Exploring the Effects of Exercise on Brown Adipose Tissue Volumes and T2 Values Using Magnetic Resonance Imaging at 7 Tesla

We aimed to evaluate the effect of exercise on brown adipose tissue (BAT) volumes and T2 values in mice. Methods Twenty-ve female Kunming mice were divided into two groups, a running group (n = 5) and a control group (n = 20). After 4 months, all magnetic resonance imaging (MRI) scans of mice were performed on a 7 Tesla (7T) MR scanner with T2-weighted imaging (T2WI) and a T2 mapping sequence. Interscapular brown adipose tissue (BAT) volumes and T2 values were measured. To reduce the impact of weight on the results, we compared the ratio of BAT volumes to body weights (V/W). The data are expressed as mean ± SD, the BAT V/W and T2 values were compared between the control and running groups using the Wilcoxon rank-sum test, P < 0.05 were considered statistically signicant. Interscapular BAT volumes of the running group (n = 5) and control group (n = 20) were (180.09ml ± 59.80 ml) and (99.98ml ± 35.05ml), respectively. The V/W ratios of the running and control groups were (3.83ml/g ± 0.78ml/g) and (2.17ml/g ± 0.56ml/g), respectively. Interscapular BAT T2 values of the running and control groups were (76.07ms ± 10.82ms) and (61.22ms ± 15.98ms), respectively. Signicant differences were found in the BAT V/W ratios (P = 0.0003, P < 0.001) and T2 values between the two groups (P = 0.0096, P < 0.05). BAT volume correlated positively with BAT T2 value (r = 0.75, p = 0.00002).


Introduction
Two main types of adipose tissues exist in mammals, brown adipose tissue (BAT) and white adipose tissue (WAT). Although they are both adipose tissues, they have different compositions. BAT cells have many small lipid droplets and mitochondria, while WAT cells contain a single large lipid droplet and fewer mitochondria. In the body, the amount of WAT is much larger than BAT. BAT has been veri ed to play an important role in energy expenditure and the maintenance of metabolic balance [1]. For human being, it was once believed that BAT, playing only a small role in adult metabolism, was primarily present in infants and gradually decreased with age [2]. In recent years, with the development of 18Fuorodeoxyglucose positron emission tomography/computed tomography ( 18 F-FDG PET-CT), researchers have found active BAT in adults [3,4]. This BAT consumed glucose and fatty acids participate in nonshivering thermogenesis to generate heat, and the investigators suggested a possible role for BAT in protecting against obesity. Since then, additional studies have found that BAT improved the metabolism of obese, diabetic and dyslipidemic patients [5,6]. Therefore, clarifying BAT mechanisms is of great importance in the treatment of these diseases. At present, assessing BAT activity is mainly based on 18 F-FDG PET-CT, histology, or serologic examination. PET-CT is the main imaging method for BAT evaluation; however, ionizing radiation limits its application, and it can only detect activated BAT. There is an urgent need for noninvasive and relatively convenient evaluation methods. Chen et al. [7] showed that BAT could be distinguished from other tissues on MRI because fat-water content and mitochondrial densities are different between WAT and BAT [1]. At present, there are many kinds of methods for BAT research, such as proton density fat fraction (PDFF) and T2* mapping, however, these methods are complicated and need expensive device that limit their application. On the contrary, T2 mapping were widely used in the clinical routine, such as heart, cartilage evaluation and so on [8,9], so we think it may be a good candidate for BAT research [10].
Currently, research regarding the factors that in uence BAT primarily focus on cold exposure and exercise. Most studies show that cold exposure promotes BAT formation [11][12][13]. When the body is stimulated by cold temperatures, BAT plays an important role in non-shivering thermogenesis to ensure body temperature stability. As we all know, exercise can accelerate glucose and fatty acid consumption, reduce weight, reduce WAT amounts, and improve metabolic conditions. However, no matter in mice or human, the relationship between exercise and BAT remains unclear. Studies looking at the effects of exercise on BAT have shown con icting results [14][15][16][17][18][19]. Therefore, in this study, T2WI and T2 mapping sequences were used to evaluate BAT volumes and T2 values, respectively, and the relationship between exercise and BAT was explored.

Materials And Methods
Animals, the exercise program, and the morphometric analyses Twenty-ve female Kunming mice weighing about 20 g (aged 5-6weeks) were included in the study. The mice were divided into two groups according to their exercise habits (each mouse was subjected to treadmill activities, and we observed a willingness to run, the observation lasted for three days, the mice those ran for more than 1 hour every day in the selection period were considered to have the willingness to run). Those that were willing to run were assigned to the running group (n = 5), and those who did not were assigned to the control group (n = 20). The mice were kept in a room with a constant temperature of 22℃, with a 12/12 h light-dark cycle and su cient food and water. Mice in the running group performed voluntary treadmill running exercises for approximately 2 hours per day. Control mice were kept in cages and were not subjected to running on a treadmill. After 4 months, MR scanning of all the mice was performed. Since the interscapular area has the largest amount of BAT in rodents [20], we focused on this area for evaluations.
After scanning, the mice were sacri ced. Interscapular BAT was immediately dissected out and stored at -80℃ for later processing. BAT tissue sections were obtained and stained with hematoxylin & eosin for morphologic assessments and morphometric analyses.

The imaging protocol and measurements
All the mice were scanned on a 7 Tesla (7T) MR scanner (BioSpec 70/30 USR, Bruker BioSpin MRI GmbH, Ettlingen, Germany) with a mini imaging gradient coil system. All animals were examined using T2WI and T2 mapping sequences. A T2WI sequence adopted rapid acquisitions with relaxation enhancements

Statistical analyses
BAT volumes could be affected by body weight, and therefore, to reduce the impact of weight on the results, we compared the ratios of BAT volumes to body weights (V/W) between the control and running groups. The data are expressed as mean ± SD, statistical analysis was performed by SPSS (version 21.0), the BAT V/W ratios and T2 values between the control and running groups were compared using the Wilcoxon rank-sum test. P-values < 0.05 were considered statistically signi cant. Spearman's correlation was performed to analyze the correlation between BAT volumes and BAT T2 values.

Results
We can see that interscapular brown adipose tissue (BAT) could be distinguished from surrounding tissue on T2 image (Fig. 1) and on T2 mapping arti cial color image (Fig. 2).
The interscapular BAT stained with a hematoxylin and eosin stain. Brown adipocytes with multiple small cytoplasmic lipid droplets were shown in Fig. 3.

Discussion
At present, the main method to evaluate BAT is with 18 F-FDG PET-CT, however, this method uses ionizing radiation, limiting its application. 18 F-FDG PET-CT detects glucose tissue uptake, which could be affected by external factors, such as serum glucose levels, diets, and environmental temperatures. All of these factors could cause BAT concentrations to be underestimated [21]. Recently, MRI begins to be used in BAT research studies as a nonradioactive imaging method. Since BAT has been shown to have different fractions of fatty acids and water content compared with those of WAT, BAT and WAT tissues can be distinguished by MRI [10].
Although many studies have been published on the relationship between exercise and BAT, this relationship is still controversial. In our study, interscapular BAT V/W ratios were higher in the running group than that in the control group, which demonstrated that running exercise promoted BAT proliferation. The result of our study was similar to those of several previous studies [19,17,14,16], whose subjects included rats and mice, and the exercises included swimming and running.
At present, the relatively accepted pathway of BAT activation during exercise is that the SNS is stimulated. The result is that target proteins, such as hormone-sensitive lipase, triglyceride lipase, and monoacylglycerol lipase, are phosphorylated, promoting the degradation of triglycerides in adipocytes and releasing free fatty acids to activate UCP1, resulting in increased thermogenesis [22][23][24]. Peres Valgas da Silva et al. showed that mature brown adipocytes from exercise-trained BAT had reduced glucose uptake in the absence of innervation, demonstrating that neural innervation is greatly signi cant for BAT functions [25].
However, some studies have shown that exercise decreases BAT activity [18,15]. In one study, male rats, subjected to moderately intense treadmill running to replicate endurance exercise, had decreased thermogenic capacities in classic BAT. In this study, the BAT of the trained rats had decreased UCP1 expression and a reduction in fatty acids, indicating that exercise decreased BAT activity. From a physiologic perspective, the authors surmised that increased BAT activity during exercise did not seem reasonable. In this theory, since BAT is a thermogenic tissue, its activation during exercise might seem like a waste of energy since exercise consumes energy and produces heat. To keep the body temperature stable, BAT activity would need to be inhibited.
In human studies looking at the impact of exercise on BAT also showed con icting results. One study evaluated BAT activity in endurance-trained athletes by using 18 F-FDG PET-CT. Compared to the control group, endurance-trained athletes showed lower cold-induced BAT [26]. However, in another 18 F-FDG PET-CT study of 40 cancer patients, increased habitual physical exercise (light exercise) led to higher BAT activity [2].
We speculate that there could be several reasons for the con icting results. First, different exercises, such as swimming and running, could have varying results. For instance, swimming, as opposed to running, could result in a loss of heat due to the body's contact with water, thereby activating the thermogenic function of BAT. Second, there could be a difference in the duration and intensity of exercise. In previous studies, both short-term exercise and long-term endurance training have been evaluated. Whether these exercises can activate BAT compensatory mechanisms remains unclear. Third, different assessment methods could lead to different results. 18 F-FDG PET-CT is one method that is affected by several variables, including ambient temperatures, blood glucose levels, and diet. If these variables are not well controlled, BAT quantities would be likely underestimated. Free fatty acids (FFA) have also been shown to be the main substrate for BAT thermogenesis [20], and, therefore, it is doubtful if 18 F-FDG PET-CT can truly re ect BAT metabolism. Moreover, the evaluation of BAT volumes using MRI is still in the exploratory stages, more studies and samples are required to standardize the evaluation methods.
In our study, we found that the mice in running group had higher interscapular BAT T2 values than those in the control group. We speculated that the increased T2 values in our study were mainly due to the increased BAT volume. That is to say, the interstitial components increased as the volume of BAT increased, leading to more water content in tissue. Secondly, due to long-term exercise stimulating BAT through the SNS, and with BAT oxygen supplies fully compensated, more oxyhemoglobin would be available in this region, results in an increase of the T2 values. Our study showed that BAT volume correlated positively with BAT T2 value. The change of T2 value may re ect the change of BAT volume to a certain extent, as the distribution of BAT in human is relatively scattered, making it di cult to measure [27].
There are some limitations to our study. First, there was a disparity in the number of mice in the running group compared with the number in the control group, which was primarily caused by a lack of preparation in the early selection of mice, with an excessive number of mice that refused to exercise. Second, our study is considered preliminary, lacking protein and gene expression analyses, which will be performed in future studies.

Conclusions
In summary, this study was a non-invasive and quantitative examination method using 7T MRI to identify BAT. Long-term running increases BAT volume and T2 value, what's more, BAT volume correlates positively with BAT T2 value.