MRI Assessment of Associations between Brown Adipose Tissue and Cachexia in Murine Pancreatic Ductal Adenocarcinoma

Objective: As the major thermogenic tissue in body, the brown adipose tissue (BAT) was recently identified as an important factor to induce the rapid weight loss and malnutrition in malignancy. Current methods for detecting and quantifying brown adipose tissue (BAT) are in limited use. The aim of this study was to evaluate the changes of BAT tissue and its function in the development of pancreatic ductal adenocarcinoma (PDAC) by using magnetic resonance imaging (MRI). Methods: Ten-week-old female C57BL/6 mice were inoculated orthotopically with Pan02 tumor cells. R2* maps and two-point Dixon MRI were performed weekly for evaluation of BAT function and volume, respectively. The T2-weighted MRI was applied weekly for monitoring tumor growth. Meanwhile, the body weight was measured daily as another indication of malnutrition. The UCP1 levels in BAT and white adipose tissue (WAT) were assessed. The serum IL-6 was also measured as the biomarker of cancer-associated cachexia. Results: T2-weighted MRI indicated the rapid tumor growth from week 3 to week 5 after tumor cell inoculation. The water-fat separated MRI could clearly identify and quantify the BAT. The function and volume of BAT could be monitored by weekly MRI measurement in tumor-bearing mice. The total body weights of PDAC tumor-bearing mice were relatively stable, however, was significantly lower than that of control C57BL/6 mice. Conclusion: The results of this study demonstrated the feasibility of detection and quantification of BAT in vivo by MRI during the development of pancreatic cancer.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is one of the most highly associated with wasting of peripheral tissues, metabolic syndrome that impairs quality of life, limits cancer therapy, leading to decreased survival rate [1]. It is reported that complex cancer-associated factors contribute to the metabolic dysfunction in PDAC patients, including circulative inflammation factors, anatomic factors (extrinsic compression from the tumor causing gastric outlet obstruction), and/or the adverse effects of chemotherapy [2]. Adipose wasting may also be caused by altered exocrine pancreatic function [3]. Although many studies have tried to reverse the symptom of cancer-associated cachexia, a complete reversal of wasting by using nutrition support or anti-cytokine treatment has not been achieved [2,4].
Brown adipose tissue (BAT) is adipose tissue that contains a high number of mitochondria within cells, making it an important thermogenic tissue for both basal and inducible energy expenditure. Primarily being found in human infants and young children, the presence of physiologically significant BAT was identified in normal adult human only less than a decade ago [5]. Besides its important role in obesity, BAT has been identified as a critical factor in metabolic dysfunction and cachexia in cancer patients due to its role in energy expenditure [1,6]. Therefore, tracking the changes of BAT is a promising biomarker for predicting the progression of PDAC and evaluating tumor response to therapy. However, the changes of BAT function and volume were not well-characterized because of limitations of non-invasive imaging methods for BAT.
Currently, 18 F-FDG positron emission tomography-computed tomography (PET/CT) is the most commonly used platform for BAT imaging. However, some limitations still exist in the application of 18 FFDG PET/CT. Firstly, the results of 18 F-FDG-PET/CT are influenced by the injection amount and activity of radioactive tracers; Secondly, there is still no widely acceptable, optimized, and validated approach for data analysis and reporting. Although a guideline currently exists for BAT evaluation in FDG-PET/CT experiments, factors such as temperature and insulin level are difficult to normalize [7]. Besides, the high cost of 18 F-FDG-PET/CT scans limits its utility in clinical work. On the other hand, magnetic resonance imaging (MRI) is a promising non-invasive technique with the advantages in lack of ionizing radiation, stability, versatility and lower-cost in identification and quantification of BAT [8,9]. Several MRI methods have been reported for analysis of BAT [9][10][11]. Compared with WAT, BAT is capillary-rich and multi-locular, indicating a higher ratio of water to fat. Therefore, the method of water-fat separated MRI, which is able to detect the fat fraction (FF) among different tissues, is an ideal method for identification of BAT. Additionally, Blood-oxygen-level-dependent (BOLD-MRI) can also been applied for BAT characterization because of the higher capillary content in BAT [8,10]. Due to the high mitochondrial density in BAT, BOLD-MRI could reflect the function of BAT, showing changes in the transverse signal decay rate (R2*). In infants and adolescents, R2* of supraclavicular BAT ranges from 39 to 84 s −1 , and subcutaneous WAT ranges from 22 to 40 s −1 , indicating R2* could be used as an indicator to evaluate the function of BAT [11,12]. Besides of the potential in measuring BAT, MRI is a well-established technique to detect and characterize neoplasia in patients, including early diagnosis and evaluation of treatment response [13]. Previously, the usefulness of T2-weighted imaging in detecting the growth of tumor has already been widely recognized [14].
In this study, the volume and function of BAT were evaluated in murine model of PDAC and compared with tumor development by MRI.

Cell culture
Panc02 cell line is derived from 3-amethylcholanthrene induced PDAC in C57BL/6 mice. Panc02 cells were cultured in RPMI 1640 medium (Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich) and penicillin and streptomycin (100 IU/ml, Sigma-Aldrich). he cells were maintained in a humidified atmosphere of 5% CO 2 at 37° C. Before tumor cell inoculation, cell viability was assessed using trypan blue staining (cell viability of >90% was confirmed).

Animals
C57BL/6 female mice (10 weeks of age, weighting between 18 and 22 gm; Charles River, Wilmington, MA) were used in the study for establishing orthotopic pancreatic cancer models. Briefly, mice were anesthetized with 2% isoflurane in oxygen at a rate of 1 l/min. 1 × 10 5 viable Panc02 cells were gently mixed with ice-cold Matrigel (Corning, NY, USA) at a ratio of 3:1 to produce a homogeneous suspension. Then the abdominal cavity was opened by a 1.5 cm longitudinal incision after local shaving and disinfection. 5 μL of Panc02 cellmatrigel mixture suspension was slowly injected into the tail of pancreas. The pancreas was placed back into the abdominal cavity, and the abdominal cavity was closed by a running two-layer silk suture. Postoperative status and wound healing were monitored every day for one week After one week, a visible nodule at the location of pancreatic tail was detected by magnetic resonance imaging in all mice (protocols described below), which indicated that the orthotopic pancreatic cancer models were established successfully. Body weight was measured and recorded daily. The mice were sacrificed on week 3, 4 and 5 respectively after Panc02 cell implantation. Our animal study was approved by the Institutional Animal Care and Use Committee (IACUC) of Northwestern University.

Magnetic resonance imaging examination
MRI examinations were performed by using a 7.0 T small-animal MRI scanner (Clinscan, Bruker BioSpin, Ettlingen, Germany) with a commercial mouse coil (ClinScan, Bruker). One week after Panc02 cell implantation, tumors were detected on MRI in all the mice to confirm the successful establishment of orthotopic pancreatic cancer models. Mice were anesthetized by inhalation of a mixture of 2% isoflurane and oxygen at 1 l/min. Then, MRI images of BAT were also performed weekly. Localization was performed using a T2-weighted sequence and was followed by shimming over the interscapular BAT. The MRI sequences for BAT were as follows: Tumor size measurement was performed using ITK-SNAP (version 3.6.0, University of Pennsylvania) software, free-hand regions of interest (ROIs) were traced along the tumor margin on each slice of axial T2-weighted images containing an orthotopic tumor, and then the three-dimensional volume was calculated. R2* maps were created using a custom script in Matlab (The Math Works, Natick, MA, USA) based on BOLD imagings, and the R2* values of BAT were analyzed.

H&E staining and immunohistochemistry
Pancreatic tumor, brown and white adipose tissue were collected separately, fixed by 10% formaldehyde and followed with 4 μm-section in thickness and stained with H&E staining. BAT and WAT sections from the same sample were processed with immunohistochemistry. Briefly, the antigen retrieval was performed in a water bath using citrate sodium buffer. After blocking for 1 h at room temperature in blocking buffer (5% goat serum, 2.5% BSA in 1 × PBS), sections were incubated with diluted anti-UCP1 antibody (Abcam, Cambridge, MA) overnight at 4° C. Then the sections were washed 3 times for 10 minutes with PBS and then incubated for 1 hour at room temperature with HRP-conjugated secondary antibody. Finally, immunostaining was detected with DAB and the slides were observed by digital camera.

Quantitative qRT-PCR
BAT RNA was extracted and purified with RNeasy Lipid Tissue Mini kits (Qiagen Inc, Netherlands) and then reverse transcribed into cDNA with a Superscript IV VILO Master Mix (Invitrogen, Waltham, MA). qRT-PCR was performed using TaqMan Fast Advanced Master Mix (Thermo Fisher, Waltham, MA) and primer probes (Thermo Fisher, Waltham, MA). Thermogenic genes in BAT, uncoupling protein 1 (UCP1) and peroxisome proliferator-activated receptor γ (PPAR-γ) were compared between tumor-bearing mice and tumor naïve controls by qRT-PCR and normalized to tissue appropriate control genes (GAPDH).

Enzyme-linked immunosorbent assay (ELISA)
The expression of IL-6 was detected by ELISA. The serum of tumorbearing mice and tumornaïve mice were collected and determined by mouse IL-6 kit (R&D bioscience, Minneapolis, MN) according to the manufacturer's protocols. The absorbance was measured at 450 nm with a micro plate reader.

Statistics
For comparison between Pan02 tumor-bearing group and control group, data were assessed by Student's t-test or analysis of variance (Prism 7.0, GraphPad Software, San Diego, CA). p<0.05 was used to assess statistical significance.

Monitoring PDAC tumor development
Representative axial T2-weighted images of the pancreatic tumors at different time points after Panc02 cell implantation ( Figure 1A). The average volumes of tumor from week 1 to week 5 were 11.01 ± 2.53, 23.8 ± 8.69, 53.21 ± 21.44, 140.20 ± 49.93 and 291.29 ± 63.03 mm 3 respectively ( Figure 1B). Histological examination showed that implanted tumor with pancreas duct-looking morphology and ragged infiltration compared with the adjacent pancreatic tissue, which confirmed the PDAC characteristics of our tumor model ( Figure  1C).

Body weight
Although the total body weight after Pan02 tumor inoculation remained stable, the tumorbearing mice gained less weight compare with the C57BL/6 mice (Data from Jackson Lab) ( Figure 2) [15].

BAT function evaluation by MRI
In this study, the mice were scanned weekly for evaluating the changes in interscapular BAT after tumor cell orthotopic inoculation. The fat fraction image was calculated based on the result of Dixon MRI, and the ROIs (regions of interest) were selected based on the FF distribution in mice cervical region. As shown in Figure 3A, the BAT region could be clearly identified, as it is distinct from surrounding tissues, such as WAT and skeletal muscle. The volume of BAT was quantified by measuring the volume of ROIs while referring to the anatomical position based on T1 and T2 images. The volume of BAT was relatively stable in first 4 weeks, a significant decrease from 28.40 ± 0.73 to 23.71 ± 1.60 mm 3 at week 5 (p=0.047, n 6-8 per group) ( Figure 3B).
In order to detect the function of BAT after in PDAC model by MRI, the R2* values in BAT were measured based on BOLD sequences. As shown in Figure 3C, the average level of R2* shows no change with the increase in tumor size ( Figure 3C).

BAT mRNA expression and serum IL-6 ANALYSIS
Recent study on the regulation of BAT development and function has led to the identification of key regulatory factors such as UCP1 and PPAR-γ. To further confirm the activity of BAT in tumor development, the mRNA expression levels of these two important functional proteins in BAT were measured. As shown in Figures 4A and 4B, both of their mRNA expression levels were significantly down regulated, a 57.55 ± 9.85 decrease of UCP1 expression (p<0.05, n=5-8 per group) and a 46.84 ± 8.54 decrease of PPAR-γ expression (p<0.05, n=5-8 per group), indicating the impairment of adipogenesis in PDAC model. IL-6, a prominent cachexia-associated factor in pancreatic cancer [16], which was remarkably increased in the serum of Pan02 tumor-bearing mice, whereas that of the tumor naive mice was below the detection limit ( Figure 4C).

Histologic characteristics and UCP1 protein expression in adipose tissue
BAT from Pan02 tumor-bearing mice displayed smaller adipocytes containing less and/or smaller lipid droplets and showed a moderate reduction in UCP1 protein levels ( Figure 5A). Histological analysis of inguinal WAT revealed the presence of smaller adipocytes with big nuclei and multilocular cytoplasm in Pan02 tumor-bearing mice, however, without changes of UCP1 protein levels ( Figure 5B).

Discussion
In this study we demonstrated the feasibility of utilizing Dixon-MRI for detection BAT tissue and evaluated the possibility to monitor the BAT thermogenic activity by BOLD-MRI in mouse PDAC model during tumor progression. The volume of BAT monitored by MRI was moderately decreased with the rapid tumor growth, while BOLD-MRI showed no changes of BAT at oxygen in the early phase of tumor. The tumor-bearing mice gained less weight in the development of tumor.
Cachexia is highly associated with cancers of the pancreas, esophagus, stomach, lung, liver and bowel [17]. In particular, PDAC presents a high penetrance of wasting, a process that seems to occur in earlier stages of tumor transformation [18]. The balances of nutrition and energy in PDAC patients are of great importance to their tolerance of treatment and overall survival. Although PDAC is one of the cancers most closely associated with cachexia, there are still limited experimental models available to quantify these changes. To this aim, we took advantage of Pan02 cells, a stable cell line derived from 3-amethylcholanthreneinduced PDAC in C57BL/6 mice [19]. While Greco and coworkers modeled cachexia by intraperitoneal injections of up to 10 million cells per mouse, thus exhibiting progressive weight loss within 2 weeks and subsequent animal death within 45 days [20]. We orthotopically inoculated only 1 × 10 5 cells (the minimal amount necessary to consistently promote tumor growth) in order to promote a slower tumor growth in pancreas, thus recapitulating the cachexia features specifically associated with PDAC at early phases. This reduced cell number resulted in 50.13 ± 20.74 mm 3 tumors at 3 weeks after injection (the time point where mice from Greco et al. already started to die). Pan02 tumor-bearing mice exhibited moderate weight loss within 1 week after tumor challenge, which may be due to operative itself. In the next 4 weeks, the total body weight after Pan02 tumor inoculation remain stable, but the tumor-bearing mice gained less weight compare with the C57BL/6 mice.
BAT plays a key role in thermogenesis and energy balance and potentially contributes to the physiologic perturbations associated with cachexia [21]. Since MRI does not use harmful ionizing radiation, as opposed to PET/CT, it is more suitable for studying BAT activity in longitudinal studies on healthy subjects. However, few studies have focused on using MR imaging technologies to detect and monitor the changes of BAT during tumor progression. We demonstrated the feasibility of measuring BAT volume in vivo using routine MRI sequences. In our results, BAT volume of Pan02 tumor-mice was slightly decreased in the first 4 weeks after tumor inoculation, and moderately decreased at 5 weeks, which confirmed the early malnutrition in PDAC. However, the R2* and R2* value showed no change during the development of tumor, despite the possibility to use BOLD-MRI for the quantification of BAT activity in mice as previously described [11]. Furthermore, in accordance with previous reports [22,23], expression of 2 genes particularly associated with BAT thermogenesis, UCP1 and PPAR-γ were decreased, which indicating the impaired function of BAT in PDAC model. Importantly, histological analysis of the adipose tissues showed significant morphologic alterations and decreased expression of UCP1 in Pan02 tumor-bearing mice. It is well known that elevated levels of IL-6 are associated with increased cachexia and decreased survival in pancreatic cancer patients [20,24]. Accordingly, we also detected an upregulation of IL-6 in Pan02 tumor-bearing mice.
The current study had several limitations. For instance, there were no significant differences of R2* with tumor growth, which are unexpected results. We are developing a new gaschallenge BOLD MRI sequence to monitor BAT function in PDAC mouse model. Furthermore, the data set presented here is the focus on evaluating the BAT changes in early PDAC mice (in 5 weeks after tumor challenge). Further investigations evaluating the effect of advanced stage of cancer on cachexia are required. In addition, recent studies on pancreatic cancer cachexia modeled in mice suggested that Pdx1Cre; KrasG12D; Tp53R172H (KPC) congenic allografts recapitulate key features of PDAC disease process and induce a wide array of cachexia features [3,22]. Further study use of KPC mouse model is needed.
In summary, we demonstrated the feasibility of detection and quantification of BAT in vivo by MRI during the development of pancreatic cancer.  BAT mRNA expression and serum IL-6 concentration measurement; (A) qPCR analysis of UCP1 and PPCR-γ expression in BAT at 5 weeks after Panc02 cell inoculation and compared with age-matched control mice; (B,C) IL-6 levels in serum after Panc02 cell inoculation (* indicates to p<0.05). Results were expressed as mean ± SEM.