Evaluating the UCP1 expression of brown adipose tissue by quantifying hepatic inflammation using synthetic magnetic resonance imaging: an experimental study with a mouse model

Introduction

Brown adipose tissue (BAT) is highly heterogeneous and can initiate thermogenesis through mitochondrial inner membrane protein uncoupling protein 1 (UCP1) (1). In recent years, BAT has emerged as a promising therapeutic target for metabolism-related diseases due to its role in maintaining metabolic balance, thereby combating obesity, insulin resistance, and other metabolic disorders (2). Studies have shown that individuals with detectable BAT activity have lower prevalence rates of cardiometabolic diseases, including type 2 diabetes, dyslipidemia, coronary artery disease, and hypertension (3). BAT activation accelerates the clearance of plasma triglyceride-rich lipoproteins, and increases high-density lipoprotein levels, contributing to an improved lipid profile (4). UCP1 expression is an essential biological indicator for evaluating the functional activity of BAT, which determines its efficacy in regulating metabolism and immune balance and resisting metabolic disorders (5). Therefore, the evaluation of BAT UCP1 expression is essential for BAT-targeted treatment efficacy and prognosis prediction in metabolic diseases (6).

To date, the clinical evaluation of BAT UCP1 expression has relied on histological examination, such as ucp1 gene expression detection or the immunohistochemical staining of UCP1 in BAT. However, biopsy is invasive and unsuitable for repeated examinations. In addition, accurate positioning and puncturing during a biopsy can be challenging because BAT is scattered in white adipose tissue (WAT) in human adults. Therefore, it can be difficult to distinguish between BAT and WAT by visual inspection or ultrasound guidance. Imaging modalities, such as positron emission computed tomography and magnetic resonance imaging (MRI), have been used in the non-invasive monitoring of BAT distribution and function (7-10); however, it is difficult to observe and measure BAT through imaging examinations directly. Moreover, BAT activity is affected by the state of the body, environmental changes, and other factors. Therefore, a non-invasive, accurate, and dynamic observation method is urgently needed to predict and evaluate BAT activity.

Recent studies revealed that BAT UCP1 expression is closely associated with the occurrence and development of liver inflammation in metabolic dysfunction-associated steatotic liver disease (MASLD) (11,12). Accordingly, we speculated that the degree of liver inflammation in MASLD could serve as an indirect indicator of BAT activity. As inflammation can prolong the longitudinal tissue relaxation time (T1 value) and shorten the hepatic transverse tissue relaxation time (T2 value), quantitative MRI can help detect and quantify liver inflammation (13-16). Synthetic magnetic resonance imaging (SyMRI) is ideal for quantitative relaxation measurements, has been successfully used in MRI of the nervous system, bones, and joints, and has shown accuracy and repeatability in quantitative diagnosis (17). However, it has not been used to evaluate liver inflammation in patients with MASLD. Therefore, this study sought to explore the association between BAT UCP1 expression and hepatic inflammation in MASLD mice, and examine the feasibility of using SyMRI to non-invasively predict BAT UCP1 expression by quantifying hepatic inflammation. The SyMRI quantitative parameters may offer a potential non-invasive and dynamic observation method for evaluating BAT activity. We present this article in accordance with the ARRIVE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1249/rc).

Methods

Protocol registration statement

A protocol was prepared before the study but was not registered.

Animals

For the sample size estimation, we assumed an area under the curve (AUC) of 80% for distinguishing between the normal control (NC) and MASLD groups. The sample size estimation was performed using MedCalc (version 19.1, MedCalc Software, Ostend, Belgium), assuming a null hypothesis value of 0.6, a sample size ratio of 1:3 for the negative to positive group, α=0.05, and β=0.2, resulting in an expected sample size of 66. Assuming 15% modeling failure and accidental death, the final estimated total sample size was 78. For grouping purposes, our final sample size was 80. The ratio of the NC group to the MASLD group was 1:3. Therefore, a total of 80 SC57/BL6 and C57 db/db male mice were used for model construction (Figure 1). The mice were randomly divided into four groups; the C57/BL6 mice (n=20) were fed a normal chow diet for 4 weeks (the NC group); the db/db mice (n=20) were fed a normal chow diet for 4 weeks; the db/db mice were fed a methionine- and choline-deficient (MCD) diet for 4 (n=10) and 8 (n=10) weeks, respectively; and the C57/BL6 mice (n=20) were fed a MCD diet for 4 weeks and injected intraperitoneally with 10% carbon tetrachloride solution (1 mL/kg, once a week).

Figure 1 Animal model construction. C57/BL6 mice (n=20) were fed a normal chow diet for 4 weeks (NC group); db/db mice (n=20) were fed a normal chow diet for 4 weeks; db/db mice were fed a MCD diet for 4 (n=10) and 8 (n=10) weeks, respectively; C57/BL6 mice (n=20) were fed a MCD for 4 weeks and were injected intraperitoneally with 10% carbon tetrachloride solution (1 mL/kg, once a week). CCL₄, carbon tetrachloride solution; MASLD, metabolic dysfunction-associated steatotic liver disease; MCD, methionine- and choline-deficient; NC, normal control; NCD, normal chow diet.

Ethical statement

All the animal experiments were approved by the Institutional Animal Care and Use Committee of The First Affiliated Hospital, Sun Yat-sen University (No. [2021]862), and performed in accordance with institutional guidelines for the care and use of animals.

Pathological analysis

After the mice were humanely euthanized, the left and lower right lobes of the liver were collected for hematoxylin and eosin (HE) and Masson staining (Figure 2). Two pathologists with 20 and 12 years of experience in pathological liver diagnosis, respectively, evaluated the liver pathology scores independently, using the clinical research network scoring standard for non-alcoholic steatosis hepatitis (Table S1). Liver steatosis and fibrosis were denoted as S and F scores, respectively. Liver inflammatory activity was denoted as A (activity), which was calculated by adding the Li (lobular inflammation) and B (hepatocyte ballooning) scores. If the two pathologists disagreed, consensus was reached through discussion.

Figure 2 Images of BAT gross specimens, BAT UCP1 immunohistochemical staining, hematoxylin and eosin staining, and Masson staining of livers of NC and MASLD mice. (A-D) Images of a case from the NC group with the following pathological scores: S=0, A=0, F=0. (E-H) Images of a case from the MASLD group with the following pathological scores: S=3, A=0, F=0. (I-L) Images of a case from the MASLD group with the following pathological scores: S=3, A=2, F=0. (M-P) Images of a case from the MASLD group with the following pathological scores: S=2, A=4, F=2. BAT in the MASLD mice had larger sizes and lower UCP1 expression levels than that in the NC mice. In the MASLD group, as the liver inflammation score (A) increased, BAT presented with a darker appearance and higher UCP1 expression. BAT, brown adipose tissue; HE, hematoxylin and eosin; MASLD, metabolic dysfunction-associated steatotic liver disease; NC, normal control; UCP1, uncoupling protein 1.

The interscapular BAT and subcutaneous WAT of the mice were resected for HE and UCP1 immunohistochemical staining (Figure 2). Three regions of the same size, avoiding the edge of the tissue, were randomly selected, and analyzed in each pathological section. The semi-quantitative analysis of UCP1 expression was performed by averaging the positive area ratios of the three selected regions, which were calculated using ImageJ software (version 1.8.0; National Institutes of Health, USA) as follows: positive area ratio = (area of UCP1 positive expression / area of selected region) ×100% (Figure S1) (18).

MR image acquisition and analysis

The MRI examinations were performed on a 3.0T scanner (SIGNA Pioneer, GE HealthCare, Chicago, IL, USA), and an 8-channel mouse-specific coil (with a diameter of 5 cm) was used for signal acquisition. Each mouse was placed in the headfirst prone position and wrapped in a cotton cloth to minimize breathing-induced motion artifacts after the intraperitoneal injection of 1% sodium pentobarbital (50 mg/kg). The major magnetic resonance (MR) sequences included the coronal, sagittal, and axial T2-weighted sequences, multiple-delay multiple-echo sequence (MDME), and iterative decomposition of water and fat with echo asymmetry and least squares estimation quantification sequence (IDEAL-IQ). The corresponding parameters are summarized in Table S2. The slice positions of these sequences were centered in the mouse liver and linked together to ensure that the images aligned. The original data obtained from the MDME sequence were processed offline using the post-processing software SyntheticMR (version 11.2.2; SyntheticMR AB, Linköping, Sweden) to generate multiple quantitative maps, including T1 and T2 mapping images. The fat fraction (FF) and R2* mapping images were obtained from the IDEAL-IQ sequence.

The ROIs were delineated in the left and lower right liver lobes, avoiding the visible vessels and bile ducts. The mean T1, T2, R2*, and FF values in the ROIs were then calculated using T1, T2, FF, and R2* mapping images (Figure 3). Notably, the left and right ROIs of one mouse liver were treated as two samples in this study. All the ROIs were delineated independently by two radiologists with 9 and 15 years of experience in MRI diagnosis, respectively, and the average values were calculated and recorded.

Figure 3 ROIs were delineated in the left and right lobes of the liver of each mouse, avoiding visible vessels and bile ducts. T1 and T2 values were obtained simultaneously from the same ROI. In addition, ROIs were delineated at the same position in the FF map and R2* map to obtain the FF and R2* values (s^–1) of the corresponding liver lobes. FF, fat fraction; ROIs, regions of interest; sT1WI, synthetic T1 weighted image; sT2WI, synthetic T2 weighted image.

Given that the different fat and iron contents in the liver of the different mouse groups might affect the prediction performance of the inflammation levels, the T1 and T2 values were corrected by the corresponding FF and R2* values using the GLM (19), which was expressed as follows:

Yi=β0+βFFXiFF+βFEXiFE+β1Xi1+β2Xi2+β3Xi3+β4Xi4+β5Xi5+εi[1]

where Y_i was the T1 or T2 value of the ith sample; X_FF and X_FE were the FF and R2* values of the ith sample, respectively; X_ij represented whether the ith sample was present (X_ij =1) in the jth group (j=1, 2, 3, 4, 5), as represented by the five groups shown in Figure 1, and if not, X_ij =0; β₀, β_FF, β_FE, and β_j represented the regression coefficients to be estimated; and ε_i was the fitting residual. The corrected T1 and T2 values (cT2_A and cT2_A, respectively) were obtained by removing the residual amount explained by the fat and iron contents for each sample: cT_1A_i = T1_i–β_FFX_iFF–β_FEX_iFE, and cT_2A_i = T2_i–β_FFX_iFF–β_FEX_iFE. Similarly, the T1 and T2 values corrected for the fat content (cT2_FF and cT2_FF) and the iron content (cT2_FE and cT2_FE) were also calculated.

Statistical analysis

The statistical analysis and graphing were performed using SPSS Statistics (version 25.0; IBM Corp., NY, USA) and MedCalc Statistical Software (version 19.1, MedCalc Software, Ostend, Belgium). Normality tests were conducted on the relevant variables to ensure that the data distribution met the assumption of normality. If the normal distribution and homogeneity of variance were satisfied, the difference between two and multiple groups was compared using an independent samples t-test and analysis of variance, respectively; otherwise, comparisons were made using the Mann-Whitney U test (two groups) or Kruskal Wallis H-test (multiple groups). For the correlation analysis, Pearson’s correlation and Spearman’s rank analyses were used for continuous variables conforming to a normal distribution and ordered variables, respectively. The inter-observer agreement for the MRI quantitative parameters measured by the two radiologists was assessed using the intraclass correlation coefficient (ICC) test. The inter-observer agreement for the pathological scores was evaluated using Kappa statistics. The performance of each SyMRI parameter in evaluating liver inflammation and fibrosis was analyzed using the pathological score as the gold standard, and compared by receiver operating characteristic (ROC) curve analysis. K-means cluster analysis was used to divide the MASLD mice into high- and low-grade BAT UCP1 expression groups. Further, the efficacy of the liver SyMRI parameters in assessing BAT UCP1 expression was evaluated by ROC curve analysis. The optimal cut-off thresholds, sensitivity, and specificity were determined using Youden’s J statistic. Pairwise comparisons of the AUCs were performed using DeLong et al.’s method (20). A P value <0.05 was considered statistically significant.

Results

Inter-observer agreement analysis

The liver pathological evaluation showed a high level of inter-observer agreement between the two pathologists in the evaluation of S (0.924, P<0.001), F (0.858, P<0.001), Li (0.883, P<0.001) and B (0.805, P<0.001) scores. The ICC test revealed good agreement between the two observers [0.782, P<0.001, 95% confidence interval (CI): 0.687–0.850].

Basic information on NC and MASLD groups

After the exclusion of mice that died during the experiment (n=8) and those with poor MR image quality (n=9), the mice were divided into the NC (n=13) and MASLD (n=50) groups according to the liver pathological scores. The mice were assigned to the MASLD group when the sum of the liver pathological scores (S, A, and F scores) was greater than zero. The liver R2* values were significantly different between the left and right liver lobes (P<0.001); however, there was no significant difference between the left and right liver lobes in terms of the other MRI parameters and pathological scores (P>0.05). The FF values measured by MRI were positively correlated with the liver steatosis score S (r=0.901, P<0.001), and the cut-off FF values for the diagnoses of mild, moderate, and severe steatosis were 4.71%, 8.57%, and 19.17%, respectively. The liver MRI parameters and BAT UCP1 expression of different groups are provided in Table S3.

UCP1 expression between the NC and MASLD groups

There was a statistically significant difference in BAT UCP1 expression between the NC and MASLD groups (P<0.001); however, there was no statistically significant difference in WAT UCP1 expression between the two groups (P=0.779). In the NC group, there was no significant correlation between the body weight of the mice and UCP1 expression in both the BAT (P=0.874) and WAT (P=0.699). In the MASLD mice, BAT UCP1 expression was moderately negatively correlated with body weight (r=0.780, P<0.001), while the expression of WAT UCP1 was not significantly correlated with body weight (P=0.221).

BAT UCP1 expression in the MASLD mice was positively correlated with liver inflammation activity

A partial correlation analysis (controlled variate: body weight) showed that BAT UCP1 expression in the MASLD mice was moderately positively correlated with the liver inflammation activity (A) score (r=0.762, P<0.001) but not significantly correlated with the liver fibrosis (F) score (P=0.865; Figure S2). However, a moderately positive correlation was found between the A and F scores (r=0.555, P<0.001).

The cT2_A value showed high diagnostic efficacy in assessing liver inflammation

The ROC curve analysis showed that the cT2_A value had the highest diagnostic efficacy in assessing liver inflammation (A>0) in MASLD mice, with an AUC of 0.717 (Figure 4A and Table 1) and a 95% CI of 0.614–0.820. In the mice without liver fibrosis (F=0), the T2 values were the best predictor of the presence of liver inflammation (AUC =0.890, 95% CI: 0.810–0.970; Figure 4B). In addition, the AUC of T2 was significantly higher than the AUC values of cT2_A (AUC =0.836, 95% CI: 0.732–0.912), cT2_FF (AUC =0.836, 95% CI: 0.731–0.912), cT1_A (AUC =0.747, 95% CI: 0.633–0.841), and cT1_FF (AUC =0.731, 95% CI: 0.615–0.828).

Figure 4 ROC curve analysis of magnetic resonance imaging quantitative parameters to assess the presence of inflammation in the liver of the MASLD mice. (A) The cT2_A value had the highest diagnostic efficacy with an AUC of 0.717 and a 95% CI of 0.614–0.820. (B) In the cases without liver fibrosis (F =0), T2 values were the best predictor of the presence of liver inflammation (AUC =0.890, 95% CI: 0.810–0.970). AUC, area under the curve; CI, confidence interval; cT1_A, iron and fat-corrected T1; cT1_FE, iron-corrected T1; cT1_FF, fat-corrected T1; cT2_A, iron and fat-corrected T2; cT2_FE, iron-corrected T2; cT2_FF, fat-corrected T2; FE, iron content; FF, fat fraction; MASLD, metabolic-associated fatty liver disease; ROC, receiver operating characteristic.

The cT2_A value was the best predictor of BAT UCP1 high expression

A partial correlation analysis controlling for body weight showed that BAT UCP1 expression in the MASLD mice was moderately positively correlated with the liver T2 (r=0.615, P<0.001), cT2_A (r=0.508, P<0.001), cT2_FF (r=0.507, P<0.001), and cT2_FE (r=0.616, P<0.001, Figure S3) values. Conversely, BAT UCP1 expression in the MASLD mice was weakly correlated with the liver cT1_A (r=0.282, P=0.005) and cT1_FF (r=0.267, P=0.007) values. However, it was not significantly correlated with the T1 (P=0.477) and cT1_FE (P=0.373) values. Additionally, BAT UCP1 expression in the NC mice was not significantly correlated with the liver SyMRI parameters (P>0.05).

The MASLD mice were divided into BAT UCP1 high- (n=29) and low-expression (n=21) groups by K-means cluster analysis (F=370.404, P<0.001; Figure S4). Pairwise comparisons of the AUCs by the DeLong test showed that cT2_A was the best predictor of BAT UCP1 high expression (AUC =0.741, cut-off value 38.71 ms, 95% CI: 0.644–0.838, Figure 5 and Table 2), and its AUC was significantly higher than the AUC values of cT1_A (P=0.016), cT2_FE (P=0.008), and T2 (P=0.008).

Figure 5 Receiver operating characteristic curve analysis showing that cT2_A was the best predictor of BAT UCP1 expression (AUC =0.741, 95% CI: 0.644–0.838) in metabolic dysfunction-associated steatotic liver disease mice. AUC, area under the curve; BAT, brown adipose tissue; CI, confidence interval; cT1_A, iron and fat-corrected T1; cT1_FE, iron-corrected T1; cT1_FF, fat-corrected T1; cT2_A, iron and fat-corrected T2; cT2_FE, iron-corrected T2; cT2_FF, fat-corrected T2; FE, iron content; FF, fat fraction; UCP1, uncoupling protein 1.

Discussion

Our results showed that UCP1 expression in the BAT of the MASLD mice was positively correlated with liver inflammatory activity scores. Further, the quantitative MRI parameter cT2_A was able to predict UCP1 expression in BAT by quantifying liver inflammation. Liver inflammation stimulates BAT to secrete neuregulin 4, which acts directly on the liver to promote the phosphorylation of signal and activator of transcription 5, weaken lipid synthesis, and regulate glucose and lipid metabolism in the liver (21). Fibroblast growth factor-21 secreted by the liver can enhance mitochondrial function in adipose tissues, and act on the central nervous system to stimulate sympathetic neurons in BAT and accelerate body heat production (22). Mills et al. reported that UCP1 in BAT regulates extracellular succinate levels through the UCP1-succinate-SUCNR1 axis; thus, regulating immune cell infiltration and liver inflammatory activities (23). Consistent with these findings, we found a positive correlation between BAT activity and liver inflammation. However, De Munck et al. showed that after four weeks of a high-fat diet, the expression of UCP1 and the genes related to the intake and consumption of free fatty acids (e.g., Cd36 and Cpt1b) in the BAT of the mice was significantly elevated, but did not increase further over 20 weeks of the high-fat diet (24). This indicated that the feedback mechanism of BAT activity mentioned above gradually reaches its peak, and shows exhaustion or inhibition under long-term chronic inflammation due to MASLD. In this study, the animal model was developed over a period of up to eight weeks while the feedback mechanism of BAT and the liver were still active. Therefore, this study showed that BAT activity was positively correlated with the degree of liver inflammation. However, the relationship between BAT activity and liver inflammation levels in MASLD mice after long-term modeling requires further exploration.

To the best of our knowledge, this was the first study to evaluate the correlation between BAT UCP1 expression and quantitative liver fibrosis scores in MASLD mice, and no significant correlation was found. Our results showed a positive correlation between liver inflammation activity and the degree of fibrosis, which is consistent with the results of previous studies (25,26), suggesting that BAT UCP1 expression may be positively correlated with liver fibrosis. However, extensive liver fibrosis may result in impaired hepatocyte function and cytokine secretion, affecting the interaction between the liver and BAT (26), and further impairing BAT activity. Therefore, there might be an indirect correlation between BAT activity and liver fibrosis in MASLD mice, which should be further investigated.

To date, no study has evaluated BAT activity by quantifying liver inflammation. The present study showed that the MRI quantitative parameter cT2_A, corrected for fat and iron, can be used to detect inflammation in the liver of MASLD mice, and has a diagnostic efficacy superior to that of cT1_A and cT2_FE. Pathophysiologic processes, such as inflammation and edema, reflect liver injury and may increase T2 values (27,28). After correcting for fat and iron, cT2_A values can provide more accurate information on liver inflammation. However, the deposition of collagen and fibrosis causes a decrease in T2 values (28), which may affect the diagnostic performance of cT2_A in evaluating liver inflammation. In this study, the T2 and cT2_A values could be used to detect liver inflammation in the cases without fibrosis; however, their diagnostic efficacy decreased when cases with fibrosis were included. The cT2_A value of the liver was found to be the best predictor of BAT UCP1 expression, as it reflected the degree of liver inflammation in the MASLD mice despite steatosis and iron deposition. Although the cT1 value has been reported to reflect liver fibrosis, its evaluation efficacy in liver inflammation is controversial (14,15). Since both inflammation and fibrosis can increase liver T1 values, the existence of fibrosis may cause an overestimation of liver inflammation, consequently affecting the efficacy of the BAT activity prediction.

The current study had several limitations. First, the association between BAT UCP1 expression and liver inflammation in the MASLD mice was only observed over a brief period (4–8 weeks); therefore, the long-term effect should be further explored. Second, the cT2_A, cT2_FF, cT2_FE, cT1_A, cT1_FF, and cT1_FE values were calculated using a mathematical-statistical model, and a validation study of this model is needed. We will also attempt to establish an intuitive quantitative map to exhibit the overall distribution of inflammation in the liver in the future. Third, the AUC of cT2_A in predicting BAT UCP1 high expression is currently not adequate for clinical application; thus, laboratory indicators and other factors will be integrated together in future research to improve the predicting efficacy. In addition, BAT UCP1 expression was only evaluated during disease progression in MASLD mice. Future studies need to be conducted to examine the efficacy and prognostic value of BAT-targeted therapy for MASLD.

Conclusions

Liver inflammation in MASLD mice can be used to evaluate UCP1 expression in BAT. Liver cT2_A values measured using SyMRI and corrected for fat and iron can quantitatively evaluate liver inflammation, and can be used to assess the functional activity of BAT in MASLD mice.

Acknowledgments

The authors thank Dr. Long Qian for help in solving MR technical problems and preparation of the manuscript.

Footnote

Reporting Checklist: The authors have completed the ARRIVE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-1249/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-24-1249/dss

Funding: This study was funded by the National Natural Science Foundation of China (Nos. 82271958 and 82001882), Natural Science Foundation of Guangdong Province (Nos. 2023A1515011097, 2024A1515011968 and 2021A1515011442), and Fund Project of Medical Science Technology Research of Guangdong Province (No. B2025025).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1249/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All animal experiments were approved by the Institutional Animal Care and Use Committee of The First Affiliated Hospital, Sun Yat-sen University (No. [2021]862) and performed in compliance with institutional guidelines for the care and use of animals.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Herz CT, Kulterer OC, Prager M, Schmöltzer C, Langer FB, Prager G, Marculescu R, Kautzky-Willer A, Hacker M, Haug AR, Kiefer FW. Active Brown Adipose Tissue is Associated With a Healthier Metabolic Phenotype in Obesity. Diabetes 2021; Epub ahead of print. [Crossref]
2. Stanford KI, Middelbeek RJ, Townsend KL, An D, Nygaard EB, Hitchcox KM, Markan KR, Nakano K, Hirshman MF, Tseng YH, Goodyear LJ. Brown adipose tissue regulates glucose homeostasis and insulin sensitivity. J Clin Invest 2013;123:215-23. [Crossref] [PubMed]
3. Becher T, Palanisamy S, Kramer DJ, Eljalby M, Marx SJ, Wibmer AG, Butler SD, Jiang CS, Vaughan R, Schöder H, Mark A, Cohen P. Brown adipose tissue is associated with cardiometabolic health. Nat Med 2021;27:58-65. [Crossref] [PubMed]
4. Yuko OO, Saito M. Brown Fat as a Regulator of Systemic Metabolism beyond Thermogenesis. Diabetes Metab J 2021;45:840-52. [Crossref] [PubMed]
5. Poekes L, Lanthier N, Leclercq IA. Brown adipose tissue: a potential target in the fight against obesity and the metabolic syndrome. Clin Sci (Lond) 2015;129:933-49. [Crossref] [PubMed]
6. Pradhan RN, Zachara M, Deplancke B. A systems perspective on brown adipogenesis and metabolic activation. Obes Rev 2017;18:65-81. [Crossref] [PubMed]
7. Hadi M, Chen CC, Whatley M, Pacak K, Carrasquillo JA. Brown fat imaging with (18)F-6-fluorodopamine PET/CT, (18)F-FDG PET/CT, and (123)I-MIBG SPECT: a study of patients being evaluated for pheochromocytoma. J Nucl Med 2007;48:1077-83. [Crossref] [PubMed]
8. Hu HH, Smith DL Jr, Nayak KS, Goran MI, Nagy TR. Identification of brown adipose tissue in mice with fat-water IDEAL-MRI. J Magn Reson Imaging 2010;31:1195-202. [Crossref] [PubMed]
9. Abreu-Vieira G, Sardjoe Mishre ASD, Burakiewicz J, Janssen LGM, Nahon KJ, van der Eijk JA, Riem TT, Boon MR, Dzyubachyk O, Webb AG, Rensen PCN, Kan HE. Human Brown Adipose Tissue Estimated With Magnetic Resonance Imaging Undergoes Changes in Composition After Cold Exposure: An in vivo MRI Study in Healthy Volunteers. Front Endocrinol (Lausanne) 2019;10:898. [Crossref] [PubMed]
10. Huo M, Ye J, Dong Z, Cai H, Wang M, Yin G, Qian L, Li ZP, Zhong B, Feng ST. Quantification of brown adipose tissue in vivo using synthetic magnetic resonance imaging: an experimental study with mice model. Quant Imaging Med Surg 2022;12:526-38. [Crossref] [PubMed]
11. Poekes L, Legry V, Schakman O, Detrembleur C, Bol A, Horsmans Y, Farrell GC, Leclercq IA. Defective adaptive thermogenesis contributes to metabolic syndrome and liver steatosis in obese mice. Clin Sci (Lond) 2017;131:285-96. [Crossref] [PubMed]
12. Mills EL, Harmon C, Jedrychowski MP, Xiao H, Gruszczyk AV, Bradshaw GA, Tran N, Garrity R, Laznik-Bogoslavski D, Szpyt J, Prendeville H, Lynch L, Murphy MP, Gygi SP, Spiegelman BM, Chouchani ET. Cysteine 253 of UCP1 regulates energy expenditure and sex-dependent adipose tissue inflammation. Cell Metab 2022;34:140-157.e8. [Crossref] [PubMed]
13. Andersson A, Kelly M, Imajo K, Nakajima A, Fallowfield JA, Hirschfield G, Pavlides M, Sanyal AJ, Noureddin M, Banerjee R, Dennis A, Harrison S. Clinical Utility of Magnetic Resonance Imaging Biomarkers for Identifying Nonalcoholic Steatohepatitis Patients at High Risk of Progression: A Multicenter Pooled Data and Meta-Analysis. Clin Gastroenterol Hepatol 2022;20:2451-2461.e3. [Crossref] [PubMed]
14. Banerjee R, Pavlides M, Tunnicliffe EM, Piechnik SK, Sarania N, Philips R, Collier JD, Booth JC, Schneider JE, Wang LM, Delaney DW, Fleming KA, Robson MD, Barnes E, Neubauer S. Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease. J Hepatol 2014;60:69-77. [Crossref] [PubMed]
15. Pavlides M, Banerjee R, Tunnicliffe EM, Kelly C, Collier J, Wang LM, Fleming KA, Cobbold JF, Robson MD, Neubauer S, Barnes E. Multiparametric magnetic resonance imaging for the assessment of non-alcoholic fatty liver disease severity. Liver Int 2017;37:1065-73. [Crossref] [PubMed]
16. Li L, Liao B, Cai H, Zhang Y, Deng K, Chen Y, Chen M, Zhou X, Tang M, Dong Z, Feng ST. Quantitative assessment of inflammation and evaluation of spatial heterogeneity for non-alcoholic fatty liver disease in mice based on iron-adjustive T1. Quant Imaging Med Surg 2024;14:219-30. [Crossref] [PubMed]
17. Kumar NM, Fritz B, Stern SE, Warntjes JBM, Lisa Chuah YM, Fritz J. Synthetic MRI of the Knee: Phantom Validation and Comparison with Conventional MRI. Radiology 2018;289:465-77. [Crossref] [PubMed]
18. Wong YL, LeBon L, Basso AM, Kohlhaas KL, Nikkel AL, Robb HM, et al. eIF2B activator prevents neurological defects caused by a chronic integrated stress response. Elife 2019;8:e42940. [Crossref] [PubMed]
19. Dukart J, Schroeter ML, Mueller K. Age correction in dementia--matching to a healthy brain. PLoS One 2011;6:e22193. [Crossref] [PubMed]
20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45.
21. Wang GX, Zhao XY, Meng ZX, Kern M, Dietrich A, Chen Z, Cozacov Z, Zhou D, Okunade AL, Su X, Li S, Blüher M, Lin JD. The brown fat-enriched secreted factor Nrg4 preserves metabolic homeostasis through attenuation of hepatic lipogenesis. Nat Med 2014;20:1436-43. [Crossref] [PubMed]
22. Owen BM, Ding X, Morgan DA, Coate KC, Bookout AL, Rahmouni K, Kliewer SA, Mangelsdorf DJ. FGF21 acts centrally to induce sympathetic nerve activity, energy expenditure, and weight loss. Cell Metab 2014;20:670-7. [Crossref] [PubMed]
23. Mills EL, Harmon C, Jedrychowski MP, Xiao H, Garrity R, Tran NV, Bradshaw GA, Fu A, Szpyt J, Reddy A, Prendeville H, Danial NN, Gygi SP, Lynch L, Chouchani ET. UCP1 governs liver extracellular succinate and inflammatory pathogenesis. Nat Metab 2021;3:604-17. [Crossref] [PubMed]
24. De Munck TJI, Xu P, Vanderfeesten BLJ, Elizalde M, Masclee AAM, Nevens F, Cassiman D, Schaap FG, Jonkers DMAE, Verbeek J. The Role of Brown Adipose Tissue in the Development and Treatment of Nonalcoholic Steatohepatitis: An Exploratory Gene Expression Study in Mice. Horm Metab Res 2020;52:869-76. [Crossref] [PubMed]
25. Troelstra MA, Witjes JJ, van Dijk AM, Mak AL, Gurney-Champion O, Runge JH, Zwirs D, Stols-Gonçalves D, Zwinderman AH, Ten Wolde M, Monajemi H, Ramsoekh S, Sinkus R, van Delden OM, Beuers UH, Verheij J, Nieuwdorp M, Nederveen AJ, Holleboom AG. Assessment of Imaging Modalities Against Liver Biopsy in Nonalcoholic Fatty Liver Disease: The Amsterdam NAFLD-NASH Cohort. J Magn Reson Imaging 2021;54:1937-49. [Crossref] [PubMed]
26. Sanyal AJ, Anstee QM, Trauner M, Lawitz EJ, Abdelmalek MF, Ding D, Han L, Jia C, Huss RS, Chung C, Wong VW, Okanoue T, Romero-Gomez M, Muir AJ, Afdhal NH, Bosch J, Goodman Z, Harrison SA, Younossi ZM, Myers RP. Cirrhosis regression is associated with improved clinical outcomes in patients with nonalcoholic steatohepatitis. Hepatology 2022;75:1235-46. [Crossref] [PubMed]
27. Chow AM, Gao DS, Fan SJ, Qiao Z, Lee FY, Yang J, Man K, Wu EX. Measurement of liver T₁ and T₂ relaxation times in an experimental mouse model of liver fibrosis. J Magn Reson Imaging 2012;36:152-8. [Crossref] [PubMed]
28. Guimaraes AR, Siqueira L, Uppal R, Alford J, Fuchs BC, Yamada S, Tanabe K, Chung RT, Lauwers G, Chew ML, Boland GW, Sahani DV, Vangel M, Hahn PF, Caravan P. T2 relaxation time is related to liver fibrosis severity. Quant Imaging Med Surg 2016;6:103-14. [Crossref] [PubMed]