Sex-dependent effects of forced exercise in the body composition of adolescent rats

Determining the body composition during adolescence can predict diseases such as obesity, diabetes, and metabolic syndromes later in life; and physical activity became an effective way to restore changes in body composition. However, current available literature assessing the body composition before, during and after adolescence in female and male rodents by in vivo techniques is scarce. Thus, by using computerized tomography, we aimed to define the baseline of the weight and body composition during the adolescence and young adulthood of female and male Sprague–Dawley rats (on P30, P60 and P90) under standard diet. Then, we determined the effect of 18 days of forced exercise on the body weight and composition during the early adolescence (P27-45). The highest percentual increments in weight, body volume and relative adipose contents occurred during the female and male adolescence. Forced running during the early adolescence decreased weight, body volume and relative adipose delta and increment values in males only. The adolescence of rats is a period of drastic body composition changes, where exercise interventions have sex-dependent effects. These results support a model that could open new research windows in the field of adolescent obesity.

cia were used in the experiment. The animals arrived at the experimental facilities on postnatal day 20 (P20). Female baseline group (FB; n = 6; litter A) and male baseline group (MB; n = 6; litter B) were housed in separate cages (3 rats per cage) connected to individual ventilation racks. Female training group (FT; n = 6; litter C and litter D), female control group (FC; n = 6; litter C and litter D) were weight-matched and composed by 3 animals from litter C and 3 animals from litter D, and housed in a separate room under the same conditions. Male training group (MT; n = 6; litter E) and male control group (MC; n = 6; litter E) were weight-matched and housed in a separate room under the same conditions. Standard size cages, of 50 × 35 × 35 cm with a 2-3 cm dry cork shavings layer, were replaced every 4-5 days. A 12:12 h dark-light cycle was set, being zeitgeber (ZT) 0 when the lights turned on (rat's passive phase), and ZT12 when the lights turned off (rat's active phase). All the experiments were performed during the rats' active phase (dark cycle) under a source of dim orange light (600 nm), at ZT14 and/or ZT20 to minimize circadian disruptions. The rooms were maintained at 22-25 °C temperature with a relative humidity of 45-60%. Standard chow diet (ENVIGO, diet 2014) and filtered water were available ad libitum.
Forced wheel protocol. The training was designed and supervised in collaboration with graduates in physical activity and sport science, certificated and with permission to work with experimental animals. The handling procedure consisted of picking up one rat at a time and petting it for 2 min. From postnatal day 26, female and male animals of the experimental groups were exposed to an 8-days exercise habituation protocol in a forced running wheel system (Campden Instruments, 80805A) as described previously (Fig. 1) 52 . The training program consisted of two sessions per day, on ZT14 and ZT20. The active period of SD rats occurs during the darkness (ZT12-ZT23). ZT14 session was defined as morning (AM) session, since it occurs in the first half of the active phase. ZT20 was defined as afternoon (PM) session, since it occurs in the second half of the active phase. On each session (AM and PM), the rats were weighted (Gram S3R-2KD) before entering the wheel system. The sessions consisted of 10-min bouts of continuous running, with 5 min of resting between bouts. The speed and/ or total duration increased every day (Fig. 1). The rotation speed of the wheels was established according to the response of the rats in preliminary experiments.
Computerized tomography. On postnatal day 30, 60 and 90 of the FB and MB groups and postnatal day 27 and 45 of the FT, FC, MT and MC groups, the rats were screened using an Albira trimodal preclinical-scanner and software (Bruker®, Billerica, MA, USA). The rats were anesthetized (MSS isoflurane Keyed Filler Cagemount, MSSVAP03) through inhalation of isoflurane (ZOETIS, IsoFlo) at 5% in a 30 × 15 × 15 cm chamber until the absence of voluntary movement. After the anesthesia induction, the rats were weighted (Gram S3R-2KD) and prepared for image acquisition. Then, the isoflurane flow was reduced to 2.5-3% and mixed with a 1.6-1.8 L/min of oxygen flow (DeVilbiss Healthcare, 5 L Oxygen Concentrator) until the end of the scanning. The image scanning process took between 25 and 40 min per rat. The x-ray intensity parameters were set to 45 kV and 0.4 mA. P27 and P30 rats were exposed to approximately 893.9 SDE mSv (or 89.4 rem) and 550 DDE mSv (or 55 rem), while P45, P60 and P90 rats received approximately 1,430.24 SDE mSv (or 143 rem) and 880 DDE mSv (or 88 rem). Three sets of 600 projections of 0.05 mm voxel size were obtained using a digital flat panel detector of 2400 × 2400 pixels and a 70 × 70 mm field of view. Once the acquisition was complete, the rat recovered from   PRE-TRAINING   HABITUATION   HABITUATION   TRAINING   TRAINING   POST-TRAINING   A   B   PND 26  2 8  3 0  3 2  3 4  3 6  3 8  4 0  4 2  4 4   PRE  TRAINING  G  N  I  N  I  A  R  T  N  O  I  T  A  U  T  I  B  A  H   5  4  P  5  3  P  4  3  P  6  2  P  5  2  P  0   www.nature.com/scientificreports/ projection algorithm (Albira reconstruction software). The researchers responsible for the image acquisition process were blind to the groups.
Image analysis. The 3D image segmentation was performed using the pMod version 3.5. With the software's reduction tool, the image resolution was halved to substantially improve the processing time. The full volume of the rat was calculated by using the "isocontour" function of the software with the following parameters: algorithm '3D'; search mode 'hot'; inner holes 'no'; multi regions 'no'; mode of searching 'automatic'; absolute threshold 'value'; threshold '-450' . The thoracoabdominal (TA) region was analyzed and defined as the contents from the intervertebral space C7-T1 (cervical and thoracic) to the intervertebral space between S2-S3 (sacral). A broader region was analyzed and defined between the posterior boundary of the olfactory bulb and the anus (olfactory bulb-anus; OBA region) ( Fig. 2A,B). The method did not permit to differentiate between white adipose tissue (WAT) or brown adipose tissue (BAT) ( Fig. 2A' ,A"), and therefore we only considered the visceral (VAT) and subcutaneous (SAT) adipose tissue. To split the VAT from the SAT, a contour was drawn manually with the software. The VAT region was defined as the contents comprised by the mentioned contour, inside of the muscle/bone plane of the thoracoabdominal wall of the TA region (Fig. 2B,B"). The adipose tissue outside the described limits was defined as SAT. To ensure the accuracy of the compartment separation, 2 contours were drawn by 3 different researchers, and then compared for differences to reach an agreement. Due to the limitations of the image quality, separating VAT from SAT was not possible in the OBA region. Therefore, the analysis www.nature.com/scientificreports/ of the OBA region does not include the separation between SAT and VAT. The sum of VAT and SAT was defined as the total adipose tissue (AT) content. Based on calibration standards that were validated previously in data from humans and rodents 43,56 , using the "segmentation" functionality of the software, three ranges of signal were obtained and considered in terms of adipose (− 200 to − 31 HU), lean (− 30 HU to 189 HU; note that lean tissue includes mainly muscle, but also includes the parenchyma and stroma of organs like kidneys, liver or intestines) or bone tissue (≥ 190 HU). The raw segmentation of the image was used in all the calculations. Therefore, no post-processing was used after the segmentation. Body composition contents were reported as the percentage of the corresponding body volume. The 3D images (Sup. 2; Sup. 5; Sup. 6) were created by loading the CT raw images in VolView 3.4 (Kitware), adjusting the radiodensity parameters until only the bone is visible, and merging it (by clicking on Plugins > Utilities > Merge volumes) with the segmented SAT and VAT volumes, prior to choosing a different color for every tissue. The 3D videos (Vids. [1][2][3][4][5][6][7][8][9][10][11][12][13][14] were created by using the VolView tools (Review > Movie > Create). The researchers responsible for the image analysis were blind to the groups.
Statistical analysis. The SPSS v25 was used to perform all the statistical analyses. First, the normality of the data was tested by using the Shapiro-Wilk test. The outliers were statistically identified by assessing the boxplots of the normality tests. Only the outliers that resulted in a change of the statistical outcome were removed. A mixed model of 2-way ANOVA (or split-plot ANOVA) was used to analyze the main effects and interactions of within-subjects and between-subjects variables regarding relative and absolute values (see results). In case of sphericity violation, Huynh-Feld test was used to establish significance. Between-subjects comparisons were then followed-up with Fisher's LSD post-hoc adjustment. Paired T-tests were used for all within-subjects paired comparisons. Delta and increment comparisons that were between-subjects (see results) were analyzed using U-Mann Whitney or T-student, in accordance to the normality of the data. Significance level was established at p < 0.05.

Results
A group of female and male SD rats was analyzed on P30, P60 and P90 to define a baseline of body weight and body composition. These key ages defined two life stages, namely adolescence (P30-60) and young adulthood (P60-90). Next, we determined the effect of a forced wheel training during the early adolescence period of female and male rats. In the results below, we show the absolute values of the whole-body weight and the TA or OBA volume of the rats. Then, we normalized the volume of the rats' adipose, lean and bone volumes, obtained by CT, to the volume of the TA region (Figs. 3, 4, 5, 6, 7) or OBA region (Sup. 1; Sup. 3; Sup. 4) and expressed it as a percentage (relative content). The delta values were reported as the difference between two age points (e.g. P30 to P90 difference = P90 value -P30 value). The percentual increment values were reported as the percentage difference between two age points (e.g. P30-P60 increment = P60 value * 100/P30 value -100). Baseline female and male SD rats experimented the highest increment in the adipose content during adolescence. In the thoracoabdominal region, the relative adipose (AT%) content -or body fat-increased with age without differences between females and males ( Fig. 4A; Sup. 2; Vids. [1][2][3][4][5][6]. No differences were found between females and males regarding the AT% delta and increment values from P30 to P90, adolescence (P30-60) or early adulthood (P60-90) ( Fig. 4B-C"). Comparing adolescence (P30-60) with young adulthood (P60-90) the AT% increment was higher during adolescence in both females and males ( Fig. 4C' ,C"). Note that the delta AT% values in the adolescence and young adulthood were similar ( Fig. 4B' ,B") (see discussion). The analysis of the OBA region revealed similar differences to the TA region, except that no difference was observed between adolescence and young adulthood OBA AT% in females (Sup. 3A-C"). In relation to the SAT%, it increased with age in both females and males ( Fig. 4D; Sup. 2A,D,G,J,M,P). In both delta and increment values, males showed higher SAT% than females from P30 to P90 (Fig. 4E,F), specifically in the adolescence (P30-60) (Fig. 4E' ,F'), but not in the young adulthood (P60-90) ( Fig. 4E",F"). Comparing adolescence (P30-60) with young adulthood (P60-90), males but not females showed higher SAT% increment during the adolescence (Fig. 4F'-F").

Baseline female and male SD rats experimented the highest increment in weight and
Regarding the VAT%, it increased with age in both females and males (Fig. 4G). In both delta and increment values, females showed higher VAT% than males from P30 to P90 (Fig. 4H,I), specifically in the adolescence (P30-60) (Fig. 4H' ,I'), but not in the young adulthood (P60-90) (Fig. 4H",I"). Comparing adolescence (P30-60) with young adulthood (P60-90), the VAT% increment in females and males was higher during adolescence www.nature.com/scientificreports/ ( Fig. 4I' ,I"). Note that the delta VAT% values of the adolescence and young adulthood were similar in both females and males (Fig. 4H' ,H") (see discussion).
In relation to the SAT%:VAT% ratio, females showed a higher ratio on P30, but lower than males on P60 and P90 (Fig. 4J). Finally, females showed lower VAT% than SAT% on P30, but higher VAT% than SAT% on P60 and P90 (Fig. 4K-K"). On the other hand, males showed higher VAT% than SAT% on P90 only (Fig. 4L-L").      www.nature.com/scientificreports/ Similar increments between adolescence and young adulthood were observed in the LEAN% and BONE%. In the TA region, the LEAN% decreased with age in both males and females (Fig. 5A). No differences were found between sexes or between life stages regarding the LEAN% delta or increment values ( Fig. 5B-C"). In contrast to the TA region, the LEAN% delta and increment values in the OBA region were lower in females than males from P30 to P90, specifically during adolescence (P30-60), but not during young adulthood (P60-90) (Sup. 4B-C"). The BONE% decreased with age in both males and females (Fig. 5D). The BONE% delta and increment values were lower in males than females from P30 to P90, specifically during adolescence (P30-P60), but not   A"" B"

AM and PM forced running during the early adolescence decreased the delta and increment values of weight and TA volume in male but not female SD rats. A group of female and male
rats was trained in a forced running wheel twice a day [at ZT14/AM and ZT20/PM during the early adolescence (P27- 44)]. The habituation (P26-34) and training (P35-44) phase consisted of a load that increased daily (Fig. 1). Control and experimental male and female rats showed similar weights before the training (P28, Fig. 6A). However, only male trained rats showed lower weight delta and increment values than their control (P44, Fig. 6A-A"). Weight differences between control and experimental male rats were observed the 3 last days of the training (Fig. 6A"'), but the weight increment shows these differences 8 days before the end of the training (Fig. 6A""). Control and experimental male, but not female rats, showed similar TA volume before the training (P27, Fig. 6B). Nonetheless, only male trained rats showed lower volume delta and increment values than their control (P45, Fig. 4B-B").
AM and PM forced running during the early adolescence decreased AT% delta and increment values in male but not female SD rats. Control and experimental female and male rats showed similar TA AT% before the training (P27, Fig. 7A; Sup. 5; Sup. 6; Vids. [7][8][9][10][11][12][13][14]. However, only male trained rats showed lower AT% delta and increment values than their control (P45, Fig. 7A'-A"). After splitting TA SAT% and VAT% contents, we observed that control and experimental male and female rats showed similar SAT% and VAT% before the training (P27, Fig. 7B,C; Sup. 5; Sup. 6). But only male trained rats showed lower SAT% and VAT% delta and increment values than their control (P45, Fig. 7B' ,B",C' ,C"). The SAT:VAT ratio was higher in experimental males, but not females, than their control after the training (Fig. 7D). Regarding the SAT% versus VAT%, only experimental males showed lower VAT% than SAT% after the training program (P45; Fig. 7E-F'). The TA LEAN% delta and increment values were higher in both female and male experimental groups on P45 (Fig. 7G'-G"). Finally, the TA BONE% delta and increment values were lower in experimental females, but higher in experimental males on P45 (Fig. 7H'-H").

Discussion
In the present research, we defined a baseline to determine the tendency of weight, body volume, and relative adipose, lean and bone contents of female and male Sprague-Dawley rats during two life stages: adolescence (P30-60) and early adulthood (P60-90). Our results highlight that, during adolescence, both female and male SD rats showed the highest increment in weight, body volume, and AT%, but also the highest decrease of BONE%. In our baseline data, males showed higher weight, body volume, SAT% and delta LEAN%, but lower delta BONE% than females. Sex-dependent differences during the adolescence were also observed in the maturation of the prefrontal cortex of Long-Evans rats 57 .
The literature assessing the body composition before, during and after adolescence in female and male rodents by in vivo techniques (e.g. MRI, CT) is scarce. Studies determining the body composition in rats, were mainly developed in young or adult periods 32,49,58,59 . The relative body fat (or adipose tissue) of SD rats determined by Schoeffner et al. 33 , with an ex-vivo technique, was lower than in our study, and did not seem to increase during the adolescence and young adulthood. These differences might be explained by the gravimetric method used to determine the lipid content, as well as the data acquisition (grams) and its normalization method 33 . In addition, ex-vivo techniques imply the utilization of cohorts, which might induce bias due to the litter differences. CT scan is appropriate to evaluate longitudinal changes in body composition, assuming that the exposure to radiation   www.nature.com/scientificreports/ during our study is not in range to produce tissue damages or morphological alterations in rats 60,61 . Nonetheless our CT images and data, normalized with the volume, indicate there are drastic changes in the relative adipose content during adolescence. Tekus et al. 49 proposed that the body composition of the region between L1-3 can be used to estimate the whole-body composition. Nonetheless, in our study, we observed some differences between the body composition in the TA region and the OBA region, especially regarding relative lean and bone contents ( Fig. 5A-F"; Sup. 3A-F"). Thus, we suggest that estimating the body composition from relatively lesser portions of the body could result in errors, since further studies are required to understand how body composition is regulated in different body regions and organs. On the other hand, many works reported their body composition data as absolute values or relative to body weight 29,30,58 . Some of these studies also include delta body weight 30 or the percentual increment 59,62 . In order to obtain the delta and incremental values, it is necessary to perform pre and post data acquisition. In our data, the absolute value (e.g. weight) or relative content (e.g. adipose tissue) highlighted the general trend of the values throughout the experiment. On the other hand, the delta value was used to determine the size of the change between two points in time (e.g. females vs. males during adolescence). However, the percentual increment better reflected the fold change from the initial to the final time points, and therefore was better for the comparison of two periods (e.g. adolescence vs. young adulthood), although it also detected the differences between conditions (e.g. experimental vs. control). As a result, the delta or the increment values were not used interchangeably. As an example the female delta VAT% was similar in adolescence and the adulthood, but the increment revealed the VAT% increased more in the adolescence (Fig. 4H' ,H",I' ,I"). We found that a forced wheel training of 18 days during the early adolescence of rats has sex-dependent effects. AM and PM forced running during the early adolescence decreased weight, body volume and relative adipose delta and increment values in males only. The sex-dependent effects in -at least-the body weight could be partially explained by experiments showing that pre-adolescent gonadectomy in rats reduced the sex-dependent differences in body weight and prefrontal cortex maturation 63,64 . In our results, males started with higher SAT% than VAT%, and exercise prevented both SAT% and VAT% gain compared to the control, with a higher effect on VAT% (Fig. 7E-F'). These results suggest that our exercise program has increased benefits in males' VAT, and therefore can potentially reduce the risk of developing pathologies related to visceral adiposity [65][66][67][68] , and preventing critical COVID-19 illness 69 . The absence of effect in the adipose delta values of female rats might suggest that the prevalence of obesity during adolescence in humans is greater in females than males. However, epidemiological studies based on body mass index suggest there is no more prevalence of overweight and obesity in girls than boys 70 . Further studies are required to understand if a different exercise load is required to decrease the adipose content of adolescent female rats. The decrease of female SAT% observed in control rats between P27 and P45 contrasts with the observed increase between P30 and P60 of the baseline rats, suggesting that SAT% in females increases mainly after P45. Regarding the lean tissue, in our study both trained females and males showed higher delta lean values. On the other hand, trained females showed lower bone gains than their control, while trained males showed higher bone gain. These results are in line with the available evidence in adolescent humans showing that exercise enhances lean body mass [71][72][73] .
Some research results suggest the sex-dependent effects of exercise on body composition might vary depending on the life stage and the voluntary or forced condition of the exercise 25,58,59 . Cortright et al. 25 observed sexdependent effects on the weight and body composition of male and female SD rats after 63 days of voluntary running from early adolescence through young adulthood (P21-84). However, only after 49 days of voluntary running the effects in weight were observed (P70, young adulthood). In Cortright's study, the body composition was determined ex-vivo on P84, and therefore the delta and percentual increment values were not obtained, but no effect was observed in females. Nonetheless, in our study we found that forced wheel running produced body weight differences in adolescent males even 10 days after exercise onset, and the body composition differences were observed in-vivo after 18 days of forced running. Complementarily, the study of Applegate et al. 58 found that after a training in forced treadmill during young adulthood (P78-100) of Osborne-Mendel rats, only males decreased in weight. In the same study, both males and females showed lower body fat after the exercise program, but the effect was higher in males than females. Finally, a voluntary running for 12 months in adult mice (1-year-old) decreased body weight, body fat and lean contents in both females and males 59 . Taken together, these data suggest that the sex-dependent results we observed with forced wheel exercise may occur mainly during the adolescence.
Changes in body composition observed after the training program (P27-P45) in our study occur when testosterone in male and estradiol in female are beginning to increase. Serum testosterone levels in SD rats are close to zero during early adolescence, with greater increase from P40 to P44. Serum estradiol levels in female SD rats are detectable during adolescence and increases gradually, reaching adult values at P48 74 . An acute increase of testosterone and growth hormone serum concentrations can be observed after a resistance exercise 75 . Further studies will need to determine if these hormonal increases occur in adolescent rats during the aerobic training and if these hormonal changes can be causally related with the changes in body composition. Also, pre-adolescent gonadectomy studies could help to clarify possible roles of testosterone and estradiol in the observed differences in body composition 63,64 .
Male decrease in body weight and fat composition, observed after our training program during early adolescence, should be analyzed in relation to neural circuits that sense energy expenditure and thermogenesis. Hypothalamic nuclei such as arcuate (through POMC and AGRP neurons) and dorsomedian, involved in the control of food intake and energy expenditure 76 are affected by the adipokine leptin, secreted by the adipose tissue. This hormone reduces food intake and increases energy expenditure, it shows higher sensitivity in females than males and was observed that exercise can decrease its concentrations in woman 77,78 . Sexual dimorphism was determined in diet-induced rodent obesity models, with females being more resistant to storing fat 79 . Also, a better improvement in food intake and weight was observed in some obesity treatments in female mice, representing a sexual dimorphism probably dependent of leptin effects 80  www.nature.com/scientificreports/ In summary, our data indicate that the most drastic changes in the body composition of SD rats occur during the adolescence when compared with the young adulthood, and that adolescent male rats showed higher susceptibility to body weight and body composition changes than females when subjected to forced wheel exercise. Also, this susceptibility seems to occur mainly in the adolescence. Future studies should analyze the molecular mechanisms which mediate the decrease of adipose tissue under physical activity in males but not females during the early adolescence.

Data availability
All the data and protocols of the procedures are available on reasonable request.