点击显示 收起
1 From the Department of Psychiatry, University of Vienna (BS, PA, and GMS-Z); Abbott Spa, Liscate, Italy (CDP); and Cross Research, Arzo, Switzerland (AA).
2 Presented at the symposium S-Adenosylmethionine (SAMe): from Molecular Mechanism to Clinical Implications, held in Santa Barbara, CA, March 7–10, 2001. 3 Address reprint requests to B Saletu, Section of Sleep Research and Pharmacopsychiatry, Department of Psychiatry, University of Vienna, Währingergürtel 18-20, A-1090 Vienna, Austria. E-mail: bernd.saletu{at}akh-wien.ac.at.
ABSTRACT
Background: S-Adenosyl-L-methionine (SAMe, or ademetionine) is a naturally occurring molecule used as both a nutraceutical and a pharmaceutical to treat depression.
Objective: The central mode of action of SAMe was investigated in 20 healthy volunteers by mapping of electroencephalograms (EEGs) and event-related potentials (ERPs) and low-resolution brain electromagnetic tomography (LORETA).
Design: In an acute and subacute, double-blind, placebo-controlled, crossover study, subjects received in random order infusions of 800 mg SAMe and placebo for 7 d, with a washout period of 3 wk between the 2 treatment periods. EEG recordings were made 0, 1, 3, and 6 h after and ERP recordings were made 0 and 1 h after the drug infusions on days 1 and 7.
Results: Multivariate analyses of variance and Hotelling T2 tests showed significant acute and subacute encephalotropic effects of SAMe compared with placebo. Acute pharmaco-EEG changes were typical of classic antidepressants of the thymoleptic type; subacute alterations were typical of cognition enhancers. Regarding ERPs, standard N1 and P2 latencies were shortened, and target P300 latencies were lengthened. N1 amplitudes increased after subacute treatment, and temporooccipital P300 amplitudes increased after the acute dose. Similar changes were described for antidepressants. LORETA showed that the N2 source strength increased in both the left and the right temporal lobes, whereas the P300 source strength increased in the dorsolateral prefrontal regions and decreased in the ventral limbic regions.
Conclusion: EEG-ERP mapping identified SAMe as an antidepressant. LORETA targeted brain regions crucial in the therapeutic efficacy of antidepressants.
Key Words: Electrophysiological neuroimaging • S-adenosyl-L-methionine • SAMe • ademetionine • electroencephalogram mapping • event-related-potentials mapping • event-related-potentials tomography • low-resolution brain electromagnetic tomography • LORETA • antidepressants
INTRODUCTION
S-Adenosyl-L-methionine (SAMe), or ademetionine, is a naturally occurring molecule that is distributed throughout virtually all human body tissues and fluids; concentrations are highest in childhood and decrease with age. SAMe serves as a crucial intermediate of 3 major pathways in all biological systems: methylation, transsulfuration, and aminopropylation. These pathways are known to be involved in the synthesis of nucleic acids, proteins, phospholipids, hormones, neurotransmitters, antioxidants, polyamines, catecholamines, and other biogenic amines (1, 2).
SAMe has been successfully used for the treatment of liver disease (3), osteoarthritis, and depressive disorders and was found to be superior to placebo and comparable with other tricyclic antidepressants (4–6). In contrast with the multitude of clinical therapeutic trials that have been conducted, only few pharmacodynamic studies have confirmed the central mode of action of SAMe (7; JP Macher, unpublished observations, 1995). Thus, the aim of the present study was to evaluate the pharmacodynamic effects of SAMe on spontaneous brain function and cognitive event-related potentials (ERPs) in 20 healthy subjects by mapping of electroencephalograms (EEGs) and ERPs (EEG-ERP mapping) and low-resolution brain electromagnetic tomography (LORETA) (8–10). These methods are used to determine whether, how, when, and at which dosage a drug acts (pharmaco-EEG). With the development of multilead and topographic analyzing methods in the past 2 decades, EEG mapping has become a readily and widely available, low-cost, noninvasive, objective, and quantitative high-time resolution method for visualizing the neurophysiology of mental disorders and their treatments (11–20). LORETA combines the high time resolution of the EEG with a source localization method that permits a truly 3-dimensional tomography of brain electrical activity. The solution space is restricted to cortical gray matter and the hippocampus, as determined in the digitized probability atlas (Brain Imaging Center, Neurologic Institute, Montreal) based on the human brain atlas of Talairach and Tournoux (10). Thus, LORETA makes it possible to identify brain target regions of mental disorders, psychotropic drugs, and other therapeutic interventions.
SUBJECTS AND METHODS
Subjects
Ten young healthy subjects (5 men, 5 women) aged 22–33 y (
In random order, the subjects received intravenous infusions of 800 mg SAMe in 250 mL isotonic solution or placebo over 30 min for 7 d, with a washout period of 3 wk between the 2 treatment periods. EEG recordings were carried out 0, 1, 3, and 6 h after the infusions, and ERP recordings were made 0 and 1 h after the infusions on days 1 (acute effect) and 7 (subacute and superimposed effects).
A 3-min vigilance-controlled EEG (V-EEG) and a 4-min resting EEG (R-EEG) were recorded from 19 leads (Fp1, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, 01, and 02 to averaged mastoids). During the V-EEG recordings the technician tried to keep the subjects alert; as soon as drowsiness patterns occurred in the recording, the subjects were aroused. During the R-EEG recordings there was no interference regarding spontaneous fluctuations of vigilance. Spectral analysis was performed for 5-s epochs (512 sample points), resulting in a frequency resolution of 0.2 Hz with the use of the fast Fourier transform technique in floating-point arithmetic to maintain precision.
Artifact-free 5-s epochs were selected after minimizing ocular artifacts by means of an automatic artifact identification method as described by Anderer et al (21). The mean spectral curves contained data from 1.3 to 35 Hz quantified into 36 EEG variables. The latter included total power, absolute power, and relative power in 12 different frequency bands [ (1.3–3.5 Hz) activity (D), (3.5–7.5 Hz) activity (T), -1 (7.5–10.5 Hz) activity (A1), -2 (10.5–13 Hz) activity (A2), ß-1 (13–16 Hz) activity (B1), ß-2 (16–20 Hz) activity (B2), ß-3 (20–25 Hz) activity (B3), ß-4 (25–30 Hz) activity (B4), ß-5 (30–35 Hz) activity (B5), combined D and T activity (DT), activity (A), and ß activity (B)]; the dominant frequency (DF) of the A activity in Hz as well as in absolute power and relative power; and the center-of-gravity frequencies (centroids; C) and the SDs (CD) of the DT, A, and B activities as well as of the total activity (CT).
For EEG mapping, 19 single values obtained from the 10–20-electrode set were mapped onto a 64 x 64 numerical matrix. Each interpolated value was based on the cubic distance from the values at the 4 nearest electrodes.
ERPs were recorded in an auditory 2-tone odd-ball paradigm without motor reaction [standard tone: loud tone burst (90 dB sound pressure level); target tone: soft tone bursts (70 dB sound pressure level, target probability 0.1)]. The subjects were asked to mentally count the soft tone bursts and report their number at the end of the experiment. Off-line data processing included ocular artifact minimization and automatic artifact rejection. Averaged ERP responses for targets and standards were digitally filtered at 20 Hz. Standard tone N1 and P2 and target tone N2 and P300 components were determined. LORETA with smoothness and neuroanatomic constraint was applied to these components, resulting in current density values in 2394 voxels with a spatial resolution of 7 mm (9, 10).
Statistics
The inferential strategy of descriptive data analysis (22), as proposed for application to the mapping situation (23, 24), was applied. Multivariate statistics included a multivariate analysis of variance and a Hotelling T2 test for the absolute power values of all 9 EEG frequency bands. The differences between SAMe- and placebo-induced changes were further evaluated with univariate, paired-samples t tests computed for EEG measures, ERP latencies and amplitudes, and log-transformed LORETA values (P < 0.05). A single null hypothesis was tested for omnibus significance by means of a binomial test at P < 0.05.
RESULTS
EEG mapping based on multivariate analyses
Maps based on Hotelling T2 tests obtained from multivariate analysis of variance on group (SAMe or placebo) by time (before or after) by variates (log-transformed power in 9 frequency bands) showed significant differences between SAMe and placebo infusions, more so in the resting than in the vigilance-controlled state (Figure 1). SAMe-induced significant effects in the first and third hours after acute drug administration and after subacute treatment (hour 0, day 7; before the daily infusion); the most pronounced effect was observed after one superimposed dose in the first hour of day 7. As on day 1, a rapid subsequent decline in the effect on the central nervous system (CNS) occurred until the sixth hour.
FIGURE 1. . Brain maps showing differences in vigilance-controlled and resting electroencephalogram (EEG) changes induced by 800 mg S-adenosyl-L-methionine (SAMe) and placebo infusions in healthy subjects (n = 20) 1, 3, and 6 h after infusion on day 1 and 0, 1, 3, and 6 h after infusion on day 7, respectively. Maps are based on Hotelling T2 tests obtained from multivariate ANOVA on absolute power changes in all 9 EEG frequency bands. Black, dark, and middle gray shadings represent significant T2 values: > 4.63 (P < 0.01), > 2.90 (P < 0.05), and > 2.27 (P < 0.10). SAMe induced more significant changes in the resting than in the vigilance-controlled EEG, predominantly in the first hour after infusion.
EEG mapping based on univariate analyses
The acute effects of 800 mg SAMe were characterized by an increase in absolute and relative D and T power as well as a decrease in A, B3, and B4 powers (Figure 2). The centroid of the DT activity slowed down, as did that of the total power spectrum. The dominant frequency and the absolute and relative powers of the dominant frequency decreased. These CNS effects were most pronounced in the first hour, declined thereafter, but remained partly significant in the sixth hour.
FIGURE 2. . Pharmaco-electroencephalogram (EEG) maps of differences in resting EEG (R-EEG) measures between changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in healthy subjects (n = 20) 1 h after the first infusion. Statistical probability maps depicting changes in 13 absolute (ABS) power measures (top), 12 relative (REL) power measures (middle), and 11 dominant frequency and centroid (C) measures (bottom) are shown (bird’s eye view, nose at the top; white dots indicate electrode positions). Compared with placebo, orange, red, and purple colors represent significant increases (P < 0.10, 0.05, and 0.01, respectively); light blue, middle blue, and dark blue colors indicate significant decreases (P < 0.10, 0.05, and 0.01, respectively). SAMe induced changes typical of imipramine-type antidepressants. See Methods for definition of abbreviations.
The subacute effects after 1 wk of daily infusions of 800 mg SAMe showed a different pattern: the total and absolute powers in all frequency bands increased (Figure 3). Concerning relative power, A activity, specifically A2, declined. The centroid of the DT and A activities slowed down.
FIGURE 3. . Pharmaco-electroencephalogram (EEG) maps of differences in resting EEG measures between changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in healthy subjects (n = 20) after 1 wk of daily infusions. For a technical description of the maps, see Figure 2. Subacute treatment with SAMe resulted in an increase in total and absolute (ABS) power in all frequency bands (except A2), a decrease in relative T and A2 power, and a deceleration of the centroids of the DT and A bands. See Methods for definition of abbreviations.
One superimposed dose in addition to subacute treatment induced the same changes, with the DT augmentation being most pronounced in the first hour, similar to the acute effects (Figure 4). In addition, the centroids of the and ß activities slowed down. These effects declined until the sixth hour, with an intermittent increase in ß power and intermittent decreases in and activities in the third hour.
FIGURE 4. . Pharmaco-electroencephalogram (EEG) maps of differences in resting EEG measures between changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in healthy subjects (n = 20) in the first hour after one superimposed infusion in addition to 7 d of treatment. For a technical description of the maps, see Figure 2. A superimposed dose of SAMe resulted in an increase in total and absolute (ABS) power in all frequency bands, an increase in relative D, a decrease in relative A2 power, and a slowing of the A and B centroids. See Methods for definition of abbreviations
ERP changes
The acute effects of 800 mg SAMe were characterized by a lengthening of P300 latency, which measures cognitive information processing speed (Figure 5). A significant augmentation of the P300 amplitude was observed in the right occipital and parietotemporal regions.
FIGURE 5. . Differences in changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in event-related potential latencies (top) and amplitudes (bottom, as statistical probability maps) in healthy subjects (n = 19) 1 h after the first infusion. The gray scale represents t values. SAMe lengthened target P300 latency and augmented the P300 amplitude occipitally.
Subacute treatment with 1 wk of daily infusions of 800 mg SAMe induced, as compared with placebo, a trend toward shortening of N1 latency, reflecting faster perception of the auditory stimuli (Figure 6). In addition, an increase in the N1 amplitude right occipitally, suggesting an increased allocation of neuronal sources for perception, was observed. One superimposed infusion in addition to subacute treatment induced, as compared with placebo, a shortening of N1 latency, a lengthening of P300 latency, and a decrease in the N2 amplitude (Figure 7).
FIGURE 6. . Differences in changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in event-related potential latencies (top) and amplitudes (bottom, as statistical probability maps) in healthy subjects (n = 19) after 1 wk of daily infusions. Subacute treatment with SAMe resulted in a trend toward a shortening of standard N1 latency and an increase in the N1 amplitude right occipitally.
FIGURE 7. . Differences in changes induced by 800 mg S-adenosyl-L-methionine (SAMe, or ademetionine) and those induced by placebo in event-related potential latencies (top) and amplitudes (bottom, as statistical probability maps) in healthy subjects (n = 19) in the first hour after one superimposed infusion in addition to 7 d of treatment. A superimposed dose of SAMe resulted in a shortening of N1 latency, a lengthening of P300 latency, and an attenuation of the N2 amplitude parietally.
Identifying target regions of SAMe by ERP-LORETA
Using LORETA, we evaluated the effects of SAMe on intracerebral current densities (Table 1).
View this table:
TABLE 1 . Comparison of the effects of S-adenosyl-L-methionine (SAMe) and placebo on the source strength of event-related potentials (ERPs) determined by low-resolution brain electromagnetic tomography1
Acute treatment with 800 mg SAMe induced an increase in N2 and P300 source strength 1 h after drug administration (Figure 8). N2 source strength increased bilaterally in the temporal lobes, ie, in brain regions where the auditory-specific endogenous depth N2b had been recorded (25). The increase was observed in a suprathreshold region involving 538 of a total of 2394 voxels, located in the temporal but also in the frontal and the right limbic lobe. P300 source strength increased prefrontally (left more than right), involving a total of 219 voxels (Table 1). Moreover, a decrease in N1 and P2 source strengths (mainly right temporally) was observed (Figure 8). However, the number of voxels involved did not constitute a suprathreshold region (ie, > 137 of 2394 voxels; P < 0.05).
FIGURE 8. . Acute effects of S-adenosyl-L-methionine (SAMe) on standard and target LORETA (low-resolution brain electromagnetic tomography) sources compared with placebo in healthy subjects (n = 19). Images are based on voxel-by-voxel t values (t > 2.10, P < 0.05). Although N1 source strength decreased mainly left prefrontally, P2 source strength decreased mainly right temporally (dotted arrows). N2 and P300 source strengths increased significantly (P < 0.05, omnibus test; solid arrows). N2 source strength showed an increase bilaterally in the temporal lobes and P300 source strength in the prefrontal cortex, left more than right.
Subacute treatment with daily infusions of 800 mg SAMe for 1 wk resulted in a significant decrease in N1 and P300 source strengths as well as an increase in P2 and N1 source strengths (Table 1), but, again, the number of voxels involved was too small for meeting the criteria of a suprathreshold region.
However, one superimposed infusion in addition to subacute treatment showed a highly significant increase in N2 source strength, predominantly in the temporal lobes (Figure 9), involving 567 voxels in the temporal, limbic, frontal, and parietal lobes. The significantly increased activity in the N2 component—observed specifically in the temporal lobe after acute, subacute, and superimposed drug administration (Figure 9)—reflects an increased allocation of attentional resources for the evaluation of possible targets.
FIGURE 9. . Differences between S-adenosyl-L-methionine (SAMe)–induced and placebo-induced changes on target N2 LORETA (low-resolution brain electromagnetic tomography) sources in healthy subjects (n = 19). Images are based on voxel-by-voxel t values (t > 2.10, P < 0.05). The significantly increased activity in the temporal lobe observed after acute, subacute, and superimposed drug administration might be interpreted as an increased allocation of attentional resources for the evaluation of possible targets.
A more detailed analysis of P300 changes 1 h after acute administration of 800 mg SAMe in the 9 elderly subjects showed a current density increase in the dorsal frontal regions (Brodmann areas 9 and 46) and the posterior cingulate cortex (Brodmann area 31) and a decrease in the ventral limbic regions (Brodmann area 25) (Figure 10).
FIGURE 10. . Similarity between changes in intracranial current density LORETA (low-resolution brain electromagnetic tomography) at target P300 latency 1 h after administration of S-adenosyl-L-methionine (SAMe) in 9 healthy, elderly subjects compared with placebo and changes in regional glucose metabolism (with [18F]fluorodeoxyglucose positron emission tomography) after 6 wk of successful antidepressant treatment with fluoxetine in 8 depressed patients (modified with permission from reference 26). In both studies, increases in dorsal frontal regions (Brodmann areas 9 and 46) and the posterior cingulate cortex (Brodmann area 31) and decreases in ventral limbic regions (Brodmann area 25) were observed.
DISCUSSION
Our pharmaco-EEG mapping studies showed that SAMe, administered as single and repeated infusions of 800 mg, had significant effects on the brain function of healthy subjects as compared with placebo—more so in the resting than in the vigilancecontrolled state.
The acute pharmaco-EEG changes observed in both the young and elderly subjects (27, 28) were typical of antidepressants of the thymoleptic type, such as imipramine and amitriptyline, as well as of doxepine, amitriptyline-N-oxide, maprotiline, binodaline, danitracene, fluvoxamine, and venlafaxine (14, 19, 29–34).
The pharmaco-EEG classification of the nutraceutical and supplementary substance SAMe as an antidepressant agrees with the findings of clinical studies performed in the past 20 y (35–44). In a study of depressed inpatients by Lipinski et al (45), either an improvement or a remission was found in > 75% of the subjects after short parenteral intravenous treatment with SAMe (100–200 mg/d). The response was rapid (within 10 d). No systemic side effects were reported by any patient on questioning or physical examination (eg, blood pressure, pulse, and respiratory rate) or from laboratory tests.
Starting with a dosage of 200 mg SAMe twice daily by mouth, which increased to up to 1600 mg/d in 19 d, Rosenbaum et al (46) stated a significant improvement and a robust response with a dosage of 1200–1600 mg/d. The efficacy of SAMe in comparison with placebo was assessed in many randomized, double-blind clinical trials. The results of a study performed by De Leo (47) showed that SAMe produced an effect superior to that of placebo and was well tolerated. All studies reporting effects on specific items of the Hamilton Depression Scale consistently found an improvement with SAMe compared with placebo in mood, suicidal tendencies, retardation, work and interests, somatic symptoms, and hypochondriasis (48–52). Patients with endogenous depression showed a much greater improvement than did those with neurotic depression. Oral administration of 1600 mg SAMe/d also proved to be more effective than placebo, and the difference in the Hamilton Depression Scale scores between the 2 treatments was significant in the second week (days 10–14) (53, 54).
In our study, the SAMe-induced quantitative EEG changes also indicated both inhibitory and excitatory drug effects, which have been described for antidepressants such as imipramine and venlafaxine (17, 29–33, 55). Generally, stands for inhibitory electrical activity, for normal adult functioning, and ß for excitatory CNS activity.
This finding is of interest because our ERP findings with SAMe showed both inhibitory and excitatory effects, specifically in regard to latencies. The latency of the automatic perceptual response, reflected by the N1 component in the acoustic ERP, was shortened, whereas information processing latency, reflected by the P300 component, was lengthened. Indeed, using somatosensory evoked potentials, we described such a shortening of the latencies of early responses and lengthening of those of late ones after imipramine and amitriptyline (56) and also found an increase in P300 latency after 20 mg citalopram (57). Concerning amplitudes, we showed an increase in N2 source strength. According to Halgren et al (25), this reflects the allocation of resources for preparation of the information processing resources used in the P300 later. Indeed, after acute administration of SAMe, we found an increase in P300 amplitudes much like that observed after administration of 20 mg of the antidepressant citalopram (57). This lengthening of the P300 may be seen as a kind of preparation process needed for the allocation of more resources. Moreover, LORETA sources involving ERP components showed both activating and inhibitory modes of action of SAMe, characterized by an increase in P300 source strength in Brodmann areas 9 and 46 and in area 31 of the posterior cingulate gyrus and a decrease in Brodmann area 25 of the subgenual cingulate gyrus. Such differential changes reflecting reciprocal limbic-cortical function were also reported by Mayberg et al (26), who described a significant change in regional glucose metabolism (measured by [18F]fluorodeoxyglucose positron emission tomography) after 6 wk of successful antidepressant treatment with fluoxetine in 8 depressed patients (Figure 10). In that and the present study, antidepressant effects were characterized by reciprocal limbic-cortical changes, ie, decreased activity in limbic and increased activity in neocortical structures.
Subacute SAMe treatment over 1 wk induced changes that were partly the same and partly different from those observed after acute administration. Specifically, there was an increase in total power as we described previously after the administration of nootropic compounds such as pyritinol (17, 18). Nootropic drugs (eg, antidementia drugs and cognition enhancers) are substances that preferentially improve intellectual and memory performance in patients with cognitive disorders. The effect of a superimposed dose in addition to 7 d of subacute administration of SAMe again resulted in a slight change in brain function toward the pharmaco-EEG profiles seen after the administration of imipramine and amitriptyline. However, this effect waned in the sixth hour after the superimposed dose.
The nootropic-like action observed after subacute treatment and the more pronounced changes in elderly subjects in the resting than in the vigilance-controlled state support the notion that SAMe has a more pronounced effect in deficit states (eg, advanced ages) than in nondeficit states (eg, young ages). Indeed, in clinical studies, elderly, depressed patients showed greater SAMe-induced changes than did younger patients, as reported recently by Delle Chiaie et al (5). Moreover, typical neurophysiological changes reflecting an improvement in brain functioning were reported by Torta et al (8) in patients with major depression but also by Coiro and Vescovi (58) in abstinent alcoholics. This evidence also supports the notion that SAMe may have a greater effect in deficit states; this notion should be confirmed in trials including both clinical evaluations and neurophysiological neuroimaging techniques.
ACKNOWLEDGMENTS
We express our thanks to M Corrado and C Scarsi (monitors), to E Grätzhofer (editorial assistant), and to the entire staff of the Section of Sleep Research and Pharmacopsychiatry for their valuable assistance in this project.
REFERENCES