Brain N-acetylaspartate is elevated in Pelizaeus–Merzbacher disease with PLP1 duplication

Patients and methods.

MRS.

In five patients and 14 healthy individuals ranging in age from 4 to 14 years (mean age, 8.5 years), proton MRS was performed with informed consent. These control individuals had normal neurodevelopmental and neurologic assessments and normal cranial MR imaging results. This study was approved by the institutional ethical standard committee on human experimentation.

We used a 1.5 Tesla apparatus (General Electric Medical Systems, Signa Horizon, Milwaukee, WI) with a standard quadrature head coil for MR imaging and proton MRS. We performed serial axial T2-weighted (4,000/100/2 [repetition time (TR)/echo time (TE)/excitation]) MR images with a slice thickness of 6 mm and a slice gap of 1.5 mm to establish a region of interest (ROI) for the proton MR spectroscopic studies. In addition to the diffuse T1 and T2 elongations in the cerebral white matter that were found in all patients, mild cerebral atrophy was observed in Patient 4 (figure 1), who has connatal PMD and microcephalus (<2 SD). No cerebral atrophy or microcephalus was observed in the other patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. T2-weighted image of Patient 4. The image shows mild cerebral atrophy, diffuse high signal intensity in the cerebral white matter, and the region of interest in the left posterior portion of the centrum semiovale.

Shimming was performed at less than 4 Hz frequency width half maximum, based on the water signal. Automated proton brain examination with the point resolved spectroscopy technique (TR and TE of 5,000 and 30 milliseconds) was performed to acquire localized spectra in the posterior portion of the centrum semiovale (see figure 1); the voxel size was 4.5 mL (15 × 15 × 20 mm); and the sum of the free induction decay was 64. To confirm that contamination of ROI with CSF did not confound our analyses, we examined MR images contiguous to the ROI. We determined the absolute metabolite concentrations using LCModel,17 which fitted spectra as a linear combination of model spectra acquired with very high signal-noise ratio and narrow line-width in vitro, known as the “basis set.” We used a basis set for TE 30 milliseconds supplied by General Electric Medical Systems, which should facilitate intersite comparisons of metabolite concentrations. To calibrate for absolute concentrations, a spectrum from a phantom containing 50 mM NAA was analyzed periodically using the LCModel. By finding the calibration factor so that the analysis of the 50 mM NAA phantom data yielded the correct concentration of 50 mM, we automatically calibrated our scanner to all of the model spectra, including Cr, Cho, and myoinositol (MI). The relaxation and concentration correction for the tissue water was unnecessary, which was essential for the internal water method. We used the Mann–Whitney U test for statistical evaluation of the data.

Results.

Proton spectra were obtained from each patient (figure 2). Although occasional baseline disturbances arose from insufficient water suppression, the fitting algorithm of the LCModel was able to generate a reasonable baseline. The NAA, Cr, Cho, and MI concentrations of the patients with PMD were compared to those of normal controls (table). Compared to age-matched controls, the patients with PMD had elevation of absolute levels of NAA by 16% (p < 0.01), Cr by 43% (p < 0.001), and MI by 31% (p < 0.01) (figure 3). There was no statistical difference in Cho concentration (see figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Proton spectra (repetition time [TR]/echo time [TE]: 5,000/30) from Patient 4. An analysis using LCModel represented N-acetylaspartate peak at 2.0 ppm (9.1 mol/L, arrow), creatine peak at 3.0 ppm (5.8 mol/L, small arrow), choline peak at 3.2 ppm (0.86 mol/L, arrowhead), and myoinositol peak at 3.6 ppm (3.8 mol/L, small arrowhead).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Absolute concentration of the brain metabolites. The mean concentrations of metabolites ±SD of the patients (○ and control subjects (•): N-acetylaspartate (NAA), 9.0 ± 0.5 and 7.8 ± 0.4; creatine (Cr), 5.4 ± 0.4 and 3.9 ± 0.3; choline (Cho) 0.97 ± 0.10 and 0.99 ± 0.10; and myoinositol (mI) 3.9 ± 0.5 and 2.9 ± 0.4 mmol/L. Patients with Pelizaeus–Merzbacher disease represent significant elevation of concentrations of NAA (p < 0.01) by 16%, Cr (p < 0.001) by 43%, and MI (p < 0.01) by 31%.

Discussion.

Proton MRS provides in vivo insights into changes in cellular metabolites in cerebral degenerative disorders. Three studies of relative proton MRS in PMD have been reported.12-14⇓⇓ However, there has been no evaluation of absolute metabolite concentrations in patients with PMD except for one case report in which no clinical and genetic information was described.18 In this study, we performed quantitative proton MRS using fully automated postprocessing technology, the LCModel,17 to determine the metabolite change in five patients with PMD with PLP1 duplications.

The concentrations of each metabolite in the control samples in this study were in agreement with previously published norms,8 suggesting that our procedure for the quantification of MR spectra was reliable. In infancy, total NAA concentrations are decreased compared to normal adult values. During childhood, NAA concentrations gradually increase in the cerebral white matter, whereas Cho concentrations gradually decrease, reflecting the maturation of the neuronal network and myelin assembly. In the white matter, NAA and Cho concentrations are 6.5 mM and 1.8 mM at 1 to 2 years of age, and shift to 8.2 mM and 1.6 mM by adulthood; the Cr concentration stays stable after 1 year of age.8

The study was performed on a reasonably uniform cohort of patients. First, patients in this study were all between 4 and 10 years old. Second, all patients had duplications of PLP1. These are particularly important because metabolic condition may change in the later stage of disease due to secondary degenerative changes. In addition, patients with PMD with PLP1 point mutations, duplications, and deletions often present with different severity and diverse degrees of dysmyelination,4 an observation suggesting the possibility for distinct pathophysiology of different mutations in PLP1.

The most significant finding of this study is the elevated concentration of NAA, an observation not detected by previous studies using the NAA/Cr ratio.12-14⇓⇓ An increased concentration of NAA has been observed in patients with Canavan disease, a deficiency of aspartoacylase, which breaks down NAA into aspartate and acetate,19 and in patients with Salla disease, a recessively inherited lysosomal storage disorder characterized by accumulation of free sialic acid (N-acetlyneuraminic acid [NANA]).20 Although we can not exclude the possibility that the PLP1 duplication results in an accumulation of metabolites whose signals overlap with NAA in MRS, it is more likely that the elevated signal results directly from increased NAA concentration because the accumulation of other metabolites has not been associated with PMD.

This moderate increase of NAA (16%) may result from altered axonal metabolism. Recent studies have proposed a hypothesis of cross communication between neurons and oligodendrocytes during the CNS development.21 Candidates for mediating this axon-myelin cross interaction include myelin-associated glycoprotein,22,23⇓ neuregulins and their receptors,24,25⇓ and integrin family members.26 Because patients with PMD cannot produce mature myelin and, consequently, proper axon–myelin interaction is not established, there may be a homeostatic change in the axon due to the lack of signals from myelin. If such signals have an inhibitory action, this may result in hyperactive axonal metabolism detected as increased NAA by quantitative MRS. Interestingly, mice with a low copy number of PLP1 transgenes have late-onset axonal degeneration,27 which suggests that, at least in mice, incomplete CNS myelination due to increased PLP1 dosage affects axons. Furthermore, mutations in the genes responsible for peripheral demyelinating diseases also affect axons.28 In summary, these findings support our hypothesis that the PLP1 duplication diminishes the axon/myelin interaction and secondarily alter axonal metabolism increasing the concentration of NAA.

Alternatively, increases in number of the oligodendrocyte progenitor cells in the brains of patients with PMD may result in the increased NAA concentration. This hypothesis is based on two observations. First, jimpy mice, a murine model of PMD carrying a plp mutation, have an increased number of oligodendrocyte progenitor cells in the brain and spinal cord, which is accompanied by extensive apoptotic cell death of differentiated oligodendrocytes.29-31⇓⇓ This increased progenitor cell population is also found in the transgenic plp mice, a model for PMD with the PLP1 duplication (personal communication, Dr. Ikenaka at National Institute for Physiologic Sciences, Aichi, Japan). Second, primary cultures of rat oligodendrocyte progenitor cells (O2A-positive cells) and mature oligodendrocytes have a NAA concentration twice as high as that observed in neurons.32,33⇓ Therefore, the elevated absolute concentration of NAA observed in patients with PMD might result, in part, from increased signal from this progenitor oligodendrocyte cell population. However, consideration has to be given to the extrapolation of in vitro cellular data to the in vivo situation, because numerous factors such as cell to cell interactions cannot be mimicked in culture, and it should be stressed that the exact mechanism of increased NAA is still unknown.

Quantitative proton MRS also revealed increased concentrations of Cr and MI. The total Cr concentration, the sum of phosphocreatine and free Cr, is a marker for energy metabolism.8,34⇓ Increased Cr concentration has been reported recently in the lesions of patients with MS.15,16⇓ In one study, the increased astrocytic activity associated with gliosis was estimated to be responsible for the elevated Cr concentration in MS.16 As the reactive astrocytic gliosis is also observed in PMD,1,2⇓ the elevated Cr peak may result from the increased astrocyte content. Supporting this hypothesis, MRS of the in vitro cultures of primary astrocytes reveal Cr concentration twice as high as those observed in neurons.32 Because of this increased concentration of Cr, the increased NAA concentration is not detected when the NAA/Cr ratio is utilized. This is probably why the increased NAA was not detected in previous studies using relative values.

The MI appears to be almost exclusively located in astrocytes, where it is now recognized as the most important osmolyte, or cell volume regulator.35 The increase of MI has been reported in MS and Krabbe disease.36-38⇓⇓ This increase in MS was more pronounced in inactive lesions with fibrillary astrocytic gliosis than in active demyelinating lesions that revealed protoplastic gliosis.36 Another study related MI increase to the acute stage of active myelin breakdown.37 Therefore, it is speculated that elevated MI in MS is derived from both the accumulation of myelin breakdown products during acute phases and astrocytic gliosis in chronic lesions.36 The former mechanism is not applicable to PMD, because it lacks in active demyelination.1,2⇓ The elevated MI in PMD, therefore, may result from the astrocytic gliosis, the same mechanism for elevated Cr. However, a cautious interpretation is suggested given the greater difficulty in quantifying MI due to side effect of water resonance, and confirmation in further studies is needed before firm conclusions are made.

Because the in vivo Cho peaks likely contain various cell membrane precursor and breakdown products, including phosphocholine, glycerophosphocholine, and phosphatidylcholine, the Cho peaks may partially represent myelin metabolism; therefore, the Cho/Cr ratio rises in the period of accelerated myelination within the first few years of life, as well as in the process of myelin disruption observed in various demyelinating disorders.9,11⇓ The elevated concentration of Cho has been utilized as a marker for demyelination in patients with white matter lesions on MR imaging.15 Biochemical studies of PMD have not identified significant alterations of the neutral phospholipid contents, including phosphatidylcholine.39 In addition, rat primary cultures of astrocytes and oligodendrocyte progenitor cells show only a little lower Cho concentrations than mature oligodengrocytes.32 Therefore, changes in the population of either cells may not affect the concentration of Cho in PMD.

Previous study of relative proton MRS with long TE (272 millliseconds) showed decreased NAA/Cr and normal Cho/Cr ratios in nine patients with PMD with PLP1 duplications.14 The former may result from larger increase of Cr than of NAA. However, the latter cannot be explained by the results in this study. We speculate that the discrepancy may be due to technical differences, such as echo time. The TE = 272 milliseconds spectrum is more T2-weighted than the TE = 30 milliseconds. It results in fewer and smaller peaks of metabolites, and also in different peak ratios, depending on the differing T2 relationships of the metabolites.40