Introduction

The classification of human and rodent adipocytes includes white, brown and beige adipocytes1,2. White adipocytes are known for their energy storage function, with a single large lipid droplet and store excess energy in the form of triglycerides when nutrients are in excess3. Brown and beige adipocytes, known as thermogenic adipocytes, have multilocular lipid droplets. Furthermore, they exhibit a high number of mitochondria that express uncoupling protein 1 (UCP-1)4. Previous studies have demonstrated that classical brown adipocytes are located within interscapular, axillary, and perirenal brown adipose tissue (BAT) depots in mice, and beige adipocytes are an inducible form of thermogenic adipocytes that develop within white adipose tissue (WAT)2. The biogenesis of beige adipocytes occurs in WAT in response to specific stimuli, including cold exposure5,6, exercise7, and treatment with β3-adrenergic receptor agonists8. This process, known as browning WAT, represents a promising therapeutic target for obesity prevention.

Magnetic resonance imaging (MRI) enables the differentiation of BAT from WAT based on morphological distinctions that do not necessitate tissue stimulation and activation9,10. The smaller lipid droplets present in brown adipocytes result in BAT that appears more hydrated than WAT10,11,12,13. BAT is highly vascularized, and the mitochondria of brown adipocytes are rich in iron5,14. Accordingly, Wu et al.9 have indicated that the iron present within the inner mitochondrial membrane and the magnetic susceptibility gradients generated at the numerous water-fat interfaces can result in rapid signal dephasing. Consequently, BAT appears to have a darker signal on T2* maps than WAT10,15. Furthermore, the differences in the microstructure of white and brown adipocytes may result in different diffusion characteristics within these cells. Diffusion-weighted imaging (DWI) provides an assessment of molecular self-diffusion, providing information on microscopic tissue compartments and structural anisotropy16,17,18,19,20. DWI is widely used to assess the diffusion of water molecules in biological tissues16,17,21.

DWI is a MRI technique based on the random Brownian motion of water molecules22. The apparent diffusion coefficient (ADC) in DWI can be calculated using two b-values and multiple motion probing gradients (MPGs). Diffusion tensor imaging (DTI) is performed using MPGs in at least six directions. The higher the number of MPGs, the higher the angular resolution for water molecule diffusion. Three eigenvalues (λ1, λ2, λ3) and eigenvectors (v1, v2, v3) are obtained from a single voxel. When the three eigenvalues are equivalent, the diffusion is isotropic; when they differ, it is anisotropic22. DTI, a type of DWI, characterizes the direction and magnitude of anisotropic diffusion, providing information about the microstructure22. DTI is used to assess microstructural changes in the brain, cancerous tissue and skeletal muscle16,17,18,19,20. ADC values depend on molecular weight, viscosity and microscopic boundaries in the medium. Therefore, changes in the mitochondria of WAT could affect ADC values. UCP-1 dissipates the mitochondrial proton gradient and converts the energy of substrate into heat23. Based on previous finding23, diffusion inside and outside the mitochondrial membrane might differ between white adipocyte, and brown and beige adipocytes with high UCP-1. In brown and beige adipocytes, the presence of smaller lipid droplets could lead to more restricted molecular diffusion. Previous studies have reported a lower ADC in BAT compared to WAT using DWI and high-resolution pulsed field gradient diffusion nuclear magnetic resonance spectroscopy24,25.

It is important to note that MRI is not yet established as a method for identifying browning WAT. Furthermore, to the best of our knowledge, T2* values and anisotropy parameters in browning WAT have not been clarified. BAT has higher density of blood vessels and nerve fibres26, and brown adipocytes are rich in iron-containing mitochondria5,14. In addition, the induction of beige adipocytes is associated with changes in tissue structure, including increased nerve fiber arborization and angiogenesis27. A previous study reported a more than threefold increase in sympathetic nerve dendritic tone in inguinal white adipose tissue (iWAT) after cold exposure28. Therefore, we hypothesized that water molecule diffusion would be more restricted in beige adipocytes compared to white adipocytes. To test this hypothesis, we investigated T2* values and anisotropy parameters (fractional anisotropy [FA], ADC, radial diffusivity [RD] and eigenvalues λ1, λ2, λ3) in BAT, iWAT and epididymal white adipose tissue (epiWAT) using a 9.4T magnetic resonance (MR) scanner, comparing the control (C) and cold exposure (CE) groups. Additionally, in vivo MRI experiments were performed in three control mice to validate T2* values and anisotropy parameters in living animals.

Materials and methods

Ethical approval

All animal studies were approved by the local animal ethics committee of Nippon Sport Science University (number: 020-A01). All methods were carried out in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines (https://arriveguidelines.org).

Animal and experimental diets

Six-week-old male C57BL/6J mice (CLEA Japan, Tokyo, Japan) were used for the present study. We housed mice individually in 22 ℃ and 55–75% humidity and in a 12 h light-dark cycle from 3 days before intervention. They had access to standard chow (CE-2, CLEA Japan, Tokyo, Japan) and water ad libitum. Six mice served as a CE group and were exposed to a cold environment at 4 ℃ for 12 h for 5 days and then 24 h for 5 days. One mouse died on day 11; therefore, we present data for 5 mice in the CE group. Six mice served as the C group and were maintained at 22 ℃ throughout the experiment (Fig. S1). The physiological effects in mice during the study period were confirmed by assessing food intake and body weight in both groups.

Ex vivo experiments

After the control and cold exposure periods, mice were sacrificed with cervical dislocation. All tissue collection procedures were performed by one researcher (Y.T.). Gastrocnemius muscle (muscle), BAT, iWAT, and epiWAT were dissected and weighed to determine their volume in sacrificed mice29. Wet weights of muscle, BAT, iWAT, epiWAT are shown in Table S1. Tissues were frozen in liquid nitrogen immediately after collection and stored at −80ºC. The tissues were thawed 10 min before the start of MR scanning. After completely thawing, the tissue samples were placed in a 4-mm square case (Fig. 1a). These samples were wrapped in a sponge and soaked in a fluorine solution (Fluorinert FC-72, Sumitomo 3 M Limited, Tokyo, Japan) during MR scanning.

Fig. 1
figure 1

Typical tissue samples (muscle, brown adipose tissue [BAT], inguinal white adipose tissue [iWAT], and epididymal white adipose tissue [epiWAT]) and magnetic resonance (MR) images. (a) tissue samples in the control (C) and cold exposure (CE) groups before MR scanning, (b) T2*map, (c) fractional anisotropy (FA) map, (d) λ1 map, (e) λ2 map and (f) λ3 map in C and CE groups. (g) The expression patterns of uncoupling protein 1 (UCP-1) and beta-tubulin in BAT, iWAT, and epiWAT.

Full size image

In vivo experiments

To corroborate the ex vivo MRI results, supplementary in vivo MRI experiments were conducted with three 8-week-old C57BL/6J male mice (CLEA Japan, Tokyo, Japan). The mice were housed under identical conditions to these of the ex vivo experiment, namely at 22°C, with a humidity range of 55–75% and a 12 h light-dark cycle. A 9.4T MR scanner was utilized, comprising a cryogenic 8-channel surface probe and a transmitter coil with an inner diameter of 86 mm (Bruker Biospin AG, Fallanden, Switzerland). For MRI scanning, the mice were anesthetized with 2% isoflurane in oxygen and maintained at 1–1.5% isoflurane throughout the scanning process. Respiratory rate was monitored throughout the procedure. The mice were positioned prone on the animal bed with their limbs secured gently. T2* maps and DTI sequences were acquired with respiratory triggering incorporated into the acquisition process. T2* maps were acquired using a multi-gradient echo sequence. The detailed T2* sequence parameters were as follows: repetition time (TR) = 800 ms, echo time (TE) range = 1.63–13.58 ms, echo spacing = 1.33 ms, echoes = 10, matrix size = 214 × 160, field of view = 32 × 24 mm2, slice thickness = 1.0 mm with gap 0.2 mm, number of averages = 8. The spin-echo echo-planar imaging sequence was used for DTI. The DTI sequence parameters were as follows: TR = 800 ms, TE = 20.6 ms, δ/Δ = 5.0/10.5 ms, b-value = 3,000 s/mm2, average = 4, field of view = 32 × 24 mm2, matrix size = 128 × 96, slice thickness = 1.0 mm with a gap of 0.2 mm, and the motion-probing gradient moment = six directions (xy, xz, yz, −xy, −xz, −yz). The imaging planes were carefully positioned to visualize the BAT in the interscapular region, iWAT in the inguinal region, and epiWAT in the epididymal region (Fig. S2).

Protein isolation and Western blot analysis

Tissues samples preparation and Western blot analysis were conducted according to a previous study’s protocol29,30. To isolate protein, tissues samples were homogenized in 300 µl of radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl pH 8.0, 150 mM sodium chloride, 1% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate) (Sigma-Aldrich, St Louis, MO) containing 1× protease inhibitor cocktail (Roche Diagnostic, Dubai, UAE). After the homogenate samples were centrifuged at 12,000 g for 10 min at 4 °C, the supernatant fluids were collected. We used the BCA protein assay kit (Thermo scientific, Rockford, IL) to measure the protein concentration. The homogenate samples (10 µg of protein) were resolved by SDS-PAGE using either a 12.5% polyacrylamide gel for UCP-1 and beta-tubulin, and then transferred onto polyvinylidene difluoride membranes (ATTO, Tokyo, Japan) at 2 mA/cm2 for 30 min. After transfer, to block nonspecific immunoreactivity, the membranes were incubated with 20 mM Tris-buffered saline (TBS) (pH 7.6) containing 1% (w/v) fat-free milk and 0.1% Tween-20 for 1 h at room temperature. The membranes were then incubated overnight at 4 °C with primary antibodies. The primary antibodies were diluted in 20 mM TBS containing 0.1% Tween-20 with 1% (w/v) fat-free milk. The primary antibodies used in the present study were as follows: anti-UCP-1 antibody (ab23841, Abcam, Cambridge, UK, diluted 1:2000) and beta-tubulin antibody (#2146, Cell Signaling Technology, Danvers, MA, diluted 1:1000). Then, they were washed with 20 mM TBS containing 0.1% Tween-20 and incubated for 1 h at room temperature with secondary antibodies. The secondary antibodies used in the present study were biotinylated anti-rabbit IgG (65-6140, Invitrogen, Camarillo, CA, diluted 1:10000). They were washed with 20 mM TBS containing 0.1% Tween-20 with 1% (w/v) fat-free milk and incubated for 1 h at room temperature with TBS containing 0.1% Tween-20 and peroxidase conjugated streptavidin horseradish (GE Healthcare, Buckinghamshire, UK). Immunoreactivity was detected by chemiluminescence using Immunostar Zeta (Wako, Osaka, Japan).

MRI

We used a 9.4T MR scanner with a cryogenic 4-channel surface probe and a transmitter coil, with an inner diameter of 86 mm. The temperature in the MR room was 20°C during the MR scan.

T2* imaging

T2*maps were acquired with a multi-gradient echo sequence. The detailed sequence parameters were as follows: TR = 400 ms, TE range = 1.99–36.49 ms (Table S2), echo spacing = 2.3 ms, flip angle = 30°, echoes = 10, matrix size = 128 × 84, field of view = 12.8 × 8.4 mm2, slice thickness = 0.6 mm with no gap; number of excitations = 8. A representative example of T2* map is shown in Fig. 1b.

DTI

The maximum magnetic field gradient amplitude was 660 mT/m. The DTI sequence was performed using a stimulated echo pulse sequence with gradient pulses16. The sequence parameters used in the present study were as follows: TR = 2,000 ms, TE = 15.9 ms, mixing time = 48.6 ms; δ/Δ = 2.1/50.7 ms; b-value = 100,000 s/mm2; average = 2; field of view = 12.8 × 8.4 mm2; matrix size = 128 × 84; number of slices = 20; slice thickness = 0.6 mm with no gap; motion-probing gradient moment = six directions.

Image analysis

T2* map processing

A radiology technician (J.H.) set each ROI to T2* maps and measured T2* value using the Image J software ver. 1.51 (NIH, Bethesda, MD, USA).

DTI processing

The DTI images were calculated using the Diffusion Toolkit ver. TrackVis 0.6 (Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Boston, MA, USA). FA and eigenvalues were calculated at different time points. A representative example of the FA map and λ maps are shown in Fig. 1c–f. ROIs were set to cover the entire tissues. A radiology technician (J.H.) set each ROI on the calculated DTI images (FA, λ1, λ2 and λ3) and measured values using the Image J software. The axial diffusivity value was equal to the λ1 value. The RD value was the mean of λ2 and λ3. ADC was calculated according to the following formula16;

ADC = (λ1 + λ2 + λ3)/3.

Statistical analysis

The results were presented as means ± standard deviations. Two-way ANOVA and Bonferroni post hoc tests used to compare the mean food intake and body weight between the groups. Unpaired t-tests were used to compare means between the groups. A one-way ANOVA and Bonferroni post hoc tests were used to compare the mean BAT in the C group and iWAT and epiWAT in both groups. Pearson’s correlation coefficient was calculated to examine the relationship between T2* value and the anisotropy parameter in adipose tissue. Statistical significance was defined as P < 0.05, and a trend was defined as P < 0.10. All statistical analysis was performed using software (IBM SPSS Statistics version 25.0, IBM Japan Corp., Tokyo, Japan).

Results

Physiological impact of the study

The body weight of the CE group was significantly decreased in comparison with that of the C group after 6 and 11 days (Fig. S3a). In addition, CE group showed higher food intake than the C group during days 3–11 (Fig. S3b). The tissue weights of the CE group’s muscle, iWAT, and epiWAT were significantly lower than the C group (Table S1). No significant intergroup difference was observed in BAT (the C group; 0.065 ± 0.007 g, the CE group; 0.068 ± 0.013 g) (Table S1). Fig. 1g shows the expression patterns of UCP-1 in BAT, iWAT and epiWAT in both groups. Adipose tissue showed different responses to cold exposure. UCP-1 expression levels were higher in iWAT of the CE group than in the C group, whereas negligible differences were observed between the groups in BAT and epiWAT.

T2* in the C and CE groups

No significant difference was observed between the C and CE groups in muscle T2* value (Fig. 2a and Table S3). However, in BAT, T2* value was significantly lower in the CE group than in the C group (Fig. 2b and Table S3, P = 0.011). T2* values for iWAT (Fig. 2c and Table S3, P = 0.085) and epiWAT (Fig. 2d and Table S3, P = 0.062) tended to be lower in the CE group than these in the C group.

Fig. 2
figure 2

Comparison of T2* values among (a) muscle, (b) brown adipose tissue (BAT), (c) inguinal white adipose tissue (iWAT) and (d) epididymal white adipose tissue (epiWAT) in the control (C) and cold exposure (CE) groups.

Full size image

Anisotropy parameters (FA, ADC, RD and eigenvalues) in the C and CE groups

The DTI sequence, which had been optimised for use in adipose tissues, could not be applied to muscle due to the occurrence of strong signal attenuation at high b-values (Fig. 1c–f). No significant differences were observed in iWAT and epiWAT between the C and CE groups for FA, ADC, RD, λ1 and λ2 (Table 1). However, iWAT (P = 0.032) and epiWAT (P = 0.014) λ3 values were significantly higher in the CE group than in the C group (Table 1).

Table 1 Comparison of fractional anisotropy (FA), apparent diffusion coefficient (ADC), radial diffusivity (RD), λ1, λ2 and λ3 between brown adipose tissue (BAT), inguinal white adipose tissue (iWAT) and epididymal white adipose tissue (epiWAT) in control (C) and cold exposure (CE) groups.
Full size table

BAT vs. iWAT vs. epiWAT in T2*

As shown in Fig. 3a, T2* values of epiWAT in the C group were significantly higher than these for BAT (P < 0.001) and iWAT (P = 0.004) in the CE group (η2 = 0.625). There were no significant intergroup or interregion differences in T2* values. In vivo additional study (n = 3), T2* values demonstrated different patterns compared to ex vivo results (Fig. 3b). In vivo T2* values were higher overall (ranging from 68.2–165.4 ms compared to 4.4–13.2 ms in ex vivo) and showed minimal differences between adipose tissues (Fig. 3b).

Fig. 3
figure 3

Comparison of T2* values (a) between the brown adipose tissue (BAT) in the control (C) group, inguinal white adipose tissue (iWAT) in the C and cold exposure (CE) groups, epididymal white adipose tissue (epiWAT) in the C and CE groups, (b) in vivo T2* values of BAT, iWAT and epiWAT in three control mice.

Full size image

BAT vs. iWAT vs. epiWAT in FA, ADC, RD and eigenvalues

In ex vivo experiments, the C group BAT, FA was significantly higher than that in the CE group iWAT (Fig. 4a, P = 0.017, η2 = 0.649). ADC value in the C group BAT was significantly lower than these in the CE group epiWAT (Fig. 4b, P = 0.047, η2 = 0.484). RD value in the C group BAT was significantly lower than these in the CE group iWAT (P = 0.033), epiWAT (P = 0.003) and the C group epiWAT (P = 0.041). RD value in the C group BAT tended to be lower than these in the C group iWAT (P = 0.061). In vivo results exhibited distinctive patterns in comparison to ex vivo results. While ex vivo results demonstrated tissue-specific variations in both FA and RD, in vivo these values showed comparable levels across BAT, iWAT and epiWAT. In vivo ADC values showed a tissue-specific pattern, with higher values in BAT compared to iWAT and epiWAT (Fig. 4e). No significant intergroup or interregional differences were observed for λ1 (Fig. 5a) and λ2 (Fig. 5b). In contrast, the λ3 values in the C group BAT was significantly lower than these in the CE group iWAT and epiWAT (P < 0.001, respectively, η2 = 0.667) (Fig. 5c). The λ3 values of the C group in BAT tended to be lower than these in the C group iWAT (P = 0.070) and epiWAT (P = 0.052) (Fig. 5c). In vivo, λ1 and λ2 values showed no significant differences between tissue types (Figs. 5d,e), while λ3 values showed a pattern opposite to that observed  ex vivo. In vivo λ3 values in BAT were higher compared to iWAT and epiWAT (Fig. 5f), which contrasts with the lower BAT values observed  ex vivo.

Fig. 4
figure 4

Comparison of (a) fractional anisotropy (FA), (b) apparent diffusion coefficient (ADC), (c) radial diffusivity (RD) between brown adipose tissue (BAT) in the control (C) group, inguinal white adipose tissue (iWAT) in the C and the cold exposure (CE) groups, epididymal white adipose tissue (epiWAT) in the C and CE groups, in vivo (d) FA, (e) ADC, (f) RD of BAT, iWAT and epiWAT in three control mice.

Full size image

The relationships between T2* values and anisotropy parameters in BAT, iWAT and epiWAT

Table 2 presents the Spearman’s correlation coefficients between T2* values and anisotropy parameters (FA, ADC, RD, λ1, λ2 and λ3) in BAT, iWAT, and epiWAT for the C and CE groups. The results showed no significant correlation between T2* values of BAT, iWAT, and epiWAT and any anisotropy parameters in both groups (Table 2).

Table 2 Pearson’s correlation coefficient between T2* values and anisotropy parameters (fractional anisotropy [FA], apparent diffusion coefficient [ADC], radial diffusivity [RD] and eigenvalues λ1, λ2, λ3) in brown adipose tissue (BAT), inguinal white adipose tissue (iWAT), and epididymal white adipose tissue (epiWAT) in the control (C) and cold exposure (CE) groups.
Full size table

Discussion

In the present study, we investigated T2* values and anisotropy parameters (FA, ADC, RD and eigenvalues) in BAT, iWAT and epiWAT using a 9.4T MR scanner, comparing C and CE groups. Furthermore, we investigated whether these T2* values and anisotropy parameters could be evaluated in vivo experiments in three control mice. The principal findings of the present study are as follows: (1) T2* values of iWAT and epiWAT in the CE group exhibited a tendency towards lower values than these observed in the C group. T2* values of epiWAT in the C group were significantly higher than these of BAT in the C group and iWAT in the CE group. (2) No significant intergroup differences were observed for FA, ADC, RD, λ1 and λ2 in iWAT and epiWAT. However, iWAT and epiWAT λ3 values were significantly higher in the CE group than in the C group. No significant intergroup or interregional differences were observed for λ1 and λ2. However, the λ3 values of BAT in the C group were observed to be lower than these of iWAT and epiWAT in both the C and CE groups. (3) Compared to ex vivo results, in vivo results showed higher overall T2* values and reversed patterns in λ3 values, with BAT exhibiting higher values than iWAT and epiWAT. The findings of present study demonstrate that cold exposure induces distinct changes in both T2* and λ3 values in iWAT and epiWAT. T2* values in iWAT and epiWAT showed a tendency to approach these observed in BAT, whereas λ3 values in iWAT and epiWAT were significantly higher in response to cold exposure.

Previous studies have confirmed that cold exposure to 4 °C causes browning in iWAT5,6,31. Lim et al.5 demonstrated that C57BL/6 mice exposed to 4°C for 7 days showed a transition from white to beige adipocytes in iWAT. We confirmed browning WAT by UCP-1 expression in iWAT (Fig. 1g). Meanwhile, epiWAT showed negligible differences in UCP-1 expression between groups (Fig. 1g). These results are in consistent with previous study32. However, the T2* values and anisotropy parameters evaluated in the present study showed similar trends for iWAT and epiWAT. The relevance of T2* values and anisotropy parameters for browning WAT requires further investigation.

Previous study showed that T2* values indicate a time constant and represent the decay of transverse magnetization in the presence of local field inhomogeneities33. In addition, T2* values have been shown to be effective in detecting iron-containing substances, including paramagnetic deoxyhemoglobin, methemoglobin, and hemosiderin deposits in tissue15. Yang et al.34 have reported that the abundance of iron-rich mitochondria and capillaries in BAT contributes to a faster dephasing of the transverse magnetisation, resulting in lower T2* values compared to these observed in WAT. Other previous studies have shown that BAT has lower T2* than WAT in mice15 and humans10. Furthermore, the existence of differences in T2* values between BAT, iWAT, and epiWAT after cold exposure was previously unidentified. We observed that after cold exposure, the T2* values of iWAT and epiWAT were similar to these of BAT in the C group (Fig. 3a). These results may indicate that, after cold exposure, the high abundance of iron-rich mitochondria and capillaries in iWAT and epiWAT is comparable to that in BAT. Tangchitphisut et al.35 reported that the minimal cell concentration that a 3.0T MR scanner can reliably detect using T2∗ map sequences are 1 × 105 cells/ml/microtube. The T2* values observed in the present study indicate that the browning WAT induced by cold exposure occurred in a wide range of adipose tissue. However, in vivo mesurements demonstrated higher T2* values and a reduced discrepancy between adipose tissue types in comparison to ex vivo mesurements. The observed discrepancies could be attributed to various factors. Firstly, blood perfusion in vivo tissues may influence local magnetic field homogeneity through the presence of paramagnetic deoxyhemoglobin and dynamic blood flow. In accordance with the findings of Chavhan et al.33, the magnetic field inhomogeneity, which has the potential to impact signal intensity and induce geometric distortion, can be categorized into two distinct types: macroscopic (intervoxel, exhibiting a consistent pattern across a voxel) and microscopic (changing within a voxel). It was reported that macroscopic inhomogeneity can be caused by deoxyhemoglobin in tiny veins, air-tissue interfaces, or metallic implants, while microscopic inhomogeneity results from paramagnetic contrast agents, blood products, or iron deposits. It can be reasonably deduced that these inhomogeneities could exert an influence on the discrepancy in T2* values observed between in vivo and ex vivo. Secondly, physiological motion, including respiratory and cardiac movements, could introduce further complexity to the measurements despite the use of respiratory gating. Thirdly, post-mortem alterations in tissue characteristics, including the cessation of blood flow and potential tissue dehydration, could result in the lower T2* values observed  ex vivo tissues. These observed results are consistent with the findings of previous studies on tumours, which have reported differences in T2* values between in vivo and ex vivo36.

To the best of our knowledge, this is the inaugural study to mesure anisotropy parameters using DTI in BAT, iWAT, and epiWAT. Previous studies applied DWI to the assessment of adipose tissues in MRI phantoms37 and humans21,24. A recent previous study reported the variation of ADC assessed by DWI due to fat content using an MRI phantom37. They observed that the ADC was significantly, negatively, and linearly associated with fat content (the linear slope ranged from − 0.005 to − 0.017, R2 = 0.925 to 0.986, all P < 0.001)37. Therefore, it is evident that alterations in the lipid content of adipocytes represent a restricting factor in diffusion. Deng et al.21 reported that BAT exhibited higher T2* (P < 0.0001) and a lower diffusion coefficient (P = 0.015) in obese children compared with normal-weight children. In the present study, the λ3 values were significantly higher in the CE group iWAT and epiWAT than in the C group BAT (Fig. 5c). Tadano et al.38 demonstrated a negative correlation between the amount of intramyocellular lipids (IMCL) and λ3 in patients with lifestyle-related diseases. The results of their study38 indicate that elevated IMCL levels are associated with reduced diffusion eigenvalues. If the results of the previous study on muscle cell38 are applicable to adipocytes, it may be posited that a reduction in lipid droplets in the adipocytes of iWAT and epiWAT after cold exposure could result in an increase in λ3. Furthermore, the values of T2*, ADC, RD, λ1, λ2 and λ3 for BAT showed a narrower distribution than these for iWAT and epiWAT in the C and CE groups (Figs. 3, 4b, 5a–c). These results could be related to the multiple layers of adipose tissue heterogeneity, including differences not only between white and brown adipocytes but also between different WAT depots and even within a single depot39,40. However, the present study did not identify the factors that affect diffusion in adipocytes. Several factors can influence adipocyte diffusion; beige adipocytes are characterized by increased mitochondrial content and multilocular lipid droplets, while structural changes in browning WAT include increased angiogenesis and sympathetic nerve fibre branching27,41. These cellular and structural changes may alter the diffusion properties of adipocytes. In the previous study, AQP 4 on the membrane of muscle cells reportedly restricts the diffusion of water molecules of muscle cells17. Lopes et al.42 reported differential AQP3, AQP5, and AQP7 gene expression between subcutaneous WAT, visceral WAT, and BAT in hamsters. In consideration of the findings of their previous study42, it might be proposed that the differential expression of AQP in adipocytes may serve to restrict the diffusion of water molecules within these cells. In addition, the different lipid droplet compositions may be related to the different molecular weights of the adipocyte types, and these may contribute to differences in water molecule diffusion between adipose tissues. A previous study reported differences in the composition of lipid droplets in white, beige, and brown adipocytes43. Machann et al.44 reported that ADC values of pure compounds decreased with increasing molecular weight by a factor of about 10 in lard at 37℃. Furthermore, they reported that ADC values could depend on the chain length of the substances for subcutaneous adipose tissue, bone marrow, IMCL, and extramyocellular lipids in humans44. However, further studies are required to invistigate the relationship between biochemical analysis, immunostaining, and MRI in order to ascertain the factors that restricting diffusion in the iWAT and epiWAT after cold exprosure. The differences between ex vivo and in vivo measurements may be related to the dynamics of adipose tissue in living mice. Previous studies have shown that in vivo BAT is characterized by higher hydration levels10,11,12,13 and high vascularity5,14. Furthermore, continuous in vivo metabolism in BAT is characterized by thermogenic activity and enhanced blood flow. These metabolically active conditions can affect the water diffusion characteristics assessed by DTI. The difference between ex vivo and in vivo ADC values has been observed in other tissues, with previous study showing that post-mortem ADC measurements in tumors were lower these measured in vivo36.

The present study has several limiting factors. We utilized b-values of 100,000 s/mm2 for ex vivo and 3,000 s/mm2 for in vivo to detect adipose tissue characteristics. However, further improvement of DTI parameters is essential for human applications, considering the technical limitations of clinical MR systems. Additionally, respiratory and cardiac motion effects in vivo parameters need to be carefully addressed in future clinical protocols. Second, although the present study demonstrated different patterns of anisotropy parameters across adipose tissue types and conditions, we could not directly demonstrate their relationship with browning WAT. Future studies are required to combine MRI parameters with detailed histological analysis of adipocyte morphology and UCP-1 expression to correlate with browning WAT under both ex vivo and in vivo.

Fig. 5
figure 5

Comparison of (a) λ1, (b) λ2, and (c) λ3 between brown adipose tissue (BAT) in the control (C) group, inguinal white adipose tissue (iWAT) in the C and cold exposure (CE) groups, epididymal white adipose tissue (epiWAT) in the C and CE groups, in vivo (d) λ1, (e) λ2, (f) λ3 of BAT, iWAT and epiWAT in three control mice.

Full size image

Conclusion

The present study investigated T2* values and anisotropy parameters (FA, ADC, and eigenvalues λ1, λ2, λ3) in BAT, iWAT and epiWAT using a 9.4T MR scanner, with comparisons between the C and CE groups. We demonstrated significant changes in T2* values and anisotropy parameters of adipose tissues in response to cold exposure. The findings indicated a tendency for T2* values in iWAT and epiWAT to approach these observed in BAT, while λ3 values were significantly higher in response to cold exposure in iWAT and epiWAT. These changes may reflect the microstructural characteristics of adipose tissue following cold exposure.  The present study demonstrated that these parameters can be successfully measured in adipose tissues under both ex vivo and in vivo. These results suggest that T2* values and anisotropy parameters may serve as a potential parameter to assess adipose tissue plasticity. However, further studies are required to investigate their utility as non-invasive indicators of browning WAT.