Quantification of N-Acetyl Aspartyl Glutamate in Human Brain using Proton Magnetic Resonance Spectroscopy at 7 T

Abstract

UDC 543.422.25:611.81 The separation of N-acetyl aspartyl glutamate (NAAG) from N-acetyl aspartate (NAA) and other metabolites, such as glutamate, by in vivo proton magnetic resonance spectroscopy at 7 T is described. This method is based on the stimulated echo acquisition mode (STEAM), with short and long echo time (TE) and allows quantitative measurements of NAAG in the parietal and pregenual anterior cingulate cortex (pgACC) of human brain. Two base- sets for the LCModel have been established using nuclear magnetic resonance simulator software (NMR-SIM). Six healthy volunteers (age 25-35 years) have been examined at 7 T. It has been established that NAAG can be separated and quantifi ed in the parietal location and does not get quantifi ed in the pgACC location when using a short echo time, TE = 20 ms. On the other hand, by using a long echo time, TE = 74 ms, NAAG can be quantifi ed in pgACC structures. LCModel. Introduction. N-Acetylaspartylglutamic acid (NAAG) is a neuropeptide that is the third most prevalent neurotransmitter in the human brain (1). NAAG is produced by the coupling between a glutamic acid (Glu) and the N-acetylaspartic acid (NAA) via a peptide bond. Therefore, it may be diffi cult to distinguish between them on account of the similarity of the chemical structure. In human studies, it has been suggested that NAAG levels are abnormal in patients with schizophrenia (2) and decreased in amyotrophic lateral sclerosis (ALS) (3). In a previous study (4), it had been diffi cult to distinguish NAAG from NAA and Glu using proton magnetic resonance spectroscopy (MRS) at 3 T. As the dominant signal of the NAAG spectrum from the N-acetyl proton resonance peak is located at 2.05 ppm, which is 0.03 ppm away from the corresponding NAA signal at 2.02 ppm (5), a high resolution of metabolite peaks can be obtained using ultrahigh-fi eld strengths such as 7 T (6, 7) or at high fi elds (2 or 3 T), with unusual shimming (8). Proton magnetic resonance spectroscopy provides markers for various human brain diseases (9, 10). Some of the important measurement characteristics of the MRS technique are the signal-to-noise ratio (SNR), spatial and spectral resolution, as well as scan time; thus, several localization methods have been proposed. Single volume spectroscopy (SVS) localization techniques, with pulse sequences based on spin echoes (PRESS) (11) or stimulated echoes (STEAM) (12), provide the highest spectral quality (13), as the lower magnetic fi eld strengths result in a peak overlap of the human brain metabolites, such as, NAA with NAAG. Besides, standard spectroscopic sequences such as STEAM or PRESS will suffer if limited transmitter voltage takes place. This results in a reduced excitation flip angle and consequently insuffi cient excitation in the volume of interest (VOI). Moreover, previous studies have shown that the increase in magnetic field strength increases the 1 H NMR spectral peak separation and SNR and thus provides higher sensitivity when compared with lower field strengths (14, 15). With standard radio-frequency (RF) amplifiers and volume RF coils at 7 T, short 180° refocusing pulses are diffi cult to achieve in most parts of the brain. Therefore, earlier, high field MRS studies frequently employed surface coils with locally higher RF fi eld amplitude (16, 17). However, not all brain areas can be adequately examined using surface coils. As the goal of this study is to quantify the NAAG concentrations in different brain structures, STEAM with short echo time benefits, from reduced signal decay, because of the J-evolution of the coupled spin system, allows a more precise quantification of the metabolites, especially with a short T2 relaxation time (18). Thus, short echo times have been used in several studies (19). In the current study, echo time (TE) was selected to be short and long 20 and 74 ms, respectively. Although earlier studies had employed STEAM with ultrashort echo times (6 ms), they were limited by the surface coils that prevented detection in the deep brain regions (20). Also, a quadrature (transmit/receive) surface RF coil or a half-volume RF coil have been used for the