Brain metabolism in patients with hepatic encephalopathy studied by PET and MR

Highlights

• Cerebral oxygen uptake and blood flow are reduced to 2/3 in cirrhotics with overt HE.

• These two measures are not changed in cirrhotic patients with minimal HE or no HE.

• Cerebral ammonia metabolism is enhanced due to increased blood ammonia in patients.

• Cerebral ammonia kinetics is not affected by hyperammonemia.

• MR demonstrates low-grade cerebral white matter oedema in cirrhotics with HE.

Abstract

We review PET- and MR studies on hepatic encephalopathy (HE) metabolism in human subjects from the point of views of methods, methodological assumptions and use in studies of cirrhotic patients with clinically overt HE, cirrhotic patients with minimal HE, cirrhotic patients with no history of HE and healthy subjects. Key results are: (1) Cerebral oxygen uptake and blood flow are reduced to 2/3 in cirrhotic patients with clinically overt HE but not in cirrhotic patients with minimal HE or no HE compared to healthy subjects. (2) Cerebral ammonia metabolism is enhanced due to increased blood ammonia in cirrhotic patients but the kinetics of cerebral ammonia uptake and metabolism is not affected by hyperammonemia. (3) Recent advantages in MR demonstrate low-grade cerebral oedema not only in astrocytes but also in the white matter in cirrhotic patients with HE.

Introduction

We review Positron Emission Tomography (PET)¹ and Magnetic Resonance (MR) studies on hepatic encephalopathy (HE) in human subjects. It is interesting to observe how unanswered questions raised by an expert commission a few years ago [1] are now solved using these imaging technologies, as for example the reversibility of the reduced cerebral oxygen consumption during HE [2]. Together with in vitro studies on basic molecular mechanisms of for example ammonia metabolism in astrocyte-neuron co-cultures or experimental animal models of hyperammonemia as discussed in other articles of this issue of ABB, PET and MR studies have helped to improve our understanding of the HE pathophysiology but also raised new questions that are fundamental for improving treatment of patients who suffer from HE.

PET and MR studies of the brain complement each other: PET provides functional molecular imaging data on blood perfusion, kinetics and rates of metabolism and MR provides data on morphology, blood perfusion and water content with high spatial resolution; MR spectroscopy (MRS) provides data on a number of relative as well as absolute metabolite concentrations. Co-registration of PET and MR images and, better, combined PET/MR cameras yield fused images for high spatial accuracy of the localization of biological processes.

PET is an advanced functional molecular imaging technology that can be used to visualize and quantify biochemical and physiologic processes in vivo using positron emitting tracers of natural substances or analogs hereof. Micro-doses of tracers labeled with positron emitting isotopes are administered intravenously or as inhalation. The absorbed radiation dose caused by radiolabeled tracer to the subject studied typically ranges from 0.5 to 10 mSv, i.e. up 3 times the yearly background radiation that in most of Europe is around 3 mSv.

Specific tracers for the processes studied are chosen, being natural substances, analogs hereof, or receptor ligands, in the present context of metabolic brain studies, ¹⁵O-water (half-life 2 min), ¹⁵O-oxygen (half-life 2 min), ¹³N-ammonia (half-life 10 min), and 2-[¹⁸F] fluoro-2-deoxy-d-glucose (¹⁸F-FDG) (half-life 109 min). Within the body, the positrons annihilate with electrons under the emission of oppositely directed pairs of gamma rays that are registered by the PET camera, being reconstructed to yield 3D images of activity concentrations. Due to the short half-life’s of ¹⁵O and ¹³N, studies of CBF, oxygen and ammonia metabolism can easily be repeated the same day for example after experimental intervention whereas the longer half-life of ¹⁸F in practice allows for only one ¹⁸F-FDG study of glucose metabolism pr. day.

For quantitative studies, tracer administration is followed immediately by dynamic PET-recording of the time-course of radioactivity concentration in the brain tissue combined with measurements of the time-course of radioactivity concentration in arterial blood samples. Dynamic PET data are traditionally analyzed using compartmental modeling of tracer transport across the blood–brain barrier and possible intracellular metabolism as illustrated below for the specific tracers used. The models are defined by rate constants for the transfer of tracer from one compartment to another, such as transport of tracer across the blood–brain barrier, and intracellular metabolic conversion of the tracer to its metabolites. The rate constants are estimated fitting by non-linear regression of the model equations to the time-course of the radioactivity concentration in tissue as output, using that in arterial blood as input.

This model is obviously not physiologically optimal as it ignores tracer concentration gradients that develop along the direction of the blood perfusion of the capillaries as tracer is taken up by the lining cells from the blood flowing through the capillary from the arterial inlet to the venous outlet. Compartmental modeling assumes uniform tracer concentration throughout the capillary equal to the arterial blood concentration and, moreover, it assumes no re-uptake from the blood of tracer that has been subject to backflux from cell-to-blood. It is a difficult and still unsolved challenge to interpret the externally recorded PET data in the concept of capillary physiology [3]. For the time being, we therefore have to stick to compartmental analysis, the use of which nevertheless has clarified important questions on the regulation of brain hemodynamics and metabolism in patients with HE.

Conventional and advanced MR techniques such as magnetization transfer MR, diffusion-weighted MR, functional MR and MR spectroscopy have provided exciting potential tools to assess patients with HE. All these techniques have the advantage that they can be easily repeated at short time intervals with no risk to the subjects. They give information about changes in brain size, function, metabolism and network integrity. In order to better illustrate the MR findings in HE patients, we will first briefly describe the MR techniques that have been more commonly employed in patients with HE.

Developed more than 20 years ago, fast fluid-attenuated inversion recovery (T2-FLAIR) imaging has become one of the cornerstones of MR protocols. T2-FLAIR imaging is based on MR sequences that suppress cerebrospinal fluid signal. As a result, T2-FLAIR renders white matter lesions, particularly those that are adjacent to the cerebrospinal fluid, much more noticeable compared with proton-density- and T2-weighted images [4].

Magnetization transfer imaging (MTI) is a more recently developed MR technique that is mainly based on the interaction between hydrogen nuclei in relatively mobile environments, such as those in body water (called bulk water protons), and hydrogen nuclei in which motion is restricted, such as those covalently or non-covalently bound to proteins, other macromolecules and membranes [5]. Most of the signal in conventional MR sequences comes from the bulk water protons, whereas low-mobility protons associated with macromolecules and membranes have an extremely short T2 and are not directly detectable with conventional MR protocols. However, the interactions between these two pools of protons result in a continuous exchange of magnetization, which is referred to as cross-relaxation or magnetization transfer. MTI is able to detect this exchange by inducing a selective saturation of the low-mobility proton pool of hydrogen nuclei. The saturation of these protons is then transferred to the bulk water protons resulting in a decrease in signal intensity in the regions where the exchange is occurring.

The difference in signal intensity with or without MT is calculated pixel by pixel and is named magnetization transfer ratio (MTR). MTR, therefore, gives an indication of the quantity of bound protons present in the tissue. Low MTRs indicate the loss of brain structures able to exchange magnetization with the surrounding bulk water protons and thus reflect myelin damage, cell loss, changes in water content but also, particularly when MTRs reductions are mild, presence of neuroinflammation and/or low-grade brain oedema. Several studies have now assessed MTR in animal models of demyelination and in patients with multiple sclerosis, HIV encephalitis and progressive multifocal leukoencephalopathy providing further evidence that demyelination and axonal loss are the main contributors to MTRs decreases [6], [7], [8]. In summary, MTI provides information relative to the macromolecular content that is not visible with conventional MR and is, therefore, potentially able to detect microstructural damage in regions that could appear normal using conventional sequences.

Diffusion tensor imaging (DTI) and diffusion tensor tractography (DTT) are the most frequently used diffusion-weighted MR techniques. They provide an indirect evaluation of neuronal connectivity in the brain by assessing integrity of tracts in the white matter. Normally water diffusion in the brain is anisotropic, as it is constrained along nerve fibers in brain tissue. Any pathological process involving the tracts leads to loss of this directionality of diffusion or anisotropy. Diffusion data may be expressed as mean apparent diffusion coefficient (ADC), which measures total molecular motion averaged over all directions, and fractional anisotropy (FA), which is a measure of the directional diffusivity of water. DTI measures the direction and magnitude of diffusivity of water molecules in tissues and can be used as an index of damage to neuronal tracts. DTI is potentially useful for the quantitative assessment of myelin versus axonal injury. The diffusion of water perpendicular to the axonal fiber (radial diffusivity) is associated with myelin integrity and therefore increases during demyelination. Conversely, axial diffusivity, which reflects microscopic water movement parallel to the axonal fiber, is decreased with loss of axonal integrity. DTT is a computational procedure that reconstructs major fiber bundles in the brain by tracking the dominant direction of water diffusion, which is assumed to correspond to the longitudinal axis of the tract.

Functional MR (fMR) detects changes in regional blood oxygenation and has been extensively used to investigate real time changes in local neural activity during sensorimotor and cognitive tests in both healthy subjects and patients with neurological disorders. Briefly, the technique uses the blood oxygenation level dependent (BOLD) effect and MR sequences, which are sensitive to changes of local blood oxygenation level as deoxyhaemoglobin, but not oxyhaemoglobin, is paramagnetic. Task-induced changes in neuronal activity lead to increased regional brain perfusion and result in raised levels of oxyhaemoglobin, which can be detected as increased signal by fMR. In more recent years, fMR has also been employed to assess brain activity in the resting state. It has been recognized that some brain regions, which are anatomically separated but functionally connected, display highly coherent spontaneous BOLD fluctuations in resting state, suggesting that there is an intrinsically organized default mode network (DMN) in the resting brain. This DMN, which is primarily made up of the posterior cingulate cortex/precuneus, the medial prefrontal cortex, the inferior parietal and temporal cortices and the parahippocampal gyrus, is thought to represent the functional substrate of the brain’s baseline activities, including consciousness but is suppressed during the performance of cognitively demanding externally clued tasks [9]. In recent years, resting state fMR has increasingly been employed to explore the functional organization of the brain and to examine how it is altered in patients with neurological and psychiatric diseases and in subjects with different degrees of impaired cognition and consciousness [10].

Finally, MR spectroscopy (MRS) uses the same basic physical principles as MR to generate a spectrum of metabolites contained in specific regions localized and sampled from MR images. The most commonly used nuclei for MRS studies are ¹H and ³¹P. ¹H MRS can be used to acquire information on brain metabolites and osmolytes such as choline, creatine, N-acetyl aspartate, glutamine, glutamate, taurine and myoinositol, whereas ³¹P MRS provides information on phosphomonoesters, inorganic phosphate, phosphodiesters, phosphocreatine, γNTP, αNTP and βNTP. The technique offers a powerful noninvasive tool to assess the regional chemical environment of the brain.