In vivo quantification of fat content in mice using the Hologic QDR 4500A densitometer

Summary

Purpose

Validation of dual-energy X-ray absorptiometry (DXA) with the Hologic QDR 4500A (QDR 4500) Fan Beam X-ray densitometer for in vivo assessment of body fat content in mice.

Methods

Precision of DXA fat measurement was assessed by repeated in vivo scanning and re-positioning of different sized mice (17.6, 24.6, 34.2 g). DXA fat and total mass measurements were correlated with dissected tissue weights in 240 female adult mice of seven strains (mean weights 21.9–26.8 g). Accuracy of DXA fat tissue measurements was assessed by chemical analysis in a subgroup of 40 female decapitated mice (mean weights 19.6–28.4 g).

Results

Precision of the DXA measurements for fat mass was dependent on body weight (mean coefficient of variation, CV, 34.2 g mouse: 7.53 ± 0.13%; 24.6 g mouse: 32.16 ± 0.17%; 17.6 g mouse: 40.64 ± 0.06%). A moderate to high correlation with the dissected fat tissue weights was found for all seven strains: r = 0.52, p ≤ 0.01 (AJ) to r = 0.83, p ≤ 0.01 (CBA, both mean weight = 22 g). The correlation of DXA measurements with the chemical analysis of the carcass was good to excellent (r = 0.80, p ≤ 0.01).

Conclusion

The results demonstrate that the QDR 4500A DXA can be utilised for in vivo measurements of fat content in mice weighing as little as 20 g, with excellent correlations between tissue dissections and chemical analysis demonstrating high consistency of the measurements. DXA values were consistently slightly lower than those by direct chemical analysis; however, the limits of agreement (mean difference 0.96 g) demonstrated good concordance between the two methods.

Introduction

Obesity is a common yet highly complex and heterogeneous disease. Genetic predisposition and our obesigenic environment play important roles in the pathogenesis of obesity, but the underlying mechanisms remain incompletely understood [1], [2]. Whilst fat accumulation around the body is a direct manifestation of obesity, it is recognised also that the pattern of regional fat distribution plays an important role in hypertension, cardiovascular disease, diabetes, schizophrenia, metabolic and respiratory disorders associated with obesity [3]. Mouse models are now used extensively in obesity research to understand the physiological and pathological mechanisms behind this disease [4], [5]. Phenotypic analysis of body composition can be performed using various techniques [6]. However, most standard laboratory methods are performed on culled animals, significantly limiting their utility for longitudinal studies of changes in body fat content and distribution [4], [7], [8], [9], [10]. Traditional methods, including adipose tissue dissection (which may involve a degree of inaccuracy if some tissue is left behind), lipid composition analysis, the underwater weighing method [11] and direct chemical carcass analysis techniques [10], require sacrificing the animal and therefore cannot be applied in vivo. This precludes the possibility of longitudinal studies in the same animal and limits the use of other analytical techniques requiring tissue sampling. Non-invasive techniques such as total body electrical conductivity (TOBEC) [12], magnetic resonance imaging (MRI) [13], quantitative magnetic resonance (QMR) and dual-energy X-ray absorptiometry (DXA) that already exist in clinical and animal applications are proving to be acceptable and increasingly useful tools in the assessment of body composition in vivo[7], [9], [14], [15], enabling longitudinal studies with interventions.

DXA instruments, specifically designed for whole body scans in humans can be applied to perform small animal scans by using modified algorithms and software analysis. Whether the DXA estimate of fat is relevant and reproducible in experimental animals such as pigs [16] and rats has already been addressed [17], [18], [19]. Whilst several murine studies have assessed the precision and/or accuracy of DXA instruments in determining bone mineral content (BMC) [20], [21], [22], [23], few have also assessed its accuracy in determining body fat content [7], [9]. Many of the studies that have investigated the precision and/or accuracy of DXA-derived body fat measures have used the less widely available “mouse” DXA machines (GE-Lunar PIXImus [9] and Norland pDXA SABRE [22]). Since the practicality and suitability of mice in bone metabolism, obesity and genetic studies is far greater than larger rodents it would be of great benefit to validate the accuracy of the more widely available human DXA for its use in determining body fat content in mice.

Whilst there are many benefits of using DXA there are also many issues associated with its application. Previous studies have attempted to validate the accuracy of DXA in measuring body composition in mice, but the instruments or software assessed in each differ [7], [9]. Since each supplier develops their own software the algorithms differ between machines. In developing these algorithms any assumptions made relating to the percent fat and lean mass and bone mineral density are not specified by the manufacturers. Unlike the human software, which applies different algorithms for different regions of the body, the small animal software utilises the same algorithms for global and sub-regional analysis. Therefore, these factors should be taken into account when interpreting all measures calculated by DXA.

The purpose of the current study was to validate the methodology for quantification of body fat composition using DXA in mice. We evaluated the use of the Hologic QDR 4500A Fan Beam X-ray densitometer for the measurement of body fat composition in seven inbred strains of mice. As outlined in the instruction manual this instrument uses ‘Rat Whole Body and Regional High Resolution’ software to measure body composition and is optimised for adult rats weighing between 200 and 750 g.