selective ion monitoring (EI- SIM) mass spectrometry (see Calder et al., 1992 and Slater et al., 1995). The fractional synthesis rate (FSR, % d-1) was calculated using the following equation: FSR = 100*(BP/FP)*1440/t, where BP is the bound pool of phenylalanine in mole percent excess (MPE), FP is the MPE of the free pool of phenylalanine measured in either plasma or tissue, and t is the time (min) between injection of phenylalanine and freezing of the tissue in minutes. The ratio between FP measured in plasma and liver or muscle was calculated to determine whether the injected phenylalanine had equally mixed with both free pools. The ratio was 1.04±0.07 and 0.99±0.08 (mean±sd) in muscle and liver respectively. Neither ratio was significantly different from unity. From these data it was concluded that the injected phenylalanine effectively equilibrated between tissue and plasma and remained so until time of death. In consequence, the FSR reported represent the FSR calculated based on plasma free phenylalanine as representative of the precursor pool. Oxidative stress in cold-exposed mice 151

Antioxidant enzyme activities & protein content Prior to enzyme activity determinations, tissue samples were homogenized by soni- cation in 20 volumes of ice cold 50 mM phosphate buffer. Following centrifugation (25 min at 3000g), the supernatant fraction was collected, divided over several tubes and stored at –80°C for enzyme activity and protein measurements. Total superoxide dismutase (SOD) activity was determined at 25°C by the inhibition of the auto-oxidation of pyrogallol by SOD in the supernatant, following the method of Marklund and Marklund (Marklund and Marklund, 1974). The reaction was followed spectrophotometrically at 420 nm in the following reaction mixture: 50 mM Tris-DTPA buffer, 15 µl supernatant and 15 µl pyrogallol in a total volume of 800 µl. Each triplicate measurement was preceded by a blank, containing only pyrogallol in Tris-DTPA buffer. One unit of SOD was defined as the amount of enzyme causing 50% inhibition of pyrogallol auto-oxidation. Glutathione peroxidase (GPx) activity was determined at 25°C via the oxidation of NADPH in the presence of reduced glutathione (GSH) and H2O2 (combining the assays of Paglia and Valentine (1967), and Lawrence and Burk (1976)). The following reaction mixture was used: 4.28 mM sodium azide (to block catalase activity), 1.07 mM EDTA, 4.286 mM GSH, 0.214 mM NADPH, 1 U ml-1 GR in ice cold 50 mM phosphate buffer. 25 µl H2O2 and 25 µl sample were added to the reaction mixture. Reactions were followed spectrophotometrically at 340 nm in a total volume of 700 µl. To correct for spontaneous oxidation reactions independent of GPx, blanks without H2O2 were measured and subtracted from the assay values. One unit of GPx was defined as the amount of enzyme that oxidized 1 µmol of NADPH per minute in the presence of reduced glutathione. Protein content of the supernatant fraction was determined using a Bradford assay (Quick start Bradford protein essay kit 2; Biorad Laboratories B.V., Veenendaal, NL).

Indirect calorimetry Prior to killing, resting metabolic rates (RMR) and daily energy expenditure (DEE) were measured for each animal in which antioxidant enzyme activity was measured, using an eight-channel indirect calorimetry system, as described by Oklejewicz et al. . (1997). The mice were put in airtight chambers where oxygen consumption (V O2, -1 . -1 l h ) and carbon dioxide production (V CO2, l h ) was measured simultaneously with ambient temperature and activity (passive infrared detectors). Oxygen and carbon dioxide concentrations of dried inlet and outlet air (drier: molecular siever 3 Å, Merck) from each chamber were measured with a paramagnetic oxygen analyzer (Servomex Xentra 4100) and carbon dioxide by an infrared gas analyzer (Servomex 1440). The system recorded the differentials in oxygen and carbon dioxide between dried reference air and dried air from the metabolic cages. The flow rate of inlet air was measured with a mass-flow controller (Type 5850 Brooks). Computerised data were collected every 10 minutes. All mice were measured for 24h at an ambient temperature of 22°C, mice from the cold group were measured for an additional 24h at 10°C. Oxygen consumption was calculated according the equation 2 of Hill 152 Chapter 9

(Hill, 1972) to correct for volume changes with respiratory quotient below 1 and expressed in standard temperature and pressure. Metabolic rate (MR, kJ h-1) was . . estimated using the following equation: MR= 16.18 x V O2 + 5.02 x V CO2 (Romijn and Lokhorst, 1961). Resting metabolic rate (RMR, kJ d-1) was defined as the lowest value of metabolic rate calculated from cumulative means over 30 minutes. Daily energy expenditure (DEE, kJ d-1) was calculated as the total metabolic rate over the 24-h measurement period.

Statistical analysis Results are reported as means ± SEM. To test for effects of treatment and/or age ANOVA models in the MIXED procedure in SAS for Windows (version 9.1) were applied. Group, age, and group x age were added as fixed factors. Factors that may have influenced the outcomes were examined by including these as covariates in the models (i.e., food intake, body mass). Adjusted means were calculated by using the least squares means command in SAS MIXED. Data on antioxidant enzymes were log10-transformed as necessary to improve normality. Significance was set at p≤ 0.05.

RESULTS

Body mass & Food intake Table 9.1 shows the characteristics of the mice housed under cold (10°C) and warm (22°C) conditions. Prior to the cold treatment there were no differences in body mass between groups (One way ANOVA: F1,29=1.71, p=0.279). Mice housed under cold conditions had lower body mass at 3 and 27 months of age and in both groups body mass increased with age. Food intake differed between groups and with age; it was increased in COLD and aged animals.

Protein synthesis Liver and muscle FSR decreased with age, i.e. between 3 and 27 months, but there was no main effect of COLD treatment (Table 9.1). In liver the age-related decrease was approximately 35% in COLD and 22% in WARM mice. These magnitudes differed as shown by a significant interaction effect between group and age. In muscle, the decreases in FSR with age averaged 26% and 39% in COLD and WARM mice, respectively. These responses were similar as shown by the lack of an interaction between group and age. Food intake can affect protein synthesis rates and was added to the models as a covariate to explore this relationship. In both liver and muscle food intake was not a significant covariate and adding food intake into the models as covariate did not substantially alter the effects of cold exposure or age. Oxidative stress in cold-exposed mice 153

Table 9.1. Effects of cold exposure on body mass, food intake, and fractional synthesis rates in liver and muscle.

3 months 27 months p-values Variable name WARM COLD WARM COLD df Group Age GxA n8856 Age (d) 98 98 833±1 836±1 Body mass (g) 28.9±1.2 26.6±1.2 37.3±1.5 30.7±1.4 1,23 0.003 <0.001 0.122 Food intake (g d-1) 3.0±0.4 5.1±0.4 5.4±0.5 6.7±0.4 1,23 <0.001 <0.001 0.303

FSR liver (% d-1) 63.6±2.0 68.0±2.0 49.8±2.5 44.5±2.3 1,23 0.841 <0.001 0.039 FSR muscle (% d-1) 4.3±0.3 4.0±0.3 2.6±0.4 2.9±0.3 1,23 0.967 <0.001 0.320

Results for ANOVA are given in addition to least square (adjusted) means±SE for all groups. Bold values represent significant results (p<0.05). One mouse from the control group at 27 months was removed, because results indicated that the phenylalanine injection was not performed properly. n= sample size per group, df= degrees of freedom, GxA= GroupxAge interaction.

Antioxidant enzyme activity & Metabolism Antioxidant enzyme activities in liver and heart of cold exposed and control mice are shown in Figures 9.1 and 9.2, respectively. The statistical analyses are shown in Table 9.2. Age affected SOD activity in both heart and liver. In both tissues, SOD activity increased between 3 and 11 months and decreased thereafter (Figure 9.1, top graph and Figure 9.2). The highest activity of hepatic GPx was observed at 3 months of age followed by a decline (Figure 9.1, bottom graph). Cold exposure did not affect SOD activity in heart or GPx activity in liver, but did affect SOD activity in liver. This effect was mainly caused by a decrease in hepatic SOD activity in COLD mice compared to WARM mice at 19 months of age (Tukey test: p<0.001). In consequence, the COLD mice showed a faster decrease in SOD activity between 11 and 19 months of age (shown by a significant interaction between group and age, Table 9.2). For WARM animals the major decrease in activity occurred between 19 and 27 months. Overall, resting metabolic rate (RMR, kJ d-1) was (mean±sd) 44.2±.5.0 and 60.8± 8.8 and daily energy expenditure (DEE, kJ d-1) was 57.1± 5.7 and 80.0± 5.7 in WARM and COLD mice respectively. RMR and DEE were highly increased at all ages in COLD mice (Two-way ANOVA, effect of group: F1,50=165, p<0.001 for RMR and F1,50=372, p<0.001 for DEE). RMR and DEE were added into the models as covariate to test whether they affected antioxidant enzyme activity. Neither RMR nor DEE predicted changes in enzyme activity within these models nor did they alter the effects of age or group. The first is also apparent in Figure 9.3 that shows the relationships between DEE and SOD activity in liver at all ages measured (none of which were significant; linear regression). Notice that variations in both metabolic rate and SOD activity were small in these inbred mice. 154 Chapter 9

250 warm cold-exposed 200

150 protein) -1 100 SOD activity (U mg 50

0 0.8

0.6 protein) -1

0.4 GPx activity 0.2 mol NADPH mg µ ( 0.0 3 11 19 27 age (month)

Figure 9.1. SOD (top graph) and GPx (bottom graph) activity in liver of mice exposed to cold (10°C) throughout their lives and of control mice (22°C) at different ages. One unit of SOD was defined as the amount of enzyme that causes 50% inhibition of pyrogallol auto-oxidation. One unit of GPx is defined as the amount of enzyme that oxidizes 1 µmol of NADPH per minute in the presence of reduced glutathione.

100 warm cold-exposed 80 protein) -1 60

SOD activity (U mg 0 3 11 19 27 age (month)

Figure 9.2. SOD enzyme activity in heart of cold-exposed (10°C) and control (22°C) mice at different ages. One unit of SOD was defined as the amount of enzyme that causes 50% inhibition of pyrogallol auto-oxidation. Values given are mean±sem. Oxidative stress in cold-exposed mice 155

Table 9.2. ANOVA on effects of group and age on antioxidant enzyme activities

Group Age Group x Age Trait N d.f. F p d.f. F p F p

SOD Liver 55 1,47 15.1 <0.001 3,47 50.7 <0.001 6.4 0.001 GPX Liver 53 1,45 0.2 0.47 3,45 18.9 <0.001 1.7 0.19 SOD Heart 57 1,49 0.5 0.68 3,49 7.2 <0.001 1.7 0.19

ANOVA models were performed in the MIXED procedure of SAS for Windows (version 9.1). Group, age and group x age were added as fixed factors. Data were log10-transformed as necessary to improve normality. N represents the total sample size and d.f. the degrees of freedom. Sample sizes were: at 3 months, n=8 and 7 respectively in warm and cold; 11 months, n=8; 19 months, n=7; 26 months, n=6. Liver samples of 2 mice were missing and for 2 other mice sample volume was too small to measure both SOD and GPx reducing the sample sizes for these measurements. Significant effects are in bold.

300 protein) -1 200

100 warm cold 3 month 11 month 19 month SOD activity liver (U mg 27 month 0 40 5060 70 80 90 100 daily energy expenditure (kJ d-1)

Figure 9.3. Relationship between daily energy expenditure and SOD activity in liver of warm (white symbols) and cold-exposed mice (black symbols) at various ages.

DISCUSSION

The primary purpose of this study was to examine the effects of cold exposure on age-related changes in protein synthesis and antioxidant enzymes in mice. In addition, the relationship between metabolic rate and antioxidant enzymes was explored. In accordance with previous reports on rodents (mice: (Blazejowski and Webster, 1983; Vaanholt et al., 2006), rats: (Lewis et al., 1985)),decreases in fractional protein synthesis rates with age in all experimental groups and both tissues were found. On average, FSR decreased approximately 30% between 3 and 27 months in both liver and muscle. Studies in humans, however, have shown conflict- 156 Chapter 9

ing effects of age on muscle protein synthesis, with either decreases (Short et al., 2004; Yarasheski et al., 1993; Young et al., 1975) or no differences (Sheffield-Moore et al., 2005; Volpi et al., 2001). Such discrepancies between studies may have many causes, including muscle type used, sex, diet and activity level of subjects (Dorrens and Rennie, 2003). The technique used may affect the response, as the large (flood) dose procedure in fasted humans may stimulate protein synthesis (Cuthbertson et al., 2005; Smith et al., 1998). In fed animals, however, such responses are not observed (Rocha et al., 1993) and this technique is routinely used in non-fasted rodent studies. Long-term cold exposure (at 10°C) did not change protein synthesis rates in muscle or liver at 3 and 27 months of age. This is in agreement with several studies that have shown no change in muscle protein synthesis rates after long-term cold- exposure (rats (McAllister et al., 2000; Samuels et al., 1996), calves (Scott et al., 1993), pigs (Lindsay et al., 1988)), but contradicts observations in chickens where muscle protein synthesis rates increased (Yunianto et al., 1997). The effects of cold exposure on protein synthesis are complex and probably depend on various factors, including the tissue measured, the intensity and the duration of the cold exposure, and the species studied. Furthermore, Scott et al (1993) have shown that (muscle) protein synthesis rates are influenced by food availability (Scott et al., 1993). In cold-exposed calves that were food restricted, protein synthesis rates in muscle and skin were reduced but these were unaffected when calves were not food restricted. In the current study with mice fed ad libitum there was no change in muscle protein synthesis in response to cold exposure. Probably the 30% increase in food intake in the cold-exposed mice was sufficient to maintain levels of protein synthesis in muscle. Age (or period of development) may also be an important factor in determining the effect of cold exposure on fractional synthesis rates. Lewis et al. (1985) studied effects of caloric restriction on whole-body protein turnover in rats at different ages and found that the most pronounced response in FSR occurred at 12 months of age (increased by 45% compared to 6% at 24 months in calorically restricted mice vs. controls) (Lewis et al., 1985). In the present study protein synthesis was only measured on 3 and 27 month old animals and at neither age was there an effect of cold exposure. Further studies would be necessary to determine if the mice were more sensitive at intermediate ages. It has been hypothesised that increased rates of tissue protein turnover may contribute to heat production in cold-exposed animals (McAllister et al., 2000) and this could have positive effects on survival. This is not supported by either previous reports (McAllister et al., 2000; Scott et al., 1993) or the current study. In contrast, protein synthesis rates in liver of the current cold-exposed animals decreased more with age than control mice kept at 22°C, and in muscle, the age-related change in protein synthesis rate did not differ between cold-exposed and control animals. Age had a strong effect on antioxidant enzyme activity of SOD and GPx in murine heart and liver. An increase in SOD activity with a peak at 11 months of age and a subsequent decline was observed in both tissues. GPx activity in liver showed Oxidative stress in cold-exposed mice 157

a steady decline in enzyme activity from the age of 3 months onwards. Conflicting data exist on the effects of age on antioxidant enzyme activities (Gunduz et al., 2004; Rao et al., 1990; Sohal et al., 1990; Tsay et al., 2000; Vaanholt et al., 2006), but comparison with other studies is complicated due to the use of different species, the choice of organs studied, and measurements at different ages (or only 2 ages). In studies where antioxidant enzyme activities were measured in mice heart and liver (see Chapter 7, Fig. 7.1 and 7.2) or rat brain (Tsay et al., 2000) at multiple ages throughout life a similar pattern with age (with a peak at 10-12 months) was found as we show here for SOD activity. These data highlight the importance of measuring antioxidant enzyme activities at various ages when exploring develop- mental responses. We believe this will resolve many of the current discrepancies that appear to exist between different studies. Long-term cold exposure increased metabolic rate by approximately 50% but did not affect antioxidant enzyme activity to early maturity (up to 11 months) in heart or liver. Data on SOD activity at 2 months of age are in agreement with those from voles of a similar age that had been bred and raised under cold conditions (~8°C) (Selman et al., 2000). In that study, GPx and CAT activity were elevated in the heart of the cold-exposed voles. Other studies provide contradictory evidence on the effects of cold-exposure (Davidovic et al., 1999; Kaushik and Kaur, 2003; Siems et al., 1999; Spasic et al., 1993) but again this may be a feature of the experimental conditions, including the tissue measured and the duration and/or intensity of the cold exposure. Age-dependent effects of cold-exposure on antioxidant enzyme activities have not been previously reported. Unexpectedly, the SOD activity in liver of cold-exposed mice (10°C) was, markedly decreased at 19 months of age compared with controls housed at 22°C. This may reflect either a higher susceptibility to oxidative stress or a lower production of free radicals at this age. Despite a 50% increase in daily energy expenditure in cold-exposed animals throughout life, no major compensatory changes were observed in the antioxidant system in heart and liver tissue in these mice. Rather, SOD activity in the liver was even decreased at 18 months of age. CAT activity was not measured, and this may show compensatory changes occurred, as is observed in other studies (Kaushik and Kaur, 2003; Selman et al., 2000). It has also been suggested that basal levels of SOD are sufficient to reduce the superoxide anion to hydrogen peroxide during moderate oxidative stress (Ji, 1999), and, if CAT was up-regulated, the available SOD and GPx might have been sufficient to cope with increases in radical production. Opposite to what we showed here (Figure 9.3), in exercising mice metabolic rate did significantly predict antioxidant enzyme activity (Vaanholt et al., 2006). The low between-individual variation for both variables in this study may explain this discrepancy. Increased protein turnover rates have been hypothesized to be an important factor contributing to the extension of life span in response to food restriction (Tavernarakis and Driscoll, 2002). The current study showed that protein synthesis and antioxidant enzyme activity decreased more steeply with age in the liver of 158 Chapter 9

cold-exposed animals. This may cause cold-exposed animals to be more susceptible to ageing, because decreased levels of antioxidants would diminish the protection against ROS while lowered protein turnover will increase the half-life of proteins, enabling damaged proteins to accumulate in the cell and cause potential malfunction. However, cold exposure had no effect on median life span in rats (Holloszy and Smith, 1986). This would suggest that other compensatory changes (e.g. uncoupling) to reduce oxidative stress in response to high metabolic rates occur in cold-exposed animals. Uncoupling of mitochondrial respiration would be beneficial, particularly in cold-exposed animals, because energy is then dissipated as heat and this would markedly reduce oxidative stress (Brand, 2000; Erlanson-Albertsson, 2003; Speakman et al., 2004). In summary, long-term cold exposure did not result in compensatory changes in antioxidant enzyme activities in the heart and liver of mice or in protein synthesis rates in liver and muscle. Age strongly affected antioxidant enzyme activities and these showed either a peak at 11 months (SOD) or a gradual decline with age (GPx). Fractional protein synthesis rates also showed a decline with age in both liver and muscle. Numerous studies have shown that oxidative damage increases with age, and a decrease in antioxidant enzyme activity and/or protein turnover could explain this effect. Metabolic rate did not predict SOD and GPx levels in the animals in this study. Cold-exposed animals may have compensated for the high metabolic rates required to maintain constant body temperature by increasing the expression of uncoupling protein, thereby dissociating oxidative respiration from ATP production and reducing the generation of free radicals. Further study is required to determine whether this is the case.

Acknowledgements We thank Suzan Anderson and David Bremner for performing the analysis of fractional synthesis rates and Annemieke Meijer for performing the antioxidant enzyme activity measurements. Serge Daan is thanked for commenting on earlier versions of the manuscript. L.M. Vaanholt was a recipient of an EU Marie Curie Training Site award to the Rowett Research Institute (Mass school), Aberdeen, Scotland. 159

Chapter10

General Perspective

Lobke M. Vaanholt 162 Chapter 10

The role of energy expenditure in ageing: “The rate of living theory”

Almost a century ago Rubner (1908) calculated the amount of food, expressed in calories, eaten over the life span by several domestic species of mammal (guinea pig, cat, dog, horse and cow). He was struck by the fact that these amounts were roughly proportional to the typical body mass of the species, so that the ratio, i.e., the caloric intake per lifetime per gram body tissue was little different between guinea pig and cow. These data led to the rate of living theory that was formulated in 1928 by Pearl based on Rubners work and on his own observations (Pearl, 1928). The theory postulates that the rate of senescence that ultimately leads to spontaneous death is positively associated with the rate of energy turnover of body tissue. In the remainder of the century an extensive debate followed and numurous studies were published inspired by this theory Interspecific comparative studies in birds and mammals have now well established that metabolic rate is indeed inversely related to life span and that the Lifetime Energy Potential (LEP) is fairly constant between species within the order of birds and within mammals (Ku et al., 1993; Speakman, 2005a). There are also experimental approaches. One line of evidence comes from experiments in exotherms. In fruitflies Drosophila melanogaster (Loeb and Northrop, 1917), houseflies Musca domestica (Ragland and Sohal, 1975) and the Nematode Caenorhabditis elegans (Van Voorhies and Ward, 1999) increasing ambient temperature (and thereby metabolic rate) results in reduced longevity. Increased activity results in decreased life span (Ragland and Sohal, 1975) and genetic factors related to longevity have been shown to decrease metabolic rate (Van Voorhies and Ward, 1999). Studies in endotherms (rodents) manipulating energy expenditure by voluntary exercise (Goodrick, 1980; Holloszy and Smith, 1987; Navarro et al., 2004), cold exposure (Johnson et al., 1963; Holloszy and Smith, 1986), caloric restriction (McCay et al., 1935; Yu et al., 1985) or evaluating interindividual variation in metabolic rate (Speakman et al., 2004) do not yield a unoform picture. The interpretation of these results is complicated because in most studies no attempt was made to accurately measure metabolic rate and/or body composition throughout the life span of the animals. Exercise studies have for instance been used to challenge the role that energy expenditure may play in the ageing process. Moderate exercise generally increases energy expenditure during the exercise bout, but does not decrease average life span: if anything, it increases the life span (in rats; (Holloszy and Smith, 1987) and mice; (Navarro et al., 2004)). It may, however, well be that animals compensate for the extra expenditure in exercise by reduced energy metabolism during subsequent rest (Deerenberg et al., 1998). The necessary measurements to determine whether (mass-specific) daily energy expenditure (DEE) is indeed increased in exercising animals, have usually not been made. General Perspective 163

Testing the rate of living theory: manipulations of energy expenditure To investigate the relationship between energy expenditure and ageing within species, we manipulated energy expenditure in two strains of mice (Hsd:ICR and C57BL6J). In the former strain we manipulated energy expenditure by exploiting the spontaneous increase in exercise resulting from selective breeding, and in the latter we decreased ambient temperature (from 22°C to 10°C). In the exercise study (Chapter 6) we used mice selectively-bred for high voluntary wheel-running activity (S+) and compared them with their random-bred controls (C+). We also studied differences between activity-selected mice housed with (S+) or without (S-) a running wheel. As expected, S+ mice had the highest daily energy expenditure throughout life (increased by average 8 kJ d-1), and C+ and S- mice had similar, low levels of DEE (average 59.0 and 61.2 kJ d-1, respectively). According to the rate of living theory one would expect C+ and S- to have similar life spans, which should exceed that of S+ mice. We did find differences in life span, but unexpectedly S+ mice and S- mice both had life spans approximately 100 days shorter than C+ mice. To accurately compare experimental groups, we calculated life-time energy potential (LEP). Mass-specific LEP, no matter whether expressed per animal, per gram body mass, dry lean mass of the carcass or metabolic organ mass, was always higher in the S+ mice compared to the other groups. In the temperature experiment (Chapter 8) we also reared three experimental groups. One group was housed at 22°C (warm, WW) throughout life, another at 10°C (cold, CC) and a third group was housed at 10°C until 15 months of age and at 22°C thereafter (coldwarm, CW). The cold mice spent approximatly 50% more energy than warm mice. The CW mice had an increased daily energy expenditure early in life, but not later in life compared to warm mice. The rate of living theory implies a cumulative effect of energy turnover on life span. We should thus expect the effects of the manipulation restricted to early life and leaving conditions late in life unchanged, to have an impact on life span as well. Strikingly, cold mice had a similar life span to warm mice throughout life, and CW mice had a ~80 days shorter median life span. LEP expressed either per gram body mass, dry lean mass or organ mass was highly increased in CC mice compared to WW and CW mice in the cold experiment. In conclusion, these two experiments manipulating energy expenditure neither showed the expected inverse relationship between metabolic rate and life span based on the rate of living theory, nor did they show a constant LEP between the groups. We emphasize that the conclusion is based on relatively large sample sizes for longevity assessments (n=60), on full-day measurements of energy metabolism in the home cages, and on careful analysis based on age-specific assays of body mass and composition,

Caloric restriction experiments and the Lifetime Energy Potential Caloric restriction (CR) is the only manipulation that increases both median and maximum life span in rodents, as first shown in rats by McCay et al. (1935). 164 Chapter 10

Because CR decreases the rate of ageing, it constitutes an excellent approach to better understand the mechanisms underlying the ageing process. The fact that CR slows the rate of ageing suggests that a reduction in some aspects of energy metabolism should be related to the rate or ageing. This idea is supported by the suppression of growth and many findings of reductions in levels of biochemical parameters in CR animals, like serum glucose, insulin, growth hormone and glucocorticoids (Masoro, 2005). In addition, oxidative damage to proteins, lipids, and DNA is reduced in CR animals compared to Ad libitum fed controls of the same age (Gredilla et al., 2001; Lopez-Torres et al., 2002; Barja, 2002a). Several investigations have reported that these decreases in oxidative damage are related to a lowering of mitochondrial free radical generation rate in various tissues of the CR animals (Sohal et al., 1994; Gredilla et al., 2001; Lopez-Torres et al., 2002). Thus, similar to what has been described for long-lived animals in comparative studies (Perez- Campo et al., 1998; Barja, 2002b), a decrease in mitochondrial free radical generation has been suggested to be one of the main determinants of the extended life span observed in CR animals (Barja, 2004a). The question whether these effects of CR can be attributed to a reduction in energy expenditure in calorically restricted animals has been the subject of consib- erable debate. Masoro, McCarter and co-workers showed that caloric intake per gram body mass is actually increased in mice subjected to CR (Masoro et al., 1982), although the total energy turnover rate decreases. There is no significant effect on 24h metabolic rate expressed per gram lean mass (McCarter et al., 1985; McCarter and McGee, 1989). Right after the start of caloric restriction, a slight decrease in 24-h metabolic rate (corrected for lean mass) was observed, but this difference diss- apeared after approximately 10 weeks (McCarter and McGee, 1989). Interpretation of the results is complicated because metabolic rate has to be corrected for body size in some way. Greenberg and Boozer (2000) have shown that the mass of the most metabolically active organs (heart, liver, kidney, brain) better explained differences in metabolic rate (see also (Daan et al., 1990). Studies measuring organ-specific metabolic rates have shown that even though the internal organs comprise only ~5% of the total body mass, they are responsible for approximately 50% of the resting energy expenditure in both humans and rats (for summary of the results see Ramsey (Ramsey et al., 2000), Table 4). Changes in organ mass that are too small to result in a significant change in total or lean body mass can thus exert large effects on the energy expenditure of the animal, and expressing energy expenditure per gram “organ mass” would thus allow a more accurate comparison between experimental groups. A problem that remains when interpreting data normalized for lean organ masses is that is assumes that all components (heart, liver, kidney and brain) have similar rates of oxygen consumption. Although this assumption is probably incorrect, there is no way of further partitioning the total metabolic rate into organ- specific contributions. General Perspective 165

Calculations based on organ metabolic rates Greenberg estimated energy expenditure based on organ metabolic rates, and found that metabolic rate per unit of organ mass was directely related to the rate of ageing in studies on CR animals, cold-exposed animals or exercising animals (Greenberg, 1999). In his calculations Greenberg used data from McCarter’s study on lifelong metabolic rate in Fischer 344 rats that were fed ad libitum or CR (60%). He found that the ratio of whole-body BMR to total body-part BMR was decreased in CR animals compared to ad libitum fed controls. He thus concluded that the rate of living theory was valid as long as one took organ tissue metabolism into account. We used data from the same study by McCarter (metabolic measurements, (McCarter and Palmer, 1992) and Yu (life span measurements, (Yu et al., 1982)) to calculate the LEP of the animals (see Table 10.1 for a summary of the data). LEPBM was much greater (+48%) in CR animals than ad lib fed animals, with 478 and 324 kJ g-1 respectively. This difference became slightly less when we calculated LEP based on organ mass, but was still much greater (+34%) in CR animals (LEPOM = 11240 vs. 8378 kJ g-1, respectively). These findings show that although there was a negative relationship between overall energy expenditure and life span, life span is not quan- titatively predictable from the hypothesis of a constant LEP as predicted by the rate of living theory, even when expressed per gram mass of the four metabolically expensive organs. This conclusion is the same as for our studies on the effects of high activity (chapter 6) and low temperature (chapter 8). Within species, the rate of living theory is thus not supported by the data in its strict formulation of constancy of lifetime mass-specific energy turnover.

Interspecific comparisons We may now also have an other look at where the rate of living theory came from: the comparison of species. As mentioned in chapter 1 interspecific comparisons yield an inverse relationship between energy expenditure and longevity. However, body mass and lack of phylogenetic control are confounding factors in such analysis (Speakman, 2005b). Energy expenditure rate per gram scales to body mass with an exponent of approximately –0.3, and life span scales to body mass with the exponent of approximately +0.3 like durations of most biological periods (Daan & Aschoff 1982). The product of the two, being the mass-specific Lifetime Energy Potential, thereby has an exponent of 0, i.e. is mass independent. This dependence is based on correlation. It does not prove that there is a causal relationship between energy turnover and lifespan. If there is a causal relationship, it does not prove that it should also cause the constancy of LEP. The association may come about only due to the causal association of both to body mass. For instance, the relationship between body mass and energy expenditure may be defined by physical constraints in relation to body size, i.e., larger animals must have lower metabolic rates per gram because of their surface area (where heat can be dissipated) is relatively smaller. Ecological constraints (e.g., predation risk) in relation to size may determine the association between body mass and life span. 166 Chapter 10

Analysis of residuals (also correcting for phylogeny, see (Speakman, 2005b)) still yield a negative association between energy expenditure and life span. This shows that it is not a simple artefact emerging from their relationship to body mass (Speakman et al., 2002; Speakman, 2005a). Nonetheless, a large part of the variation around the regression line remains unexplained. For instance, birds expend typically in the order of 1.47 times as much energy as a mammal of similar size in BMR and 1.59 in DEE (Daan et al., 1991), but have greater longevity (Holmes et al., 2001; Brunet-Rossinni and Austad, 2004). Marsupials spend less energy than would be expected for a certain mass and have shorter life spans than other mammals (Austad and Fischer, 1991). These discrepancies cannot be explained by the rate of living theory, and may be related to physiological or ecological differences that have occurred during evolution between larger taxa.

The role of free radicals & defense against them in ageing: “The free radical theory of ageing”

Reactive Oxygen Species We have shown that within species life span is not solely dependent on the rate of energy expenditure throughout life and that Lifetime Energy Potential is not necessarily a constant, even if expressed in the proper way, as the total turnover per gram of the high-energy turnover organs (Chapter 6 and 8). This does not imply that metabolic rate is not involved in ageing, and that increased energy turnover does not speed up the ageing process. The free radical theory of Denham Harman (Harman, 1956) proposed that ageing was caused by the accumulation of oxidative damage by oxygen free radicals (reactive oxygen species, ROS) produced during aerobic metabolism. Free radicals influence cellular function because they can cause damage to macromolecules: DNA, lipids and proteins. With age this damage accumulates and eventually results in death (Harman, 1956). Numerous studies have investigated the involvement of free radicals in the ageing process, and is clear that they play an important role (Beckman and Ames, 1998; Sohal et al., 2002; Barja, 2002b). Theoretically a higher metabolism would result in a higher production of reactive oxygen species (ROS) and thus in more rapid ageing and a shorter life span. The relationship between metabolic rate and the production of ROS is not linear. The amount of ROS that is produced is thought to depend on the state of respiration in mitochondria (state 3 or 4), and also on the activity of uncoupling proteins (UCP) (Brand, 2000). When non-limiting amounts of ADP are available, mitochondria are in state 3 respiration. When ADP is absent there can be no ATP production and proton transduction mechanisms become backed up: state 4 respiration. During state 3 respiration there is an abundant flow of protons across the inner membrane, while during state 4 respiration no protons flow through complex IV, and free radical production is expected to be higher (Brand, 2000). During state 4 oxygen con- General Perspective 167

sumption is reduced (leak) and in state 3 the demand for energy and O2 consumption is the highest. UCP’s uncouple oxidative phosphorylation from ATP generation and generate heat instead. When uncoupling occurs, the respiratory chain is speed- ed up, and less ROS are produced per unit O2 consumed. That this can have a positive effect on life span has been shown in a study by Speakman and Selman (2004) where mice with more uncoupling (and higher metabolic rate) had longer life spans. Once ROS are produced there are several protection mechanisms that reduce the damage they can cause. First, antioxidant enzymes can scavenge ROS. The main endogenous antioxidant is superoxide dismutase (SOD) which catalyzes the dismu- tation of superoxide (O2·) into oxygen (O2) and hydrogen peroxide (H2O2). Hydrogen peroxide is still a ROS and can fall apart in hydroxyl radicals (OH·). Catalase (CAT) and glutatione peroxidase (GPx) can prevent this by catalyzing the decomposition of hydrogen peroxide to the harmless water (H2O) and oxygen. Antioxidant enzymes thus reduce the amount of ROS and the damage they can cause. Studies on mice with the genes for such enzymes knocked out have supported the important role of these enzymes. Mice lacking manganese superoxide dismutase die within 10 days (Li et al., 1995). When damage does occur, DNA repair and protein turnover can counteract the damage to macromolecules before it causes permanent loss of function. These defense mechanisms (uncoupling, antioxidant enzymes and repair) influence the relationship between metabolism and ageing (see Figure 1.1). The fact that exercising mice generally spend more energy, but do not show a reduced (or even an increased) life span (Holloszy and Smith, 1987; Navarro et al., 2004), could be explained by changes in the systems described above. For instance, exercise moves mitochondrial respiration toward state 3, which may reduce the ROS produced. Also, voluntary exercise increases the production of antioxidant enzymes and cellular protection against cellular damage (Powers et al., 1999; Kakarla et al., 2005).

Changes in defense systems and life span To see whether changes in these mechanisms could explain the differences we observed in the life span in our experimental groups we measured UCP mRNA expression in muscle, brown adipose tissue and white adipose tissue, SOD and GPx activity in heart and liver, SOD, CAT and GPx mRNA expression in liver and muscle, and protein turnover in liver and muscle (Chapter 7 and 9, for summary see Table 10.2). No significant differences in UCP expression were found between the groups in any of the tissues (data for exercise experiment not shown, for cold experiment see Figure 9.5, for summary see Table 10.1). In the cold experiment, UCP expression was slightly increased in all tissues and at both ages measured, indicative of an increase in uncoupling. Several other studies have reported increases in UCP mRNA expression in response to cold exposure (Carmona et al., 1998; Simonyan et al., 2001; von Praun et al., 2001), but others did not (Boss et al., 1998; 168 Chapter 10

Table 10.1. Life-time energy potential in calorically restricted (CR) animals.

Ad lib CR Ref

DEE (kJ d-1) 183 105 McCarter (1992) 90% survival (d) 800 1240 Yu (1985), from graph Body mass (g) 451 272 McCarter (1992) Lean mass (g) 355 232 McCarter (1992) Organ mass (g) 17.4 11.6 Yu (1985), Greenberg (1999) -1 -1 DEEBM (kJ g d ) 0.40 0.39 -1 -1 DEEOM (kJ g d ) 10.5 9.1 -1 LEPBM (kJ g ) 324 478 -1 LEPOM(kJ g ) 8378 11240

McCarter (1992) measured body mass, lean mass and 24h metabolic rate in ad libitum fed and calorically restricted Fisher rats at 6, 12, 18 and 24 months of age. The graph shows the average value for these measurements. Data on 90% survival were obtained from the survival curve in Yu et al. (1985). Estimates of organ masses were taken from Greenberg (1999), based on measurements by Yu.

von Praun et al., 2001). It remains unclear whether increases in mRNA UCP levels do result in an increase in mitochondrial uncoupling. Differences in uncoupling, antioxidant defense and/or protein synthesis could not explain the difference we observed in life span. No major differences were found between the groups in our exercise or cold experiment for either variable.

The role of endocrine factors in ageing

The difference in life span we observed between control and activity-selected mice may be due to factors that have unintentionally been influenced during the selection procedure. This could potentially include any physiological or behavioural trait that has previously been shown to differ between the lines, such as for instance corticosterone levels (Girard and Garland, Jr., 2002; Malisch et al., 2006). In a study on rats, Cavigelli (2003) showed that neophobic rats (inactive in new environment) had increased basal levels of corticosterone throughout life, and that they had a 60% higher chance to die relative to neophilic animals (active in new environment) at all ages. The median life span was 100 days shorter for the neophobic rats. We measured basal corticosterone levels in our mice at various ages, but found no differences in basal corticosterone between the groups in the exercise (Chapter 3, Figure 3.2) experiment. A previous study did show an increase in basal corticosterone at 2 months of age, mainly in female activity-selected mice (Malisch et al., 2006). Based on our results we may presume that these differences become smaller with age (when differences in wheel-running activity also become smaller, see Chapter 4 and 6). Also, when measured in an elevated-plus maze, no differences in anxiety between control and activity-selected mice were observed (personal obser- General Perspective 169

vation, data not shown). Differences in corticosterone or response to a novel situa- tion thus do not appear to have had a role in differences in life span between the control and selected groups. In Chapters 4 and 5, we have shown that mice selected for high wheel-running activity show some interesting adaptations to their active phenotype. The most intriguing was the observation that plasma adiponectin levels are significantly increased in S mice compared to the levels found in their random-bred controls (Chapter 4, Figure 4.2). This increase was found in S mice irrespective of the availability of running wheels and occurred in all of the separate selection lines. This suggests that the increase in adiponectin is a trait genetically co-segregated with selection for increased wheel-running activity, instead of being mediated via increased physical activity per se. There are several reasons to believe that increased circulating adiponectin levels might contribute to the different phenotypes seen in S mice compared to C mice. For example, Fruebis et al (Fruebis et al., 2001) found that chronic administration of gAcrp30 (i.e., adiponectin) caused weight loss in mice despite the fact that food consumption was unaffected. This is a phenotype which appears homologous to the one found in activity-selected animals (Swallow et al., 1999; Swallow et al., 2001). Fruebis et al. (2001) attributed the effect of adiponectin on body mass to increased fat oxidation, specifically in liver and muscle, and this was confirmed in subsequent studies (Berg et al., 2002; Yamauchi et al., 2002; Bruce et al., 2005). In addition, we observed a decrease in the respiratory quotient (RQ) of S mice compared to C mice measured over a period of 24 hours in non-fasted animals (see Chapter 5), which indeed indicates higher levels of fat oxidation in selected mice. We speculate that an increased capacity to downgrade lipids in muscular tissue contributes the increased physical activity displayed by activity-selected mice. The effects of adiponectin on fat oxidation are believed to arise through stimulation of AMP-activated protein kinase (AMPK)(Berg et al., 2001; Yamauchi et al., 2002). Zhang et al. (2006) have shown increased levels of phosphorylated AMPK in the aorta of male activity-selected mice compared with controls, which is consistent with the observed elevated levels of adiponectin in our study (Chapter 4). Another interesting finding was that the distribution of the fat over the body differ between control and selected male mice, even though the total amount of fat did not. Activity-selected mice stored fat more viscerally and one can imagine that this shift in fat distribution is beneficial for mice that run intensively. The important role of adiponectin in insulin sensitization (Yamauchi et al., 2001; Baratta et al., 2004; Schondorf et al., 2005) makes the selected mice an interesting model to study factors related to the metabolic syndrome. In chapter 5 we investigated whether selected mice were less prone to develop diet-induced obesity. We found the most striking differences between control and selected females. Selected females on the fat diet did not develop obesity. Control mice reduced their food intake, because their food efficiency was higher on the fat diet, but selected females showed opposite results. Selected males did show an increase in fat mass, similar to that of control males on the fat diet. In the plasma levels of several metabolic hor- 170 Chapter 10

mones, we also observed significantly different responses between control and selected males. On the fat diet, both selected males and females had increased adiponectin levels. Because of their abnormal response to the fat diet, which does not lead to develop obesity, these mice are an interesting model to further study the metabolic syndrome. Future studies should establish whether increased adiponectin levels protect the selected mice against developing insulin resistance on a high-fat diet. We also measured hormone levels in the cold-exposed mice (see Box 4.1). Cold- exposure is perhaps experienced by the mice as a chronic stressor. Chronic stress is known to have adverse effects on health and longevity (Paré, 1965). We measured basal corticosterone levels in the mice at various ages and found no differences between the groups, indicating that the HPA-axis was not upregulated in response to the cold exposure. Plasma leptin and adiponectin were significantly decreased in cold-exposed mice relative to warm mice throughout life (see Box 4.1). When corrected for fat mass the decrease in leptin, but not adiponectin remained significant. No studies investigating the direct effect of adiponectin or leptin on ageing have been undertaken. Adiponectin is known to improve factors associated with the metabolic syndrome, which is quickly becoming one of the most important factors compromising human health. Circulating adiponectin levels are reduced in obese humans compared with lean individuals (Arita et al., 1999) and this is associated with increases in cardiovascular risk factors such as insulin resistance and athero- genic lipid profiles. Adiponectin protects against vascular diseases by inhibiting local proinflammatory signals, preventing preatherogenic plaque formation, and by impeding arterial wall thickening (Schondorf et al., 2005). Also in nonobese, healthy adults hypoadiponectinemia results in increased cardiovascular risk factors (Im et al., 2006). Opposite to adiponectin, leptin levels in obese humans are increased in compared with lean individuals (Park et al., 2004). Leptin may play an important role in the pathogenesis of hypertension related to obesity and metabolic syndrome. Furthermore, the lipotoxic effect of leptin resistance may cause insulin resistance and _ cell dysfunction, increasing the risk of type 2 diabetes and leptin has also been shown to possess proliferative, pro-inflammatory, pro-thrombotic, and pro-oxidative actions (Correia and Rahmouni, 2006). Low levels of leptin and high levels of adiponectin thus seem to protect against developing several pathologies. Differences we observed in adiponectin and/or leptin levels between the groups (see Table 10.2), may thus have had a role in protecting or making animals more vulnerable to develop associated pathologies. If so, these effects can not explain the differences we observed in life span between the groups, since selected mice (with high adiponectin levels, and low leptin) actually lived shorter than control mice. In the cold-exposed mice the picture is also complex since they had low adiponectin and low leptin levels. General Perspective 171

Table 10.2. Overview of our results.

S+ vs. S- Table/Fig Cold vs. Warm Table/Fig

Energy expenditure + T6.4, F6.4 + T8.5, F8.4 Life span = T6.1, F6.3 = T8.1, F8.1

LEPBM + T6.4 + T8.5 LEPOM = T6.4 + T8.5 Body compostion: Body mass - T6.2 - T8.2 Dry lean mass + T6.2 + T8.2 Fat content - T6.2 - T8.2 Organ mass = T6.2 + T8.2 Defense systems: Antioxidant enzyme activity = F7.1-F7.3 =- F9.1, F9.2 Protein synthesis = T7.1 = T9.1 Uncoupling proteins = Not shown = F8.5 Hormones: Corticosterone = F4.2 = Box 4.1 Leptin = F4.2 - Box 4.1 Adiponectin = F4.2 - Box 4.1

Overview of the results from the ageing study in exercising and cold-exposed mice. We compare mice that have been selected for high wheel-running activity that were housed with a running wheel (S+) with their sedentary controls (S-). We do not include te comparison between C+ and S+ mice, because in this comparison, differences between the animals that have occurred during the selection process are a confounding factor. In addition, we compared cold- exposed mice to warm mice. + indicates an increase in S+/Cold mice compared to S-/warm mice, - indicates a decrease in S+/Cold mice compared to S-/warm mice, and = indicates no differences between the groups. The Table/Fig. columns show in which tables and figures of the thesis the results are shown.

The role of fat in ageing

One property common to exercising, cold-exposed and calorically-restricted mice is that they all have a significantly reduced fat content compared to their respective controls. Ad libitum fed animals in captivity usually develop obesity because of their sedentary life style, which may have adverse affects on their health (e.g., developing metabolic syndrome and associated diseases, see Chapter 4 and 10.3) and reduce their life span. A negative relationship between body mass and life span (Miller et al., 2002) is an indicator of these effects. Fat is a major source of nitric oxide (NO) stimulated by leptin, and as fat stores increase, leptin and NO release increase in parallel. NO is highly toxic and can cause damage to macromolecules. NO may be responsible for increased coronary heart disease as obesity progresses (McCann et al., 2005). It has further been shown that fat cells increase carcinogenesis in mice (Lu et al., 2006). In this study, voluntary wheel running activity stimulated UVB-induced apoptosis in the epidermis and in tumours of mice. This effect was related to the reduced fat content in exercising 172 Chapter 10

animals; removal of the parametrial fat pads (partial lipectomy) 2 weeks before UVB irradiation also enhanced UVB-induced apoptosis. The enhancement of apoptosis eventually resulted in a 30% lower incidence of tumours in the lean exercising mice. Fat cells may thus secrete substances that inhibit apoptosis in cells with DNA damage, and possibly also in tumours, thereby increasing the incidence of cancer (Lu et al., 2006). It is known in humans that the incidence of certain cancers increases in obese subjects (Bray and Bellanger, 2006). The reduced fat content in cold-exposed, exercising and calorically restricted animals compared to their controls, may thus counteract the effects of their energy expenditure, and thereby result in a longer life span than expected based on predictions from the rate of living theory. Indeed all three factors have been shown to reduce tumour growth and the incidence of cancer: cold-exposure (Holloszy and Smith, 1986), exercise (Kritchevsky, 1990; Lu et al., 2006) and caloric restriction (Kritchevsky, 2001). Whether there is a direct link between fat content and tumour incidence under these conditions needs to be examined.

In summary

Between species there is a general pattern showing an inverse relationship between energy expenditure and longevity. We show that within a species (mice), manipulating energy expenditure by increased activity or decreasing ambient temperature, did not affect life span in the direction predicted, and the quantitative expectation of constancy of the Lifetime Energy Potential was not upheld. These results may refute the strict rate of living theory. Yet they are consistent with a causal influence of energetics on ageing, and they highlight the importance of understanding how different systems work together in ageing animals. ROS that are produced during normal aerobic metabolism are known to have an important role in ageing, because they can cause damage to DNA, lipids and proteins. There are different ways to prevent ROS damage to cellular macromolecules, i.e., by reducing ROS production (by reducing metabolic rate or increasing uncoupling), by reducing the amount that can cause damage (antioxidant enzymes), and by repairing damage that does occur (DNA repair, protein turnover). We showed large differences in daily energy expenditure between exercising (+14%) and cold-exposed mice (+50%), but we did not find strong evidence that differences in uncoupling, antioxidant enzyme activity or protein synthesis occurred. The observed differences in life span between the groups could thus not be explained by differences in these processes. Also, differences in hormone levels could not explain differences in longevity between the groups. Other factors involved in ageing, may have enabled the mice to expend large amounts of energy without a change in life span. One of these factors may be the reduced fat content in exercising and cold-exposed mice, since it seems to protect against developing tumours. General Perspective 173

Ageing from an evolutionary perspective

In their natural environment, animals do not survive long enough to reach a very long life span due to extrinsic mortality (i.e., predation, starvation, disease). For this reason it is suboptimal to invest indefinitely in maintaince of the body and protection against ageing. There is a trade-off in the investment of resources between reproduction and survival. This is described by the disposable soma hypothesis (Kirkwood, 2002). The theory states that ageing results from the twin principles that (1) the force of natural selection declines with age, and (2) longevity requires investments in somatic maintenance and repair that must compete with investments in growth, reproduction and activities that enhance fitness. Animals with low extrinsic rates of mortality are thus expected to invest more in maintenance than animals with high extrinsic mortality rates (e.g., predation). During evolution, animals have developed various systems that can protect them against damage that occurs due to aerobic metabolism. The disposable soma hypothesis should predict that species will adjust their investment in these systems to the forces of natural selection, and this may lead to differences in life span. Birds on average expend energy at a rate on average 1.5 times faster than a mammal of similar mass (Daan et al., 1991), but have yet greater longevity. Has the specific way of living of birds, possibly associated with reduced risks resulted in increased investment in maintenance? There is evidence that birds have evolved mitochondria that produce less ROS per ml oxygen consumed, are better protected against ROS and have lower oxidative damage (Barja et al., 1994; Herrero and Barja, 1998; Herrero and Barja, 1999). Comparing long- and short-lived mammalian species, the long-lived species also generally have lower levels of ROS production and higher protection against ROS (Barja, 1998; Perez-Campo et al., 1998; Barja, 2002b). During evolution differences in ROS production at a certain metabolic rate and in the capacity to defend against oxidative stress have thus emerged. Also within species, manipulations that increase life span like caloric restriction enhance maintenance processes (e.g., protein turnover, antioxidant enzymes, DNA repair). Aging could thus be defined as the failure of maintenance and repair. Different maintenance mechanisms exist and most of them have been shown to decrease with age. They depend on many genes and a considerable investment of metabolic resources is necessary to keep up their activity. Individual theories of ageing revolve around the failure of given maintenance systems, and highlight different aspects of a complex process rather than being mutually exclusive explanations. Ageing does not result in a given cause of death; the system that fails first is largely a matter of chance. Ageing should be viewed as a multi-factorial process. No manipulation will affect only one factor of this process at the same time and this causes complications when testing a single theory such as the rate of living theory. Most manipulations of energy expenditure usually lead to changes in body mass (which by itself affects life span) and in body composition. The expression of energy expenditure then requires a correction for body size. The best way to do this is 174 Chapter 10

e.g. fat survival

e.g. ROS

energy expenditure

Figure 10.1. The relationship between energy expenditure and survival. The graph shows the presence of two separate processes that occur with changing energy expenditure and can affect life span. An increase in energy expenditure can have deleterious effects on survival, e.g. via the increased production of ROS, shown as a negative relationship. On the other hand, a decrease in energy expenditure may also have deleterious effects on survival, for instance because it causes excess body fat, which is a health risk and will negatively affect life span (positive relationship). These two opposing processes then result in an optimal survival at a certain energy expenditure (striped line). See text for further explanation.

still under debate (Ramsey et al., 2000; Speakman, 2005a). Changes occur in numerous physiological parameters (e.g., hormone levels, fat content, antioxidants, protein turnover) that may affect life span. This makes the attribution of differences in life span to a single overall process such as energy expenditure nearly futile. Our studies reject the rate of living theory in its simplest quantitative form that states that the lifetime energy turnover per gram (whether body, lean or organ mass) remains constant when the instantaneous rate of turnover changes. They do not prove the absence of a negative effect of energy turnover on life span. In fact too many studies have demonstrated such an effect in one form or an other. Taking different consequences of energy turnover into consideration implies that the relationship between energy turnover and lifespan can not be a simple unidirec- tional process. In Figure 10.1 we conclude with a more realistic, if schematical proposition. Energy turnover increases deleterious effects (e.g. ROS) as proposed in the rate of living theory that will cause a decline in expected lifespan. Simulta- neously, a decrease in energy turnover rate may cause other negative effects mediated by conditional problems, as excess fat content in rodents in captivity. Together these two processes will generate an intermediate optimum as far as the maximiza- tion of life span is concerned (This needs not be the same energy expenditure that maximizes individual fitness, since fitness, i.e. the expected rate of gene propaga- tion to the next generation, includes the additional component of reproductive output). On this basis of opposing processes we should not even expect to find evidence for the rate of living theory across the whole range of energy metabolic rates. 175 176 References 178 References

A Adage, T., Roorda, B., Visser, G. H., Daan, S., and Scheurink, A. J. W. (2002) Hormonal, metabolic and behavioural consequences of high intensity workloads in rats. 6th Endo-Neuro meeting Doorwerth. Ahren, B. and Scheurink, A. J. (1998) Marked hyperleptinemia after high-fat diet associated with severe glucose intolerance in mice. Eur J Endocrinol 139, 461-467. Arita, Y., Kihara, S., Ouchi, N., Takahashi, M., Maeda, K., Miyagawa, J. I., Hotta, K., Shimomura, I., Nakamura, T., and Miyaoka, K. (1999) Paradoxical Decrease of an Adipose-Specific Protein, Adiponectin, in Obesity. Biochem.Biophys.Res.Commun. 257, 79-83. Arnold, J., LeBlanc, J., Cote, J., Lalonde, J., and Richard, D. (1986) Exercise suppression of thermoregulatory thermogenesis in warm- and cold-acclimated rats. Can.J Physiol Pharmacol. 64, 922-926. Aschoff, J. (1962) Spontane lokomotorische aktivitat. In: Handbuch der Zoologie (ed. Walter de Gruyter),Berlin. Austad, S. N. and Fischer, K. E. (1991) Mammalian aging, metabolism, and ecology: evidence from the bats and marsupials. J.Gerontol. 46, B47-B53. Baratta, R., Amato, S., Degano, C., Farina, M.G., Patane, G., Vigneri, R., and Frittitta, L. (2004) Adiponectin relationship with lipid metabolism is independent of body fat mass: evidence from both cross-sectional and intervention studies. J.Clin.Endocrinol.Metab 89, 2665-2671.

B Barja, G. (1998) Mitochondrial free radical production and aging in mammals and birds. Ann.N.Y.Acad.Sci. 854, 224-238. Barja, G. (2002b) Rate of generation of oxidative stress-related damage and animal longevity. Free Radic.Biol.Med. 33, 1167-1172. Barja, G. (2004a) Aging in vertebrates, and the effect of caloric restriction: a mitochondrial free radical production-DNA damage mechanism? Biol.Rev.Camb.Philos.Soc. 79, 235-251. Barja, G. (2002a) Endogenous oxidative stress: relationship to aging, longevity and caloric restriction. Ageing Res.Rev. 1, 397-411. Barja, G. (2004b) Free radicals and aging. Trends Neurosci. 27, 595-600. Barja, G., Cadenas, S., Rojas, C., Perez-Campo, R., and Lopez-Torres, M. (1994) Low mitochondrial free radical production per unit O2 consumption can explain the simultaneous presence of high longevity and high aerobic metabolic rate in birds. Free Radic.Res. 21, 317-327. Barnett, S. A. (1965) Adaptation of mice to cold. Biol.Rev. 40, 5-51. Barnett, S. A. and Widdowson, E. M. (1965) Organ-weights and body-composition in mice bred for many generations at -3 degrees C. Proc.R.Soc.Lond B Biol.Sci. 162, 502-516. Baudinette, R. V., Halpern, E. A., and Hinds, D. S. (1993) Energetic cost of locomotion as a function of ambient temperature and during growth in the marsupial Potorous tridactylus. J.Exp.Biol. 174, 81-95. Bautista, L. M., Tinbergen, J., Wiersma, P., and Kacelnik, A. (1998) Optimal foraging and beyond: How starlings cope with changes in food availability. Am Nat. 152, 543-561. Beckman, K. B. and Ames, B. N. (1998) The free radical theory of aging matures. Physiol Rev. 78, 547-581. Belter, J. G., Carey, H. V., and Garland, T., Jr. (2004) Effects of voluntary exercise and genetic selection for high activity levels on HSP72 expression in house mice. J Appl.Physiol 96, 1270- 1276. Bennett, A. F. (1990) Thermal dependence of locomotor capacity. Am J Physiol 259, R253-R258. Berg, A. H., Combs, T. P., Du, X., Brownlee, M., and Scherer, P. E. (2001) The adipocyte-secreted protein Acrp30 enhances hepatic insulin action. Nat.Med. 7, 947-953. Berg, A. H., Combs, T. P., and Scherer, P. E. (2002) ACRP30/adiponectin: an adipokine regulating glucose and lipid metabolism. Trends Endocrinol.Metab 13, 84-89. References 179

Biolo, G., Maggi, S. P., Williams, B. D., Tipton, K. D., and Wolfe, R. R. (1995) Increased rates of muscle protein turnover and amino acid transport after resistance exercise in humans. Am.J.Physiol 268, E514-E520. Blazejowski, C. A. and Webster, G. C. (1983) Decreased rates of protein synthesis by cell-free preparations from different organs of aging mice. Mech.Ageing Dev. 21, 345-356. Boden, G. (1997) Role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes 46, 3-10. Boss, O., Samec, S., Kuhne, F., Bijlenga, P., Assimacopoulos-Jeannet, F., Seydoux, J., Giacobino, J. P., and Muzzin, P. (1998) Uncoupling protein-3 expression in rodent skeletal muscle is modulated by food intake but not by changes in environmental temperature. J.Biol.Chem. 273, 5-8. Brand, M. D. (2000) Uncoupling to survive? The role of mitochondrial inefficiency in ageing. Exp.Gerontol. 35, 811-820. Bray, G. A. and Bellanger, T. (2006) Epidemiology, trends, and morbidities of obesity and the metabolic syndrome. Endocrine. 29, 109-117. Bronikowski, A. M., Carter, P. A., Swallow, J. G., Girard, I. A., Rhodes, J. S., and Garland, T., Jr. (2001) Open-field behavior of house mice selectively bred for high voluntary wheel-running. Behav.Genet. 31, 309-316. Bronikowski, A. M., Morgan, T. J., Garland, T., Jr., and Carter, P. A. (2006) The evolution of aging and age-related physical decline in mice selectively bred for high voluntary exercise. Evolution 60, 1494-1508. Bronikowski, A. M., Morgan, T. J., Garland, T., Jr., and Carter, P. A. (2002) Antioxidant gene expression in active and sedentary house mice (Mus domesticus) selected for high voluntary wheel-running behavior. Genetics 161, 1763-1769. Bronson, F. H. (1987) Susceptibility of the fat reserves of mice to natural challenges. J Comp Physiol [B] 157, 551-554. Bruce, C. R., Mertz, V. A., Heigenhauser, G. J., and Dyck, D. J. (2005) The stimulatory effect of globular adiponectin on insulin-stimulated glucose uptake and fatty acid oxidation is impaired in skeletal muscle from obese subjects. Diabetes 54, 3154-3160. Brunet-Rossinni, A. K. (2004) Reduced free-radical production and extreme longevity in the little brown bat (Myotis lucifugus) versus two non-flying mammals. Mech.Ageing Dev. 125, 11-20. Brunet-Rossinni, A. K. and Austad, S. N. (2004) Ageing studies on bats: a review. Biogerontology. 5, 211-222.

C Carbone, C. (2005) How far do animals go? Determinants of day range in mammals. Am Nat. 165, 290-297. Carmona, M. C., Valmaseda, A., Brun, S., Vinas, O., Mampel, T., Iglesias, R., Giralt, M., and Villarroya, F. (1998) Differential regulation of uncoupling protein-2 and uncoupling protein-3 gene expression in brown adipose tissue during development and cold exposure. Biochem.Biophys.Res.Commun. 243, 224-228. Caro, J. F., Sinha, M. K., Kolaczynski, J. W., Zhang, P. L., and Considine, R. V. (1996) Leptin: the tale of an obesity gene. Diabetes 45, 1455-1462. Carroll, L., Voisey, J., and van, D. A. (2004) Mouse models of obesity. Clin.Dermatol. 22, 345-349. Carroll, S. and Dudfield, M. (2004) What is the relationship between exercise and metabolic abnormalities? A review of the metabolic syndrome. Sports Med. 34, 371-418. Cavigelli, S. A. and McClintock, M. K. (2003) Fear of novelty in infant rats predicts adult corticosterone dynamics and an early death. Proc.Natl.Acad.Sci.U.S.A 100, 16131-16136. Chappell, M. A., Garland, T., Jr., Rezende, E. L., and Gomes, F. R. (2004) Voluntary running in deer mice: speed, distance, energy costs and temperature effects. J.Exp.Biol. 207, 3839-3854. 180 References

Chennaoui, M., Gomez-Merino, D., Lesage, J., Drogou, C., and Guezennec, C. Y. (2002) Effects of moderate and intensive training on the hypothalamo-pituitary-adrenal axis in rats. Acta Physiol Scand. 175, 113-121. Chesley, A., MacDougall, J. D., Tarnopolsky, M. A., Atkinson, S. A., and Smith, K. (1992) Changes in human muscle protein synthesis after resistance exercise. J.Appl.Physiol 73, 1383- 1388. Cheuvront, S. N. and Haymes, E. M. (2001) Thermoregulation and marathon running: biological and environmental influences. Sports Med. 31, 743-762. Cnop, M., Havel, P. J., Utzschneider, K. M., Carr, D. B., Sinha, M. K., Boyko, E. J., Retzlaff, B. M., Knopp, R. H., Brunzell, J. D., and Kahn, S. E. (2003) Relationship of adiponectin to body fat distribution, insulin sensitivity and plasma lipoproteins: evidence for independent roles of age and sex. Diabetologia 46, 459-469. Cook, R. D. and Weisberg, S. (1999) Applied Regression including computing and graphics. New York: Wiley. Correia, M. L. G. and Rahmouni, K. (2006) Role of leptin in the cardiovascular and endocrine complications of metabolic syndrome. Diabetes, Obesity and Metabolism 8, 603-610. Coutinho, A. E., Fediuc, S., Campbell, J. E., and Riddell, M. C. (2006) Metabolic effects of voluntary wheel running in young and old Syrian golden hamsters. Physiol Behav. 87, 360-367. Cuthbertson, D., Smith, K., Babraj, J., Leese, G., Waddell, T., Atherton, P., Wackerhage, H., Taylor, P. M., and Rennie, M. J. (2005) Anabolic signaling deficits underlie amino acid resistance of wasting, aging muscle. FASEB J 19, 422-424.

D Daan, S., Deerenberg, C., and Dijkstra, C. (1996) Increased daily work precipitates natural death in the Kestrel. Journal of Animal Ecology 65, 539-544. Daan, S., Masman, D., and Groenewold, A. (1990) Avian basal metabolic rates: their association with body composition and energy expenditure in nature. Am.J.Phys.[B] 259, R333-R340. Daan, S., Masman, D., Strijkstra A, and Kenagy, G. J. (1991) Daily energy turnover during reproduction in birds and mammals: its relationship to basal metabolic rate. Acta XX congressus internationalis ornithologici 1976-1987. Daan, S., Masman, D., Strijkstra, A., and Verhulst, S. (1989) Intraspecific allometry of basal metabolic rate: relations with body size, temperature, composition and circadian phase in the Kestrel, Falco tinnuculus. J.Biol.Rhythms 4, 267-283. Daan, S. and Tinbergen, J. M. (1997) Adaptation of life histories. In: Behavioural Ecology: an Evo- lutionary Approach (eds. Krebs, J. R. and Davies, N. B.), pp. 311-333. Blackwell Science Ltd. Davidovic, V., Djokic, I., Petrovic, N., Durasevic, S., and Cvijic, G. (1999) Activity of antioxidant enzymes in rat skeletal muscle and brown fat: effect of cold and propranolol. J.Therm.Biol. 24, 385-389. Davies, K. J., Quintanilha, A. T., Brooks, G. A., and Packer, L. (1982) Free radicals and tissue damage produced by exercise. Biochem.Biophys.Res Commun. 107, 1198-1205. Davies, S. J. (1982) Behavioural adaptations of birds to environments where evaporation is high and water is in short supply. Comp Biochem.Physiol A 71, 557-566. Day, D. E. and Bartness, T. J. (2001) Effects of foraging effort on body fat and food hoarding in Siberian hamsters. J.Exp.Zool. 289, 162-171. Deerenberg, C., Arpanius, V., Daan, S., and Bos, N. (1997) Reproductive effort decreases antibody responsiveness. Proc.R.Soc.Lond B Biol.Sci. 264, 1021-1029. Deerenberg, C., Overkamp, G. J. F., Visser, G. H., and Daan, S. (1998) Compensation in resting metabolism for experimentally increased activity. J Comp Physiol [B] 168, 507-512. Dorrens, J. and Rennie, M. J. (2003) Effects of ageing and human whole body and muscle protein turnover. Scand.J.Med.Sci.Sports 13, 26-33. References 181

Drent, R. and Daan, S. (1980) The prudent parent: energetic adjustments in avian breeding. Ardea 68, 225-252. Droste, S. K., Gesing, A., Ulbricht, S., Muller, M. B., Linthorst, A. C., and Reul, J. M. (2003) Effects of long-term voluntary exercise on the mouse hypothalamic-pituitary-adrenocortical axis. Endocrinology 144, 3012-3023. Dumke, C. L., Rhodes, J. S., Garland, T., Jr., Maslowski, E., Swallow, J. G., Wetter, A. C., and Cartee, G. D. (2001) Genetic selection of mice for high voluntary wheel running: effect on skeletal muscle glucose uptake. J.Appl.Phys. 91, 1289-1297.

E Economos, A. C., Miquel, J., Ballard, R. C., Blunden, M., Lindseth, K. A., Fleming, J., Philpott, D. E., and Oyama, J. (1982) Effects of simulated increased gravity on the rate of aging of rats: implications for the rate of living theory of aging. Arch.Gerontol.Geriatr. 1, 349-363. Erlanson-Albertsson, C. (2003) The role of uncoupling proteins in the regulation of metabolism. Acta Physiol Scand. 178, 405-412. Everson, C. A. (1995) Functional consequences of sustained sleep deprivation in the rat. Behav.Brain Res 69, 43-54.

F Fallo, F., Scarda, A., Sonino, N., Paoletta, A., Boscaro, M., Pagano, C., Federspil, G., and Vettor, R. (2004) Effect of glucocorticoids on adiponectin: a study in healthy subjects and in Cushing's syndrome. Eur J Endocrinol 150, 339-344. Ferguson, M. A., White, L. J., McCoy, S., Kim, H. W., Petty, T., and Wilsey, J. (2004) Plasma adiponectin response to acute exercise in healthy subjects. Eur.J.Appl.Physiol 91, 324-329. Fruebis, J., Tsao, T. S., Javorschi, S., Ebbets-Reed, D., Erickson, M. R., Yen, F. T., Bihain, B. E., and Lodish, H. F. (2001) Proteolytic cleavage product of 30-kDa adipocyte complement-related protein increases fatty acid oxidation in muscle and causes weight loss in mice. Proc.Natl. Acad.Sci.U.S.A 98, 2005-2010.

G Garland, T. (1983) Scaling the ecological cost of transport to body-mass in terrestrial mammals. Am Nat. 121, 571-587. Garland, T., Jr. (2003) Selection experiments: an under-utilized tool in biomechanics and organis- mal biology. In: Vertebrate biomechanics and evolution (eds. Bels, V. L., Gasc, J.-P., and Casinos, A.), pp. 23-56. Oxford, U.K.: BIOS ScientificPublishers. Garlick, P. J., McNurlan, M. A., and Preedy, V. R. (1980) A rapid and convenient technique for measuring the rate of protein synthesis in tissues by injection of [3H]phenylalanine. Biochem.J. 192, 719-723. Gelegen, C., Collier, D. A., Campbell, I. C., Oppelaar, H., and Kas, M. J. (2006) Behavioral, physiological, and molecular differences in response to dietary restriction in three inbred mouse strains. Am.J.Physiol Endocrinol.Metab 291, E574-E581. Gessaman, J. A. and Nagy, K. A. (1988) Energy metabolism: errors in gas-exchange conversion factors. Phys.Zool. 61, 507-513. Gil-Campos, M., Canete, R., and Gil, A. (2004) Adiponectin, the missing link in insulin resistance and obesity. Clin.Nutr. 23, 963-974. Girard, I. and Garland, T., Jr. (2002) Plasma corticosterone response to acute and chronic voluntary exercise in female house mice. J.Appl.Physiol 92, 1553-1561. Girard, I., Rezende, E. L., and Garland, J. T. (2006) Leptin levels and body composition of mice selectively bred for high voluntary activity. Under Review. 182 References

Girard, I., Swallow, J. G., Carter, P. A., Koteja, P., Rhodes, J. S., and Garland, T. (2002) Maternal- care behavior and life-history traits in house mice (Mus domesticus) artificially selected for high voluntary wheel-running activity. Behav.Processes 57, 37-50. Goodrick, C. L. (1980) Effects of long-term voluntary wheel exercise on male and female Wistar rats. I. Longevity, body weight, and metabolic rate. Gerontology 26, 22-33. Goszczynski, J. (1986) Locomotor activity of terrestrial predators and its consequences. Acta Theriologica 31, 79-95. Gredilla, R. and Barja, G. (2005) Minireview: the role of oxidative stress in relation to caloric restriction and longevity. Endocrinology 146, 3713-3717. Gredilla, R., Barja, G., and Lopez-Torres, M. (2001) Effect of short-term caloric restriction on H2O2 production and oxidative DNA damage in rat liver mitochondria and location of the free radical source. J.Bioenerg.Biomembr. 33, 279-287. Greenberg, J. A. (1999) Organ metabolic rates and aging: two hypotheses. Med.Hypotheses 52, 15- 22. Greenberg, J. A. and Boozer, C. N. (2000) Metabolic mass, metabolic rate, caloric restriction and aging in male Fischer 344 rats. Mech.Ageing Dev. 113, 37-48. Gunduz, F., Senturk, U. K., Kuru, O., Aktekin, B., and Aktekin, M. R. (2004) The effect of one year's swimming exercise on oxidant stress and antioxidant capacity in aged rats. Physiol Res 53, 171-176.

H Halaas, J. L., Gajiwala, K. S., Maffei, M., Cohen, S. L., Chait, B. T., Rabinowitz, D., Lallone, R. L., Burley, S. K., and Friedman, J. M. (1995) Weight-reducing effects of the plasma protein encoded by the obese gene. Science 269, 543-546. Hambly, C. and Speakman, J. R. (2005) Contribution of different mechanisms to compensation for energy restriction in the mouse. Obes Res 13, 1548-1557. Hammond, K. A. and Diamond, J. (1997) Maximal sustained energy budgets in humans and animals. Nature 386, 457-462. Harman, D. (1956) Aging: a theory based on free radical and radiation chemistry. J.Gerontol. 11, 298-300. Havel, P. J. (2001) Peripheral Signals Conveying Metabolic Information to the Brain: Short-Term and Long-Term Regulation of Food Intake and Energy Homeostasis. Exp.Biol.Med. 226, 963- 977. Hayase, K. and Yokogoshi, H. (1992) Effect of exercise on tissue protein synthesis in rats. Biosci.Biotechnol. Biochem. 56, 1637-1639. Hernandez, J. M., Fedele, M. J., and Farrell, P. A. (2000) Time course evaluation of protein synthesis and glucose uptake after acute resistance exercise in rats. J.Appl.Physiol 88, 1142-1149. Herrero, A. and Barja, G. (1998) H2O2 production of heart mitochondria and aging rate are slower in canaries and parakeets than in mice: sites of free radical generation and mechanisms involved. Mech.Ageing Dev. 103, 133-146. Herrero, A. and Barja, G. (1999) 8-oxo-deoxyguanosine levels in heart and brain mitochondrial and nuclear DNA of two mammals and three birds in relation to their different rates of aging. Aging (Milano.) 11, 294-300. Hickey, M. S., Israel, R. G., Gardiner, S. N., Considine, R. V., McCammon, M. R., Tyndall, G. L., Houmard, J. A., Marks, R. H., and Caro, J. F. (1996) Gender differences in serum leptin levels in humans. Biochem.Mol.Med. 59, 1-6. Hill, R. W. (1972) Determination of oxygen consumption by use of the paramagnetic oxygen analyzer. J Appl Physiol 33, 261-263. Holloszy, J. O. (1988) Exercise and longevity: studies on rats. J.Gerontol. 43, B149-B151. Holloszy, J. O. and Schechtman, K. B. (1991) Interaction between exercise and food restriction: effects on longevity of male rats. J Appl Physiol 70, 1529-1535. References 183

Holloszy, J. O. and Smith, E. K. (1986) Longevity of cold-exposed rats: a reevaluation of the "rate- of-living theory". J Appl Physiol 61, 1656-1660. Holloszy, J. O. and Smith, E. K. (1987) Effects of exercise on longevity of rats. Fed.Proc. 46, 1850- 1853. Holloszy, J. O., Smith, E. K., Vining, M., and Adams, S. (1985) Effect of voluntary exercise on longevity of rats. J Appl Physiol 59, 826-831. Holmes, D. J., Fluckiger, R., and Austad, S. N. (2001) Comparative biology of aging in birds: an update. Exp.Gerontol. 36, 869-883.

I Im, J. A., Kim, S. H., Lee, J. W., Shim, J. Y., Lee, H. R., and Lee, D. C. (2006) Association between hypoadiponectinemia and cardiovascular risk factors in nonobese healthy adults. Metabolism 55, 1546-1550. Imai, J., Katagiri, H., Yamada, T., Ishigaki, Y., Ogihara, T., Uno, K., Hasegawa, Y., Gao, J., Ishihara, H., Sasano, H. et al. (2006) Cold exposure suppresses serum adiponectin levels through sym- pathetic nerve activation in mice. Obesity (Silver.Spring) 14, 1132-1141. Iossa, S., Lionetti, L., Mollica, M. P., Crescenzo, R., Botta, M., Barletta, A., and Liverini, G. (2003) Effect of high-fat feeding on metabolic efficiency and mitochondrial oxidative capacity in adult rats. Br.J Nutr. 90, 953-960.

J Jen, K. L., Buison, A., Pellizzon, M., Ordiz, F., Jr., Santa, A. L., and Brown, J. (2003) Differential effects of fatty acids and exercise on body weight regulation and metabolism in female Wistar rats. Exp.Biol.Med.(Maywood.) 228, 843-849. Ji, L. L. (1999) Antioxidants and oxidative stress in exercise. Proc.Soc.Exp.Biol.Med. 222, 283- 292. Johnson, H. D., Kintner, L. D., and Kibler, H. H. (1963) Effects of 48°F (8.9°C) and 83°F (28.4°C) on longevity and pathology of male-rats. J.Gerontol. 18, 29-36. Jurimae, J., Purge, P., and Jurimae, T. (2005) Adiponectin is altered after maximal exercise in highly trained male rowers. Eur.J.Appl.Physiol 93, 502-505.

K Kakarla, P., Vadluri, G., and Reddy, K. S. (2005) Response of hepatic antioxidant system to exercise training in aging female rat. J Exp.Zoolog.A Comp Exp.Biol. 303, 203-208. Katzeff, H. L., Ojamaa, K. M., and Klein, I. (1994) Effects of exercise on protein synthesis and myosin heavy chain gene expression in hypothyroid rats. Am.J.Physiol 267, E63-E67. Kaushik, S. and Kaur, J. (2003) Chronic cold exposure affects the antioxidant defense system in various rat tissues. Clin.Chim.Acta 333, 69-77. Kibler, H. H. and Johnson, H. D. (1961) Metabolic rate and aging in rats during exposure to cold. J Gerontol 16, 13-16. Kibler, H. H., Silsby, H. D., and Johnson, H. D. (1963) Metabolic trends and life span of rats living at 9°C. and 28°C. J Gerontol 18, 235-239. Kirkwood, T. B. (2002) Evolution of ageing. Mech.Ageing Dev. 123, 737-745. Kirkwood, T. B. and Shanley, D. P. (2005) Food restriction, evolution and ageing. Mech.Ageing Dev. 126, 1011-1016. Klaus, S., Heldmaier, G., and Ricquier, D. (1988) Seasonal acclimation of bank voles and wood mice: nonshivering thermogenesis and thermogenic properties of brown adipose tissue mitochondria. J Comp Physiol [B] 158, 157-164. Koteja, P., Carter, P. A., Swallow, J. G., and Garland, T., Jr. (2003) Food wasting by house mice: variation among individuals, families, and genetic lines. Physiol Behav. 80, 375-383. 184 References

Koteja, P., Swallow, J. G., Carter, P. A., and Garland, T., Jr. (1999) Energy cost of wheel running in house mice: implications for coadaptation of locomotion and energy budgets. Physiol Biochem. Zool. 72, 238-249. Koteja, P., Swallow, J. G., Carter, P. A., and Garland, T. (2001) Maximum cold-induced food consumption in mice selected for high locomotor activity: implications for the evolution of endotherm energy budgets. J.Exp.Biol. 204, 1177-1190. Kozlowski, J. and Weiner, J. (1996) Interspecific allometries are by-products of body size opti- mization. Am Nat. 149, 352-380. Krebs, C. J. (1994) Ecology: The experimental analysis of distribution and abundance. New York: Harper Collins College publishers. Kritchevsky, D. (1990) Influence of caloric restriction and exercise on tumorigenesis in rats. Proc.Soc.Exp.Biol Med 193, 35-38. Kritchevsky, D. (2001) Caloric restriction and cancer. J Nutr.Sci Vitaminol.(Tokyo) 47, 13-19. Krol, E., Redman, P., Thomson, P. J., Williams, R., Mayer, C., Mercer, J. G., and Speakman, J. R. (2005) Effect of photoperiod on body mass, food intake and body composition in the field vole, Microtus agrestis. J.Exp.Biol. 208, 571-584. Ku, H. H., Brunk, U. T., and Sohal, R. S. (1993) Relationship between mitochondrial superoxide and hydrogen peroxide production and longevity of mammalian species. Free Radic.Biol.Med. 15, 621-627.

L Lambert, A. J. and Merry, B. J. (2000) Use of primary cultures of rat hepatocytes for the study of ageing and caloric restriction. Exp.Gerontol. 35, 583-594. Lawrence, R. A. and Burk, R. F. (1976) Glutathione peroxidase activity in selenium-deficient rat liver. Biochem.Biophys.Res Commun. 71, 952-958. Levine, J. A. (2004) Non-exercise activity thermogenesis (NEAT). Nutr.Rev. 62, S82-S97. Lewis, S. E. M., Goldspink, D. F., Phillips, J. G., Merry, B. J., and Holehan, A. M. (1985) The Effects of Aging and Chronic Dietary Restriction on Whole-Body Growth and Protein- Turnover in the Rat. Exp.Gerontol. 20, 253-263. Li, Y., Huang, T. T., Carlson, E. J., Melov, S., Ursell, P. C., Olson, J. L., Noble, L. J., Yoshimura, M. P., Berger, C., Chan, P. H. et al. (1995) Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat.Genet. 11, 376-381. Lin, S., Thomas, T. C., Storlien, L. H., and Huang, X. F. (2000) Development of high fat diet- induced obesity and leptin resistance in C57Bl/6J mice. Int.J Obes Relat Metab Disord. 24, 639-646. Lindsay, D. B., Dauncey, M. J., Barker, P. J., and Ingram, D. L. (1988) Protein-Synthesis and Degradation After Acclimation to Different Environmental Temperatures and Energy Intakes. J.Therm.Biol. 13, 79-83. Lobley, G. E. (1990) Energy metabolism reactions in ruminant muscle: responses to age, nutrition and hormonal status. Reprod.Nutr.Dev. 30, 13-34. Loeb, J. and Northrop, J. H. (1917) On the influence of food and temperature upon duration of life. J.Biol.Chem. 32, 103-121. Lopez-Torres, M., Gredilla, R., Sanz, A., and Barja, G. (2002) Influence of aging and long-term caloric restriction on oxygen radical generation and oxidative DNA damage in rat liver mitochondria. Free Radic.Biol.Med. 32, 882-889. Lu, Y. P., Lou, Y. R., Nolan, B., Peng, Q. Y., Xie, J. G., Wagner, G. C., and Conney, A. H. (2006) Stimulatory effect of voluntary exercise or fat removal (partial lipectomy) on apoptosis in the skin of UVB light-irradiated mice. PNAS 103, 16301-16306. Lyman, C. P., O'Brien, R. C., Greene, G. C., and Papafrangos, E. D. (1981) Hibernation and longevity in the Turkish hamster Mesocricetus brandti. Science 212, 668-670. References 185

Lynn, W. S. and Wallwork, J. C. (1992) Does food restriction retard aging by reducing metabolic rate? J.Nutr. 122, 1917-1918.

M MacMillen, R. E. and Dawson, R. J. (1986) Energy and water metabolism of the kowari, Dasyuroides byrnei (Marsupialia: Dasyridae), while resting and running. Australian Mammal 9, 87-95. Mahoney, S. A. (1980) Cost of locomotion and heat balance during rest and running from 0 to 55 degrees C in a patas monkey. J.Appl.Physiol 49, 789-800. Makinen, T., Rintamaki, H., Hohtola, E., and Hissa, R. (1996) Energy cost and thermoregulation of unrestrained rats during exercise in the cold. Comp Biochem.Physiol A Physiol 114, 57-63. Malisch, J. L., Saltzman, W., Gomes, F. R., Rezende, E. L., Jeske, D. R., and Garland, T., Jr. (2006) Basal and stress-induced plasma corticosterone concentrations of mice selectively bred for high voluntary wheel running. Physiol Biochem Zool. In Press. Marklund, S. and Marklund, G. (1974) Involvement of the superoxide anion radical in the autoxi- dation of pyrogallol and a convenient assay for superoxide dismutase. Eur.J Biochem. 47, 469- 474. Masoro, E. J. (2005) Overview of caloric restriction and ageing. Mech.Ageing Dev. 126, 913-922. Masoro, E. J. (1996) Possible mechanisms underlying the antiaging actions of caloric restriction. Toxicol.Pathol. 24, 738-741. Masoro, E. J., Yu, B. P., and Bertrand, H. A. (1982) Action of food restriction in delaying the aging process. Proc.Natl.Acad.Sci.U.S.A 79, 4239-4241. McAllister, T. A., Thompson, J. R., and Samuels, S. E. (2000) Skeletal and cardiac muscle protein turnover during cold acclimation in young rats. Am.J.Physiol Regul.Integr.Comp Physiol 278, R705-R711. McCann, S. M., Mastronardi, C., de, L. A., and Rettori, V. (2005) The nitric oxide theory of aging revisited. Ann.N.Y.Acad.Sci 1057, 64-84. McCarter, R., Masoro, E. J., and Yu, B. P. (1985) Does food restriction retard aging by reducing the metabolic rate? Am.J.Physiol 248, E488-E490. McCarter, R. J. and McGee, J. R. (1989) Transient reduction of metabolic rate by food restriction. Am.J.Physiol 257, E175-E179. McCarter, R. J. and Palmer, J. (1992) Energy metabolism and aging: a lifelong study of Fisher 344 rats. Am.J.Phys. 263, E448-E452. McCay, C. M., Crowell, M. F., and Maynard, L. A. (1935) The effect of retarded growth upon the length of life span and upon the ultimate body size. J.Nutr. 10, 63-79. McEwen, J. E., Zimniak, P., Mehta, J. L., and Reis, R. J. (2005) Molecular pathology of aging and its implications for senescent coronary atherosclerosis. Curr.Opin.Cardiol. 20, 399-406. McMurray, R. G. and Hackney, A. C. (2005) Interactions of metabolic hormones, adipose tissue and exercise. Sports Med. 35, 393-412. Mecocci, P., Fano, G., Fulle, S., MacGarvey, U., Shinobu, L., Polidori, M. C., Cherubini, A., Vecchiet, J., Senin, U., and Beal, M. F. (1999) Age-dependent increases in oxidative damage to DNA, lipids, and proteins in human skeletal muscle. Free Radic.Biol.Med. 26, 303-308. Melov, S., Coskun, P., Patel, M., Tuinstra, R., Cottrell, B., Jun, A. S., Zastawny, T. H., Dizdaroglu, M., Goodman, S. I., Huang, T. T. et al. (1999) Mitochondrial disease in superoxide dismutase 2 mutant mice. Proc.Natl.Acad.Sci.U.S.A 96, 846-851. Melov, S., Doctrow, S. R., Schneider, J. A., Haberson, J., Patel, M., Coskun, P. E., Huffman, K., Wallace, D. C., and Malfroy, B. (2001) Lifespan extension and rescue of spongiform encephalopathy in superoxide dismutase 2 nullizygous mice treated with superoxide dismutase-catalase mimetics. J Neurosci. 21, 8348-8353. Merry, B. J. (2005) Dietary restriction in rodents--delayed or retarded ageing? Mech.Ageing Dev. 126, 951-959. 186 References

Miller, R. A., Harper, J. M., Galecki, A., and Burke, D. T. (2002) Big mice die young: early life body weight predicts longevity in genetically heterogeneous mice. Aging Cell 1, 22-29. Moller, D. E. and Kaufman, K. D. (2005) Metabolic syndrome: a clinical and molecular perspective. Annu.Rev.Med. 56, 45-62. Morens, C., Keijzer, M., de, V. K., Scheurink, A., and van, D. G. (2005) Effects of high-fat diets with different carbohydrate-to-protein ratios on energy homeostasis in rats with impaired brain melanocortin receptor activity. Am.J Physiol Regul.Integr.Comp Physiol 289, R156-R163. Morgan, T. J., Garland, T., Jr., and Carter, P. A. (2003) Ontogenies in mice selected for high voluntary wheel-running activity. I. Mean Ontogenies. Evolution 57, 646-657. Mosoni, L., Valluy, M. C., Serrurier, B., Prugnaud, J., Obled, C., Guezennec, C. Y., and Mirand, P. P. (1995) Altered response of protein synthesis to nutritional state and endurance training in old rats. Am.J.Physiol 268, E328-E335. Mount, L. E. (1966) Basis of heat regulation in homoeotherms. Br.Med.Bull. 22, 84-87. Murray, H. A., Jr. (1926) Physiological ontogeny: A. Chicken embyos. XII. The metabolism as a function of age. J.Gen.Physiol. 10, 337-343.

N Navarro, A., Gomez, C., Lopez-Cepero, J. M., and Boveris, A. (2004) Beneficial effects of moderate exercise on mice aging: survival, behavior, oxidative stress, and mitochondrial electron transfer. Am.J.Physiol Regul.Integr.Comp Physiol 286, R505-R511. Newsholme, E. A. (1987) Substrate cycles and energy metabolism: their biochemical, biological, physiological and pathological importance. In: Energy metabolism of farm animals (eds. Moe, P. W., Tyrell, H. F., and Reynolds, P. J.), pp. 174-186. Totowa, NJ: Rowman and Littlefield. Novelli, M., Pocai, A., Skalicky, M., Viidik, A., Bergamini, E., and Masiello, P. (2004) Effects of life-long exercise on circulating free fatty acids and muscle triglyceride content in ageing rats. Exp.Gerontol. 39, 1333-1340.

O Oklejewicz, M., Hut, R. A., Daan, S., Loudon, A. S., and Stirland, A. J. (1997) Metabolic rate changes proportionally to circadian frequency in tau mutant Syrian hamsters. J Biol.Rhythms 12, 413-422. Orr, W. C. and Sohal, R. S. (1994) Extension of life-span by overexpression of superoxide dismutase and catalase in Drosophila melanogaster. Science 263, 1128-1130.

P Paglia, D. E. and Valentine, W. N. (1967) Studies on the quantitative and qualitative characteriza- tion of erythrocyte glutathione peroxidase. J Lab Clin.Med 70, 158-169. Paladino, F. V. and King, J. R. (1984) Thermoregulation and oxygen consumption during terrestrial locomotion by white-crowned sparrows Zonotrichia leucophrys gambelii. Physiol Zool. 57, 226-236. Paré, W. P. (1965) The effect of chronic environmental stress on premature aging in the rat. J Gerontol 20, 78-84. Park, K. G., Park, K. S., Kim, M. J., Kim, H. S., Suh, Y. S., Ahn, J. D., Park, K. K., Chang, Y. C., and Lee, I. K. (2004) Relationship between serum adiponectin and leptin concentrations and body fat distribution. Diabetes Res Clin.Pract. 63, 135-142. Pearl, R. (1928) The rate of living. New York: Albert Knopf, Inc. Pellizzon, M., Buison, A., Ordiz, F., Jr., Santa, A. L., and Jen, K. L. (2002) Effects of dietary fatty acids and exercise on body-weight regulation and metabolism in rats. Obes Res 10, 947-955. Perez-Campo, R., Lopez-Torres, M., Cadenas, S., Rojas, C., and Barja, G. (1998) The rate of free radical production as a determinant of the rate of aging: evidence from the comparative approach. J.Comp Physiol [B] 168, 149-158. References 187

Perrigo, G. (1987) Breeding and feeding strategies in deer mice and house mice when females are challenged to work for their food. Anim.Behav. 35, 1298-1316. Perrigo, G. and Bronson, F. H. (1983) Foraging effort, food intake, fat deposition and puberty in female mice. Biol.Reprod. 29, 455-463. Perrigo, G. and Bronson, F. H. (1985) Behavioral and physiological responses of female house mice to foraging variation. Physiol Behav. 34, 437-440. Phillips, S. M., Tipton, K. D., Aarsland, A., Wolf, S. E., and Wolfe, R. R. (1997) Mixed muscle protein synthesis and breakdown after resistance exercise in humans. Am.J.Physiol 273, E99- 107. Pohl, H. and West, G. C. (1973) Daily and seasonal variation in metabolic response to cold during rest and forced exercise in the common redpoll. Comparative Biochemistry and Physiology Part A: Physiology 45, 851-867. Powers, S. K., Ji, L. L., and Leeuwenburgh, C. (1999) Exercise training-induced alterations in skeletal muscle antioxidant capacity: a brief review. Med.Sci.Sports Exerc. 31, 987-997. Puerta, M., Abelenda, M., Rocha, M., and Trayhurn, P. (2002) Effect of acute cold exposure on the expression of the adiponectin, resistin and leptin genes in rat white and brown adipose tissues. Horm.Metab Res. 34, 629-634.

R Ragland, S. S. and Sohal, R. S. (1975) Ambient temperature, physical activity and aging in the housefly, Musca domestica. Exp.Gerontol. 10, 279-289. Ramsey, J. J., Harper, M. E., and Weindruch, R. (2000) Restriction of energy intake, energy expenditure, and aging. Free Radic.Biol.Med. 29, 946-968. Ramsey, J. J., Roecker, E. B., Weindruch, R., Baum, S. T., and Kemnitz, J. W. (1996) Thermogenesis of adult male rhesus monkeys: Results through 66 months of dietary restriction. Faseb Journal 10, 4198. Rao, G., Xia, E., and Richardson, A. (1990) Effect of age on the expression of antioxidant enzymes in male Fischer F344 rats. Mech.Ageing Dev. 53, 49-60. Rattan, S. I. (1996) Synthesis, modifications, and turnover of proteins during aging. Exp. Gerontol. 31, 33-47. Rennie, M. J., Edwards, R. H., Krywawych, S., Davies, C. T., Halliday, D., Waterlow, J. C., and Mill- ward, D. J. (1981) Effect of exercise on protein turnover in man. Clin.Sci.(Lond) 61, 627-639. Rezende, E. L., Kelly, S. A., Gomes, F. R., Chappell, M. A., and Garland, T., Jr. (2006) Effects of size, sex, and voluntary running speeds on costs of locomotion in lines of laboratory mice selectively bred for high wheel-running activity. Physiol Biochem.Zool. 79, 83-99. Rhodes, J. S., Gammie, S. C., and Garland, T., Jr. (2005) Neurobiology of mice selected for high voluntary wheel-running activity. Integr.Comp.Biol. 45, 438-455. Rhodes, J. S., Hosack, G. R., Girard, I., Kelley, A. E., Mitchell, G. S., and Garland, T., Jr. (2001) Differential sensitivity to acute administration of cocaine, GBR 12909, and fluoxetine in mice selectively bred for hyperactive wheel-running behavior. Psychopharmacology (Berl) 158, 120- 131. Rhodes, J. S., Koteja, P., Swallow, J. G., Carter, P. A., and Garland, J. T. (2000) Body temperature of house mice artificially selected for high voluntary wheel-running behavior: repeatability and effect on genetic selection. J.Therm.Biol. 25, 391-400. Richter, C. P. (1922) Behaviouristic study of the activity of the rat. Comp.Phsychol.Monogr. 1. Rocha, H. J., Nash, J. E., Connell, A., and Lobley, G. E. (1993) Protein synthesis in ovine muscle and skin: sequential measurements with three different amino acids based on the large-dose procedure. Comp Biochem.Physiol B 105, 301-307. Romijn, C. and Lokhorst, W. (1961) Some aspects of energy metabolism in birds. In: Proceedings second symposium on energy metabolism (eds. Brouwer, E. and van Es, A. J. H.), pp. 49-58. Wageningen: EAAP. 188 References

Rowland, N. (1984) Effects of chronic cold exposure on wheel running, food intake and fatty acid synthesis in Syrian hamsters. Physiol Behav. 33, 253-256. Rubner, M. (1908) Das Problem der Lebensdauer und seine Beziehungen zum Wachstum und Ernahrung. Munich: Oldenburg. Ryan, A. S., Berman, D. M., Nicklas, B. J., Sinha, M., Gingerich, R. L., Meneilly, G. S., Egan, J. M., and Elahi, D. (2003) Plasma adiponectin and leptin levels, body composition, and glucose utilization in adult women with wide ranges of age and obesity. Diabetes Care 26, 2383-2388. Ryazanov, A. G. and Nefsky, B. S. (2002) Protein turnover plays a key role in aging. Mech.Ageing Dev. 123, 207-213.

S Samuels, S. E., Thompson, J. R., and Christopherson, R. J. (1996) Skeletal and cardiac muscle protein turnover during short-term cold exposure and rewarming in young rats. Am.J.Physiol 270, R1231-R1239. Schondorf, T., Maiworm, A., Emmison, N., Forst, T., and Pfutzner, A. (2005) Biological background and role of adiponectin as marker for insulin resistance and cardiovascular risk. Clin.Lab 51, 489-494. Scott, S. L., Christopherson, R. J., Thompson, J. R., and Baracos, V. E. (1993) The effect of a cold environment on protein and energy metabolism in calves. Br.J.Nutr. 69, 127-139. Selman, C., McLaren, J. S., Collins, A. R., Duthie, G. G., and Speakman, J. R. (2002) Antioxidant enzyme activities, lipid peroxidation, and DNA oxidative damage: the effects of short-term voluntary wheel running. Archives of Biochemistry and Biophysics 401, 225-261. Selman, C., McLaren, J. S., Himanka, M. J., and Speakman, J. R. (2000) Effect of long-term cold exposure on antioxidant enzyme activities in a small mammal. Free Radic.Biol.Med. 28, 1279- 1285. Selman, C., Phillips, T., Staib, J. L., Duncan, J. S., Leeuwenburgh, C., and Speakman, J. R. (2005) Energy expenditure of calorically restricted rats is higher than predicted from their altered body composition. Mech.Ageing Dev. 126, 783-793. Sheffield-Moore, M., Paddon-Jones, D., Sanford, A. P., Rosenblatt, J. I., Matlock, A. G., Cree, M. G., and Wolfe, R. R. (2005) Mixed muscle and hepatic derived plasma protein metabolism is differentially regulated in older and younger men following resistance exercise. Am J Physiol Endocrinol Metab 288, E922-E929. Sheffield-Moore, M., Yeckel, C. W., Volpi, E., Wolf, S. E., Morio, B., Chinkes, D. L., Paddon-Jones, D., and Wolfe, R. R. (2004) Postexercise protein metabolism in older and younger men following moderate-intensity aerobic exercise. Am.J.Physiol Endocrinol.Metab 287, E513-E522. Shi, Z. Q., Nelson, A., Whitcomb, L., Wang, J., and Cohen, A. M. (1998) Intracerebroventricular administration of leptin markedly enhances insulin sensitivity and systemic glucose utilization in conscious rats. Metabolism 47, 1274-1280. Short, K. R., Vittone, J. L., Bigelow, M. L., Proctor, D. N., and Nair, K. S. (2004) Age and aerobic exercise training effects on whole body and muscle protein metabolism. Am.J.Physiol Endocrinol.Metab 286, E92-101. Shringarpure, R. and Davies, K. J. (2002) Protein turnover by the proteasome in aging and disease. Free Radic.Biol.Med. 32, 1084-1089. Siems, W. G., Brenke, R., Sommerburg, O., and Grune, T. (1999) Improved antioxidative protection in winter swimmers. QJM: Monthly Journal Of The Association Of Physicians 92, 193-198. Simonyan, R. A., Jimenez, M., Ceddia, R. B., Giacobino, J. P., Muzzin, P., and Skulachev, V. P. (2001) Cold-induced changes in the energy coupling and the UCP3 level in rodent skeletal muscles. Biochim.Biophys.Acta 1505, 271-279. Smith, K., Reynolds, N., Downie, S., Patel, A., and Rennie, M. J. (1998) Effects of flooding amino acids on incorporation of labeled amino acids into human muscle protein. Am J Physiol Endocrinol Metab 275, E73-E78. References 189

Sohal, R. S. (2002) Role of oxidative stress and protein oxidation in the aging process. Free Radic.Biol.Med. 33, 37-44. Sohal, R. S., Agarwal, A., Agarwal, S., and Orr, W. C. (1995) Simultaneous overexpression of cop- per- and zinc-containing superoxide dismutase and catalase retards age-related oxidative damage and increases metabolic potential in Drosophila melanogaster. J Biol.Chem. 270, 15671-15674. Sohal, R. S., Arnold, L. A., and Sohal, B. H. (1990) Age-related changes in antioxidant enzymes and prooxidant generation in tissues of the rat with special reference to parameters in two insect species. Free Radic.Biol.Med. 9, 495-500. Sohal, R. S. and Buchan, P. B. (1981) Relationship between physical activity and life span in the adult housefly, Musca domestica. Exp.Gerontol. 16, 157-162. Sohal, R. S., Ku, H. H., Agarwal, S., Forster, M. J., and Lal, H. (1994) Oxidative damage, mitochondrial oxidant generation and antioxidant defenses during aging and in response to food restriction in the mouse. Mech.Ageing Dev. 74, 121-133. Sohal, R. S., Mockett, R. J., and Orr, W. C. (2002) Mechanisms of aging: an appraisal of the oxidative stress hypothesis(1,2). Free Radic.Biol.Med. 33, 575. Spasic, M. B., Saicic, Z. S., Buzadzic, B., Korac, B., Blagojevic, D., and Petrovic, V. M. (1993) Effect of long-term exposure to cold on the antioxidant defense system in the rat. Free Radic.Biol. Med. 15, 291-299. Speakman, J. R. (2005b) Correlations between physiology and lifespan--two widely ignored problems with comparative studies. Aging Cell 4, 167-175. Speakman, J. R. (2005a) Body size, energy metabolism and lifespan. J.Exp.Biol. 208, 1717-1730. Speakman, J. R., Ergon, T., Cavanagh, R., Reid, K., Scantlebury, D. M., and Lambin, X. (2003a) Resting and daily energy expenditures of free-living field voles are positively correlated but reflect extrinsic rather than intrinsic effects. Proc.Natl.Acad.Sci U.S.A 100, 14057-14062. Speakman, J. R. and Krol, E. (2005) Limits to sustained energy intake IX: a review of hypotheses. J Comp Physiol [B] 175, 375-394. Speakman, J. R. and Selman, C. (2003) Physical activity and resting metabolic rate. Proc.Nutr.Soc. 62, 621-634. Speakman, J. R., Selman, C., McLaren, J. S., and Harper, E. J. (2002) Living fast, dying when? The link between aging and energetics. J.Nutr. 132, 1583S-1597S. Speakman, J. R., Talbot, D. A., Selman, C., Snart, S., McLaren, J. S., Redman, P., Krol, E., Jackson, D. M., Johnson, M. S., and Brand, M. D. (2004) Uncoupled and surviving: individual mice with high metabolism have greater mitochondrial uncoupling and live longer. Aging Cell 3, 87-95. Speakman, J. R., van Acker, A., and Harper, E. J. (2003b) Age-related changes in the metabolism and body composition of three dog breeds and their relationship to life expectancy. Aging Cell 2, 265-275. Spinu, M., Benveneste, S., and Degen, A. A. (2003) Effect of density and season on stress and behaviour in broiler breeder hens. Br.Poult.Sci. 44, 170-174. Stadtman, E. R. (2004) Role of oxidant species in aging. Curr.Med.Chem. 11, 1105-1112. Stearns, S. C. (2000) Life history evolution: successes, limitations, and prospects. Naturwissen- schaften 87, 476-486. Stiegler, P. and Cunliffe, A. (2006) The role of diet and exercise for the maintenance of fat-free mass and resting metabolic rate during weight loss. Sports Med 36, 239-262. Storlien, L. H., James, D. E., Burleigh, K. M., Chisholm, D. J., and Kraegen, E. W. (1986) Fat feeding causes widespread in vivo insulin resistance, decreased energy expenditure, and obesity in rats. Am.J Physiol 251, E576-E583. Surwit, R. S., Feinglos, M. N., Rodin, J., Sutherland, A., Petro, A. E., Opara, E. C., Kuhn, C. M., and Rebuffe-Scrive, M. (1995) Differential effects of fat and sucrose on the development of obesity and diabetes in C57BL/6J and A/J mice. Metabolism 44, 645-651. 190 References

Swallow, J. G., Carter, P. A., and Garland, T., Jr. (1998) Artificial selection for increased wheel running behavior in house mice. Beh.Gen. 28, 227-237. Swallow, J. G., Koteja, P., Carter, P. A., and Garland, T. (1999) Artificial selection for increased wheel-running activity in house mice results in decreased body mass at maturity. J.Exp.Biol. 202 ( Pt 18), 2513-2520. Swallow, J. G., Koteja, P., Carter, P. A., and Garland, T., Jr. (2001) Food consumption and body composition in mice selected for high wheel- running activity. J.Comp Physiol [B] 171, 651- 659. Swallow, J. G., Rhodes, J. S., and Garland, T. Jr. (2005) Phenotypic and evolutionary plasticity of organ masses in response to voluntary exercise in house mice. Integr.Comp.Biol. 45, 426-437.

T Tavernarakis, N. and Driscoll, M. (2002) Caloric restriction and lifespan: a role for protein turnover? Mech.Ageing Dev. 123, 215-229. Taylor, C. R., Heglund, N. C., and Maloiy, G. M. (1982) Energetics and mechanics of terrestrial locomotion. I. Metabolic energy consumption as a function of speed and body size in birds and mammals. J.Exp.Biol. 97:1-21., 1-21. Taylor, C. R., Schmidt-Nielsen, K., and Raab, J. L. (1970) Scaling of energetic cost of running to body size in mammals. Am.J Physiol 219, 1104-1107. Tieleman, B. I., Williams, J. B., and Buschur, M. E. (2002) Physiological adjustments to arid and mesic environments in larks (Alaudidae). Physiol Biochem.Zool. 75, 305-313. Tolmasoff, J. M., Ono, T., and Cutler, R. G. (1980) Superoxide dismutase: correlation with lifespan and specific metabolic rate in primate species. Proc.Natl.Acad.Sci.U.S.A 77, 2777-2781. Trayhurn, P., Duncan, J. S., Nestor, A., Thomas, M. E. A., and Rayner, D., V (1994) Chemilu- minescent Detection of mRNAs on Northern Blots with Digoxigenin End-Labeled Oligo- nucleotides. Analytical Biochemistry 222, 224-230. Tsay, H. J., Wang, P., Wang, S. L., and Ku, H. H. (2000) Age-associated changes of superoxide dismutase and catalase activities in the rat brain. J Biomed.Sci. 7, 466-474. Tschop, M. and Heiman, M. L. (2001) Rodent obesity models: an overview. Exp.Clin.Endocrinol Diabetes 109, 307-319. Tyler, D. D. (1975) Role of superoxide radicals in the lipid peroxidation of intracellular membranes. FEBS Lett. 51, 180-183.

V Vaanholt, L. M., Garland, T., Jr., Daan, S., and Visser, G. H. (2006) Wheel-running activity and energy metabolism in relation to ambient temperature in mice selected for high wheel-running activity. J.Comp Physiol [B] 177(1), 109-118. Van Dijk, G., Donahey, J. C., Thiele, T. E., Scheurink, A. J., Steffens, A. B., Wilkinson, C. W., Tenenbaum, R., Campfield, L. A., Burn, P., Seeley, R. J. et al. (1997) Central leptin stimulates corticosterone secretion at the onset of the dark phase. Diabetes 46, 1911-1914. Van Dijk, G., Seeley, R. J., Thiele, T. E., Friedman, M. I., Ji, H., Wilkinson, C. W., Burn, P., Campfield, L. A., Tenenbaum, R., Baskin, D. G. et al. (1999) Metabolic, gastrointestinal, and CNS neuropeptide effects of brain leptin administration in the rat. Am.J.Physiol 276, R1425- R1433. Van Remmen, H. et al. (1995) Gene expression and protein degradation. In: Handbook of Physiology, Section 11: Aging. Oxford University Press, New York (ed. Masoro, E. J.), pp. 171- 234. Van Voorhies, W. A. and Ward, S. (1999) Genetic and environmental conditions that increase longevity in Caenorhabditis elegans decrease metabolic rate. Proc.Natl.Acad.Sci.U.S.A 96, 11399-11403. Visser, G. H. and Schekkerman, H. (1999) Validation of the doubly labeled water method in References 191

growing precocial birds: the importance of assumptions concerning evaporative water loss. Physiol Biochem.Zool. 72, 740-749. Volpi, E., Sheffield-Moore, M., Rasmussen, B. B., and Wolfe, R. R. (2001) Basal muscle amino acid kinetics and protein synthesis in healthy young and older men. JAMA 286, 1206-1212. Von Praun, C., Burkert, M., Gessner, M., and Klingenspor, M. (2001) Tissue-specific expression and cold-induced mRNA levels of uncoupling proteins in the Djungarian hamster. Physiol Biochem.Zool. 74, 203-211.

W Wallace, D. C. (2005) A mitochondrial paradigm of metabolic and degenerative diseases, aging, and cancer: a dawn for evolutionary medicine. Annu.Rev.Genet. 39, 359-407. Ward, W. F. (2000) The relentless effects of the aging process on protein turnover. Biogerontology. 1, 195-199. Weindruch, R., Walford, R. L., Fligiel, S., and Guthrie, D. (1986) The retardation of aging in mice by dietary restriction: longevity, cancer, immunity and lifetime energy intake. J.Nutr. 116, 641-654. Wester, T. J., Lobley, G. E., Birnie, L. M., Crompton, L. A., Brown, S., Buchan, V., Calder, A. G., Milne, E., and Lomax, M. A. (2004) Effect of plasma insulin and branched-chain amino acids on skeletal muscle protein synthesis in fasted lambs. Br.J.Nutr. 92, 401-409. Westerterp, K. R. (2001) Limits to sustainable human metabolic rate. J.Exp.Biol. 204, 3183-3187. Wiersma, P., Salomons, H. M., and Verhulst, S. (2005) Metabolic adjustments to increasing foraging costs of starlings in a closed economy. J.Exp.Biol. 208, 4099-4108. Wiersma, P. and Verhulst, S. (2005) Effects of intake rate on energy expenditure, somatic repair and reproduction of zebra finches. J.Exp.Biol 208, 4091-4098. Winzell, M. S. and Ahren, B. (2004) The high-fat diet-fed mouse: a model for studying mechanisms and treatment of impaired glucose tolerance and type 2 diabetes. Diabetes 53 Suppl 3, S215-S219. Wolf, T. J. and Schmid-Hempel, P. (1989) Extra loads and foraging life span in honeybee workers. J.Anim.Ecol. 58, 943-954. Woods, S. C., D'Alessio, D. A., Tso, P., Rushing, P. A., Clegg, D. J., Benoit, S. C., Gotoh, K., Liu, M., and Seeley, R. J. (2004) Consumption of a high-fat diet alters the homeostatic regulation of energy balance. Physiology & Behavior 83, 573-578. Wunder, B. A. (1970) Temperature regulation and the effects of water restriction on Merriams's chipmunk, Eutamias merriami. Comparative Biochemistry and Physiology 33, 385-403.

Y Yamamoto, T., Ohkuwa, T., Itoh, H., Sato, Y., and Naoi, M. (2002) Effect of gender differences and voluntary exercise on antioxidant capacity in rats. Comp Biochem.Physiol C.Toxicol.Pharmacol. 132, 437-444. Yamauchi, T., Kamon, J., Minokoshi, Y., Ito, Y., Waki, H., Uchida, S., Yamashita, S., Noda, M., Kita, S., Ueki, K. et al. (2002) Adiponectin stimulates glucose utilization and fatty-acid oxidation by activating AMP-activated protein kinase. Nat.Med. 8, 1288-1295. Yamauchi, T., Kamon, J., Waki, H., Terauchi, Y., Kubota, N., Hara, K., Mori, Y., Ide, T., Murakami, K., Tsuboyama-Kasaoka, N. et al. (2001) The fat-derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity. Nat.Med. 7, 941-946. Yan, L. J. and Sohal, R. S. (2000) Prevention of flight activity prolongs the life span of the housefly, Musca domestica, and attenuates the age-associated oxidative damamge to specific mitochondrial proteins. Free Radic.Biol.Med. 29, 1143-1150. Yarasheski, K. E. (2003) Exercise, aging, and muscle protein metabolism. J.Gerontol.A Biol.Sci. Med.Sci. 58, M918-M922. 192 References

Yarasheski, K. E., Zachwieja, J. J., and Bier, D. M. (1993) Acute effects of resistance exercise on muscle protein synthesis rate in young and elderly men and women. Am.J.Physiol 265, E210- E214. Young, V. R., Steffee, W. P., Pencharz, P. B., Winterer, J. C., and Scrimshaw, N. S. (1975) Total Human Body Protein-Synthesis in Relation to Protein Requirements at Various Ages. Nature 253, 192-194. Yu, B. P. (1996) Aging and oxidative stress: Modulation by dietary restriction. Free Radical Biology and Medicine 21, 651-668. Yu, B. P., Masoro, E. J., and McMahan, C. A. (1985) Nutritional influences on aging of Fischer 344 rats: I. Physical, metabolic, and longevity characteristics. J.Gerontol. 40, 657-670. Yu, B. P., Masoro, E. J., Murata, I., Bertrand, H. A., and Lynd, F. T. (1982) Life span study of SPF Fischer 344 male rats fed ad libitum or restricted diets: longevity, growth, lean body mass and disease. J.Gerontol 37, 130-141. Yunianto, V. D., Hayashi, K., Kaneda, S., Ohtsuka, A., and Tomita, Y. (1997) Effect of environmental temperature on muscle protein turnover and heat production in tube-fed broiler chickens. Br.J.Nutr. 77, 897-909.

Z Zhang, Y., Lee, T. S., Kolb, E. M., Sun, K., Lu, X., Sladek, F. M., Kassab, G. S., Garland, T., Jr., and Shyy, J. Y. (2006) AMP-activated protein kinase is involved in endothelial NO synthase activation in response to shear stress. Arterioscler.Thromb.Vasc.Biol. 26, 1281-1287. Zoladz, J. A., Konturek, S. J., Duda, K., Majerczak, J., Sliwowski, Z., Grandys, M., and Bielanski, W. (2005) Effect of moderate incremental exercise, performed in fed and fasted state on car- dio-respiratory variables and leptin and ghrelin concentrations in young healthy men. J Physiol Pharmacol. 56, 63-85. List of abbreviations 193

List of abbreviations

ACT = Costs of Activity (DEE-RMR) (kJ d-1) MEI = Metabolisable Energy Intake (kJ d-1) AN(C)OVA = Analysis of (Co-)Variance MPE = Mole percent excess BM = Body mass (g) MR = Metabolic rate (kJ d-1) BP = Bound pool of phenylalanine (mpe) n = Sample size C+ = Control mice with wheels N = Size of the body water pool (moles) CAT = Catalase OM = Organ mass (g) CC = Mice housed at 10°C p = Propability COT= Costs of Transport (kJ km-1) PIR = Passive infrared CTRL = Control mice rCO2 = Rate of carbondioxide production CW = Mice housed at 10°C and 22°C RMR = Resting Metabolic Rate (kJ d-1) DEE = Daily Energy Expenditure (kJ d-1) ROS = Reactive oxygen species DL = Dry lean mass (g) RQ = Respiratory Quotient DLW = Doubly labeled Water RWA = Wheel-running activity (km d-1) ERS = Event Recording System S- = Selected mice without wheels FE = Food efficiency (g kJ-1) S+ = Selected mice with wheels FFM = Fat free mass (g) SD = Standard deviation FP = Free pool of phenylalanine (mpe) SEL= Selected mice FSR = Fractional synthesis rate (% d-1) SEM = standard error of the mean GC/MS = Gas Chromatography / Mass SOD = Superoxide Dismutas spectrometry SSA = Sulphosalisylic acid GEI = Gross Energy Intake (kJ d-1) t = Time (s) GLM = General Linear Model V = Running speed (km h-1) -1 GPx = Glutathione Peroxidase V˙ CO2 = Production of carbondioxide (l h ) -1 -1 kd = Fractional turnover rate of 2H (d ) V˙ O2 = Oxygen consumption (l h ) ko = Fractional turnover rate of 18O (d-1) W = Waste (kJ d-1) LEP = Lifetime energy potential (kJ Live-1) WW = Mice housed at 22°C 194 Samenvatting – Dutch summary 196 Samenvatting

Wordt de levensduur bepaald door de energieomzetting?

De “Rate of Living” theorie Wij weten allemaal dat een auto of een elektrisch apparaat dat vaker en intensiever gebruikt wordt eerder kapot gaat. Hetzelfde zou van toepassing kunnen zijn op de levensduur van mensen en dieren. De opvatting dat de levensduur samenhangt met de hoeveelheid verbruikte energie, is bekend als de “rate of living” (=snelheid van leven) theorie. Aan het begin van de vorige eeuw ontdekte Max Rubner dat de hoeveelheid voedsel die verschillende dieren eten (per gram lichaamsgewicht), lager was in bij soorten met een langere levensduur. Hij berekende de energie inname per gram lichaamsgewicht gedurende het gehele leven – wat nu de life-time energy potential (LEP) genoemd wordt – voor 5 verschillende diersoorten (cavia, hond, kat, koe en paard) en vond dat deze ongeveer gelijk was in alle soorten. Per gram lichaamsgewicht verbruikten een cavia en een paard dus een gelijke hoeveelheid energie in hun leven. Als dat binnen een soort ook op zou gaan, zou dit betekenen dat een dier dat meer energie verbruikt dan een soortgenoot een korter leven zal hebben. Een verklaring voor het verband tussen energie verbruik en levensduur wordt gegeven in de “free radical” (=vrije radikalen) theorie van Harman (1956). Deze theorie stelt dat veroudering en dood veroorzaakt worden door de gevolgen van toxische bij-produkten die vrijkomen tijdens de energie omzetting (metabolisme). Energie omzetting vindt plaats in onderdelen in de cel, de mitochondria, waar de ingeadem- de zuurstof en bepaalde substraten omgezet worden in kooldioxide en water, waarbij ATP gevormd wordt. ATP is als het ware een klein energie bommetje dat alle cel- len in het lichaam van energie voorziet. Bij de omzetting van zuurstof in het mito- chondrion worden zeer kortstondig de vrije radikalen gevormd. Als deze radikalen niet onschadelijk gemaakt worden, kunnen ze schade veroorzaken aan DNA, eiwitten en vetten. Tijdens de veroudering zou deze schade zich opstapelen, wat uitein- delijk tot het falen van de cel en celdood kan leiden. Er bestaan enkele mechanis- men die bescherming bieden tegen de effekten van deze vrije radikalen, waaronder antioxidant enzymen, eiwit synthese en DNA reparatie. Antioxidant enzymen reageren met de vrije radikalen waardoor zij onschadelijk worden. De vrije radikalen die niet worden weggevangen kunnen nog steeds schade veroorzaken, maar de schade kan deels gerepareerd worden door het vervangen en nieuw produceren van eiwitten (eiwit turnover) of via DNA reparatie. De rate of living theorie veronderstelt in zijn scherpste formulering dat de levenslange energieomzet per gram lichaamsgewicht min of meer een constante is. Een verhoging van de energieomzet zou dan leiden tot een voorspelbare reductie van de levensduur. Een argument hiervoor is het reeds een halve eeuw bekende feit dat ratten en muizen die op rantsoen gezet zijn veel langer leven dan hun soortge- noten die zich onbeperkt aan voedsel tegoed kunnen doen. Dat heeft tot uitgeberei- de discussies geleid. Tegenstanders van de rate of living theorie wijzen er op dat de dieren op rantsoen veel lichter zijn, en dat hun energieverbruik per gram lichaamsgewicht niet gereduceerd is. De theorie zou dus in het geheel geen langere levens- Dutch summary 197

duur voorspellen. Daar is weer tegen in gebracht dat de stofwisseling zeer uiteen- loopt tussen verschillende delen van het lichaam, en dat lichaamsgewicht dus eigenlijk niet de relevante maat is. Men zou het gewicht moeten weten van de orga- nen in het lichaam die de hoogste stofwisseling hebben, met name hersenen, hart, nieren, lever, en niet het gewicht van huid en skelet meerekenen, weefsels die nau- welijks meedoen aan de stofwisseling. Voor een complete toetsing van de theorie zou het daarom noodzakelijk zijn zowel de energieomzet als de lichaamssamenstelling te kennen van dieren waarbij het energieverbruik verhoogd of verlaagd wordt en de levensduur bepaald. In dit proefschrift wordt precies dit gedaan, in twee omvangrijke experimenten. In het ene wordt getracht de energieomzet van muizen te verhogen langs genetische weg, in het andere door de temperatuur te veranderen. Deze experimenten hebben elk circa 3 jaar geduurd, tot de laatste muis spontaan gestorven was. Ze vinden hun neerslag in Deel II (Energieverbruik en Veroudering) van het proefschrift. Daarnaast wordt in deel I Activiteit & Energieverbruik een aantal experimenten beschreven waarin de relatie tussen energieverbruik en activiteit onderzocht wordt. Gebruik is dus gemaakt van muizenlijnen die zijn geselekteerd op een hoge loopwielactiviteit. Wanneer deze geselekteerde muizen op een leeftijd van 6-8 weken de beschikking krijgen over een loopwiel in hun kooi, dan lopen zij gemiddeld 10 km per dag, dat is 2,7 keer zo ver als controle (= niet-geselekteerde) muizen op deze leeftijd. Ze zijn kleiner en hebben minder vet dan controle muizen. Wel hebben ze een hogere voedselopname. Naast de relatie tussen energieverbruik en activiteit is gezocht naar verdere aanpasssingen die zijn ontstaan in deze op activiteit geselekteerde muizen. Heeft de selektie op een hoge loopwielactiviteit ook geleid tot een verhoging van de efficiëntie van het lopen en verbruiken geselekteerde muizen dus minder energie per afgelegde kilometer (Hoofdstuk 2)? Zijn zij beter voorbereid op een voedselsituatie waarin zij grotere afstanden moeten afleggen voor eenzelfde hoeveelheid voer (Hoofdstuk 3)? Zijn er verschillen in metabole hormo- nale factoren tussen de geselekteerde en controle muizen (Hoofdstuk 4)? En zo ja, zijn zij hierdoor ook beter beschermd tegen het ontwikkelen van het metabole syndroom (Hoofdstuk 5)? In beide hoofdexperimenten hebben wij het energieverbruik beïnvloed (door de activiteit te verhogen of de omgevingstemperatuur te verlagen) en de effecten hiervan op de levensduur, voedselopname, loopwiel activiteit, energieverbruik en lichaamssamenstelling gemeten (Hoofdstuk 6 en 8). In beide experimenten is bovendien onderzocht of muizen met een hoog energieverbruik zich beter bescher- men tegen vrije radikalen door een verhoogde antioxidant enzym activiteit en/of eiwit synthese snelheid (Hoofdstuk 7 en 9)? 198 Samenvatting

Activiteit en Energieverbruik

Hoofdstuk 2 beschrijft welke veranderingen optraden in loopwielgedrag en energieverbruik bij muizen gehuisvest bij verschillende omgevingstemperaturen. Daarnaast is onderzocht of muizen geselekteerd op een hoge loopwielactiviteit een andere reactie laten zien op de omgevingstemperatuur. Het zou gunstig kunnen zijn om bij lage temperaturen veel te rennen als de warmte die hierbij geproduceerd wordt gebruikt kan worden voor thermoregulatie. Controle en geslekteerde muizen werden blootgesteld aan drie temperaturen (10, 20 en 30°C) en hun loopwielactiviteit en energieverbruik werd gemeten. Gemiddeld hadden de geselekteerde muizen door hun hogere activiteit ook een hoger energieverbruik dan de controle muizen. Bij de laagste temperatuur liepen beide groepen minder dan bij de twee hogere temperaturen. De relatie tussen loopsnelheid en energieverbruik was gelijk bij alle temperaturen. Dit bewijst dat de muizen geen energie besparen door de overtollige warmte tijdens het rennen in de kou te gebruiken voor warmteregulatie. Er waren geen verschillen in de kosten van het lopen (in kilojoule per kilometer) tussen geselekteerde en controle dieren. De geselekteerde muizen rennen dus niet effciienter dan de andere. Zijn de selectielijnen wel beter aangepast aan een situatie waarin zij grotere afstanden moeten lopen om aan voedsel te komen? Dit is onderzocht in Hoofdstuk 3. Loopwielen werden aangesloten op voedselverdelers die een voerbrokje van 45 mg loslieten wanneer het dier een bepaald aantal rondjes had gelopen in het wiel. Dat aantal rondjes per ‘pellet’ werd langzaam opgevoerd en veranderingen in lichaamsgewicht, voedselopname, stress hormoon en energie budget werden gemeten. Het lichaamsgewicht nam af en het corticosteron niveau nam toe (wat duidt op stress) zowel bij controle als bij geselekteerde muizen. Beide typen waren in staat hetzelfde maximale loopniveau te bereiken. Op dit niveau lieten ze ook dezelfde veranderingen ten opzichte van het basisniveau zien in lichaamsgewicht, hormoon niveau’s en energiebudget. Op het hoge werkniveau was er een positieve relatie tussen de afgelegde afstand en het energieverbruik, maar er waren geen significante verschillen meetbaar tussen controle en geselekteerde dieren. Dieren die veel ren- den, hadden in rust een lager metabolisme, maar gemiddeld over de dag een hoger energieverbruik. Metabole hormonen spelen een belangrijke rol in het handhaven van de energie- regeling in het lichaam. In Hoofdstuk 4 is onderzocht of er verschillen zijn ontstaan in de niveaus van de hormonen leptine en adiponectine tussen controle en geselekteerde muizen en of de aanwezigheid van een loopwiel deze niveau’s beinvloedt. Bij de geselekteerde muizen (met of zonder loopwiel) vonden wij verhoogde niveau’s van adiponectine. De leptine in het bloed (gecorrigeerd voor vet massa van de dieren) verschilde niet tussen de groepen. De hogere adiponectine in het bloed van op aktiviteit geselecteerde muizen zou betrokken kunnen zijn bij het bepalen van het lagere gewicht bij een verhoogde voedselopname. Het is bekend dat adiponectine injecties bij muizen leiden tot gewichtsafname bij een gelijke voedselopname. Ook Dutch summary 199

speelt adiponectine een belangrijke rol bij de insuline resistentie. Insuline resistentie ontstaat met name in mensen die leiden aan obesitas en veroorzaakt een aantal problemen voor de gezondheid. De geselekteerde muizen met hun hoge adiponectine niveau zijn dus wellicht beter beschermd tegen het ontwikkelen van insuline resistentie en andere sympto- men van het metabole syndroom. Om dit te testen hebben wij in Hoofdstuk 5 geselekteerde vrouwen en mannen blootgesteld aan een vet dieet en het effect hiervan op voedselopname, lichaamsgewicht en metabole hormonen onderzocht. Op het vet dieet lieten de vrouwtjes in de geselekteerde lijn geen gewichtstoename zien, terwijl de mannen en vrouwen van de andere groepen wel zwaarder werden. Ook was de voedselopname in de geselekteerde vrouwen verhoogd op het vette dieet, terwijl de voedselopname lager was in de andere groepen. Dit wijst erop dat de voedseleffi- ciëntie verlaagd was bij de vrouwtjes. Deze waren dus beter beschermd op het vet dieet. Ze waren ook de enige groep die moeite had met het wegwerken van glucose op dit dieet. Juist doordat deze dieren anders reageren op het vet dieet zijn zij een goed model om het metabole syndroom te bestuderen.

Metabolisme en Veroudering

Voor het toetsen van de rate of living theorie heb ik dus gebruik gemaakt van de op aktiviteit geselecteerde muizen. Deze selectie is door Garland al 31 generaties vol- gehouden. Muizen uit deze lijn leggen inmiddels een ruim twee keer grotere afstand afleggen per dag in een loopwiel dan controle muizen. In het tweede experiment heb ik het energieverbruik verhoogd door muizen bloot te stellen aan kou (10°C) en te vergelijken met een controle groep die werd gehuisvest bij 22°C. In deze verouderingsexperimenten bestond elke experimentele groep uit 100 dieren. Zestig hiervan werden regelmatig gewogen en voor elk van hen werd de dag geno- teerd dat zij dood gingen. Bij de overige 40 muizen werden op verschillende leeftijden (2-3, 10-11, 18-19 en 26-27 maanden) voedselopname, stofwisseling (met behulp van de dubbel gemerkt water methode) en lichaamssamenstelling bepaald – op elke leeftijd bij ~8 muizen.. Ook werden monsters van de lever, hart, spier en vetweefsel ingevroren voor analyses van antioxidant enzymen, ontkoppelings eiwitten en eiwit synthese snelheid. In het eerste experiment (Hoofstukken 6 en 7) hebben wij gebruik gemaakt van drie experimentele groepen: 1. Controle muizen met loopwiel (C+), 2. Geselek-teerde muizen met loopwiel (S+), en 3. Geselekteerde muizen zonder loopwiel (S-). Ook in het tweede experiment (Hoofstukken 8 en 9) hebben wij metingen verricht aan drie groepen muizen: 1. Dieren gehuisvest bij 22°C (warm, WW), 2. Dieren gehuisvest bij 10°C (koud, CC) en 3. Dieren die het eerste deel van hun leven (tot 15 maanden) gehuisvest waren bij 10°C en vervolgens bij 22°C (koud-warm, CW). Deze laatste groep test de voorspelling van de rate of living theorie dat een hoog energieverbruik vroeg in het leven ook een effect heeft op de uiteindelijke levensduur. 200 Samenvatting

In beide experimenten was het energieverbruik sterk verhoogd in de tweede groep (S+ en CC) in vergelijking met de andere groepen in het experiment (Hoofdstukken 6 en 8). Bij de actieve muizen was het energieverbruik verhoogd met 14% en in de muizen in de kou met 46%. Toch leidden deze verschillen niet tot verschillen in levensduur. De geselekteerde muizen gehuisvest met loopwiel hadden een gemiddelde levensduur (704 dagen) die niet verschilde van die van de muizen zonder loopwiel (711 dagen). Ook de dieren in de kou hadden een gemiddelde levensduur (798 dagen) nagenoeg gelijk aan die in de warmte (801 dagen) en de dieren die eerst in de kou en later in de warmte waren gehuisvest (768 dagen). Er bestond dus geen directe relatie tussen de verbruikte energie en de levensduur. De totale energie verbruikt over het leven (LEP) was geen constante voor de soort, ook niet als we die LEP uitdrukken per gram orgaangewicht. Wij vonden geen grote verschillen in de activiteit van antioxidant enzymen of sterke veranderingen in eiwitsynthese (Hoofdstukken 7 en 9), die misschien zou- den verklaren dat de snelheid van veroudering niet toeneemt. Bij de actieve dieren was de eiwitsynthese iets verhoogd in de spieren op jonge leeftijd (als zij veel rennen). Op latere leeftijd, als dit proces belangrijker wordt om de accumulatie van beschadigde eiwitten te voorkomen, vonden wij geen verschillen tussen actieve en niet actieve dieren. In de lever was er op beide leeftijden geen verschil (Hoofdstuk 7). Ook de antioxidant enzym activiteiten waren niet verschillend in hart, lever of spier in de geselekteerde dieren versus de controles. Wanneer dit werd uitgedrukt in de bescherming per verbruikte kilojoule (energieverbruik) bleek de bescherming tegen vrije radikalen ook gelijk in beide groepen. In de koude dieren was de antioxidant enzym activiteit verlaagd in 19 maanden oude dieren vergeleken met de warme dieren op deze leeftijd. Door hun veel hogere energieverbruik was de resulterende bescherming door antioxidanten verlaagd (Hoofdstuk 9). Ondanks deze schijnbaar lagere bescherming in koude dieren, werd de levensduur niet beïnvloed. In beide experimenten was de hoeveelheid lichaamsvet verlaagd in de groepen waarin het energieverbruik verhoogd werd (de S+ en CC muizen). Zowel de actieve geselekteerde dieren, en de dieren gehuisvest in de kou hadden minder vet en leef- den langer dan wij verwachtten gebaseerd op de “rate of living” theorie. Vet kan het ontstaan van tumoren bevorderen. Bij muizen met hoge activiteit of die in de kou zitten is aangetoond dat zij minder tumoren ontwikkelen. Mogelijk wordt dus een negatief effect van energieverbruik op levensduur te niet gedaan door het lage vetgehalte van deze dieren.

Conclusies

Muizen die zijn geselekteerd op een hoge loopwielactiviteit waren niet beter aangepast aan een situatie waarin zij waren gehuisvest in de kou of wanneer zij moesten werken voor hun voedsel en lieten geen andere veranderingen in energieverbruik en lichaamssamenstelling zien dan controles. Wel bleken de vrouwtjes beter bestand Dutch summary 201

tegen de negatieve effecten van een vet dieet zoals het ontwikkelen van obesitas. In geen van beide grote experimenten leidde verhoogd energieverbruik tot een kor- tere levensduur. Dit kon niet verklaard worden door compenserende veranderingen in ontkoppelingseiwit, antioxidant enzym activiteit of eiwit synthese snelheid. Wellicht spelen veranderingen in metabole hormonen als adiponectine in actieve geselekteerde muizen een rol,. Tevens zou het lage vetgehalte van actieve en koude muizen beschermend kunnen werken en de negatieve effecten van hun hogere energieverbruik te niet kunnen doen. Veroudering is een complex proces met vele oorzaken. Terwijl energie zeer goed een rol hierbij kan spelen zijn er vele andere aspecten van verhoogd energieverbuik die een simpele voorspelling als in de rate of living theorie doorkruisen. De Life Time Energy Potential is daardoor geen constante voor de soort, hoe we hem ook uitdrukken, per muis, per gram muis, of per gram van de energetisch duren orga- nen. 202 Addresses of co-authors 203

Addresses of co-authors

Serge Daan: Department of Behavioural Biology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Gertjan Van Dijk: Department of Endoneurocrinology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Mark Doornbos: Department of Endoneurocrinology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Theodore Garland Jr: Department of Biology, University of California, Riverside, 109 University Lab Building, Riverside, CA 92521, USA, [email protected]

Izabella Jonas: Department of Endoneurocrinology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Berber De Jong: Department of Behavioural Biology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Gerald E. Lobley: Metabolic Health Group, Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen, AB21 9SB, Scotland (UK), [email protected]

Peter Meerlo: Department of Molecular Neurobiology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

Kristin Schubert: Department of Behavioural Biology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands, [email protected]

John R. Speakman: Aberdeen Centre for Energy Regulation and Obesity, University of Aberdeen, School of Biological Sciences, Tillydrone Ave, Aberdeen, AB24 2TZ, Scotland (UK), [email protected]

G. Henk Visser: Department of Behavioural Biology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands and Centre for Isotope Research, Nijenborg 4, 9747 AG, Groningen, The Netherlands, [email protected] 204 Dankwoord – Acknowledgements 206 Dankwoord

In de afgelopen jaren heb ik de wondere wereld van de wetenschap heel wat beter leren kennen. Hoewel deze wereld prachtig is en op verschillende momenten euphorie opwekt, heeft het ook zijn duistere kanten. Op al deze momenten is de steun en luisterend oor van collega’s en vrienden onmisbaar en in dit dankwoord wil ik iedereen bedanken die op wat voor manier dan ook heeft bijgedragen aan mijn boekje. Een aantal mensen wil ik graag persoonlijk bedanken. Als eerste wil ik Henk bedanken voor alle praktische en theoretische input. Ik zal nooit de (brakke) zater- dag- en zondagochtenden vergeten waarop we samen met wat muizen in de kelder stonden om bloed monsters te nemen voor dubbel gelabeld water metingen. Je per- soonlijke betrokkenheid bij alles wat ik deed, heb ik erg gewaardeerd. Je hebt me erg laten schrikken toen je vertelde dat je kanker had en ik heb veel bewondering voor de manier waarop je hiermee om bent gegaan. Het is ongelovelijk hoe snel je elke keer weer op het lab verscheen alsof er niks gebeurd was en hoe positief en vol vertouwen je houding was. Ben erg blij dat het er nu op lijkt dat je de strijd voor- goed hebt gewonnen. Als tweede wil ik Serge bedanken. Met name in de laatste fase van mijn proefschrift en toen Henk ziek werd, heb je de begeleiding van hem overgenomen. Op het moment dat ik het overzicht een beetje kwijt was, heb je me geholpen een goed plan te maken en ondanks de strakke deadlines heeft dit goed uitgepakt. Bedankt voor alle vruchtbare discussies en het verbeteren van mijn manuscripten. I also would like to thank John Speakman and Gerald Lobley for making it possible for me to come over to the Rowett institute in Aberdeen to do measurements on the samples I had collected in my ageing experiments. John, thank you for the hospitality you showed me every time I visited Aberdeen. I really enjoyed sharing your office with you all those times, and even though not everybody agrees I still believe your idea to save journal space by organising references differently deserves attention. Gerald, thanks for your statistical insights and the enlightning talks about science and my future. Ted Garland thank you for sending us the ‘runner’ mice I worked with and for inviting me over in Riverside to teach me the tricks of SAS. Also thanks for all the useful discussions about my work. Verder wil ik mijn paranimfen Kristin en Daan bedanken. Kristin, vanaf onze eerste ontmoeting hebben we veel inhoudelijke discussies gehad over experimenten en andere zaken, en bovenal erg veel lol. Ongeveer een week nadat je bij me in huis kwam wonen, brak ik mijn enkel en ik wil je ook erg bedanken voor de goede zorg die je me toen hebt gegeven. Kristin rocks! Daan, toen ik bij gedragsbiologie begon kende ik je al via de duikvereniging, maar onze vriendschap is daarna alleen maar sterker geworden. Ik hoop dat we onze gezellige avondjes in de kroeg vol wetenschap en levensbeschouwingen voort kunnen zetten in de pub! Daan rules! Natuurlijk wil ik ook mijn overige (aio-)collega’s bedanken voor alle leuke avondjes uit en nog veel meer: Peter (de rasoptimist), Kamiel, Ate, Roelof (stoere man), Arjen, Barbara, Nicolaus, Wendt, Bernd (lekker kippetje), Marian, Vivian, Acknowledgements 207

Margriet, Ralf, Martijn, Egbert, Sandra, Sarah, Izabella, Kristina, Ton, Cor, Simon, Mareike, Martha en Domien. De dierverzorgers, Monique, Roelie, Sjoerd en met name Saskia bedankt voor de goede zorgen voor mijn dieren. Gerard, bedankt voor alle hulp; zonder jou zijn we allemaal de weg kwijt. Jackie, David, Fiona, Suzan and my other colleagues at the Rowett for making my time there a pleasure. En ook mijn fantastische master studenten; Alinde, Mark, Berber, Jan-Albert en Annemieke. Dan wil ik Peter M. bedanken. De stage die ik bij jou deed in Chicago heeft voor mij de doorslag gegeven om aio te worden. Nou bedankt daarvoor he! Ook bedankt voor alle hulp bij het verzamelen van bloed en breintjes en voor de sarcastische noot zo af en toe en voor alle kopjes koffie, oh nee, thee. Gertjan v.D., bedankt voor de leuke samenwerking. Dick Visser, mijn boekje ziet er echt prachtig uit, bedankt daarvoor! Ook al hebben zij niet direct bijgedragen aan mijn boekje, zij zorgden voor de juiste afleiding die nodig was om alles tot een goed einde te brengen en dus wil ik al mijn vrienden bedanken voor alles: oa. Gertjan (hij deugt) voor de heerlijk ont- spannen stapavondjes; Suuz, Mayo en Mareike voor alle leuke avontuurtjes; Judith en Johan, voor de gezelligheid en alle keren dat jullie op mijn poezen hebben gepast en mij naar het vliegveld hebben gereden als ik weer eens naar Aberdeen ging; Daan en Ilse voor alle gezelligheid (Ilse ook voor het aanhoren van alle science talk); Henk en Manon; Hiske; Anke; Claes; Dorris; Irene; Iris; Rudolf en al mijn andere vriendjes van biologie, de duikvereniging en onderwaterhockey. Then I’d like to thank Dave for opening up his house to me all those times I visited Aberdeen, and Paula for showing me around the lab and taking me to the underwaterhockey club. Verder wil ik mijn ouders bedanken voor het onvoorwaardelijke vertrouwen in mij. Het is fijn dat jullie proberen me te helpen voorkomen dat ik in dezelfde val- kuilen val als jullie en ook al is dat niet altijd succesvol, jullie steun maakt veel goed. Dan mijn lieve broers Remco en Tijs en hun lieftallige schone dames Patricia en Mathanje. Ik hoop dat jullie de komende jaren net zo van Schotland gaan houden als ik. En natuurlijk Marijn, mijn lieve neefje, die op ieders gezicht een glimlach tovert (ook al huilt ie zelf). One of the best things science brought me may be meeting you in a pool in Aberdeen. Jon thanks for your unconditional love.

“I cannot control the truth of death, whatever my desperation. I can only make certain that those moments of my life I have remaining are as rich as they can be”. Drizzt Do’Urden in The Icewind Dale Trilogy – The Halflings Gem – Book 1: Halfway to Everywhere by R.A. Salvatore

CARPE DIEM!! 208