In Vivo Phenotyping of the ob/ob Mouse by Magnetic Resonance Imaging and 1H-Magnetic Resonance Spectroscopy

Authors


  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Department of Morphological-Biomedical Sciences, Human Anatomy and Histology Section, Medical Faculty, University of Verona, Strada Le Grazie 8, 37134, Verona, Italy. E-mail: laura@anatomy.univr.it

Abstract

Objective: We studied ob/ob and wild-type (WT) mice to characterize the adipose tissues depots and other visceral organs and to establish an experimental paradigm for in vivo phenotyping.

Research Methods and Procedures: An in vivo evaluation was conducted using magnetic resonance imaging and 1H-magnetic resonance spectroscopy (1H-MRS). We used T1-weighted images and three-dimensional spin echo T1-weighted images for the morphological analysis and 1H-MRS spectra on all body mass, as well as 1H-MRS spectra focalized on specific lipid depots [triglyceride (TG) depots] for a molecular analysis.

Results: In ob/ob mice, three-dimensional evaluation of the trunk revealed that ∼64% of the volume consists of white adipose tissue, which is 72% subcutaneous and 28% visceral. In vivo 1H-MRS showed that 20.00 ± 6.92% in the WT group and 58.67 ± 6.65% in the ob/ob group of the total proton content is composed of TG protons. In in vivo-localized spectra of ob/ob mice, we found a polyunsaturation degree of 0.5247 in subcutaneous depots. In the liver, we observed that 48.7% of the proton signal is due to water, whereas in the WT group, the water signal amounted to 82.8% of the total proton signal. With the sequences used, the TG amount was not detectable in the brain or kidneys.

Discussion: The present study shows that several parameters can be obtained by in vivo examination of ob/ob mice by magnetic resonance imaging and 1H-MRS and that the accumulated white adipose tissue displays low polyunsaturation degree and low hydrolipidic ratio. Relevant anatomical alterations observed in urinary and digestive apparatuses should be considered when ob/ob mice are used in experimental paradigms.

Introduction

ob/ob Mice are a well-known animal model for obesity and diabetes (1). The genetic defect at chromosome 4 induced by the replacement of an arginine at position 105 to a stop codon (2), which gives the phenotype of the ob/ob mouse, has been known since 1950. In 1994, the mouse obese gene (ob) was isolated using classic positional cloning techniques (3). Its product, leptin, is a 167-amino acid protein, whose expression and secretion by adipose tissue is closely correlated with body fat mass and adipocyte size in rodents and humans (4, 5). The leptin gene is expressed in adipose tissue, predominantly in mature adiposities, but also in newborns (6). Leptin receptors, members of the interleukin-6 cytokine family of receptors (7), were originally found in the brain, and many of leptin's effects on energy balance are thought to be mediated centrally through interaction with several hypothalamic nuclei (8). However, this does not seem to be the only function of leptin. Similar to other cytokines, leptin has a role in hematopoiesis, immune function, fertility, and development.

The homozygote ob/ob mice leptin mRNA expression is extinguished, and there is a non-sense mutation that leads to production of a truncated, inactive form of leptin with a compensatory increase of leptin mRNA (Big Brain). Ob/ob mice are widely used in biomedical research, and we have found more than 2000 papers on Medline dealing with this animal model for disease. There are several reasons for performing in vivo phenotyping of ob/ob mice using magnetic resonance imaging (MRI)1 and proton magnetic resonance spectroscopy (1H-MRS). Both MRI and 1H-MRS provide information on the total amount of the fat (9, 10), and MRI also describes fat distribution in the different depots. 1H-MRS can evaluate the polyunsaturation degree (PUD) of the lipid chain (11) and evaluate the presence of lipids in different organs.

In the present study, we performed in vivo evaluations using MRI and 1H-MRS with the goal of characterizing morphological and biochemical aspects of adipose tissues and other visceral organs and establishing an experimental paradigm for in vivo phenotyping. In general, in vivo phenotyping of the ob/ob mouse could expand our knowledge of this model and could represent a powerful tool for comparing different animal models of obesity and diabetes. More specifically, the three-dimensional (3D) and spectroscopic data sets were used for the evaluation of six parameters: total amount of fat in the trunk (using MRI); total amount of triglyceride (TG) on all body mass (using 1H-MRS) (it is important to specify that the analysis is about liquid lipid, lipids that have fast movement and not lipids that compose fixed structure-like membrane) (12); 3D morphology and volume of single fat depots (to evaluate whether the fat hypertrophy involved all of the depots); the PUD of adipose tissue using localized 1H-MRS; the degree of liquid lipid accumulated in the liver, brain, and kidneys (using localized 1H-MRS); and the presence of any lesions or deformations in visceral organs.

Research Methods and Procedures

Animals

Two groups of ob/ob and wild-type (WT) mice, obtained from Harlan Italy (Indianapolis, IN) were used. All mice were housed at constant temperature (20 °C to 24 °C) in 12-hour light/dark cycles and were fed standard mouse chow ad libitum. The male mice in the ob/ob group (n = 5) weighed ∼40 grams, whereas the males in the WT group (n = 6) weighed ∼25 grams. The experiments were conducted following the principles of the NIH Guide for the Use and Care of Laboratory Animals and the European Community Council (86/609/EEC) directive. All efforts were made to minimize the number of animals used and to avoid causing suffering.

Imaging and Spectroscopy

For MRI acquisition, anesthesia was induced by inhalation of a mixture of oxygen and 5% isoflurane and maintained by a mixture of oxygen containing 1% to 2% isoflurane. Mice were placed in a supine position in a birdcage transmitter/receiver coil with 3.5-cm inner diameter. All MRI experiments were carried out using a Biospec Tomograph System (Bruker, Karlsruhe, Germany) equipped with a 4.7-Tesla horizontal magnet (Oxford Ltd, Oxford, United Kingdom) having a 33-cm bore. The tomograph was equipped with a Bruker BGA29 gradient insert (maximum gradient intensity, 20 Gauss/cm). A Silicon Graphics O2 computer (Silicon Graphics Inc., Mountain View, CA) was used to analyze the data.

T1-weighted (T1W) images were used to study the distribution of fat stores in ob/ob mice. After a pilot acquisition, T1W transversal and coronal slices were acquired using the following parameters: TR/TE, 1009/18 ms; slice thickness, 2 mm; and NEX, 1. Other parameters were FOV, 10 × 5 cm2, matrix size, 256 × 128, and FOV, 4.5 × 4.5 cm2, matrix size, 256 × 256 for sagittal and transversal acquisitions, respectively. Animals were positioned in the radio frequency (r.f.) coil in such a way that the kidneys were in the approximate isocenter of the coil (and magnet). Three axial slices at the abdominal level and five coronal slices were acquired from each mouse.

For 1H-MRS, spectra of all body mass were acquired using a single r.f. pulse sequence with TR = 5000-ms r.f. pulse duration of 50 μs and NEX = 16. The obtained data allowed calculation of TG and water protons from the whole body (13).

1H-localized spectra were acquired using a stimulated echo sequence with TR, 2500 ms; TM, 8.9 ms; TE, 22 ms; voxel size, 3 × 3 × 3 mm3; and NEX, 128. The spectra were defined as localized because a voxel was placed on the liver, subcutaneous abdominal area, kidneys, and brain.

For analysis of the spectral data, we defined a reference peak that was either the water peak or the peak of the double bond at 5.5 ppm (−CH = CH−). We fixed the integral value of this reference peak at one, and integral values of all TG recognized peaks were normalized to it. Analysis of the integral values of TG peaks in localized spectra allowed calculation of the PUD index (11). The PUD index was defined by PUD = G/(2/3)A, where G and A indicate the areas of peaks ∼2.8 and 0.88 ppm, respectively (integral value after baseline correction). The degree of hydration of the fat stores was evaluated in the spectral data by using a hydrolipidic ratio (HR), calculated using the following formula:

image

HR expresses the percentage of water protons over total protons in the tissue (water plus TG protons). Every time we use the term water proton or TG proton, we consider a percentage value proportional to the integral intensity of the water peak or the TG peak, respectively.

To test the validity of the data from the 1H-localized spectra of the liver, subcutaneous abdominal area, kidneys, and brain, we used phantoms of vegetable oil prepared according to Lunati et al. (14) and Poon et al. (15).

The aim was to investigate whether 1H-localized spectra could quantify lipid amount in a reliable way. Therefore, we analyzed, with 1H-localized spectroscopic sequences, suspensions with known percentages of lipid and water, and we evaluated the correlation between known percentages of fat into suspensions and those calculated by our analysis of the spectra.

The phantoms were made by mixing known amounts of water (doped with 0.2 mM MnCl2) and vegetable oil (extra virgin olive oil; Conad, Bologna, Italy) containing both 2% and 5% of Tween-80 (polyoxyethylene sorbitan monooleate; Sigma-Aldrich, St. Louis, MO), in relation to the proportion between oil and water. The mixture homogeneously mixed by ultrasonic disruption [Sonopuls ultrasonic homogenizer, with a main voltage of 50 or 60 Hz that transforms into high-frequency voltage of 20 kHz, BANDELIN by Elettrofor, Borsea (RO), Italy] contained 20%, 40%, 80%, and 100% oil by volume (Figure 3). Suspensions, in 15-mm diameter plastic tubes, were placed longitudinally into the magnet. We analyzed phantoms with the same experimental protocol used in vivo; after a spin echo sequence to localize the phantom position, we acquired a transversal T1W image to localize the voxel useful for the 1H-MRS-localized spectra acquisition, which was a stimulated echo sequence with TR, 2500 ms; TM, 8.9 ms; TE, 22 ms; voxel size, 3 × 3 × 3 mm3; and NEX, 128. All acquired spectra were analyzed and interpreted in accordance with Zancanaro et al. (16).

Figure 3.

Localized spectra in phantoms of oil-in-water suspension at different percentage of oil in volume. (A) Spectrum of suspension for 20% oil. (B) Spectrum of suspension for 40% of oil in water. (C) Plot representing the relationship between HR, calculated from the integral value of peaks of spectra, and percentage of oil in water suspensions. (D) Spectrum of suspension for 80% oil in water. (E) Spectrum of suspension for 100% oil in water.

Histology

A new group of animals, four ob/ob and three WT mice, were perfused transcardially with phosphate-buffered saline, followed by freshly prepared 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4); brains, kidneys, and livers were dissected and cut with a microtome into 30-μm thickness slices from the coronal plane of a single organ. The slices were incubated in 2% (w/v) Oil Red O reagent (an indicator of intracellular lipid accumulation) for 5 minutes at room temperature. Excess stain was removed by washing with 70% ethanol, followed by several changes of distilled water. The slices were counterstained for 2 minutes with hematoxylin. After this staining protocol, the slices were observed with the microscope and photographed.

Results

In Vivo Morphology of the Adipose Organ

3D reconstructions (Figure 1A) showed that the largest fat depots were in the subcutaneous tissue. At this location, a rostral mass and a caudal mass were visible. The rostral portion was in continuity with axillary fat and with interscapular brown adipose tissue (BAT) depot. The caudal portion of the subcutaneous fat was in continuity with the inguinal fat pad. Curiously, the ventral portion of the subcutaneous tissue, in the abdominal region, was almost devoid of adipose tissue, whereas in the visceral space, there was a plentiful store (Figure 1B).

Figure 1.

3D reconstructions of whole trunk and selected structures in an ob/ob mouse.

In ob/ob mice, 3D evaluation of the trunk by Amira software (ZIB, Berlin, Germany) revealed that ∼64% of the volume was made up of adipose tissue. With WT mice, using the 3D evaluation, it was not possible to obtain reliable 3D reconstruction because too many biases can be introduced with segmentation and smoothing processes. This computerized analysis allowed us to calculate that 72% of the total amount of white fat is subcutaneous, and 28% is visceral.

3D reconstruction gave us some information about particular lipid amount. We can see, for example, in the thorax region, which is marked cranially, a small cervical amount of BAT that emitted white adipose tissue-like signal intensity and has the characteristic lateral wings, which are thick and extended ventrally, merging with subcutaneous fat. In this region, there is a high lipid infiltration, and the thymus is not visible. Caudally lipid depot is mainly subcutaneous. The liver is very large, with high signal intensity. The enlarged liver probably deformed the stomach and caudally displaced the right kidney (Figure 1A).

Total Body Spectra

In vivo 1H-MRS allowed a quantification of the liquid lipid accumulation degree on the whole body. In both of the groups, the total body spectra showed two peaks: one of water at 4.7 ppm, and the −(CH2)n− chain peak at 1.2 ppm. As expected, the integral values of the TG peaks were significantly higher in the ob/ob mice (Figure 2A) than in the WT group (Figure 2B).

Figure 2.

1H-MRS of the whole body of ob/ob (A) and WT (B) mouse. Note the peak representative of the acid fat chains (1.2 ppm) and the peak of the water (4.7 ppm). Water peaks were at the same value to mark the different amounts of fat.

The ob/ob group had an integral value of the TG peak of 1.476 ± 0.445 [mean of the five animals ± standard deviation (SD)], and the WT group had an integral value of the TG peak of 0.2574 ± 0.0996 (mean of the six animals ± SD). Statistic valuation with Student's t test produced a p = 0.00015. The HR of the ob/ob group (mean ± SD of the five animals) was 41.3 ± 6.7, and the HR of the WT group was 79.9 ± 7.0 (mean ± SD of six animals).

From these data, it can be calculated that, in the total body of the WT group, 20.00 ± 6.92% of the total proton content was composed of TG protons. In the ob/ob group, 58.67 ± 6.65% of the total proton content was composed of TG protons.

Localized Spectra

Localized spectra obtained in phantoms of oil-in-water suspension at different percentages of oil in volume are reported in Figure 3. Figure 3C reports the relationship between the HR value (calculated through localized spectroscopy) and the known percentage of oil in water and shows that there is a good correlation.

In vivo-localized spectroscopy was used to obtain some biochemical information about the liquid lipid depot in the ob/ob mice. 1H-MRS sequences used are specific to detect liquid lipid (TG), specific for lipid in fast movement; all lipids that are localized in fixed structure-like membranes are not recognized.

The localized spectra for subcutaneous adipose tissue (Figure 4B) were calculated on a selected region of interest (ROI; Figure 4A). Because of the small size of the subcutaneous lipid depot, it was not possible to calculate spectra in WT mice. All spectra were analyzed in accordance with Zancanaro et al. (16).

Figure 4.

(A) Axial abdominal section of ob/ob mice; a selected ROI of subcutaneous fat is shown. (B) 1H-localized spectrum of subcutaneous fat. Many peaks are recognizable, in accordance with Zancanaro et al. (16). (A) 0.9-ppm terminal CH3. (B and C) 1.3-ppm CH2 of aliphatic chain. (D) 1.6-ppm CH2 β to carboxyl. (E) 2.0-ppm CH2 α to double bond. (F) 2.2-ppm CH2 α to carboxylM. (G) 2.8-ppm diallylic CH2, between two double bonds. (I and H) 4.2 to 4.3 ppm α to carbonyl and β to carboxyl. (L) 5.3-ppm CH = CH, double bond.

PUD Estimation

On localized spectra of subcutaneous depot, we have calculated the PUD of the lipid tissue according to Lunati et al. (11). In WAT of non-obese rodents, the PUD index ranged from 0.1 to 1.3 (11), and it was >1 in long-chain fatty acids due to the high number of double bonds reflecting different polyunsaturated fatty acid contents. In ob/ob mice, using the mean of the integral values of the localized spectra of subcutaneous area, we obtained a PUD value = 0.5247. This means that 9.5353% of the total subcutaneous lipid store is constituted by polyunsaturated fatty acids. It is a low PUD value, compared with Lunati et al. (11) data that found PUD = 0.68 for subcutaneous tissue in non-obese rats. The assignment of the peaks was in accordance with Zancanaro et al. (16).

Liver

In the liver, the HR value for the ob/ob group was 48.7 ± 5.1%, which corresponded to 48.7% water in the total proton signal vs. the HR of the WT group, which was 82.8 ± 4.4%. A statistical valuation between the integral values of the two groups with Student's t test gave a highly significant value (p = 2.02 × 10−9) (Figure 5).

Figure 5.

(A) Abdominal axial section showing the liver of a representative case of ob/ob group (left) and WT group (right). 1H-localized spectra were calculated into the selected ROIs. (B) 1H-localized spectra of a ROI in the liver of ob/ob (left) and WT (right) groups. Water peak on the left and fat peak on the right of both spectra marked the different proportions between water and fat amount in the hepatic tissue.

Brain and Kidney

No significant amount of liquid lipid was found in the brains or kidneys. Parallel morphological evaluation by T1W images did not reveal significant lesions in these organs in the examined ob/ob mice. We have confirmed by matched anatomical investigation with a specific staining (Oil Red O) of cerebral and renal tissues that there are no lipid amounts (Figure 6).

Figure 6.

Liver (A and B), brain (C and D), and kidney (E and F) from WT (A, C, and E) and ob/ob (B, D, and F) mice groups were dissected with a microtome from the coronal plane of a single organ and cut into 30-μm thickness slices. The slices were incubated in 2% (w/v) Oil Red O reagent (an indicator of intracellular lipid accumulation) for 5 minutes at room temperature. The slices were counterstained for 2 minutes with hematoxylin. After the staining protocol, slices were microscopically photographed. In panels A to F, scale bar = 100 μM.

Discussion

The ob/ob mouse represents a unique model to study the adipose organ under extreme hypertrophy. In ob/ob mice, all of the lipid depots are hypertrophic and hyperplastic (17). The present work demonstrates that several parameters can be obtained by in vivo examination of adipose organs. Both MRI and 1H-MRS can evaluate the degree of lipid accumulation in the whole body, and further biochemical data, such as PUD, can be calculated.

From 1H-MRS in vivo evaluation, it is possible to calculate that the TG protons represent ∼58.6% in ob/ob mice and only 22.0% in WT mice. Data have been demonstrated by parallel measures of spectra of phantoms of oil-in-water suspension with the same sequences for in vivo study. These data are in accordance with gravimetric findings (17), which, after necropsy, revealed that, in ob/ob mice, the weight of the adipose organ is three times that in lean animals. The spectroscopic data are also in good agreement with findings obtained by Amira software that, in 3D evaluation of the trunk, revealed ∼64% of the volume was formed by adipose tissue. The small mismatch between the data obtained by 1H-MRS and 3D-MRI is probably due to the inclusion of the head in the volume spectroscopically analyzed. Therefore, ob/ob mice seem to represent an extreme model of obesity described in small laboratory animals.

The 3D in vivo method of analysis allowed us to calculate that the total amount of white fat is 72% subcutaneous and 27.9% visceral. These findings are in agreement with necropsy data demonstrating that the two subcutaneous depots (17) represent ∼70% of the total weight of the organ.

In ob/ob mice, spectroscopy data also demonstrate a low PUD of the subcutaneous lipids depot [0.53 vs. 0.48 of WAT calculated in non-obese rats by Lunati et al. (11)] and a low degree of hydration as demonstrated by an HR of 7%. Morphological phenotyping obtained by MRI data demonstrates that the anatomy of the adipose organ has undergone a marked modification that does not, however, homogeneously involve all of the lipid depots. The superficial adipose tissue was mainly abundant dorsally and was virtually absent in a large region of the ventral abdominal wall. This is an interesting finding considering that the abdominal fat is the tissue most involved in the production of leptin. In general, the BAT depots almost disappeared, and the regions where they are usually located emit a WAT-like high-intensity signal. This high-intensity signal is consistent with the presence of WAT-like monolobular adipocytes in the BAT (17).

The study also provided some information about internal organs that may be useful when ob/ob mice are used as an animal model of disease. The liver is the organ most involved in lipid accumulation. In its parenchyma, the lipid depot appears to be homogeneous, and it is quantified by in vivo-localized spectroscopy and by anatomical surveys. The gross anatomy of this organ is modified with a partial divarication of different lobes due to infiltration of extrahepatic adipose tissue. Considering that the ob/ob mice are a model for hepatic steatosis, 1H-MRS seems to be a promising tool to follow the time course of lipid accumulation.

In the lymphatic system, the thymus is not detectable, and we can suppose that the spleen is compressed by the high visceral lipid infiltration. These aspects seem to be the morphological counterpart of the modified immunologic status described in ob/ob mice that have small hypocellular thymuses and impaired cellular immunity (18).

In the brains and kidneys, we have not found signs of liquid lipid accumulation by MRI or1 H-MRS or anatomical surveys; therefore, eventual altered functionality in these organs could be related to other pathogenic mechanisms. However, the liver enlargement displaces the kidneys, in particular the right one, and a mechanical deformation of the urinary way seems probable.

In conclusion, this work suggests that, in the ob/ob mouse, the anatomy of some organs is changed with respect to the WT mouse. The change could be due to simple displacement due to massive lipid accumulation (i.e., kidneys), external compression (i.e., kidneys, lungs), structural modification (i.e., stomach), or massive intraparenchymal lipid accumulation (i.e., liver). A review of the literature demonstrates that authors regularly focus on the metabolic aspect, omitting these features which could be relevant when dealing with this model.

Acknowledgement

This study was partially supported by a donation from Merrill Lynch International.

Footnotes

  • 1

    Nonstandard abbreviations: MRI, magnetic resonance imaging; 1H-MRS, proton magnetic resonance spectroscopy; PUD, polyunsaturation degree; 3D, three-dimensional; TG, triglyceride; WT, wild type; T1W, T1 weighted; r.f., radio frequency; HR, hydrolipidic ratio; BAT, brown adipose tissue; WAT, white adipose tissue; SD, standard deviation; ROI, region of interest.

Ancillary