Alzheimer-like brain metabolic and structural features in cholesterol-fed rabbit detected by magnetic resonance imaging

Background Hypercholesterolemia is known to increase the risk of AD in later life, the purpose of this study is to illustrate brain metabolic and structural changes in a cholesterol-fed rabbit model of Alzheimer’s Disease (AD) by using clinical 3 T Magnetic Resonance Imaging (MRI). Methods The Institutional Animal Care and Use Committee of Zhejiang Chinese Medical University approved the study. Totally 16 Japanese White Rabbits (JWR) were randomly divided into 2 groups including normal control group fed with routine diet (group NC) and high cholesterol diet group (group CD) fed a 2% cholesterol diet with 0.24 ppm copper in the drinking water for 12 weeks. Magnetic resonance spectroscopy (MRS) and structural image of rabbit brain were performed by using a 3 Tesla (T) MRI scanner with an 8 channel Rabbit coil. The chemical metabolites were identified by LC Model including N-acetylaspartate (NAA), creatine (Cr), glutamate (Glu), glutamine (Gln), Glycerophosphatidylcholine (GPC), phosphorylcholine (PCH), and myoinositol (MI). The relative concentrations (/Cr) were analyzed. Additionally, Amyloid-β (Aβ) accumulation in the brain was measured postmortem. For comparisons of MR and Aβ data between groups, two-tailed t-tests were performed. Results The ratio of NAA/Cr (0.76 ± 0.10) and Glu/Cr (0.90 ± 0.14) in group CD were lower than those in the group NC (0.87 ± 0.06, 1.13 ± 0.22, respectively, P <  0.05). Compared to the group NC (2.88 ± 0.09 cm3, 0.63 ± 0.08 cm3, respectively), the cortical and hippocampal volumes (2.60 ± 0.14 cm3 and 0.47 ± 0.07 cm3, respectively) of rabbits brain decreased in the group CD while the third and lateral ventricular volumes enlarged (44.56 ± 6.01 mm3 vs 31.40 ± 6.14 mm3, 261.40 ± 30.98 mm3 vs 153.81 ± 30.08 mm3, P <  0.05). These metabolic and structural changes were additionally accompanied by the significant increase of Aβ1–42 in the cortex and hippocampus (163.60 ± 16.26 pg/mg and 215.20 ± 69.86 pg/mg, respectively, P <  0.05). Conclusion High cholesterol diet can induce the brain metabolic and structural changes of the rabbit including lowered level of NAA and Glu and the atrophy of the brain which were similar to those of human AD.


Background
Alzheimer's disease (AD), the most common form of dementia, is characterized by the accumulation of βamyloid (Aβ) plaques and neurofibrillary tangles (NFT) composed of tau amyloid fibrils [1]. AD is commonly associated with loss of neural synapse [2] and neurodegeneration leading to memory impairment and other cognitive ailments. Epidemiological studies have revealed the development of AD cannot be traced to a single root cause [3] and no known effective treatment seems to markedly slow the progression of AD [4]. As the processing of profound studies on the role of diet and nutraceuticals, the association of the high cholesterol diet or hypercholesterolemia with neurocognitive dysfunction, cardiovascular and other metabolic diseases has been reported [5,6]. A growing body of evidence supports the claim that cholesterol metabolism plays a leading role in AD pathogenesis [7][8][9]. In addition to participating in the formation and degradation of Aβ, Cholesterol also modulates the function and aggregation of Aβ and influences its neurotoxic effects [10][11][12]. Owing to this, many cholesterol-fed animal models of AD have been established to study the pathophysiology of AD [13][14][15].
The hallmark of AD on magnetic resonance imaging (MRI) is hippocampal atrophy [16][17][18]. The workup of AD relies heavily on magnetic resonance imaging (MRI) which not only provides quantitative assessment of neural atrophy via structural magnetic resonance imaging (sMRI), but also detects changes of brain metabolism by proton magnetic resonance spectroscopy ( 1 H MRS) [19,20]. As a technique that can evaluate brain metabolites in living individuals non-invasively, 1 HMRS has been used to assess neurodegenerative diseases through N-acetylasparate (NAA), glutamate (Glu), glutamine (Gln), myoinositol (MI), choline (Cho) levels. NAA levels are correlated to the regeneration of neurons, while MI is an indicator of gliocyte activity. Glu is the main excitatory neurotransmitter in the central nervous system and Cho acts as an indicator of cell membrane metabolism.
The goal of this study was to report the brain metabolic alterations of the cholesterol-fed JWR rabbit by MRS and show its brain structural changes via sMRI. Furthermore, we planned to explore the correlation between the Aβ accumulation and the data of brain metabolism and structure.

Animal
A total of 16 male JWR rabbits 3-4 months of age and weighting approximately 2 kg upon arrival were housed individually, with free access to rabbit chow and water. Animals were maintained on a 12-h light/12-h dark cycle and experiments were performed according to the Guidelines proposed by the Laboratory Animal Research Center of Zhejiang Chinese Medical University (SYXK, Hangzhou, 2013-0164, China); approved by the Institutional Animal Care and Use Committee of the Zhejiang Chinese Medical University. 8 JWR were fed 120 g/day of rabbit chow supplemented with 2% cholesterol. Eight additional age-matched rabbits were fed routine diet and served as controls. Rabbits were kept on their respective diets for 12 weeks.

Blood biochemical index measurement
After an abrosia period of 12 h, blood samples were drawn from the JWR rabbits via the Auricular marginal veins. The serum total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) and Triglycerides (TG) levels were analyzed by an auto-biochemical analyzer (7020, ITACHI, Japan) using the corresponding kits (Shanghai Shenneng-DiaSys Diagnostic Technology Co., Ltd., China).

MRS and sMRI analyses
The spectra of the routine diet and 12-week cholesterol diet rabbits were analyzed with LC Model (Version 6.3, Provencher SW.) [21]. The metabolites were identified including N-acetylaspartate (NAA), creatine (Cr), glutamate (Glu), Glycerophosphatidylcholine (GPC), phosphorylcholine (PCH), and myoinositol (MI). Well fitted metabolic peaks were included in the analyses when LC Model reported Cramer-Rao minimum variance bounds of less than 20%. The relative neurochemical concentrations (/Cr) including ratios of NAA/Cr, Glu/Cr, MI/ Cr and (GPC + PCH)/Cr were conducted automatically.
The cortex, hippocampus and ventricles of the rabbits were identified with reference to the rabbit brain atlas [22][23][24] (Fig. 2). The morphological measurement of the cortex and hippocampus were performed on the 3D SPGR images. Once the whole region of cortex or hippocampus was identified manually, the volume was calculated automatically with a standard volume measurement tool equipped in the workstation of MR system (ADW 4.6, GE, USA). Due to the thinner and irregular shape, the third and lateral ventricular volumes were calculated by multiplying the areas of each slice by the thickness of slice, and summing the volumes from each slice, And the areas of each slice was measured manually with a distance measurement tool in the ADW4.6 workstation of MRI. The atrophy rates of the cortex and hippocampus and the enlargement rates of the third and lateral ventricles were calculated.

Aβ 1-42 level of cortex and hippocampus
Animals were anesthetized using 3% pentobarbital sodium 30 mg/kg injected via the marginal ear vein. The cortex and hippocampus of the rabbits were separated after the brains were extracted. A 10% w/v rabbit cortex and hippocampus homogenate were made in pre-chilled physiological saline using an IKA ULTRA-TURRAX homogenizer (IKA®-werke Gmbh & Co., KG, German), the obtained homogenate was centrifuged at 4500 rpm, 4°C for 10 min. The supernatant was extract and the Aβ1-42 level was assayed immediately using homologous specific ELISA kits (Nanjing Jiancheng Bioengineering Institute, China), according to the manufacturer's instructions. Protein concentration was determined by Coomassie brilliant blue method (Nanjing Jiancheng Bioengineering Institute, China).

Statistical analysis
For comparisons of MR and Aβ data between cholesterolfed and control group two-tailed t-tests were performed. The correlation between the Aβ accumulation and the data of brain metabolism and structure were explored by using Pearson correlation analyses. For all tests, the nominal level of significance was P < 0.05. GraphPad Prism 4.0a (GraphPad Software Inc., San Diego, CA) was used for statistical analysis.

General characteristics of two JWR rabbits groups
Eight JWR were fed a 2% cholesterol diet for 12 weeks to establish an AD model. Compared to the NC group, cholesterol-fed animals in CD group had significantly higher TC, HDL-C and LDL-C levels (P < 0.01, Table 1).

The metabolites in MRS
The values of NAA/Cr (0.76 ± 0.10) and Glu/Cr (0.90 ± 0.14) in high cholesterol diet (CD) group were lower than those in the NC group (0.87 ± 0.06, 1.13 ± 0.22, respectively, P < 0.05). On the other hand, there were no statistical significance on the values of MI/Cr, and (GPC + PCH)/Cr between two groups (Fig. 3).

Aβ 1-42 levels in cortex and Hippocampus
The Aβ 1-42 levels differed between group CD and group NC. Compared to group NC, rabbits of group CD had higher cortical and hippocampal Aβ 1-42 levels (149.60 ± 16.26 pg/mg vs 114.5 ± 5.86 pg/mg  and 212.20 ± 18.86 pg/mg vs 132.3 ± 2.40, respectively, P < 0.05) (Fig. 6). The Aβ 1-42 levels of cortex and hippocampus had significant correlations with the values of NAA/Cr and Glu/Cr and the volumes of the cortex and hippocampus (Fig. 7).

Discussion
While many causes of AD are known, hypercholesterolemia is specifically known to drastically increase the risk of AD in later life. In our study we have established a cholesterol-fed rabbit model with significantly higher TC, TG, HDL-C and LDL-C levels. Furthermore, compared to the group NC, these rabbits in group NC were characterized by the significant increase of Aβ1-42 in the cortex and hippocampus. As the previous studies indicated [7][8][9], cholesterol metabolism plays a leading role in AD pathogenesis. Increased cholesterol levels can enhance the β-secretase pathway of APP to promote the formation of Aβ via lipid rafts [25] and reduce the degradation of Aβ by affecting the function of enzymes participating in the regulation of Aβ. These enzymes include insulin degrading enzyme (IDE) [26]. In addition, cholesterol also modulates the function and aggregation of Aβ and influences its neurotoxic effects [11].
We detected the variation of brain metabolites with 3.0 T MR Scanner. The ratio of NAA/Cr and Glu/Cr in cholesterol-fed JWR rabbits in the current study was significantly lower than the control group. This finding parallels previous clinical data regarding markers of AD which have reported that the ratio of NAA/Cr and Glu/Cr significantly decreases in AD patients compared to normal individuals [27,28]. The ratio of NAA/Cr and Glu/Cr in our result is also in accord with the previous in vitro 1 H MRS study in rats and mouse model of AD [29].
NAA mainly exists in neurons and is synthesized by chondriosome from acetyl-coenzyme A and aspartate. Although NAA remains a mysterious molecule, it is considered as a marker of neuronal vigour and number, the decreased level of NAA in MRS can be used as a representation of neuronal dysfunction or death [30].
Glu is the most abundant amino acids in mammalian brain. Converted from glutamine by neuron, Owe to the decreased level of Glu reflects the destruction of glutamatergic synapses in AD, some researchers pointed out the Glu may be more significant than NAA/Cr alone in representing the progress of AD [31,32].
As the main excitatory neurotransmitter in the central nervous system, Glu could selectively active the Nmethyl-D-asperate receptor (NMDAR), which is widely distributed in the brain. NMDAR is one of the most important excitatory amino acids receptor, and of great significance in the physiology and disease states. On the other hand, via upregulation of NMDAR, the excess Glu can causes the Ca 2+ influx and induce excitatory neural toxicity leading to apoptosis [33]. As a previous study [34] showed indicated, cholesterol could increase the activity of NMDAR to intensify the neuronal excitability of Glu. As mentioned above, cholesterol could induce neuronal apoptosis not only through Aβ but also via NMDAR.
MI exists only in gliocyte, and is used as a marker of gliocyte health, with the rise of MI levels often used as an indicator of gliocyte proliferation [35].The activation of gliocytes is highly involved in the pathogenesis of AD. Activated gliocytes secrete immune molecules and activate complements to trigger the immune inflammatory reaction, resulting in neuronal damage in AD [36,37]. Researchers investigated the levels of MI/Cr in AD have demonstrated varied conclusions. While some researches have suggested the level of MI/Cr [38] increases during AD, more studies pointed out the difference was not statistically significant between AD and normal control group [39,40]. Our study showed the (GPC + PCH)/Cr ratio was not statistically different between the two groups.
Another finding of our study was that JWR rabbits fed a 2% cholesterol diet showed significant cortical and hippocampal atrophy, in accordance with current clinical standards for AD [16][17][18]41].Similar to previous studies based on New Zealand White rabbit model of AD [42], cholesterol-fed JWR rabbits provided support for one of the most consistent findings in AD-an ventricular enlargement. In our study, the atrophy of the hippocampus is more evident than the cortex in cholesterol-fed rabbits, and the enlargement of the lateral ventricle adjacent to the hippocampus is more evident than the third ventricle. These result implied that hippocampus is affected earlier than other parts of brain by the hyperlipidemia. This brain characteristic in cholesterol-fed rabbit was similar to that of AD [43,44]. As the hypothesis posed by Smith describes [45], in AD patients, the atrophy resulted from the formation of NFT begins in the hippocampus. With the loss of the input from the projection neurons in the hippocampus, the activity of neurons in neocortex begins to decrease, resulting in the decrease of the local cerebral blood flow and the formation of NFT.
Besides verifying cortical and hippocampal atrophy with ventricular enlargement in JWR rabbits fed a 2% cholesterol diet, our study detected the change of brain metabolites in cholesterol-fed JWR rabbit model using MRS, adding metabolic evidence to the feasibility of the cholesterol-fed rabbit model of AD. Yet, this study is not without limitations. Firstly, this study did not include any behavioral tests to verify clinical symptoms of memory loss. Secondly, a test for tau protein, which is important in AD, was not included in our study.

Conclusion
Rabbits fed with cholesterol diet displayed Alzheimerlike signs of AD development, not only the cortical and hippocampal atrophy with ventricular enlargement, but also the alterations of brain metabolites such as lowered ratios of NAA/Cr and Glu/Cr.