Original Paper Received: October 14, 2014 Accepted: December 14, 2014 Published online: March 25, 2015
Eur Neurol 2015;73:247–256 DOI: 10.1159/000371575
Diffuse Brain Abnormalities in Myotonic Dystrophy Type 1 Detected by 3.0 T Proton Magnetic Resonance Spectroscopy Yuhei Takado a, b Kenshi Terajima a, b, e Masaki Ohkubo a, d Kouichirou Okamoto a, c Takayoshi Shimohata b Masatoyo Nishizawa a, b Hironaka Igarashi a Tsutomu Nakada a, f
a Center for Integrated Human Brain Science and Departments of b Neurology and c Neurosurgery, Brain Research Institute, and d Graduate School of Health Sciences, University of Niigata, and e Department of Medical Informatics, University of Niigata Medical and Dental Hospital, Niigata, Japan; f Department of Neurology, University of California, Davis, Calif., USA
Abstract Patients with myotonic dystrophy type 1 (DM1) (n = 14) were compared with healthy controls (n = 13) using 3.0 T proton magnetic resonance spectroscopy (1H-MRS) to investigate brain pathophysiology. 1H-MRS imaging revealed reduced N-acetylaspartate to creatine ratio (NAA/Cr) in multiple brain regions (average 24%), suggesting diffuse brain abnormalities among patients with DM1. Single-voxel 1H-MRS among patients with DM1 showed (1) reduced NAA in both the frontal cortex (23%) and frontal white matter (31%) and unaltered myo-inositol, suggesting neuronal abnormalities without significant gliosis; and (2) elevated glutamine in the frontal cortex (36%) and reduced glutamate in the frontal white matter (20%) among patients with DM1, suggesting abnormalities in the glutamatergic system in the brain of patients with DM1. We consider that these results reflect brain abnormalities that cannot be detected by neuropathological assessment in patients with DM1. © 2015 S. Karger AG, Basel
© 2015 S. Karger AG, Basel 0014–3022/15/0734–0247$39.50/0 E-Mail [email protected] www.karger.com/ene
Introduction
Myotonic dystrophy type 1 (DM1), which is caused by an expanded cytosine-thymine-guanine (CTG) repeat [1], is a multisystem disease characterized by progressive muscular weakness and extramuscular manifestations, which include central nervous system (CNS) impairment [2]. There are several magnetic resonance imaging (MRI) [3, 4] and diffusion tensor imaging (DTI) [5, 6] studies in vivo, showing diffuse brain damage among patients with DM1. Moreover, the involvement of cortical and subcortical regions among patients with DM1 is in agreement with pathological studies documenting neuronal loss and intracellular neurofibrillary tangles in the frontal and temporal lobes, thalamus, basal ganglia, and brainstem [7–10]. However, despite the numerous findings from neuroimaging and neuropathological studies, the pathophysiology of CNS impairment in DM1 remains unclear. Proton magnetic resonance spectroscopy (1H-MRS) is a powerful noninvasive method for analyzing brain metabolites in patients [11]. Even though 3.0 T 1H-MRS has a lower resolution than optical microscopy, it offers some advantages over the pathological assessments at autopsy. Hironaka Igarashi MD, PhD Center for Integrated Human Brain Sciences, Brain Research Institute University of Niigata, 1–757 Asahimachi Niigata 951–8585 (Japan) E-Mail: higara @ bri.niigata-u.ac.jp
Downloaded by: University of Tokyo 157.82.153.40 - 5/15/2015 9:59:37 AM
Key Words Myotonic dystrophy · MRS · Cerebral metabolism · Glutamate · Frontal lobe
Methods Participants and Design Fourteen patients with DM1 (male/female, 4/10; age, 43.6 ± 12.6) were compared with 13 healthy age-matched controls (male/ female, 4/9; age, 42.6 ± 14.2) (table 1). Written informed consent was obtained from all participants in accordance with the human research guidelines of the Internal Review Board of the University of Niigata. None of the patients or control subjects had a past medical history of other neuromuscular or CNS disorders. All patients with DM1 had undergone DNA analysis for CTG repeats [1]. Single-voxel 1H-MRS (FWM and ACG) and single-slice MRSI in two slices (the basal ganglia level or the upper lateral ventricles level) were performed in patients with DM1 and control subjects. Two DM1 patients could not tolerate the study and were unable to complete the entire protocol. One patient accomplished only single-voxel 1H-MRS, whereas the other accomplished only MRSI. The concentrations of metabolites were measured in the FWM,
248
Eur Neurol 2015;73:247–256 DOI: 10.1159/000371575
Table 1. Profile of patients with DM1 and age matched control
subjects
DM1 (n = 14)
Control subjects (n = 13)
M/F Age Disease duration, y CTG repeat FAB MMSE
4/10 43.6 (12.6) 11.8 (9.0) 685 (462) 14.0 (3.3) 27.2 (3.2)
4/9 42.6 (14.2) – – – –
Values are expressed as mean (SD). M = Male; F = female; y = year; CTG = cytosine-thymine-guanine; FAB = frontal assessment battery; MMSE = mini-mental state examination.
corona radiata, parietal white matter (PWM), putamen, thalamus, insular cortex, and posterior limb of the internal capsule (PLIC) using MRSI, and in the FWM and ACG using 1H-MRS to enable more accurate comparisons between gray and white matter in the frontal lobe. Metabolite levels determined by 1H-MRS were then analyzed for possible correlations with clinical parameters, such as the frontal assessment battery (FAB) [27] and Mini-Mental State Examination (MMSE) [28]. MRI Protocol MRI/1H-MRS/MRSI were performed using a 3.0 T system (Signa LX; General Electric, Waukesha, Wisc., USA). A standard head coil with quadrature detection was used for signal transmission. Using axial fast spin-echo (FSE) images that were parallel to the plane aligned to the anterior commissure and the posterior commissure [repetition time (TR), 6,000 ms; effective echo time (TE), 24 ms; 18–20 slices; thickness, 5.0 mm; gap, 2.5 mm; field of view (FOV), 20 cm; matrix, 512 × 256], MRSI data using the point-resolved spectroscopy (PRESS) sequence on two 10-mm thick axial slices passing through the basal ganglia or the upper lateral ventricles were acquired (fig. 1a, b). FSE images were also used to confirm that patients with DM1 and control subjects did not have cerebrovascular damage in the brain. The acquisition parameters were an FOV of 18 × 18 cm2 or 20 × 20 cm2, 24 phase-encoding steps in each direction, a TE of 144 ms, and a TR of 1,500 ms. For single-voxel 1H-MRS, two volumes of interest (VOIs) were localized in the frontal cortex and right FWM (fig. 1c, d). The size of both VOIs was 16 × 16 × 16 mm3. A PRESS sequence with chemical-shift-selective water suppression was used with the following parameters: TR, 1,500 ms; TE, 30 ms; number of acquisitions, 128; data points, 2 K; and spectral width, 5,000 Hz. The VOIs were followed by an identical acquisition (only 8 acquisitions) with water suppression turned off. MRS Post-Processing Post-processing for MRSI was performed using SAGE software (Spectroscopy Analysis GE; General Electric Medical Systems) that included spatial apodization (Fermi diameter = 80%, Fermi transition = 10%), followed by a Fourier transformation to spatial and frequency dimensions. Spectral analysis of MRSI and 1H-MRS
Takado/Terajima/Ohkubo/Okamoto/ Shimohata/Nishizawa/Igarashi/Nakada
Downloaded by: University of Tokyo 157.82.153.40 - 5/15/2015 9:59:37 AM
First, the ability to take noninvasive measurements in patients enables a nearly real-time correlation of clinical symptoms with brain metabolites, which can provide information about neuropathological findings, such as gliosis and neuronal loss [12, 13]. Second, because 1H-MRS can cover almost entire regions in a targeted brain slice simultaneously, this can be advantageous compared with microscopic investigation of a small part of specimens for detecting pathological changes in the broad region in the brain. Thus, 1H-MRS can be a valuable means for various brain diseases, including neurodegenerative diseases [14], reversible conditions associated with ischemia [15, 16], and toxic insults [17], as well as functional abnormalities in psychiatric disorders such as schizophrenia [18]. Using 1H-MRS, changes in the brain metabolite profile of patients with DM1 were observed in previous studies [19–22]; however, these results have been controversial, given the previous studies that suggested the involvement of synaptic dysfunction among patients with DM1 [23– 25]. The controversies in previous 1H-MRS studies might be attributed to the limited examined metabolites, as well as to the limited spatial information achieved only by single-voxel 1H-MRS. Thus, in this study, we aimed at investigating the pathophysiology in brains of patients with DM1 by using 1H-MRS and 1H-MRS imaging (MRSI) at 3.0 T. Using 1H-MRS, we studied metabolites in the anterior cingulate gyrus (ACG), which is known to be involved in cognitive and emotional processing [26], as well as frontal white matter (FWM) to investigate gray and white matter differences. Using MRSI, we studied multiple brain regions to investigate spatial information related to brain abnormalities among patients with DM1.
Color version available online
a
b
Fig. 1. Typical voxels of interest for MRSI c
d NAA
e
4.0
3.0
2.0 ppm
f 4.0
3.0
2.0 ppm
was performed by using LCModel (Stephen Provencher Inc., Oakville, Ontario, Canada), which is a user-independent, frequency-domain spectral-fitting program [29, 30]. Sixteen metabolites were included in this LCModel basis set: alanine, aspartate, creatine (Cr), γ-aminobutyric acid, glucose, glutamine (Gln), glutamate (Glu), glycerophosphocholine (GPC), phosphocholine (PC), lactate, myo-inositol (MI), N-acetylaspartate (NAA), N-acetylaspartylglutamate (NAAG), scyllo-inositol, taurine, and guanine. Only results with Cramér-Rao lower bounds (CRLB)
Eur Neurol 2015;73:247–256 DOI: 10.1159/000371575
Diffuse Brain Abnormalities in Myotonic Dystrophy Type 1 Detected by 3.0 T Proton Magnetic Resonance Spectroscopy Yuhei Takado a, b Kenshi Terajima a, b, e Masaki Ohkubo a, d Kouichirou Okamoto a, c Takayoshi Shimohata b Masatoyo Nishizawa a, b Hironaka Igarashi a Tsutomu Nakada a, f
a Center for Integrated Human Brain Science and Departments of b Neurology and c Neurosurgery, Brain Research Institute, and d Graduate School of Health Sciences, University of Niigata, and e Department of Medical Informatics, University of Niigata Medical and Dental Hospital, Niigata, Japan; f Department of Neurology, University of California, Davis, Calif., USA
Abstract Patients with myotonic dystrophy type 1 (DM1) (n = 14) were compared with healthy controls (n = 13) using 3.0 T proton magnetic resonance spectroscopy (1H-MRS) to investigate brain pathophysiology. 1H-MRS imaging revealed reduced N-acetylaspartate to creatine ratio (NAA/Cr) in multiple brain regions (average 24%), suggesting diffuse brain abnormalities among patients with DM1. Single-voxel 1H-MRS among patients with DM1 showed (1) reduced NAA in both the frontal cortex (23%) and frontal white matter (31%) and unaltered myo-inositol, suggesting neuronal abnormalities without significant gliosis; and (2) elevated glutamine in the frontal cortex (36%) and reduced glutamate in the frontal white matter (20%) among patients with DM1, suggesting abnormalities in the glutamatergic system in the brain of patients with DM1. We consider that these results reflect brain abnormalities that cannot be detected by neuropathological assessment in patients with DM1. © 2015 S. Karger AG, Basel
© 2015 S. Karger AG, Basel 0014–3022/15/0734–0247$39.50/0 E-Mail [email protected] www.karger.com/ene
Introduction
Myotonic dystrophy type 1 (DM1), which is caused by an expanded cytosine-thymine-guanine (CTG) repeat [1], is a multisystem disease characterized by progressive muscular weakness and extramuscular manifestations, which include central nervous system (CNS) impairment [2]. There are several magnetic resonance imaging (MRI) [3, 4] and diffusion tensor imaging (DTI) [5, 6] studies in vivo, showing diffuse brain damage among patients with DM1. Moreover, the involvement of cortical and subcortical regions among patients with DM1 is in agreement with pathological studies documenting neuronal loss and intracellular neurofibrillary tangles in the frontal and temporal lobes, thalamus, basal ganglia, and brainstem [7–10]. However, despite the numerous findings from neuroimaging and neuropathological studies, the pathophysiology of CNS impairment in DM1 remains unclear. Proton magnetic resonance spectroscopy (1H-MRS) is a powerful noninvasive method for analyzing brain metabolites in patients [11]. Even though 3.0 T 1H-MRS has a lower resolution than optical microscopy, it offers some advantages over the pathological assessments at autopsy. Hironaka Igarashi MD, PhD Center for Integrated Human Brain Sciences, Brain Research Institute University of Niigata, 1–757 Asahimachi Niigata 951–8585 (Japan) E-Mail: higara @ bri.niigata-u.ac.jp
Downloaded by: University of Tokyo 157.82.153.40 - 5/15/2015 9:59:37 AM
Key Words Myotonic dystrophy · MRS · Cerebral metabolism · Glutamate · Frontal lobe
Methods Participants and Design Fourteen patients with DM1 (male/female, 4/10; age, 43.6 ± 12.6) were compared with 13 healthy age-matched controls (male/ female, 4/9; age, 42.6 ± 14.2) (table 1). Written informed consent was obtained from all participants in accordance with the human research guidelines of the Internal Review Board of the University of Niigata. None of the patients or control subjects had a past medical history of other neuromuscular or CNS disorders. All patients with DM1 had undergone DNA analysis for CTG repeats [1]. Single-voxel 1H-MRS (FWM and ACG) and single-slice MRSI in two slices (the basal ganglia level or the upper lateral ventricles level) were performed in patients with DM1 and control subjects. Two DM1 patients could not tolerate the study and were unable to complete the entire protocol. One patient accomplished only single-voxel 1H-MRS, whereas the other accomplished only MRSI. The concentrations of metabolites were measured in the FWM,
248
Eur Neurol 2015;73:247–256 DOI: 10.1159/000371575
Table 1. Profile of patients with DM1 and age matched control
subjects
DM1 (n = 14)
Control subjects (n = 13)
M/F Age Disease duration, y CTG repeat FAB MMSE
4/10 43.6 (12.6) 11.8 (9.0) 685 (462) 14.0 (3.3) 27.2 (3.2)
4/9 42.6 (14.2) – – – –
Values are expressed as mean (SD). M = Male; F = female; y = year; CTG = cytosine-thymine-guanine; FAB = frontal assessment battery; MMSE = mini-mental state examination.
corona radiata, parietal white matter (PWM), putamen, thalamus, insular cortex, and posterior limb of the internal capsule (PLIC) using MRSI, and in the FWM and ACG using 1H-MRS to enable more accurate comparisons between gray and white matter in the frontal lobe. Metabolite levels determined by 1H-MRS were then analyzed for possible correlations with clinical parameters, such as the frontal assessment battery (FAB) [27] and Mini-Mental State Examination (MMSE) [28]. MRI Protocol MRI/1H-MRS/MRSI were performed using a 3.0 T system (Signa LX; General Electric, Waukesha, Wisc., USA). A standard head coil with quadrature detection was used for signal transmission. Using axial fast spin-echo (FSE) images that were parallel to the plane aligned to the anterior commissure and the posterior commissure [repetition time (TR), 6,000 ms; effective echo time (TE), 24 ms; 18–20 slices; thickness, 5.0 mm; gap, 2.5 mm; field of view (FOV), 20 cm; matrix, 512 × 256], MRSI data using the point-resolved spectroscopy (PRESS) sequence on two 10-mm thick axial slices passing through the basal ganglia or the upper lateral ventricles were acquired (fig. 1a, b). FSE images were also used to confirm that patients with DM1 and control subjects did not have cerebrovascular damage in the brain. The acquisition parameters were an FOV of 18 × 18 cm2 or 20 × 20 cm2, 24 phase-encoding steps in each direction, a TE of 144 ms, and a TR of 1,500 ms. For single-voxel 1H-MRS, two volumes of interest (VOIs) were localized in the frontal cortex and right FWM (fig. 1c, d). The size of both VOIs was 16 × 16 × 16 mm3. A PRESS sequence with chemical-shift-selective water suppression was used with the following parameters: TR, 1,500 ms; TE, 30 ms; number of acquisitions, 128; data points, 2 K; and spectral width, 5,000 Hz. The VOIs were followed by an identical acquisition (only 8 acquisitions) with water suppression turned off. MRS Post-Processing Post-processing for MRSI was performed using SAGE software (Spectroscopy Analysis GE; General Electric Medical Systems) that included spatial apodization (Fermi diameter = 80%, Fermi transition = 10%), followed by a Fourier transformation to spatial and frequency dimensions. Spectral analysis of MRSI and 1H-MRS
Takado/Terajima/Ohkubo/Okamoto/ Shimohata/Nishizawa/Igarashi/Nakada
Downloaded by: University of Tokyo 157.82.153.40 - 5/15/2015 9:59:37 AM
First, the ability to take noninvasive measurements in patients enables a nearly real-time correlation of clinical symptoms with brain metabolites, which can provide information about neuropathological findings, such as gliosis and neuronal loss [12, 13]. Second, because 1H-MRS can cover almost entire regions in a targeted brain slice simultaneously, this can be advantageous compared with microscopic investigation of a small part of specimens for detecting pathological changes in the broad region in the brain. Thus, 1H-MRS can be a valuable means for various brain diseases, including neurodegenerative diseases [14], reversible conditions associated with ischemia [15, 16], and toxic insults [17], as well as functional abnormalities in psychiatric disorders such as schizophrenia [18]. Using 1H-MRS, changes in the brain metabolite profile of patients with DM1 were observed in previous studies [19–22]; however, these results have been controversial, given the previous studies that suggested the involvement of synaptic dysfunction among patients with DM1 [23– 25]. The controversies in previous 1H-MRS studies might be attributed to the limited examined metabolites, as well as to the limited spatial information achieved only by single-voxel 1H-MRS. Thus, in this study, we aimed at investigating the pathophysiology in brains of patients with DM1 by using 1H-MRS and 1H-MRS imaging (MRSI) at 3.0 T. Using 1H-MRS, we studied metabolites in the anterior cingulate gyrus (ACG), which is known to be involved in cognitive and emotional processing [26], as well as frontal white matter (FWM) to investigate gray and white matter differences. Using MRSI, we studied multiple brain regions to investigate spatial information related to brain abnormalities among patients with DM1.
Color version available online
a
b
Fig. 1. Typical voxels of interest for MRSI c
d NAA
e
4.0
3.0
2.0 ppm
f 4.0
3.0
2.0 ppm
was performed by using LCModel (Stephen Provencher Inc., Oakville, Ontario, Canada), which is a user-independent, frequency-domain spectral-fitting program [29, 30]. Sixteen metabolites were included in this LCModel basis set: alanine, aspartate, creatine (Cr), γ-aminobutyric acid, glucose, glutamine (Gln), glutamate (Glu), glycerophosphocholine (GPC), phosphocholine (PC), lactate, myo-inositol (MI), N-acetylaspartate (NAA), N-acetylaspartylglutamate (NAAG), scyllo-inositol, taurine, and guanine. Only results with Cramér-Rao lower bounds (CRLB)
