This protocol describes the use of a body composition analyzer and metabolic animal monitoring system to characterize body composition and metabolic parameters in mice. An obesity model induced by high-fat feeding is used as an example for the application of these techniques.
Alterations to body composition (fat or lean mass), metabolic parameters such as whole-body oxygen consumption, energy expenditure, and substrate utilization, and behaviors such as food intake and physical activity can provide important information regarding the underlying mechanisms of disease. Given the importance of body composition and metabolism to the development of obesity and its subsequent sequelae, it is necessary to make accurate measures of these parameters in the pre-clinical research setting. Advances in technology over the past few decades have made it possible to derive these measures in rodent models in a non-invasive and longitudinal fashion. Consequently, these metabolic measures have proven useful when assessing the response of genetic manipulations (for example knockout or transgenic mice, viral knock-down or overexpression of genes), experimental drug/compound screening and dietary, behavioral or physical activity interventions. Herein, we describe the protocols used to measure body composition and metabolic parameters using an animal monitoring system in chow-fed and high fat diet-fed mice.
Metabolism underpins many aspects of normal cellular, organ, and whole-body physiology. Consequently, in the setting of various pathologies, alterations to metabolism may directly contribute to the underlying condition or may be adversely impacted as a side-effect of the pathology. Traditionally, metabolic research and studies into energy balance have been concentrated on the field of obesity and related conditions such as insulin resistance, pre-diabetes, glucose intolerance, cardiovascular disease, and diabetes. This research is warranted given the escalating prevalence of such conditions worldwide and the individual, societal, and economic costs these conditions inflict. As such, the development of prevention strategies and new therapeutics to target obesity is a continuing goal in research laboratories around the world and preclinical mouse models are heavily relied upon for these studies.
While weighing mice provides a reliable assessment of weight gain or loss, it does not provide a breakdown of the different components that make up whole-body composition (fat mass, lean mass, free water as well as other components such as fur and claws). The weighing of fat pads at the completion of studies once the mouse is deceased provides an accurate measure of different fat depots but can only provide data for a single time point. As a consequence, it is often necessary to enroll multiple cohorts to investigate the development of obesity over time, significantly increasing animal numbers, time, and costs. The use of dual-energy X-ray absorptiometry (DEXA) provides an approach to assess body fat and lean tissue contents and allows the researcher to obtain data in a longitudinal fashion. However, the procedure requires mice to be anesthetized1, and repeated bouts of anesthesia may impact the accumulation of adipose tissue or impact other aspects of metabolic regulation. EchoMRI utilizes nuclear magnetic resonance relaxometry to measure fat and lean mass, free water, and total water content. This is achievable due to the creation of contrast between the different tissue components, with differences in the duration, amplitude and spatial distribution of generated radio frequencies allowing the delineation and quantification of each tissue type. This technique is advantageous as it is non-invasive, quick, simple, requires no anesthesia or radiation, and, importantly, has been positively validated against chemical analysis2.
A key consideration of obesity and related research is the energy balance equation. While fat accumulation is more complicated than purely energy in (food intake) versus energy out (energy expenditure), they are vital factors to be able to measure. Daily energy expenditure is the total of four different components: (1) basal energy expenditure (resting metabolic rate); (2) the energy expenditure due to the thermic effect of food consumption; (3) the energy required for thermoregulation; and (4) the energy spent on physical activity. As energy expenditure generates heat, measuring heat production by an animal (known as direct calorimetry) can be used to assess energy expenditure. Alternatively, measurement of inspired and expired concentrations of O2 and CO2, allowing for determination of whole-body O2 consumption and CO2 production, can be utilized as a way to indirectly measure (indirect calorimetry) heat production and consequently calculate energy expenditure. An increase in food intake or a decrease in energy expenditure will predispose mice to weight gain and observations of changes in these parameters can provide useful information of likely mechanisms of action in particular models of obesity. A related metabolic parameter of interest is the respiratory exchange ratio (RER), an indicator of the proportion of substrate/fuel (i.e., carbohydrate or fat) that is undergoing metabolism and being utilized to produce energy. Consequently, measurement of food intake (energy consumed) combined with physical activity levels, O2 consumption, RER, and energy expenditure can provide a broad understanding of an organism's metabolic profile. One method to gather such data is to use a comprehensive laboratory animal monitoring system (CLAMS), which is based on the indirect calorimetry method to measure energy expenditure and has the added capabilities of determining physical activity levels (beam breaks) and food intake via scales incorporated into the measurement chamber.
In this protocol we provide a straight-forward description of the use of a body composition analyzer to assess body composition in mice and a metabolic animal monitoring system to measure aspects of metabolism. Considerations and limitations for these techniques will be discussed as well as suggested methods of analysis, interpretation, and data representation.
All experiments described were approved by the Alfred Medical Research Education Precinct Animal Ethics Committee (AMREP AEC) and mice were provided humane care in line with the National Health and Medical Research Council (NHMRC) of Australia Guidelines on Animal experimentation. Animals were administered their prescribed diet and water ad libitum and housed in a temperature-controlled environment (~21 – 22 °C) with a 12 h light and 12 h-dark cycle. Seven week old male mice (on a C57Bl/6J background) were fed either regular normal chow diet (energy content 14.3 MJ/kg, consisting of 76% of kJ from carbohydrate, 5% fat, 19% protein; see Table of Materials) or for the high fat-feeding group, a high fat diet (HFD) (energy content 19 MJ/kg, consisting of 36% of kJ from carbohydrate, 43% fat, 21% protein, Specialty Feeds) for 3 weeks. Body weight and body composition measurements using an EchoMRI machine were made weekly while the metabolic monitoring analysis took place in a CLAMS after 3 weeks of the diet.
1. Body Composition Analyzer Procedure
Note: To function optimally, the EchoMRI 4-in-1 used in this protocol should be contained within a room where the air temperature is stable and does not fluctuate. Ideally this should be constantly monitored. Moving of the machine and interruptions to power should also be avoided if possible. If the power supply has been interrupted and the system has to be restarted, allow at least 2 – 3 h for the machine to warm up before using it again. Before starting, ensure that you are wearing correct personal protective equipment.
2. Metabolic Animal Monitoring System Procedure
NOTE: The system requires ~2 h to warm up and stabilize. If the machine has been turned off, it must be switched on to allow the Zirconia cell to be heated to 725 °C. Also we generally place mice in the body composition analyzer a day prior to entering the animal monitoring system to avoid any issues with restraint stress.
The results seen in Figure 3 display a typical change in body composition parameters upon high fat feeding, as measured via EchoMRI. At baseline there was no difference in any parameter measured (Figure 3A-F). However, after just 1 week of high fat feeding, there was a significant increase in body weight, fat mass, and fat mass percentage in the HFD group (Figure 3A,B,D). The magnitude of the differences between the two groups for these measures continued to increase over the 3 week dietary intervention. Lean mass, free water, and total water content (Figure 3C,E,F) did not differ between the groups at any time point. It can also be seen that the chow fed mice continued to put on weight over the study period (Figure 3A) and that this was due to an increase in lean mass (Figure 3C) rather than a fat mass increase (Figure 3B).
As can be seen in Figure 4, three weeks of high fat feeding led to a number of changes as detected in the metabolic animal monitoring system. VO2 when not adjusted for body weight was significantly higher in the heavier high fat fed mice (Figure 4A). Notably, normalization of VO2 via two different factors resulted in two different outcomes. Normalization to total body weight led to no difference in VO2 between the standard chow fed and high fat fed mice, while normalization to lean mass produced a significant difference (Figure 4B,C). These results demonstrate that normalization of VO2 data by dividing by mass variables significantly affected the results, and caution should be exercised when interpreting VO2 data when it is expressed in such a way. For a detailed discussion of how to express VO2 data and the effects of normalizing to different parameters see the excellent discussion in Tschop, et al.5 In their guide to the analysis of mouse energy metabolism, Tschop and colleagues suggest the use of analysis of co-variance (ANCOVA) to statistically interrogate the effects of body weight or body composition on energy expenditure and food intake data. In this case, performing an ANCOVA on the data shown in Figure 4A, using body weight as the covariate, reveals that no statistically significant difference exists between normal chow and HFD, thus indicating that once accounting for body weight, there is no difference in oxygen consumption between the groups. This result can be easily visualized when plotting VO2 against body weight as a scatterplot as shown in Figure 4D. Plotting VO2 against body weight (Figure 4D) demonstrates that the VO2 data lie on a common line in relation to body weight, with the heavier animals consuming more oxygen. Of note, plotting VO2 against lean mass demonstrates that the VO2 data lie on two distinct lines in relation to lean mass (Figure 4E).
RER was significantly lower in the high fat fed mice, indicating fat utilization over carbohydrate utilization when fed the high fat diet (Figure 5A). Energy expenditure (heat) without normalization was increased in the heavier animals, likely due to the animals having more metabolically active tissue (Figure 5B), with this difference being lost once normalized to body weight (Figure 5C). Also note the increases in VO2, RER, and energy expenditure in the dark cycle as compared to the light cycle when the mice are more active. These differences represent the classical daily alterations in metabolism that occur in mice. While in this example, we have divided the data into 12 h blocks, division of the data further into smaller time epochs can also be useful. Physical activity levels are also a factor that contribute to energy expenditure. These were not different between groups, suggesting that a decrease in movement was not the driver of the obese phenotype in the high fat fed mice (Figure 5D).
The other side of the energy balance equation is the amount of energy that is consumed and enters the body. To look at this aspect of metabolism we analyzed the amount of food that the mice consumed while in the metabolic animal monitoring system. As can be observed in Figure 6A, the mice ate the same quantity of food as measured by weight or when normalized to body weight (Figure 6B). (ANCOVA can again be used to assess the impact of body weight on food intake.) Normalization of food intake to body weight may be an important step to consider if energy expenditure has also been normalized to weight, thus keeping each side of the energy equation in balance. While the mice ate the same quantity of food, it is important to account for the energy density of each of the diets used. When taking this factor into account, we observe the mice on the HFD consuming more energy (Figure 6C) and from these experiments it is likely that this is driving the obese phenotype. That is, since the mice are taking in more energy, but they are not proportionally expending more energy, their obesity can be attributed to energy storage.
통계학
All data in this paper are presented as mean ± standard error of the mean (SEM). Statistical significance was set at p <0.05. * indicates p <0.05, ** indicates p <0.01, *** indicates p <0.001, and n = 6 per group unless indicated. Investigators were unable to be blinded to the dietary group intervention due to a difference in color of the diets. The mice were randomly chosen as to which diet they were given.
Figure 1: Correct placement of mouse COSTS and small animal specimen holder containing mice within the body composition analyzer. To perform a system test using a calibration standard (COSTS) or for scanning of mice within the small animal specimen holder, place each inside the gantry of system. The red arrows indicate the cylinder in which the mice will be contained entering the gantry of the machine.
Figure 2: Assembly of individual chambers. A) Place the food hopper in the center of the balance. B) Insert the platform into each chamber and place chamber over the hopper. C) Place mice in the chambers individually and secure lid. D) Position the water bottle and fasten.
Figure 3: Body composition analysis over 3 weeks of a high fat diet. A) Body weight, B) fat mass, C) lean mass, D) fat mass percentage, E) free water content, and F) total water content. Circles represent normal chow diet, squares represent HFD. Please click here to view a larger version of this figure.
Figure 4: Metabolic parameters obtained from metabolic animal monitoring system experiments after 3 weeks of the respective diets. Mice were housed in the chambers for 48 h with the first 24 h acting as familiarization. The data obtained from the second 24 h was analyzed and presented in these figures. A) Raw VO2 rates, B) VO2 normalized to body weight C) VO2 normalized to lean mass, D) scatterplot for unadjusted VO2 (total 24h period) to body weigh,t and E) unadjusted VO2 to lean mass. A-C White bars represent normal chow diet, black bars represent high fat diet. D-E Circles represent normal chow diet, squares represent HFD. Please click here to view a larger version of this figure.
Figure 5: A) Respiratory exchange ratio (RER), B) heat (energy expenditure), and C) heat normalized to body weight. D) Activity levels calculated as the sum of the ambulatory X and Y beam breaks and Z beam breaks. White bars represent normal chow diet; black bars represent HFD.
Figure 6: Food intake data obtained in the system for the final 24 h. A) Food intake in grams, B) food intake normalized to body weight and C) calculated energy intake. n = 4 – 5 (3 mice were excluded due to making a large mess with their food). White bars represent normal chow diet; black bars represent HFD. Please click here to view a larger version of this figure.
Critical steps
The protocols described herein provide an example of ways in which to measure body composition and various metabolic parameters in mice using a body composition analyzer and a metabolic animal monitoring system. For both techniques, it is critically important to ensure that the machines are working optimally, and to do this, it is imperative that the researcher performs a system test for the body composition analyzer and calibrates to a known gas composition for the metabolic animal monitoring system prior to use of the equipment. This will ensure greater consistency of results and the opportunity to detect any potential issues with the machinery.
The way in which data is normalized for the metabolic animal monitoring experiments is also vitally important to ensure the validity of the results obtained from the technique. As indicated in our representative results (Figure 4A-E) VO2 can be reported in a number of different ways: its absolute rate (L/min), relative to the body mass of the mouse (mL/kg*min), or relative to lean body mass (mL/kgLBM*min) if that data is available (for example obtained from a body composition analyzer). Depending on the phenotype, it may be more appropriate to normalize the values a particular way to rule out any potential bias. For example, if an animal has increased body mass, they have more tissue that is available and able to consume oxygen and naturally their energy expenditure is higher. Normalizing to total body mass may not be the best option as it will bias towards the observation of a decrease in oxygen consumption per unit of mass, even though the oxygen consumption of the tissues may not be different. As an alternative to normalizing to body weight, one can normalize to the lean body mass of the mouse. As lean tissue mass is primarily responsible for oxygen consumption, and lean mass is typically unaltered or only modestly different between experimental groups, normalization in this manner may be a more representative way of expressing VO2 data. It should be noted that the lean mass compartment is comprised of many different tissues, all with different metabolic rates, and consequently normalization in this manner may not be appropriate or provide any insight into which lean mass component is driving the change. Also, it rules out the contribution of the fat mass component on metabolism.
Given these issues, an alternative statistically-based method has also been proposed5,6. Analysis of covariance (ANCOVA) is a statistical test that allows the comparison of a variable (e.g., energy expenditure) across multiple groups while correcting for other factors or variables termed covariates. In this manner factors such as body weight, fat mass and lean mass can be included as variables that influence energy expenditure, but even this method has its own specific assumptions6, including the fact that using multiple variables in ANCOVA is likely to invalidate it unless the variables are independent of each other. Given there seems to be no perfect or universally agreed single way to normalize and present VO2 or energy expenditure data, it may be appropriate to display and present the data in a number of ways to give the clearest picture of the phenotype to the reader. Physical activity levels can increase oxygen consumption, and so in animals which have activity phenotypes (an increase or decrease), it may also be necessary to account/normalize for changes in movement to determine if this can account completely or partially account for any change in VO2.
Modifications and troubleshooting
The representative results displayed in this protocol were obtained from experiments conducted at a room temperature of 21 – 22 °C. The thermoneutral zone of a mouse is approximately 30 °C, so in a traditional animal house with its temperature set to 20 – 22 °C for human comfort, a mouse is put under thermal stress. To counter this, non-shivering thermogenesis is activated at these colder temperatures, resulting in an up to a 2-fold increase in energy expenditure between mice housed at 20 °C compared with those housed at 30 °C7. The environmental housing of mice is an important consideration for these experiments as it has been shown that housing of mice at thermoneutrality can potentiate the development of some conditions such as atherosclerosis8 and high fat diet-induced non-alcoholic fatty liver disease (NAFLD) pathogenesis9. Environmental temperature is therefore also an important consideration when conducting experiments in a metabolic animal monitoring system, as a phenotype can be present at certain temperatures but not at others, which could point to a potential mechanism of action. One such scenario could be a phenotype that involves the activation of recruited beige fat whereby a greater quantity of this tissue allows for a larger increase in thermogenesis under cooler conditions10. Thus, it may be necessary to modify the environmental set up that was described in these current experiments and to conduct experiments under multiple environmental temperatures to get an accurate depiction of the true metabolic status of the model. For troubleshooting due to technical errors, it may be necessary to contact the manufacturers directly for instruction. If there are issues with this type of body composition analyzer it is recommended to perform a Repeat Scans test, which runs 25 scans against the COST. The company will need this information for diagnostics. Similarly with the metabolic animal monitoring system, if issues arise, collect the data files from the last time the system worked well and the files from when the issues arose so that support can make a likely diagnosis.
Limitations
While the body composition analyzer provides excellent data on whole-body fat accumulation, it doesn't allow for the determination of regional adipose depots. This is important in the field of obesity research, as not all fat is the same, with the location that the fat has accumulated and its functional properties being particularly important. Indeed, the protective effects of subcutaneous fat depots (or metabolically healthy fat) have been described11. Micro-computed tomography (micro-CT) can discriminate between subcutaneous and visceral fat12, as can magnetic resonance imaging (MRI) analysis13. Use of these techniques can provide further information on the site of adipose accumulation. The metabolic animal monitoring system also has its limitations. While total daily energy expenditure can be measured, the system is not capable of discerning between the different components that make up energy expenditure. A further limitation of the system is that it is possible that obesity can develop without a measurable decrease in energy expenditure detected via these types of systems, even independently of food/energy intake alterations. Studies have shown that small decreases in energy expenditure, which are substantial enough to cause significant weight gain over the long term, cannot be robustly detected in such metabolic systems over the short term14,15,16. While we have used an n of 6 per group in the current study to demonstrate this methodology as an example study, to detect small differences in energy expenditure that could contribute to obesity likely requires many more mice5. Advancements in the resolution of detection in these systems and the ability to perform these types of studies over a longer time period will aid in the ability to detect these smaller but significant changes. With regards to the measurement of food intake, we have typically observed that 24 h food intake in mice housed within the metabolic animal monitoring system is lower than would be observed in the home cage, likely due to the reasons discussed above. Therefore, in addition to monitoring food intake in this system, we additionally assess food intake in the home cages of mice. While this can only be done in a situation where mice from particular experimental groups are housed separately, it has the advantage of allowing near continuous daily assessment. The investigator simply weighs the amount of food in the hopper at a specified time of day, always accounting for food scattered throughout the cage, and then divides this total amount of food consumed by the number of mice present in the cage.
Future applications
While within this review we have used obesity acquired via high fat feeding as an example of a disease state where measuring body composition and metabolic parameters are useful, the use of this equipment is far from confined to this research field. The use of these techniques is also valuable when studying diseases such as diabetes, cardiovascular disease, age-related sarcopenia, frailty, cancer-cachexia, muscular dystrophies, and lipodystrophy. While the initial costs of purchasing such infrastructure is considerable, the ability to use the equipment across multiple and diverse fields of medical research mitigates this initial cost. Furthermore, ongoing reagent and consumable costs are minimal for these machines; however, preventative maintenance and servicing must be considered and budgeted for.
Just as lean mass obtained via body composition analysis may be an important normalization factor for oxygen consumption derived from the metabolic animal monitoring system, lean mass determination can also be used to normalize drug/test dosages. For instance, in metabolic studies, it is commonplace to perform an intraperitoneal or oral glucose tolerance test (GTT), or an intraperitoneal insulin tolerance test (ITT). These tests examine the ability of a mouse to dispose of a glucose load or respond to insulin. Alterations in blood glucose levels in response to these tests provides information on the level of whole-body glucose and insulin tolerance in the model. Traditionally, the glucose and insulin bolus administered in these tests is dosed as per the body weight of the mouse. However, as models of obesity accumulate fat mass over lean mass, dosing per body weight could bias the heavier model towards glucose intolerance in a GTT as they receive more glucose. This is due to the fact that the liver, skeletal muscle, and brain, organs that dispose of the majority of glucose in the post-prandial state17, are components of the lean mass measurement and rarely or mildly change in most models. Conversely, in an ITT when dosed to body weight, a heavier model which would receive more insulin may appear more sensitive to the glucose lowering effects of insulin purely because it has received a greater quantity. Therefore, if the investigator has access to body composition data, the lean mass may be the most appropriate measure, as opposed to the whole-body mass, for such dosage calculations18. Taking this further, lean mass data obtained from body composition analysis could also be used to dose experimental drugs if the need arose to account for lean mass at the exclusion of fat mass. Another application of the metabolic animal monitoring system that has not been discussed or demonstrated in this manuscript is attaching an enclosed motorized treadmill to the system so that the metabolic parameters discussed herein can also be measured during exercise.
The procedures described in this review can be used to characterize body composition and various metabolic parameters in mice. These measures are applicable to a wide-range of research fields and can provide important information for the characterization of a phenotype. Data derived from these methods can also provide evidence towards underlying mechanisms driving a particular metabolic phenotype. Further development and refinement of these technologies will enable researchers to advance their findings towards therapeutic outcomes.
The authors have nothing to disclose.
We thank the staff from the Alfred Medical Research and Education Precinct Animal Services (AMREP AS) team for their assistance and care of the mice used in this study and for the support of the Operational Infrastructure Support scheme of the Victorian State Government.
4 in 1 system | EchoMRI | 4 in 1 system | Whole body composition analyser |
Canola oil test sample (COSTS) | EchoMRI | Mouse-specific (contact company for cat number) | |
Animal specimen holder | EchoMRI | 103-E56100R | |
Delimiter | EchoMRI | 600-E56100D | |
12 chamber system | Columbus Instruments | Custom built | Metabolic Caging System; includes control program |
Drierite | Fisher Scientific | 238988 | CLAMS consumable |
Calibration gas tank | Air Liquide | Mixed to order | Gas calibration (0.5% CO2, 20.5% O2, balance nitrogen). |
Normal chow diet | Specialty Feeds | Irradiated mouse and rat diet | |
High fat diet | Specialty Feeds | SF04-001 | |
Balance | Mettler Toledo | PL202-S | Balance for weighing mice |
TexQ Disinfectant spray | TexWipe | ||
Hydrogen Peroxide cleaning solution | TexWipe | TX684 |