Distinctly abnormal brain metabolism in late-onset ornithine transcarbamylase deficiency

Patients and methods.

Results.

MRI revealed T1 and T2 prolongated round lesions in the deep white matter in Patient 1 (figure 1A), with no change in the abnormalities on the first and second scans. Spotty T1 and T2 prolongations in the subcortical white matter were observed in Patient 3. Parenchymal lesions and cerebral atrophy were not observed in the other four patients or control subjects. We examined MR images contiguous to the ROI and confirmed that there was no contamination of the ROI with CSF and abnormal T2 prolongation. Only Patient 1 had minimal CSF and abnormal T2 prolongation in the ROI.

• Download figure
• Open in new tab
• Download powerpoint

Figure. (A) T2-weighted MR image of Patient 1. The image showed small high signal–intensity lesions in the cerebral white matter, and the region of interest in the left centrum semiovale, which involved faint T2 prolongation and minimal CSF. (B) Proton spectra (repetition time 5000/echo time 30) from Patient 1 with clinical stage 3, before liver transplantation. An analysis using an LC Model represented normal N-acetyl aspartate (NAA) peak at 2.0 ppm (8.5 mM) and creatine (Cr) peak at 3.0 ppm (4.5 mM), decreased choline (Cho) peak at 3.2 ppm (0.73 mM, arrowhead), and increased glutamine plus glutamate (Glx) peak at 2.1 to 2.5 ppm (18.2 ppm). No myo-inositol (mI) peak at 3.6 ppm could be seen. (C) Proton spectra from Patient 1 with clinical stage 0, after liver transplantation. An analysis using an LCModel revealed normalization of all metabolites; NAA, 8.6 mM; Cr, 4.3 mM; Cho, 1.11 mM; mI, 3.1 mM; Glx, 8.6 mM.

Although occasional baseline disturbances arose from insufficient water suppression, the fitting algorithm of an LCModel was always able to generate a reasonable baseline. The NAA, Cr, Cho, mI, and Glx concentrations in the patients with OTCD and control subjects are given in table 2 (abnormal data are in bold letters). Compared with age-matched control subjects, the NAA and Cr concentrations remained normal in all patients with OTCD. The Glx concentration increased in four patients in proportion to the clinical stage. mI could not be detected in five patients with recent clinical manifestations. A decreased Cho concentration was only detected in the two stage 3 patients (first scan before liver transplantation in Patient 1 [see figure 1B] and second scan in Patient 2). ¹HMRS normalized after liver transplantation in Patient 1 (see figure 1C).

Discussion.

In patients with late-onset OTCD, we found mI depletion and Glx accumulation followed by Cho depletion. As ammonium appears to be the only cause of the acute encephalopathy in patients with OTCD, we hypothesize that these metabolic abnormalities, measured with ¹HMRS, are associated with hyperammonemia. In the clinical model of hyperammonemic encephalopathy during a 24-hour period in awake primates, gross neuropathologic changes included brain swelling, flattening of cortical gyri, and herniation of cerebellar tonsils.17 Light and electron microscopy revealed astrocyte swelling with pleomorphic mitochondria. Of particular im-portance was the absence of pathologic changes in neurons, axons, dendrites, oligodendroglia, and synapses. It is apparent that brain swelling often leading to increased intracranial pressure is a primary physiologic response to hyperammonemia, and that neurologic symptoms occur in the absence of neuronal pathology.1 It may be inferred that brain swelling is due to the astrocytic swelling as astrocytes are the only type of brain cell affected. Because of the intimate relationship of the astrocyte processes with cerebral capillaries and venules, it is not surprising that alterations of cerebral blood flow are commonly found in experimental hyperammonemia.18,19⇓

We performed ¹HMRS in the normal-appearing posterior centrum semiovale in five of the six patients with OTCD. We could not completely exclude contamination of faint T2 prolongation and minimal CSF in Patient 1. MRI in patients with OTCD may show the presence of localized edema in the acute phase and damage to cortex and underlying white matter in the chronic phase.12,20⇓ ¹HMRS of diffusely T2 prolongated white matter in the chronic phase of OTCD revealed very low NAA, Cr, and Cho signals consistent with a major loss of all cell types.12 This was probably due to hyperammonemia causing cerebral edema and subsequent brain damage. Therefore, contamination of the ROI with a T2 lesion or CSF in Patient 1 may cause a reduction of NAA, Cr, and Cho concentration. Because the concentration of NAA and Cr before the liver transplantation and all metabolites after the procedure were normal and contamination with T2 prolongation and CSF was minimal, we did not correct the concentration of metabolites in Patient 1 for these factors.

One of the most prominent ¹HMRS findings in this study was the markedly increased Glx concentration in four of the six patients with OTCD. Although there have only been a few reports of ¹HMRS in patients with OTCD, increased Glx was also ob- served in three of four patients previously reported.11,12⇓ It has become increasingly evident that the cell volume is determined by intracellular organic osmolyte metabolism.21 Because astrocytes are rich in Gln synthetase, it has been suggested that ammonia-induced astrocyte Gln accumulation creates an osmotic gradient that causes a shift of water into the astrocytes, resulting in astrocyte swelling, cerebral edema, and increased intracranial pressure.19,22⇓ It was also demonstrated that the cerebral edema associated with hyperammonemia can be prevented by preventing Gln accumulation in the brain, suggesting that hyperammonemia is necessary but not sufficient to produce cerebral edema.23 Supporting this hypothesis, the cerebrospinal Gln concentrations in patients with OTCD are extraordinarily high during hyperammonemic encephalopathy.1 Although Gln and Glu are indistinguishable at 1.5 Tesla, the increase of Glx concentration on ¹HMRS in OTCD likely reflects accumulation of Gln in astrocytes with hyperammonemia.

In this study, the Glx concentration increased in proportion to the clinical stage of OTCD. The Glx concentration was normal in two clinically mild patients (less than stage 1). In Patient 2, the Glx concentration increased more than 50% after a 1-year interval, during which her clinical condition advanced from stage 2 to stage 3. A stepwise increase in Glx with increasing clinical severity has also been observed in hepatic encephalopathy,7 which makes it possible to discriminate asymptomatic patients with liver cirrhosis from healthy control subjects and patients with overt hepatic encephalopathy. This suggests that the Glx concentration may be a better indicator of metabolic abnormalities in patients with OTCD as well as hepatic encephalopathy.

The second prominent ¹HMRS finding was the marked decrease in mI in five of six patients with clinical manifestations of OTCD. Glycine and mI both contribute to the peak at 3.6 ppm. Glycine depletion may lead to a reduction in the intensity of the resonance at 3.6 ppm, when large doses of benzoic acid are given as treatment for OTCD (benzoate+ glycine→hippuric acid), as in our six patients. However, depletion of the peak at 3.6 ppm in patients with OTCD might result from mI depletion itself, not from a decreased glycine concentration, because the actual cerebral concentration of glycine is probably no more than 20% that of mI.13

The role of mI in the brain is not completely understood. In vitro studies have shown that astrocytes contain high amounts of mI in contrast to neurons.3 Accordingly, mI is markedly increased in AD, HIV encephalopathy, MS, and Pelizaeus–Merzbacher disease—disorders known to be associated with astrocyte proliferation and hyperplasia.3,16⇓ It has also been shown that mI may serve as an osmolyte and plays an important role in the volume regulation of astrocytes. The role of mI as an osmolyte in the human brain is supported by the observation that hypernatremia, which is expected to lead to cell shrinkage and to favor osmolyte accumulation in brain cells, is indeed associated with an increased mI signal on ¹HMRS.5 In addition, the results of in vitro cell studies have demonstrated osmotic regulation of the expression of genes encoding mI transporters in the plasma membrane.24 A more likely explanation is that the markedly decreased mI in the early clinical stage of OTCD reflects volume-regulatory mI release in response to ammonia-induced astrocyte Gln accumulation. We hypothesize that the depleted mI and normal Glx concentrations in Patient 5 may be due to the earlier depletion of mI, before the accumulation of Gln. ¹HMRS findings in OTCD also support the hypothesis that increased Glx and decreased mI concentrations observed with hepatic encephalopathy are in fact due to hyperammonemia rather than other neurotoxins.

Decreased Cho is one component of the triad of ¹HMRS abnormalities in hepatic encephalopathy.4-10⇓⇓⇓⇓⇓⇓ Patients with liver disease without hepatic encephalopathy only showed a reduction of Cho, whereas patients with subclinical or clinical hepatic encephalopathy exhibited a notable reduction in mI and increased Glx in addition to lower Cho.4 These findings strongly suggest a progression of the neurochemical events of hepatic encephalopathy, with Cho depletion followed by mI depletion and Gln accumulation. Conversely, decreased Cho was only shown in two of our clinically advanced patients with OTCD (Patients 1 and 2). In Patient 2, initial ¹HMRS showed a normal Cho concentration at stage 2, and revealed decreased Cho at stage 3 after a 1-year interval. These findings suggested an inverse temporal pattern of metabolic abnormalities in OTCD compared with that seen in hepatic encephalopathy, with mI depletion and Gln accumulation followed by Cho depletion. Although the role of Cho in the brain is not completely understood, especially in hyperammonemia, these differences in Cho may be due to neurotoxins other than ammonium.

¹HMRS, when performed serially, may document disease severity and may be used to follow the effect of treatment in patients with OTCD. Though there are currently no definite clinical or laboratory criteria for liver transplantation in patients with late-onset OTCD, ¹HMRS should be developed further as noninvasive method for evaluating brain metabolism. ¹HMRS may be particularly helpful in liver transplantation programs to help identify patients with OTCD requiring transplantation and to follow their response to clinical interventions designed to minimize the progression of their disease.