Decreased frontal lobe phosphocreatine levels in methamphetamine users

Abstract

BACKGROUND

Mitochondria-related mechanisms have been suggested to mediate methamphetamine (METH) toxicity. However, changes in brain energetics associated with highenergy phosphate metabolism have not been investigated in METH users. Phosphorus-31 (³¹P) magnetic resonance spectroscopy (MRS) was used to evaluate changes in mitochondrial high energy phosphates, including phosphocreatine (PCr) and β-nucleoside triphosphate (β-NTP, primarily ATP in brain) levels. We hypothesized that METH users would have decreased high-energy PCr levels in the frontal gray matter.

METHODS

Study participants consisted of 51 METH (age=32.8±6.7) and 23 healthy comparison (age=31.1±7.5) subjects. High-energy phosphate metabolite levels were compared between the groups and potential gender differences were explored.

RESULTS

METH users had lower ratios of PCr to total pool of exchangeable phosphate (PCr/TPP) in the frontal lobe as compared to the healthy subjects (p=0.001). The lower PCr levels in METH subjects were significantly associated with lifetime amount of METH use (p=0.003). A sub-analysis for gender differences revealed that female METH users, who had lower daily amounts (1.1±1.0 gram) of METH use than males (1.4±1.7 gram), had significantly lower PCr/TPP ratios than male METH users, controlling for the amount of METH use (p=0.02).

CONCLUSIONS

The present findings suggest that METH compromises frontal lobe high-energy phosphate metabolism in a dose-responsive manner. Our findings also suggest that the abnormality in frontal lobe high-energy phosphate metabolism might be more prominent in female than in male METH users. This is significant as decreased PCr levels have been associated with depressive symptoms, and poor responses to antidepressant treatment have been reported in those with decreased PCr levels.

1. INTRODUCTION

Methamphetamine (METH) abuse is a disorder characterized by compulsive METH-craving and consumption despite an apparent awareness of serious negative consequences. METH use has been linked to the emergence of psychotic symptoms (Iyo et al., 2004; London et al., 2004; Winslow et al., 2007) as well as morphological, functional, and neurochemical abnormalities in multiple brain areas (Bae et al., 2006; Chung et al., 2007; Ernst et al., 2000; Hwang et al., 2006; Oh et al., 2005). Alterations in monoaminergic neurotransmission (Kokoshka et al., 1998; Ricaurte et al., 1980; Robinson and Berridge, 1993) in the frontal lobe have been related to cognitive impairments in METH users because prefrontal cortical neural networks play a central role in impaired decision-making and inhibitory control (Lubman et al., 2004; Yucel and Lubman, 2007).

In METH toxicity, multiple lines of evidence suggest that dysfunctional energy metabolism plays an important role. For instance, 1) impairments of mitochondrial function have been reported after administration of METH to animals including impairments in mitochondrial electron transport chain enzyme complexes (Brown et al., 2005; Burrows et al., 2000a; Klongpanichapak et al., 2006); 2) METH toxicity involves a depletion of energy stores, evidenced by synergistic metabolic inhibition by METH resulting in depletion of striatal dopamine content (Burrows et al., 2000b); and 3) ex vivo METH exposure has been associated with oxidative cell injury and apoptosis in rat cortical neuron and undifferentiated pheochromocytoma (PC12) cells (Cunha-Oliveira et al., 2006; Oliveira et al., 2002). Considering the evidence for mitochondrial involvement in the potential pathophysiology of METH toxicity, it is not surprising that proton (¹H) magnetic resonance spectroscopy (MRS) studies have consistently reported that METH users, relative to healthy comparison subjects, have decreased levels of total creatine (phosphocreatine plus creatine) as well as decreased N-acetylaspartate (NAA, a marker of neuronal viability or integrity (Moffett et al., 2007)) levels (Chang et al., 2005; Ernst et al., 2000; Nordahl et al., 2002; Sailasuta et al., 2010b; Sekine et al., 2002; Smith et al., 2001; Sung et al., 2007; Taylor et al., 2007). As NAA synthesis occurs primarily in the mitochondria (Patel and Clark, 1979), decreased NAA levels in METH users are potentially consistent with compromised brain energetics.

Phosphocreatine (PCr) and adenosine triphosphate (ATP) make up the PCr-ATP energy buffering system in neuronal cells that have high and fluctuating energy demands. The enzyme creatine kinase controls the transfer of a phosphate group from PCr to ADP, thereby replenishing brain ATP. It is reported that PCr serves as a buffer to maintain constant ATP levels so that ATP levels remain relatively stable at the cost of PCr expenditure (Schlattner et al., 2006). In the fluctuating energy requirements of neurons, mitochondrial dysfunction may lead to decreased formation of phosphocreatine mediated by the mitochondrial creatine kinase isoenzyme (Dolder et al., 2001; Wallimann et al., 1998). Regarding region specific deficits, frontal hypometabolism has been reported in METH users using positron emission tomography (PET; Kim et al., 2005; London et al., 2005, 2004). These findings would be consistent with the potential mitochondrial abnormality and decreased PCr levels in the frontal lobe of METH users.

Published neuroimaging studies have reported gender differences in METH toxicity with favorable outcomes in female METH users in terms of frontal glucose metabolism and white matter hyperintensities/integrity (Bae et al., 2006; Chung et al., 2007; Kim et al., 2005), probably reflecting a protective effect of estrogen (Dhandapani and Brann, 2002). However, female psychostimulant users are more sensitive to the reinforcing effects than male users (Anker and Carroll, 2011; Carroll et al., 2004). Also, female METH users have a higher incidence of depression and more severe depressive symptoms than male METH users (Dluzen and Liu, 2008; Hser et al., 2005; Kalechstein et al., 2000). In non-METH users, depression severity has been significantly associated with decreased brain PCr levels (Kato et al., 1992), but it is not known what the effects of METH use will be on brain high energy phosphate metabolism in female METH users.

Phosphorus-31 (³¹P)-MRS provides a unique method to evaluate changes in high-energy phosphate metabolites such as PCr and beta-nucleoside triphosphate (β-NTP), which arises primarily from ATP in brain (Renshaw et al., 2001). To date, there have been no reports measuring ³¹P-MRS metabolite levels in METH users relative to healthy subjects. In this study, we aimed to investigate whether METH use significantly altered high energy phosphate metabolism. It was hypothesized that first, METH use would be associated with decreased highenergy PCr levels by ³¹P-MRS in the frontal cortex, and second, that altered PCr levels will be significantly correlated with lifetime amount of METH use. In addition, we explored possible gender differences in the phosphorus metabolite levels in METH users.

2. METHODS

2.1 Subjects

This was a cross-sectional study in which phosphorus MRS data was acquired to examine brain metabolite alterations related to METH use. The study participants consisted of 51 METH dependent subjects (age=32.8±6.7, female=23) and 23 healthy comparison subjects (age=31.1±7.5, female=11). Each individual underwent two dimensional phosphorus magnetic resonance spectroscopic imaging (2D ³¹P-MRSI) as well as assessment of clinical and drug abuse history. METH-dependent subjects were evaluated for the severity of their lifetime METH use (METH amount, frequency, duration, and abstinence). Inclusion criteria for METH subjects were as follows: (1) age 18–55 years, (2) subjects who met diagnostic criteria for current methamphetamine abuse or dependence as their preferred drug of abuse as determined by the Structured Clinical Interview for DSM-IV (SCID-IV). A board certified psychiatrist took complete medical histories and physical examinations of the subjects. The SCID-IV was administered by a trained psychologist (ECM), and (3) METH use within the past six months. Exclusion criteria for METH subjects included (1) major medical or neurological disorders, including HIV seropositivity; (2) comorbid psychiatric disorders including schizophrenia, bipolar disorder, and use of other illicit drugs as preferred drug of abuse; (3) major sensorimotor handicaps (e.g., deafness, blindness, and paralysis), full scale IQ <70 or learning disabilities; and (4) contraindications to magnetic resonance imaging.

Healthy comparison (HC) subjects were recruited with the inclusion criteria (1) age 18 to 55 years, (2) no dependence or abuse of alcohol, METH, amphetamine, cocaine, heroin, alcohol, and cannabis, and (3) no psychiatric, neurologic, and medical disease identified by physical examination. The healthy subjects were matched on age and gender with METH users. The healthy subjects had a slightly higher level of education than METH users and this difference was controlled for statistically. Exclusion criteria for the HC subjects were the same as for the METH users. The study protocol was approved by the Institutional Review Boards of the University of Utah and the Department of Human Services of the State of Utah. Written informed consent was obtained from all study subjects before participation.

2.2 Magnetic Resonance Imaging and Spectroscopy Data Acquisition and Processing

2.2.1 Structural MR Images

Brain MR imaging was performed using a 3 Tesla Siemens scanner (Trio, Siemens AG, Erlangen, Germany) and a ³¹P/¹H double-tuned volume head coil (Clinical MR Solutions LLC, Brookfield, WI) for transmission and reception. To obtain high resolution T1-weighted anatomical images for tissue segmentation and positioning MRS grids, a three dimensional chemical Magnetization Prepared Rapid Acquisition Gradient Echo (MPRAGE) pulse sequence was used. The parameters for the structural MRI were as follows: T1 weighted image: TR=2000 ms, TE=3.37 ms, TI=1100 ms, average number=1, flip angle=8°, FOV=256 mm, matrix 256×192×144, bandwidth=300 Hz/pixel, slice thickness 1.0 mm and no gap.

2.2.2 ³¹P Magnetic Resonance Spectroscopy

Phosphorus spectra were acquired on the same 3 Tesla Siemens system using the ³¹P/¹H double-tuned volume head coil. The spectra were obtained using a two dimensional-chemical shift imaging (2D-CSI) free induction decay (FID) pulse sequence with TR= 3000, TE=2.3 ms, average number=36, flip angle=90°, vector size=1024, FOV=200×200mm², slice thickness=2.5 cm, matrix 8×8, sampling bandwidth=2.5 kHz, and voxel dimension=2.5×2.5 cm². The 2D-CSI grid was positioned covering an axial brain slice just above an imaginary line connecting the anterior commissure and posterior commissure. The acquisition matrix of the 2D-CSI grid was 8x8 with the slice thickness 25mm (Figure 1). Shimming was performed over the excited brain volume. Since MRS data is significantly affected by magnetic field inhomogeneity, high order advanced shimming routine on the Siemens system was used to achieve linewidths of less than or equal to 15 Hz for the unsuppressed water signal. The proton channel was used for shimming, localization, and anatomic imaging. As an a priori region of interest (ROI), frontal lobe spectroscopic data were quantified. Also, as a control region for the comparison, temporoparietal lobe and occipital lobe were included in the spectroscopic and statistical analyses. The frontal lobe was selected as an active ROI because of prior reports suggesting frontal neurochemical abnormalities and hypometabolism in METH users (Kim et al., 2005; London et al., 2005; London et al., 2004; Sailasuta et al., 2010a)

Figure 1.

Illustration of two dimensional MRSI grid placement: (A) Sagittal and (B) Axial views. Dotted red line in the figure shows frontal lobe region of interest. (C) A representative ³¹P spectrum is displayed in frequency domain with 10 Hz exponential filtering. Abbreviations: PCr, phosphocreatine; PME, phosphomonoester; PDE, phosphodiester; α-, β-, γ-NTP (nucleoside triphosphate).

2.2.3 ³¹P-MRS Data Analysis

Before performing 2D Fast Fourier Transform (FFT) on raw data, a Hamming filter was applied to reduce the effect of the point-spread-function, and each free induction decay (FID) was line-broadened with 10 Hz of apodization. After Fourier transformation and frequency shift correction, zero-/first-order phase correction and baseline correction with polynomial interpolation were applied. Metabolite location error resulting from different chemical shift displacement was corrected along in-plane readout and phase-encoding directions. Following that, spatial filtering with a Hamming window function was implemented to reduce the signal contamination from neighboring voxels. The preprocessed ³¹P-MRSI data was fitted using jMRUI software (Naressi et al., 2001) with the Advanced Method for Accurate, Robust and Efficient Spectral fitting (AMARES) algorithm (Vanhamme et al., 1997). Metabolites of interest were PCr, β-NTP, and their ratio (Figure 1C). Each metabolite concentration was expressed relative to the total pool of exchangeable phosphate (TPP; Blumberg et al., 1999). Referencing ³¹P metabolites to TPP facilitated effective evaluation of high energy phosphate metabolism (Amess et al., 1997; Cady et al., 2008; Iwata et al., 2008).

From the registered anatomical images, tissue segmentation was performed using FSL (FMRIB's Software Library) software so that cerebrospinal fluid (CSF)-corrected metabolite concentrations as well as gray matter percentage in each voxel could be used as covariates in the statistical analysis. No significant difference in CSF amount was detected in the frontal voxels between METH and control subjects (t₇₂=0.56, p=0.58).

2.2.4 Statistical Analysis

Collected demographic and clinical data were managed using Research Electronic Data Capture (REDCap) electronic data capture tools hosted at the University of Utah, which is a secure, web-based application designed to support data capture for research studies (Harris et al., 2009). Analysis of variance (ANOVA) was used for between-group comparisons involving continuous demographic data. Generalized Estimating Equations regression modeling was used to evaluate group differences in metabolite levels controlling for age, sex, education level, and tissue partial volume effects (i.e. gray matter, white matter, CSF). Fisher's Exact Test compared groups on categorical variables. To determine the clinical relevance of metabolite changes in METH users, robust multiple regression analyses (White, 1980) were performed to assess the relationship between PCr levels and lifetime amount of METH use or days of abstinence of METH. Partial correlation coefficients were calculated covarying for age and sex. Models with all pair-wise and three-way interactions between covariates were first considered, and interactions not significant were removed from the models. Statistical significance was defined at an alpha level of p=0.05, two-tailed. Stata for Unix, version 12.1 (StataCorp, College Station, TX) was used for all computations.

3. RESULTS

3.1 Demographics and clinical characteristics

There were no significant differences between the METH and the HC groups in terms of age (32.8±6.7 vs. 31.1±7.5 respectively, t₇₂=0.95, p=0.35) and sex (45% vs. 47% female ratio respectively, Fisher's Exact Test, p=0.5). The METH users, however, had lower education level than the HC subjects although this difference was modest (12.7±1.8 vs. 13.9±1.3, respectively, p<0.05). Fifty METH users met lifetime METH dependence DSM-IV criteria and one participant was diagnosed with METH abuse. Table 1 presents study subject characteristics and other demographics in detail. Table 2 presents METH and other drug use history for all METH users.

Table 1.

Clinical characteristics and demographics

	Methamphetamine				Healthy subjects				Significance
	Total	Female	Male		Total	Female	Male
Subject demographics	(n=51)	(n=23)	(n=28)	p-value	(n=23)	(n=11)	(n=12)	p-value
Education (years)	12.7 (1.8)*	12.6 (2.0)	12.8 (1.6)	.65	13.9 (1.3)*	13.9(1.2)	13.9(1.5)	.95	t72=2.88, p=.005*
Handedness (right/left)	43/8	20/3	23/5	.47	22/1	10/1	12/0	.48	FET, p=.16
Ethnicity (white/others)	45/6	20/3	25/3	.57	21/2	10/1	11/1	.74	FET, p=.52
Onset age of METH use (yrs)	21.7 (6.6)	20.4 (5.9)	22.8 (7.0)	.20	-	-	-	-	-
Duration of METH use (yrs)	11.1 (7.0)	9.3 (5.8)	12.5 (7.6)	.10	-	-	-	-	-
Abstinence of METH use (days)	38.1 (42.5)	46.3(54.1)	31.4(29.2)	.21	-	-	-	-	-
Daily dosage of METH use (g)	1.3 (1.4)	1.1(1.0)	1.4(1.7)	.46	-	-	-	-	-

Data are expressed with mean (SD);

^*Indicates statistical significance between methamphetamine and healthy group;

Abbreviations: METH, methamphetamine; FET, Fisher's Exact Test

Table 2.

Profiles of drug use history

Substance	Total Number of Subjects	Mean Age of First Use (yrs)	Mean Amount of Substance per day	Frequency Used	Total Number of Subjects Currently Using	Total Lifetime Amounta
Alcohol	51	13	65 ounces	Occasionally to Weekly	17	34 drink-years b
Amphetamine	26	19	Varies	Occasionally	1	Varies
Methamphetamine	51	22	1.81 grams	Daily	30	1923 grams
Cocaine	48	20	1.49 grams	Occasionally	1	1656 grams
Opiates	22	25	0.69 grams	Occasionally	0	351 grams
Hallucinogens	36	17	Varies	Rarely to Occasionally	0	Varies
Cannabis	50	14.5	2.42 grams	Weekly	12	3721 grams
Nicotine	47	14	11 cigarettes	Daily	41	10 pack-yearsc
Caffeine	50	9	28 ounces	Daily	49	27 cups-years d

^aTotal lifetime drug amount was estimated using "average daily dosage×frequency×duration."

^bTotal intake calculated in drink-years (1 drink/day in 1 year).

^cA pack-year is defined as twenty cigarettes smoked everyday for one year (1 pack/day in 1 year).

^dA cups-year is defined as five cups of coffee everyday for one year.

3.2 Group differences in Metabolite levels

PCr levels in METH users were significantly reduced in the frontal lobe compared with those in HC group but not β-NTP/TPP levels (PCr/TPP, t₇₂=−3.46, p=0.001; β-NTP/TPP, t₇₂=0.56, p=0.58, Figure 2). The statistically significant reduction in PCr levels was maintained after controlling for potential confounding variables such as age, sex, education level, and tissue segmentation (t₆₈=−2.99, p=0.004). As there was a high collinearity between group variable and alcohol amount/nicotine amount consumed, we did not conduct regression analysis adjusting for alcohol, nicotine and caffeine.

Figure 2.

Comparison of frontal lobe phosphorus metabolites levels in methamphetamine (METH, n=51) dependent subjects compared to healthy controls (HC, n=23). Phosphocreatine levels were significantly reduced in METH users compared to HC (p<0.001). There were no significant differences in β-NTP levels between the HC and METH groups. Error bars represent 95% confidence intervals. * Indicates statistically significant difference. Abbreviations: TPP, total pool of exchangeable phosphate.

To ensure that decreased PCr/TPP did not result from increased total TPP values, we repeated our statistical analysis using β-NTP as a denominator, with findings remaining the same. Specifically PCr/β-NTP levels were significantly decreased in METH users compared to the HC subjects (t₇₂=−2.08, p=0.04). Also, to ensure that the MR system did not drift across the study period, absolute measurements (arbitrary institutional unit) of total phosphorus were compared between the groups, with no group differences in TPP levels (p=0.10). Further, to assess test-retest reliability of our ³¹P-MRS protocol, we scanned nine healthy subjects twice using the exact same protocol. The intraclass correlation coefficient (ICC; Shrout and Fleiss, 1979) for PCr levels was calculated to be 0.72, which suggests good reliability.

Other metabolite levels such as phosphomonoester, phosphodiester, and inorganic phosphate were not significantly different between the groups (see Table 3). We did not find a statistically significant difference in PCr/TPP levels in other brain regions such as temporoparietal lobe (t₇₂=−1.71, p=0.10) and occipital lobe (t₇₂=1.60, p=0.12) between METH and HC subjects.

Table 3.

Phosphorus metabolite levels in the frontal lobe

	Methamphetamine			Healthy			p-value
	Total			Total
Metabolites†	(n=51)	Female	Male	(n=23)	Female	Male
PCr	.212 (.021)*	.205 (.019)**	.217 (.021)**	.230 (.016)*	.225 (.018)	.231 (.014)	.001*
β-NTP	.192 (.022)	.195 (.025)	.189 (.019)	.189 (.020)	.187 (.027)	.190 (.013)	.58
PME	.215 (.037)	.220 (.041)	.210 (.033)	.218 (.022)	.215 (.027)	.221 (.018)	.68
PDE	.286 (.036)	.291 (.036)	.281 (.037)	.292 (.034)	.285 (.034)	.298 (.034)	.47
Pi	.076 (.012)	.080 (.012)	.074 (.013)	.074 (.012)	.073 (.013)	.075 (.012)	.38

Data are expressed with mean (SD);

^†metabolite ratios over total pool of exchangeable phosphate;

^*Indicates statistically significant PCr differences between methamphetamine and healthy group (p=.001);

^**Indicates significant PCr differences between female and male methamphetamine users (p=.04). For other metabolites, there were neither group differences nor within-group gender differences.

Abbreviations: PCr, phosphocreatine; β-NTP, beta-nucleoside triphosphate; PME, phosphomonoester; PDE, phosphodiester; Pi, inorganic phosphate.

3.3 PCr levels in female vs. male METH users

Frontal PCr/TPP levels in female METH users were significantly lower (5.3%) than those in male METH users (t₄₉=−2.06, p=0.04, Figure 3). Female METH users reported lower amounts of daily METH use (1.1±1.0 gram) compared to male METH users (1.4±1.7 gram) (t₄₉=−0.75, p=0.46). After adjusting for the daily amount of daily METH use, the significance for betweengroup gender difference of PCr/TPP levels was maintained (t₄₈=−2.41, p=0.02). In the healthy subjects, there were no significant gender differences in either PCr (t₂₁=−0.90, p=0.38), β-NTP (t₂₁=−0.90, p=0.38), or other metabolites (Table 3)

Figure 3.

Comparison of gender difference in PCr levels. (A) In METH-dependent subjects, female METH users (n=23) had significantly lower PCr levels compared to male METH users (n=28). (B) Healthy subjects did not show significant gender difference. Error bars represent 95% confidence intervals. * Indicates statistically significant difference. Abbreviations: TPP, total pool of exchangeable phosphate.

3.4 Relationship between Frontal lobe PCr levels and lifetime METH amount

There was significant relationship between frontal lobe PCr levels and total amount of lifetime METH use (Figure 4), before (t₄₉=−3.17, p=0.003), and after (t₄₇=−2.87, p=0.006) controlling for age and education.

Figure 4.

Significant relationship (p=0.003) between PCr/TPP levels and total amount of lifetime METH use (gram) in METH-dependent subjects (regression analysis using robust estimator of variance (White, 1980). Gray line and area denote predicted 95% confidence interval. Abbreviations: TPP, total pool of exchangeable phosphate.

3.5 Relationship between Frontal lobe PCr levels and duration of abstinence from METH

There was a trend toward a negative relationship between frontal lobe PCr levels and duration of METH abstinence but this was not statistically significant (t₄₉=−1.49, p=0.14).

4. DISCUSSION

In this study, we found that METH users have abnormal PCr levels in the frontal lobe compared with the HC subjects. To the best of our knowledge, this is the first ³¹P-MRS study reporting abnormal brain cellular energetics in human METH users.

Our findings are consistent with the prior neuroimaging reports regarding metabolic hypofrontality in METH users. For example, ¹⁸F-fluorodeoxyglucose (FDG) PET studies have noted lower FDG uptake in the frontal lobe and cingulate (Kim et al., 2005; London et al., 2005; 2004) and in the striatum (Volkow et al., 2001a). Also, single photon emission computed tomography (SPECT) studies have reported decreased frontal blood flow in METH users (Hwang et al., 2006).

Of particular relevance to the decreased PCr levels and mitochondrial dysfunction in METH users, several ³¹P-MRS studies of individuals with primary mitochondrial disease have reported decreased PCr levels in the resting state of muscles (Argov et al., 1987; Arnold et al., 1985; Hoang et al., 1998) as well as in affected brain regions (Barbiroli et al., 1995,1993). Importantly, when there was no clinical brain involvement in the mitochondrial disease, patients had normal PCr levels (Rango et al., 2001). The abnormal PCr levels in our METH users are consistent with results observed in opioid-dependent subjects. For example, heroin-dependent subjects were reported to have decreased PCr levels (−15.3%) at the onset of methadone maintenance treatment (Silveri et al., 2004). Notably, decreased PCr levels in opiate-dependent polydrug abusers tended to recover after methadone maintenance therapy (Kaufman et al., 1999).

The alteration of PCr levels only in the frontal lobe of METH users suggest that the damage induced by METH may be region-specific rather than global in terms of brain energetics. The frontal lobe is more metabolically active than other brain regions (Ivancevic et al., 2000; Loessner et al., 1995) and frontal lobe dysfunction is commonly associated with chronic METH use (Henry et al., 2010). We previously reported that compromised frontal executive function was significantly correlated with reduced frontal glucose metabolism in METH dependent subjects (Kim et al., 2005). Therefore, it is possible that the reduced PCr levels may be a ³¹PMRS metabolic biomarker of hypofrontality in METH users.

The present phosphorus MRS findings are consistent with the existing proton MRS literature. For instance, reduced NAA levels, which are known to reflect the functional status of mitochondria have been associated with human METH use (Chang et al., 2005; Ernst et al., 2000; Nordahl et al., 2002; Sailasuta et al., 2010b; Sekine et al., 2002; Smith et al., 2001; Sung et al., 2007; Taylor et al., 2007). Of note, NAA is generated by L-aspartate N-acetyltransferase, an enzyme predominantly located in the mitochondria. Further, decreased PCr levels in our ³¹P-MRS findings are in line with the prior ¹H-MRS reports of decreased total creatine plus phosphocreatine levels in METH users (Ernst et al., 2000).

Constant brain adenosine triphosphate levels are critical for cell survival and proper human brain functioning (Niizuma et al., 2009). In high energy phosphate metabolism, PCr-ATP buffering and the creatine kinase system (PCr²⁺ + ADP- ↔ Cr + ATP²⁺) play important roles in maintaining constant ATP levels. No significant differences in the β-NTP levels between METH users and HC subjects suggest a possible compensatory or homeostatic mechanism in maintaining ATP levels in METH users. Although the precise mechanisms remain unclear, creatine and PCr have been reported to have neuroprotective effects in various brain disorders including ischemic stroke and Alzheimer’s disease (Beard and Braissant, 2010). Therefore, our finding of decreased PCr levels may indicate an increased susceptibility to neurotoxic changes in METH users. This suggests a potential treatment target, as exogenous creatine supplementation, which increases PCr levels in healthy volunteers (Lyoo et al., 2003), might provide beneficial effects directly targeting the underlying pathophysiology of METH-induced mitochondrial toxicity. However, further study would be required to confirm the possible therapeutic effects of oral creatine in METH users.

Gender differences in METH toxicity have been reported in previous neuroimaging studies of diffusion tensor imaging and white matter hyperintensities (Bae et al., 2006; Chung et al., 2007). Male METH users had lower fractional anisotropy values and greater severity of white matter hyperintensites than female METH users. In the present ³¹P-MRS study, however, decreased PCr levels were more prominent in females. The exact mechanism responsible for lower PCr levels in female METH users is unclear and has yet to be elucidated. However, since lower brain levels of PCr have also been associated with severe depression (Kato et al., 1992) and worse outcomes in response to antidepressant treatment (Iosifescu et al., 2008), our findings may be related to the higher incidence of depression in female METH users (Semple et al., 2007). Indeed, METH abuse is associated with increased risk of depression and suicide attempts compared to rates observed in the general population (Glasner-Edwards et al., 2009). Also, METH abusers have higher female/male ratios (46%) than heroin (11%) or cocaine (29%) abusers (Cohen et al., 2007; Holdcraft and Iacono, 2004; Hser et al., 2008). Therefore, the relatively large reduction of PCr levels in female METH users highlights the potential consequences of bioenergetic impairment that merit special clinical attention in female subjects. Another possible underlying mechanism for lower PCr in female METH users may be related to the effect of gonadal hormones on brain PCr levels. For example, estrogen has been reported to stimulate brain-specific cytosolic creatine kinase activity (Kaye, 1983). Therefore, the higher estrogen levels in females may contribute to the decreased PCr levels through the forward (i.e. regenerating ATP) creatine kinase activity in the PCr-ATP buffer system. Although it was not statistically significant, we found slightly lower PCr levels in female healthy subjects. Interestingly, a recent PET study demonstrated decreased prefrontal metabolism in females but not in male cocaine users (Volkow et al., 2011). This observation suggests that gonadal hormones may be associated with increased risk of abnormal high energy phosphate metabolism. However, further study will be required to clarify the relationship between hormone levels and brain high energy phosphate changes.

The significant relationship between total amount of METH use and the PCr levels illustrates that heavy METH users might manifest as an abnormality in high energy phosphates in a dose-responsive manner. These findings raise the possibility that heavy METH users may benefit from agents which facilitate the recovery of frontal lobe PCr levels. A non-significant negative association between frontal lobe PCr levels and duration of METH abstinence suggests that recovery of PCr levels may not occur during the moderate levels of simple abstinence (95% confidence limits: 26 to 50 days). Indeed, the negative trend suggests that longer abstinence may result in PCr levels that remain low or even decline further. These findings are consistent with persistently lower striatal, but not thalamic, FDG metabolism in METH users compared to HC even after 12 to 17 months of abstinence (Wang et al., 2004), as well as significantly reduced striatal dopamine transporter levels in METH users compared to HC even after 11 months of abstinence (Volkow et al., 2001b).

4.1 Study Limitations

When considering our findings, several factors should be taken into consideration. The ³¹PMRS data are reported by ratio to TPP with no absolute ³¹P metabolite concentration. Although this is a standard practice in the MRS literature, decreased PCr/TPP may have originated from increased TPP levels (Jansen et al., 2006; Jayakumar et al., 2010). However, even when we use a different denominator such as β-NTP, we observe similar results. Also, metabolite ratio measures have been compared favorably with absolute measures (Klunk et al., 1994). Therefore, decreased PCr/TPP ratios are likely to reflect decreased PCr levels.

Female subjects were not scanned at the same phase of the menstrual cycle in the present study. Since prior published studies suggest hormonal levels are related to subjective reports of stimulation after D-amphetamine administration (Reed et al., 2010; White et al., 2002), phosphorus metabolite levels may also fluctuate across the menstrual cycle in the females, placing the lack of the control of menstrual phase as a limitation of this study. However, since the female subjects were randomly scanned for both controls and METH users, we believe that the variance in PCr levels was similar between groups, and hence the menstrual phase variation did not significantly affect our findings.

Multiple comparisons were performed in statistical analysis for our metabolites of interest, PCr and β-NTP. As we used an a priori hypothesis in our research design, corrections for multiple comparisons were not attempted. Therefore, readers should be aware of possible type I errors arising from our analysis.

The METH subject enrollment criteria involved the inclusion of subjects who identified METH as their drug of choice. Therefore, the characteristics of our cohort may not represent general METH users and our findings may not be generalized to other METH-dependent populations. On the other hand, all METH users abused other drugs to some degree and decreased PCr levels in METH users may be, in part, due to polysubstance abuse. Self report of drug use history in METH subjects might have resulted in some degree of uncertainty including underreport of drug use behaviors.

Because the location of our 2D-CSI grid was immediately superior to the AC-PC line, we could not include subgenual cingulate or cerebellum as regions-of-interest to evaluate whether these brain regions have abnormalities in high energy phosphorus metabolism. Since the subgenual cingulate is reported to be associated with depression (Drevets et al., 2008; Mayberg et al., 2005), a study design including this brain region will be important in future studies.

5. CONCLUSION

Overall, the present study provides evidence of altered high-energy phosphate metabolism in METH users. Our findings suggest that (1) METH use is associated with dose-dependent decreases in high-energy PCr levels, which may imply a decreased energetic buffer due to mitochondrial dysfunction; and (2) female METH users may be more vulnerable than male METH users in terms of high-energy phosphate neurochemistry. Further study is warranted to explore the relationship between the altered phosphorus metabolism and cognitive as well as psychiatric symptoms. Also, a longitudinal study design would shed light on within-subject changes in high energy phosphate metabolism following long-term abstinence and/or treatment. Efforts to characterize the relationship between the changes in high energy phosphates and psychiatric symptomatology have the potential to generate novel treatment strategies for METH toxicity.