Congenital portosystemic shunts (PSS) in mice occurs in the widely used strain of laboratory mice, i.e. the C57BL/6J strain. Recently, we’ve shown that the striking abnormal neurochemical “phenotype”, e.g. high glutamine and low myo-inositol levels in the C57BL/6J mice, is associated with sporadic congenital PSS (1). Therefore, non-invasive diagnosis of such vascular abnormality in murine liver would help to improve the reliability of this model for neurobiology. The traditional gold standard test to confirm PSS is contrast radiographic portography, which requires catheterization of a branch of the portal vein and injection of a contrast agent to illustrate portal vascular structures. Alternatively, MR angiography (MRA) is capable of imaging vascular structures with endogenous blood not only in human but also in mice.

Thus, the objective of this study was to illustrate that MRA with respiration gating allows non-invasive diagnosis of congenital PSS in C57BL/6J mice. We further evaluated potential neurochemical alterations in dorsal hippocampus of PSS mice.

Two of each healthy and abnormal neurochemical “phenotype” C57BL/6J mice (PSS mice) were obtained from the Charles River directly, and identified in a horizontal 14T magnet, as previously (1,2). A home-built birdcage coil (50-mm-OD-diameter, 32-mm-long) was used for transmitting and receiving RF (3). In order to demonstrate the feasibility of MRA on detection of PSS, five more PSS mice and their controls, including two litter-mates, were studied in a blind fashion.

In the magnet, animals were lying prone with a face mask that delivered 1-2% isoflurane mixed with 0.8L/min air and oxygen (1:1) for anesthesia. Rectal temperatures (35-37ºC) and breathing patterns (90-120bpm) of mice were examined through an MR compatible monitor system, which also delivered the desired respiration trigger signal to all imaging sequences. Throughout the entire study, animals were kept warm using a heating pad, which was regulated by a DC-voltage controller and did not interfere with the RF. A 3D GRE (GE3D) image sequence was applied without any trigger and with the respiration triggers on one phase encoding dimension. Identical FOV (30×24×15mm³) with the same resolution (RO×PE×PE2=128×96×96) was acquired. The maximum intensity projection (MIP) images were generated from the 3D MRI volume images.

Localized ¹H MRS (SPECIAL, TE/TR=2.8/4000ms, 240 averages) was applied on dorsal hippocampus (1.8×1.2×1.6mm³) of all animals as previously described (1). A ¹H quadrature coil was used. Quantification was done by LCModel referencing the endogenous water from the identical volume (80% of brain content, 1).

Without any trigger, GE3D images (15° flip angle, sw=100kHz, TE/TR=1.9/3.8ms, nt=8, 4.5 minutes) were acquired with some motion artifacts (Figure 1A). Further GE3D acquisition with the respiration triggering signals on one phase encoding dramatically improved the quality of images (Figure 1B) with a slight increase in the acquisition time, i.e. 6.4-8 minutes when the respiration was in the range of 120-90bpm. Such quality images preserved anatomical structures of liver and clearly show vascular structures, including inferior vena cava (IVC), aorta, portal vein (PV) and its branches. Most strikingly, in animals with PSS, shunt (Sh) was clearly visible in both 3D MRI volume and maximum intensity projection (MIP) images (Figure 2B). When we measured healthy mice, PV with branches to liver was obviously shown in the MIP images (Figure 1B). In PSS mice, the branches originated from the PV were absent (Figure 2B). The mean shunt length was 2.1±0.7mm (1.3-3.2mm) and the maximum shunt width was 4.2±0.7mm (2.7-4.8mm). The aorta showed much lower signal intensity when comparing to both IVC and PV (Figure 1-2). This might due to the saturation scheme here tended to enhance vessels originated from the dorsal and remained to be evaluated. Nonetheless, congenital PSS can be identified using MR angiography.

When we further evaluated the neurochemical profile of dorsal hippocampus in both groups, the strikingly neurochemical alterations were observed in dorsal hippocampus of all PSS mice when compared to their controls (Figure 3), i.e. elevated glutamine (from 3.2±0.3µmol/g in controls to 6.9±1.3µmol/g in PSS mice, student t-test p<0.0001) and reduced myo-inositol (from 5.9±0.4µmol/g in controls to 4.0±0.4 µmol/g in PSS mice, p<0.0001), reproducing previous observations (1,3). In addition, total choline (p=0.004) and taurine (p=0.019) were also altered in the dorsal hippocampus of PSS mice (Figure 3&4). This is highly consistent with that portosystemic shunting has an important impact in ammonia detoxification, which can occur in the astrocytic-specific enzyme glutamine synthase and lead to accumulation of glutamine in PSS mouse brain (1).

We conclude that mouse liver vasculature can be imaged with minimal respiration motion to identify the potential abnormality, e.g. congenital PSS in the experimental C57BL/6J mice.