Glial Dysfunction in Abstinent Methamphetamine Abusers

Abstract

Persistent neurochemical abnormalities in frontal brain structures are believed to result from methamphetamine use. We developed a localized ¹³C magnetic resonance spectroscopy (MRS) assay on a conventional MR scanner, to quantify selectively glial metabolic flux rate in frontal brain of normal subjects and a cohort of recovering abstinent methamphetamine abusers. Steady-state bicarbonate concentrations were similar, between 11 and 15 mmol/L in mixed gray-white matter of frontal brain of normal volunteers and recovering methamphetamine-abusing subjects (P>0.1). However, glial ¹³C-bicarbonate production rate from [1-¹³C]acetate, equating with glial tricarboxylic acid (TCA) cycle rate, was significantly reduced in frontal brain of abstinent methamphetamine-addicted women (methamphetamine 0.04 μmol/g per min (N=5) versus controls 0.11 μmol/g per min (N=5), P=0.001). This is equivalent to 36% of the normal glial TCA cycle rate. Severe reduction in glial TCA cycle rate that normally comprises 10% of total cerebral metabolic rate may impact operation of the neuronal glial glutamate cycle and result in accumulation of frontal brain glutamate, as observed in these recovering methamphetamine abusers. Although these are the first studies to define directly an abnormality in glial metabolism in human methamphetamine abuse, sequential studies using analogous ¹³C MRS methods may determine ‘cause and effect’ between glial failure and neuronal injury.

Introduction

The neuropathologic presentation of substance abuse and addiction, a serious health, social, and economic problem in the United States, involves abnormalities of the dopaminergic and glutamatergic systems leading to neuronal–glial dysfunction (Stephans and Yamamoto, 1994; Volkow et al, 2001, 2002). It has been commonly observed that objective neurochemical changes persist beyond the period of active drug use, inviting the hypothesis that persistent neurochemical abnormalities contribute to addictive behavior and relapse (Ernst et al, 2000; Sekine et al, 2002). The timing and possible role of glial injury in addiction and relapse is poorly understood but may have implications for the design of pharmaceuticals for future treatment of addiction and relapse prevention. Evolving methods of noninvasive brain analysis, in particular ¹³C magnetic resonance spectroscopy (MRS) with ¹³C glucose and ¹³C acetate, respectively (Bachelard, 1998; Bluml et al, 2001, 2002; Cerdan and Seelig, 1990; Gruetter et al, 1993; Lin et al, 2003; Mason et al, 1999), offer new opportunities to examine the impact of drugs of abuse on neurons and glia separately. This pilot study was undertaken to establish the feasibility of applying [1-¹³C]acetate MRS to quantify glial metabolic rate in frontal brain structures implicated in drug abuse, reward, and relapse (Gass and Olive, 2008). Future studies, in which the starting substrate would be [2-¹³C] glucose, could then be applied to probe the impact on neuronal metabolism directly (Gropman et al, 2009).

Glutamate (Glu) is an amino acid intimately involved in neurotransmission and a marker of mitochondrial oxidative status and tricarboxylic acid (TCA) cycle function. The glutamate–glutamine cycle links glia with neurons in the mechanism of glutamate neurotransmission, the neurotransmitter pathway overall believed to account for 80% of brain neurotransmitter function. The functions of glial cells in neurotransmission are often viewed as merely supportive; however, failure of glial reuptake of glutamate released at synapses can result in accumulation of extracellular glutamate, leading to neurotoxicity. Glutamate is crucial in regulating both the development and the expression of addictive behaviors, which requires glutamate receptor stimulation in the ventral tegmental area and is associated with enhanced glutamate release in the prefrontal cortex (Kalivas et al, 2009; Rebec and Sun, 2005). Using the single voxel proton MRS method, which determines steady-state concentration of brain metabolites, we found that preliminary data provide more direct evidence of elevated frontal brain glutamate concentration during the cerebral adaptation to methamphetamine abuse (Abulseoud et al, 2009). This is in addition to earlier reports of persistent neurochemical abnormalities; elevated myoinositol, a glial marker, and reduced N-acetylaspartate, a neuronal marker, observed in the frontal cortex of abstinent methamphetamine users (Ernst et al, 2000; Sung et al, 2007). Together these results indicate that neuronal injuries from drug abuse are likely accompanied by significant abnormalities in glial function and metabolism.

The aim of this pilot study was to determine the extent of glial metabolic abnormalities in abstinent methamphetamine-dependent (AMD) subjects using MRS. ¹³C MRS techniques, which permit TCA cycle rate determination directly and further allow the dissection of glial from neuronal pathways of glutamate metabolism based on the selective uptake and metabolism of acetate by glia (Muir et al, 1986) and glucose by neurons have recently been explored in intact brains of animals and humans (Bluml et al, 2002; Cerdan, 2003; Lebon et al, 2002; Ross et al, 2003; Rothman et al, 2003; Shen and Rothman, 2002). A limitation of previous studies in relation to the present purpose of drug abuse has been the need to confine human ¹³C MRS assays to posterior brain regions. This brain region is distant from heat-sensitive optical structures but purportedly of minimal relevance to the neuropsychopathology of drug abuse, which most investigators locate in frontal and limbic pathways (London et al, 2000). These technical limitations have recently been overcome (Sailasuta et al, 2008) with two distinct and robust in vivo ¹³C MRS techniques adapted to the human frontal lobe. First, the use of ¹³C isotopes of which the products can be readily observed with minimal or absent proton decoupling. Second, the development of low-power nuclear overhauser effects can now be used for the same purpose. Details of these technical advances and their efficacy in human studies using a conventional low field 1.5 T MR scanner have been published (Li et al, 2007; Sailasuta et al, 2008). In this report we describe the first such method designed to separately interrogate glial glutamate formation and specifically, to quantify the glial oxidative rate based on the appearance of ¹³C label from the unique glial precursor, acetate, into its final product bicarbonate (HCO₃⁻). We have determined the rate at which [1-¹³C]acetate appears in cerebral bicarbonate [H¹³CO₃] in normal frontal brain, and shown a marked reduction in the contribution of glial acetate oxidation to the TCA cycle in frontal brain of AMD subjects.

Materials and methods

Subjects

A total of 21 in vivo ¹³C MRS studies were performed. Eleven subjects participated in this study approved by institutional review board and HIPAA, of whom 6 were normal, age- and gender-matched subjects. The six healthy control subjects (mean age 26 ± 1 years, one man) were recruited from the local community. Three of the control subjects were scanned twice on separate days to compare ¹³C acetate metabolism in frontal and posterior brain. Five female AMD subjects (mean age 31 ± 7 years) recruited from local drug rehabilitation centers and counseling centers within the Los Angeles County were examined only once, in the frontal brain region. All five subjects reported methamphetamine as the drug of choice and met Diagnostic and Statistical Manual of Mental Disorders, fourth edition criteria for lifetime methamphetamine dependence from the Structured Clinical Interview (First et al, 1994). Self-report by study participants indicated the average period of methamphetamine use was 74 days and each had been abstinent from methamphetamine or other illicit drugs for at least 7 days before the studies. The study protocols were approved by the institutional review board of the Huntington Memorial Hospital and the University Hospital at University of Southern California. Written informed consent was obtained from all participants.

Carbon MRS

The techniques for quantification of glial metabolic rate of in vivo human brain ¹³C MRS using [1-¹³C]acetate as substrate have been described in detail previously (Bluml et al, 2002). Significant modifications necessary for safe transfer of the technique to frontal brain for this study were as follows:

• (i)

  To establish metabolic rates for [1-¹³C]acetate for the frontal lobe of human brain, not hitherto examined, we developed low rf power ¹³C MRS (Sailasuta et al, 2008). To minimize RF heating in a clinical magnetic resonance imaging system, all subjects underwent ¹³C MRS at 1.5 T GE clinical scanner (GE Healthcare, Milwaukee, WI, USA) equipped with broadband exciter, a stand-alone proton decoupler, and a vector signal generator (Agilent Technology E4433B, Santa Clara, CA, USA). A half-helmet head coil was used with the subject lying supine within the MR scanner. Its construction was a modification of the half-head coil used in the posterior brain ¹³C MRS study from this laboratory (Bluml et al, 2001). The coil consisted of a 4-inch proton loop and a 3.5-inch ¹³C loop for transmission and reception. Fast spin echo imaging scan was used to confirm the consistency of the observed volumes in all patients. For the frontal brain study, the half-helmet coil placement was such that the anterior portions of the frontal lobe were mainly included in the field of view and similarly for the posterior brain where the posterior region of the brain was included in the field of view (Gropman et al, 2009). The coil was tuned to best match both proton and carbon channels. Power deposition was continuously monitored and remained below Federal guidelines for Specific Absorption Rate.

• (ii)

  Because the method used is new, to confirm that metabolic rates for [1-¹³C]acetate in human frontal lobe are comparable to posterior parietal brain, values previously reported using [1-¹³C]acetate by us, and also recently confirmed using [2-¹³C] acetate by Lebon et al (2002), assays of ¹³C enrichment in the posterior brain were performed in four normal subjects. To allow for complete clearance of ¹³C-enriched metabolites from the posterior brain for three of the four normal subjects, this examination was followed a minimum of 48 h later by examination of the frontal brain region in each of the three normal subjects.

Magnetic resonance image was acquired at the outset of each ¹³C MRS examination, indicating regions of interest of mixed white and gray matter of posterior parietal or frontal and prefrontal brain structures, respectively, for ¹³C MRS acquisition. Brain volumes, approximately 100 cm³ in each protocol, included similar proportions of white and gray matter, were comparable and have been previously published (Gropman et al, 2009; Sailasuta et al, 2008).

Determinations of Cerebral Bicarbonate Concentration from Natural Abundance ¹³C MRS and Maximum Rate of HCO₃⁻ (^maxV_HCO3) Production

Natural abundance ¹³C MRS was acquired for 20 mins before the start of the acetate infusion to show adjacent, well-resolved individual resonances for bicarbonate and for Cr + PCr (total creatine, tCr) as previously reported by Bluml et al (2001). Because tCr has been used successfully as an internal metabolite reference in several proton MRS studies in abstinent methamphetamine users (Nordahl et al, 2005, 2002; Salo et al, 2007) as well as other human studies (see review, Xu et al, 2008), and does not become enriched during the time scale of the present studies, we have used this as an internal reference to quantify HCO₃⁻ and its subsequent enrichment from [1-¹³C]acetate. To improve signal to noise (natural abundance ¹³C = only 1.1% of total) and thereby improve quantification, we acquired additional natural abundance ¹³C brain spectra during separate imaging sessions not associated with ¹³C infusions. The timed intravenous infusion of 99% ¹³C-enriched [1-¹³C]acetate infusion was started and brain spectra were acquired continuously thereafter for a minimum of 2 h. The 6½ min blocks of low-power noise nuclear overhauser effects (0.9 W for frontal brain and 2.5 W for the posterior brain) ¹³C MRS spectra acquired were stored and summed, as previously described (Sailasuta et al, 2008). ¹³C MRS examination times of 120 mins or more were well tolerated by all subjects with no observed adverse reactions. Maximum rate of HCO₃⁻ production (^maxV_HCO3⁻) was calculated from the natural abundance bicarbonate concentration at baseline and at the end of the infusion estimated from the ratio of the areas under the two resonances, HCO₃⁻ (161 p.p.m.) and tCr (158 p.p.m.) using a Marquardt fitting algorithm and assuming tCr concentration 11 μmol/g in the brain.

Infusion Protocol

The 99% ¹³C-enriched [1-¹³C]acetate solution was prepared under good manufacturing practice by the Cambridge Isotope Laboratories (Andover, MA, USA) and shipped in bulk to a licensed pharmacist in Minnesota, registered and approved by the US Food and Drug Administration to ship sterile, pyrogen-free reagents to California. The reagent was prepared on a named-subject basis and used within 48 h after preparation. This study was performed under the Food and Drug Administration Investigational New Drug Application no. 59690. To establish similar metabolic status, subjects were fasted for 6 h before the study. Intravenous administration (through an arm vein) of 120 to 300 ml of 0.4 mol/L sterile, pyrogen-free solution of [1-¹³C]acetate 99% enrichment was started (OA) at an infusion rate of 3 mg/kg per min for 60 mins.

Fractional Enrichment of Serum [1-¹³C]Acetate

Because earlier preclinical and clinical studies have established that ¹³C acetate consumption proceeds rapidly in the brain and the [1-¹³C]acetate resonance cannot readily be distinguished from its primary metabolic product, 5-¹³C glutamate on the basis of near-identical chemical shift, plasma ¹³C acetate enrichment was used as a surrogate marker of the fractional ¹³C enrichment (E) of stable isotope availability for glial oxidation. Serial blood samples were collected using heparinized syringe before infusion and at 15 min intervals until the end of the study (at 120 mins). Blood samples were centrifuged and the resulting protein-free plasmas were stored frozen until analysis. Plasma samples were deproteinized using 2:1 volume ratio of 99.9% ethanol to plasma and dried under vacuum. Samples were dissolved in 99.9% D₂O before proton NMR measurements using a Varian 300 MHz spectrometer. Fractional enrichment of plasma acetate was determined under fully relaxed conditions (repetition time 18 s) from the ratio of the relative peak areas of the ¹³C satellites to the total area of the corresponding acetate H2 resonance at 1.9 p.p.m. (Moreno et al, 2001).

Data Processing and Fractional Enrichment Calculation of Cerebral Metabolites

The principal metabolic end point of this study was the quantification of intrinsic cerebral bicarbonate (HCO₃⁻), together with the progressive increase in fractional ¹³C enrichment (%E) of the HCO₃⁻ pool as the result of complete oxidation of the sole ¹³C-enriched precursor of glial TCA cycle activity, [1-¹³C]acetate. Because ¹³C creatine and phosphocreatine do not become enriched from ¹³C acetate (Cerdan, 2003; Chapa et al, 1995) and resonate distinctly in a portion of the brain spectrum unencumbered by overlying resonances, we used ¹³C total signal intensity for [Cr + PCr] to quantify the initial brain bicarbonate concentration in nonenriched baseline ¹³C MR scans from each subject. We assume there are no significant differences in T1 or T2 relaxation times between the two. Fractional enrichment of HCO₃⁻ is then defined as the increase in intensity relative to that of [Cr + PCr] which is not enriched from the infused source of [1-¹³C]acetate. Progressively increasing intensity of ¹³C HCO₃⁻ resonance was recorded throughout and the relative rate of enrichment was expressed as percent over time, slope of initial increase, and final %E bicarbonate. The other anticipated cerebral metabolites of [1-¹³C]acetate, glutamine C1 and C5, and glutamate C1 and C5, were also readily observed. An observer-independent automatic IDL-based software developed in our laboratory (Shic and Ross, 2003) and SAGE (GE Healthcare, Milwaukee, WI, USA)/IDL (ITT Visual Information Solutions, Boulder, CO, USA), a commercially available software, were used for data processing (NS, KH). Data blocks (6½ min) were first zero filled to 16K data points, applied 7 Hz Lorenzian to Gaussian lineshape transformation, Fourier-transformed, and phase-corrected. Owing to the absence of overlapping resonances, relative contributions of ¹³C singlet of HCO₃⁻ (161 p.p.m.), C1-acetate and C5-Glu (182 p.p.m.), and C5-Gln (178 p.p.m.) were readily determined from peak amplitudes of the observed resonances. In the case of C1-acetate and C5-Glu, no separation of the resonances could be achieved under near-steady-state conditions. Based on earlier studies, the contribution of [1-¹³C]acetate cerebral flux, %E ¹³C brain was considered approximately %E acetate of blood.

¹³C fractional enrichments (E) of metabolic products were calculated from:

$E_{\mathrm{m}\mathrm{e}\mathrm{t}}=S_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}/S_{\mathrm{p}\mathrm{r}\mathrm{e}}\times 1.1\mathrm{\%}$

where S_post and S_pre are the signal intensities of the resonances in the baseline scan and in scans after infusion start, and 1.1% is the natural abundance of ¹³C.

Determination of the Glial Acetate Oxidation Rate (V_{g ac})

For consistency with earlier studies, results are expressed as glial acetate oxidation rate (V_{g ac}). Acetate has been shown to be metabolized exclusively in glia, probably as a result of the unique membrane transporter present in glia but lacking in neurons (Muir et al, 1986; Waniewski and Martin, 1998). During the first turn of the TCA cycle, C1 label (from C1-acetate) is transferred to form glial C5 and C1 glutamate and transformed by glial glutamine synthase to C5 and C1 glutamine. C1 label CO₂ (HCO₃⁻) is released in the second turn of the TCA cycle (Cerdan et al, 1990). Calculation of oxidative rate from [1-¹³C]acetate to the final oxidative product H¹³CO₃ therefore involves no assumptions concerning neuronal compartmentation or flux as proved necessary for calculations of [2-¹³C]acetate metabolism (Lebon et al, 2002). Nevertheless, the two methods appear to yield comparable estimates for average glial metabolic rate at approximately 10% to 15% of the rate of normal neuronal TCA cycle (Shen et al, 1999). In this study we used the approach described previously by Bluml et al (2002) (and Figure 1): assuming that metabolic pools that accumulate ¹³C label other than bicarbonate have constant ¹³C fractional enrichment and the total influx of ¹³C label metabolites into the glial and neuronal TCA cycles is equal to the total outflux into the bicarbonate pool, the glial acetate oxidation rate, V_{g ac}, was determined using the following formula:

$V_{\mathrm{g}\mathrm{a}\mathrm{c}}=V_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}\times (\mathrm{\%}{E}_{\mathrm{H}\mathrm{C}{\mathrm{O}}_3}-1.1)/(\mathrm{\%}{E}_{Ac}-1.1)$

The total bicarbonate production, V_total, is given by

V_tot = 3 × V_tca + 2 × V_{g ac}, where V_tca is the rate of glucose oxidation in neuron and glia (V_{n glc} + V_{n glc}) (Figure 1). Because V_{g ac} < < V_tca, V_tot is approximately 3 × V_tca.

Figure 1

Glial versus neuronal metabolism: pathway of ¹³C label from [1-¹³C]acetate to tricarboxylic acid (TCA) cycle in human brain. Selective metabolism of acetate and glucose occurs in glia and neurons, respectively. Only the relevant intermediates are shown. Glia selectively transports acetate which is then metabolized through acetate thiokinase and reactions of the mitochondrial tricarboxylic acid (TCA) cycle forming ¹³C glutamate and glutamine, enriched in the C1 and C5 positions, before final oxidation and release as ¹³C bicarbonate (CO₂ = HCO₃⁻). Five of the relevant cerebral metabolic products are showed in sequential ¹³C brain spectra of Figure 3. The alternative metabolic pathways of ¹³C glucose metabolism, predominantly in neurons, but not contributing to this study in which only the single substrate, [1-¹³C]acetate is provided. Ac, acetate; Glc, glucose; Glu, glutamate; Gln, glutamine; CO₂, carbon dioxide; Pyr, pyruvate; V_{g glc}, glial glucose oxidation; V_{g ac}, glial acetate oxidation; α-KG, α-ketoglutarate; V_glu, glutamate synthesis rate; V_gln, glutamine synthesis rate. Glucose is metabolized in neuron and glia whereas acetate is metabolized exclusively in glia.

An additional method of calculating maximum glial TCA cycle rate was applied, determining brain bicarbonate in natural abundance and ¹³C-enriched spectra with reference to the adjacent Cr/PCr resonance (see above).

Statistical Analysis

Statistical analyses were performed with StatView software (SAS Institute, Cary, NC, USA). For continuous variables, the means and standard deviations were calculated separately for the AMD and the normal subjects. An analysis was performed to show that variance in the two groups was the same using F test, followed by analysis with the two-sample t-test. All comparisons yielding a P-value of less than 0.05 were considered statistically significant. All data are expressed as mean ± standard deviation (s.d.) in the text.

Results

Determination of Cerebral Bicarbonate Concentration from Natural Abundance ¹³C MRS

Cerebral [HCO₃⁻] calculated in a contiguous volume of mixed white and gray matter of the human frontal lobe was found to be between 11 and 15 mmol/L (Figure 2), values consistent with literature data and comparable to the those obtained in earlier direct and indirect assays in simian and animal brains (Siesjo, 1964). Results obtained by direct ¹³C MRS assay in ‘frontal brain’ were also indistinguishable from those obtained in ‘posterior brain’ in the same group of control subjects (Table 1). Results for cerebral bicarbonate concentration in the cohort of recovering methamphetamine-abusing subjects (Table 2) fell within this ‘normal’ range indicating that metabolic pH homeostasis is broadly normal in the recently abstinent methamphetamine-abuse human brain.

Figure 2

Human ¹³C MR frontal brain spectra acquired during [1-¹³C]acetate infusion: carbon-13 magnetic resonance spectroscopy (MRS) at baseline and 120 mins after the infusion (154 to 164 p.p.m.) in a control and an abstinent methamphetamine-dependent (AMD) subject. ¹³C MRS spectra show HCO₃⁻ and total Cr (tCr) chemical shift region (154 to 164 p.p.m.) at baseline and at the end of the infusion study in a normal control subject (A) and an AMD subject (B). The lower, natural abundance spectra for the normal control (left) indicate that cerebral bicarbonate concentration is very similar to that of [Cr+ PCr], i.e., 11 mmol/L. The resting cerebral bicarbonate concentration in a representative AMD subject (right) was indistinguishable from the control (HCO₃ = 11 mmol/L). After 2 h of [1-¹³C]acetate infusion, substantial but different increases in cerebral ¹³C bicarbonate intensity were observed. Again, by comparing the adjacent, unenriched Cr + PCr, the rate of enrichment of cerebral bicarbonate was substantially lower in the AMD (right) than in the normal control (left). These representative data are seen to be statistically significantly different when normal and AMA subjects are compared in Tables 1 and 2. In the control subject, the concentrations of bicarbonate determined in this way were 15 μmol/g at baseline and 49 μmol/g at the end of the 2 h infusion study (Cr does not become enriched) whereas in the AMD subject initial bicarbonate was 10 μmol/g at baseline but reached only 25 μmol/g after 2 h.

Table 1

Fractional bicarbonate enrichment and glial metabolic rate for frontal and posterior brain in controls

ID	[HCO3] (brain, μmol/g)	%EAc (plasma)	Initial slope a	%E HCO3− at 60 mins (brain)	%E HCO3− at 80 mins (brain)	%E HCO3− at 120 mins (brain)	Vg ac (μmol/g per min) b	Vg ac from Bluml et al c (μmol/g per min)
Control posterior brain (N = 4)	12 ± 1.3	78 (N = 1)	0.11 ± 0.06	4.74 ± 1.1	5.2 ± 1.3	6.2 ± 1.6	0.09 ± 0.03	0.13 ± 0.03 (N = 3)
Control frontal brain (N = 5)	NA	78 (N = 1)	0.10 ± 0.02	5.1 ± 0.7	5.9 ± 0.8	7.0 ± 1.0	0.11 ± 0.01 d
P (frontal versus posterior brain)		NA	NS	NS	NS	NS	NS

NA, not available; NS, not significant.

Values are mean ± standard deviation (N = number of examinations).

a

Because the time course of the appearance of HCO₃⁻ signal amplitude was linear from the start of the infusion to 40 mins, the slope of the straight line was calculated and reported here.

b

Determined using %E at 60 mins and total TCA cycle rate of 0.7 μmol/g per min (Mason et al, 1995).

c

Value reported by Bluml et al (2002) determined at 60 mins after start of infusion in the posterior brain.

d

V_{g ac} calculated at 80 and 120 mins are 0.13 and 0.16 μmol/g per min.

Table 2

Fractional bicarbonate enrichment and glial metabolic rate for frontal brain in control and AMD subjects

Subjects	Mean age/gender	[HCO3] (brain, μmol/g)	%EAc (plasma)	Initial slope a	%E HCO3 at 60 mins (brain)	%E HCO3 at 80 mins (brain)	%E HCO3 at 120 mins (brain)	Vg ac ( μmol/g per min) b
Control frontal brain (N = 5)	26 ± 1	NA	78 (N = 1)	0.10 ± 0.02	5.1 ± 0.7	5.9 ± 0.8	7.0 ± 1.0	0.11 ± 0.01
	5 F
AMD frontal brain (N = 5)	30 ± 7	10 c	83 (N = 1)	0.05 ± 0.02	2.7 ± 0.5	3.14 ± 0.5	3.79 ± 0.6	0.04 ± 0.01
	5 F
P (control versus AMD)	0.12	—	NA	0.002	< 0.001	< 0.001	< 0.001	< 0.001

AMD, abstinent methamphetamine-dependent; F, female; NA, not available.

Values are mean ± standard deviation (N = number of examinations).

a

Because the time course of the appearance of HCO₃⁻ signal amplitude was linear from the start of the infusion to 40 mins, the slope of the straight line was calculated and reported here.

b

Determined using %E at 60 mins and total TCA cycle rate of 0.7 μmol/g per min (Mason et al, 1995).

c

Natural abundance [HCO₃⁻] could not be determined with accuracy (low signal-to-noise ratio) from individual AMD subjects, the reported value was determined from summed spectra from all five AMD subjects, and appeared to be indistinguishable from controls.

Determination of the Maximum Rate of Bicarbonate Production (^maxV_HCO3) in Human Brain

In contrast, using time course data in the two groups of subjects, it was observed that the rate at which [1-¹³C]acetate was incorporated into the final oxidation product bicarbonate was markedly different between drug-abusing and normal subjects. Figure 2 compares ¹³C MRS spectra (154 to 164 p.p.m.) from a representative control subject (A) and a methamphetamine abstinent subject (B). From the start (natural abundance) to the end of the 2 h data acquisition, there was almost fivefold increase in ¹³C incorporation (fully ¹³C-enriched) in the control, whereas in the AMD subject, the full extent of ¹³C enrichment was closer to threefold. Natural abundance bicarbonate concentration at baseline and at the end of the infusion estimated was similar in the two subjects. In the control subject, the concentrations of bicarbonate determined in this way were 15 μmol/g at baseline and 49 μmol/g at the end of the 2 h infusion study (Cr does not become enriched) whereas in the AMD subject initial concentration of bicarbonate was similar, 10 μmol/g, at baseline but reached only 25 μmol/g after 2 h. This was less than 50% of the enrichment measured in the normal control. The baseline bicarbonate concentration is within the range of normal values calculated from Henderson–Hasselbach equation, pH = pK_a + log ([HCO₃]/[CO₂]) (Agabiti et al, 1973) using brain pH of 6.99 (Bluml et al, 1998), pK_a = 6.1 and pCO₂ of 35 to 48 mm Hg in normal adults brain of 11.2 to 19.9 μmol/L. The maximum rate of bicarbonate enrichment (^maxV_HCO3⁻) in the frontal brain in a control was 0.28 μmol/g per min compared with an AMD subject of 0.13 μmol/g per min. This result is highly suggestive of a persistent major abnormality in glial oxidative mechanism in frontal brain of abstinent methamphetamine subjects and is consistent with previously published work. For a precise evaluation of the impact of AMD on glial metabolic rate, we applied the previously used method for calculation of V_{g ac}.

Comparison of Fractional Enrichment of Bicarbonate and Glial Acetate Oxidation Rate in the Posterior and Frontal Brain of Normal Subjects

Five control women were scanned using the frontal brain protocol and three of these women and one man were scanned using the posterior brain protocol. Summary of the results is shown in Table 1. Initial slope of H¹³CO₃ appearance in brain was similar for frontal and posterior brain (P = 0.25), as were fractional enrichment of bicarbonate at 60, 80, and 120 mins of observation. The calculated V_{g ac} from the normal posterior brain (0.09 ± 0.03 μmol/g per min) was in excellent agreement with value previously obtained by us. Finally, the calculated rate of frontal brain [1-¹³C]acetate oxidation (0.11 ± 0.01 μmol/g per min) was indistinguishable from posterior brain.

Appearance of ¹³C-Enriched Metabolites of [1-¹³C]Acetate in the Frontal Brain of Normal and AMD Subjects

Figure 3 is a time course of the appearances of 5-¹³C glutamate and [1-¹³C]acetate (A), 5-¹³C glutamine (B), and HCO₃⁻ (C) resonances in a control subject. Similar plots were obtained for AMD (not shown). Rapidly increased to peak amplitudes for the enriched acetate and 5-¹³C glutamate as well as 5-¹³C glutamine as were observed at 56 mins and slowly eliminated by approximately 120 mins. Bicarbonate signal amplitude reached a maximum and remained essentially constant after 60 mins. However, the pseudo-steady state for 5-¹³C glutamate could only be indirectly verified because of the co-resonance of the precursor [1-¹³C]acetate. Nevertheless, these observations confirmed that all of the ¹³C label entering the glial TCA cycle is being transferred to bicarbonate and is not sequestered in other metabolic pools.

Figure 3

Time courses of glutamate C5 and acetate (A), glutamine C5 (B), and HCO₃⁻ (C) from a healthy subject. Signal intensities are shown in arbitrary units. Note that glutamine C5, glutamate C5, and bicarbonate become increasingly enriched over time, consistent with the expected path of glial metabolism of ¹³C acetate shown in Figure 1. After significant enrichment up to 60 to 80 mins, pools of glutamate C5 and glutamine decline, presumably as they are in turn metabolized to the final product, ¹³C HCO₃⁻. No significant sequestration in glutamate and glutamine occurs.

Figure 4 shows stack plots of labeled metabolites comparing results obtained in a normal control (A) with those from a representative AMD subject (C), over time and below, and the summed spectra of the total brain enrichment at the end of the infusion protocol for control (B) and AMD (D). C5 glutamate (and acetate), C5 glutamine, C1 glutamate–glutamine and bicarbonate resonances are clearly visible. The strikingly reduced appearance of ¹³C label in bicarbonate is evident in Figure 2D, in the AMD subject compared with control in Figure 2B. This result was confirmed in the entire group of AMD. The enrichment of HCO₃⁻ estimated from the time course of the fractional enrichment of HCO₃⁻ was significantly decreased in AMD (Figure 5) and the rate of acetate oxidation in glial, V_{g ac}, as well as %E HCO₃⁻ was quantified from the initial slope of bicarbonate enrichment versus time, or from %E at 60, 80, and 120 mins (compare Table 2 with Table 1). Comparison of bicarbonate productions between normal and AMD subjects is shown in Figure 6. Assuming the overall TCA cycle rate of 0.7 μmol/g per min for the intact human brain (Mason et al, 1995), the average rate of glial acetate oxidation in the frontal brain was 0.04 for AMD subjects compared with 0.11 μmol/g per min in normal subjects, a reduction of 60% in glial metabolism rate for acetate oxidation. There were no significant differences (P = 0.12) between gender and age in the two subject groups, normal controls and AMD subjects.

Figure 4

Effect of methamphetamine abuse on cerebral [1-¹³C]acetate metabolism. Frontal brain ¹³C metabolites observed during [1-¹³C]acetate infusion: figure shows carbon-13 MRS (150 to 195 p.p.m.) in a control and an abstinent methamphetamine-dependent subject after enrichment with [1-¹³C]acetate. Stacked sequential carbon-13 spectra and spectra summed for the entire infusion protocol (120 mins, bottom) are shown for a control (A, B) and an abstinent methamphetamine-dependent (AMD) subject (C, D). The spectra identified enriched cerebral metabolites glutamate C5 +C1 acetate (182 p.p.m.), glutamine C5 (178 p.p.m.), glutamate C1 +glutamine C1 (175 p.p.m.), and the final oxidation product, bicarbonate (161 p.p.m.).

Figure 5

Effect of methamphetamine abuse on cerebral metabolism of ¹³C acetate. Time courses of appearance of the end product of [1-¹³C]acetate oxidation, HOC₃⁻ are compared. Low-power noise nuclear overhauser effects (NOE) spectra were acquired in 6.5 min blocks for up to 120 mins after enrichment by i.v. acetate (3 mg/kg body weight) from frontal brain in five abstinent methamphetamine-dependent (AMD) subjects and five control subjects. Logarithmic fits of the averaged percent fractional enrichment of bicarbonate (%E) were plotted as a function of time after start of the infusion. %E in AMD (lower trace) is shown to be lower than controls (upper trace). Standard deviations are also shown as cross-bars.

Figure 6

Comparison of methods to express effect of abstinent methamphetamine-dependent (AMD) in ¹³C HCO₃⁻ production from [1-¹³C]acetate in human frontal brain. Results shown in Figure 5 are quantified. Means and standard deviation of percent fractional bicarbonate enrichment calculated at three different time points (60, 80, and 120 mins) and the initial slope of bicarbonate production from control and AMD groups. P-value in each group comparison is < 0.001.

Discussion

In this pilot study we confirm the unique value of low-power ¹³C MRS applied for the first time to methamphetamine abusers in frontal regions of the human brain. We have discovered a hitherto unsuspected 50% reduction in oxidative rate in the recovering brain after methamphetamine abuse. Although it has been known for many years that acetate, which is a preferred glial respiratory fuel, can be applied to defining glial TCA cycle rate of the normal-functioning human brain in vivo, to the best of our knowledge, this study represents the first systematic attempt to define alterations in glial metabolic rate as a result of neuropathology. Several earlier studies have used ¹³C MRS with ¹³C glucose to explore human brain neurochemical abnormality. [1-¹³C]glucose, a predominant neuronal respiratory fuel, has revealed significant abnormalities of glutamate enrichment and TCA cycle rate in diseases as diverse as Alzheimer's, hepatic encephalopathy, mitochondrial disorders, and epilepsy (Bluml et al, 2001; Lin et al, 2003; Ross et al, 2003). However, in each of these examples the neuron was the target of investigation. Furthermore, the region of brain examined was the posterior parietal lobe, and suboptimal for frontal lobe study based on safety considerations dictated by Federal guidelines for Specific Absorption Rate. The present report appears more relevant both as a study focused directly on the question of deficiency in glial oxidative metabolism and by exploring a brain region implicated in drug abuse abnormality by almost all independent neuroscience techniques including executive function and neuropsychologic metrics. Acetate is metabolized exclusively in glia (Muir et al, 1986; Waniewski and Martin, 1998); it is reasonable to assume a single-compartment model for cerebral bicarbonate production, that is, the rate of bicarbonate production is equal to the rate of acetate influx into the glial TCA cycle. Intravenous infusion of [1-¹³C]acetate rapidly produced cerebral C5 glutamate in the first turn of the TCA cycle where it is co-resonant and indistinguishable from the acetate at 182 p.p.m. Glutamine C5 (at 178 p.p.m.) appeared in the second turn of TCA cycle and glutamate C1 and glutamine C1 (both resonates at 175 p.p.m.) in a subsequent turn. Carbon dioxide is the final product of the TCA cycle that appeared as bicarbonate at 161 p.p.m. Because of our low [1-¹³C]acetate infusion dosage, steady states were not convincingly reached in our experimental studies, therefore to determine fractional enrichment of bicarbonate, an exponential function was fitted to the time course of the production of bicarbonate.

The glial bicarbonate production rate was subsequently calculated from exponential fitting of %E at 60 mins and therefore represents the glial TCA cycle rate. The glial TCA cycle rate determined this way represents approximately 10% of the total TCA cycle rate in the brain of fasted normal subjects. This rate is similar to the previously reported value by us in the parietal brain region using similar [1-¹³C]acetate protocol (Bluml et al, 2002).

Information regarding abnormalities of glial cells can typically be obtained with proton MRS. There are few MRS studies (nine studies from 2000 to 2008) on methamphetamine abuse that reveal abnormalities in metabolite concentrations in frontal cortex (Ernst et al, 2000), frontal white matter (Ernst et al, 2000; Nordahl et al, 2002), and basal ganglia (Sekine et al, 2002). Proton MRS studies show that methamphetamine-dependent subjects had decreased N-acetylaspartate, higher myoinositol, and elevated glutamate in the frontal white matter. Possible interpretations of these metabolite abnormalities in these subjects include local loss or dysfunction of neurons, and glial abnormalities (Abulseoud et al, 2009). From ¹³C MRS, there is a significant reduction of glial TCA cycle rate in AMD subjects compared with non-drug users, further emphasizing the impact of methamphetamine use on glial function.

Possible limitations to ¹³C MRS: as with all ‘isotope’ techniques, there is an unknown fraction of cerebral metabolism that will continue to consume (unenriched) endogenous respiratory fuels, in this case predominantly C-12 glucose. This may ‘dilute’ the resulting ¹³C bicarbonate pool derived from ¹³C-enriched acetate. One correction has been applied, as suggested by Bluml et al (2002) in expressing the ¹³C HCO₃⁻ as a fraction of the total bicarbonate produced in unit time. Although there is no way with the present technique to ensure that overall cerebral metabolic rate is normal in AMD (additional studies using ¹³C glucose will clarify the point further), nevertheless, it is difficult to envisage a more than 50% change in intrinsic glucose metabolic rate that would be necessary to explain the present findings. Earlier positron emission tomography studies of AMD argued against any significant increase in neuronal metabolic rate in frontal brain (Kim et al, 2005, 2009; London et al, 2005). Contrary results (small with reference to the present results) were obtained by Wang et al (2004). There are other limitations to our study. First, the study sample is small; though the effect is large and statistically robust. This is partly a consequence of the costs ([1-¹³C]acetate infusion) and technical difficulty of the present procedures. In the future, both aspects promise to be greatly simplified allowing large-scale replication of the present findings in this and other studies. Another possible limitation of the study is the assumption that fractional enrichment of [1-¹³C]acetate is similar between the brain and blood (Bluml et al, 2002). Note that the enrichments of Glu1, Gln1, as well as Gln5 were also observed but under our non-steady-state experimental condition, we did not generate kinetic data for metabolic rate calculations used in more complex theoretical models. Alternative, steady-state protocols, involving larger doses of [2-¹³C] acetate and longer protocols may lend themselves to resolving any remaining kinetic issues. Finally, the study of short-term abstinence may limit the generalizability of results to abuse versus craving and relapse in study of methamphetamine. We propose that this successful demonstration of a hitherto unrecognized but sizeable neurometabolic adaptation is worthy of further consideration and more sophisticated experimental design that may reveal novel mechanisms relevant to human drug abuse and many other conditions currently ascribed to glial dysfunction.

In this preliminary study, we used a novel method of frontal lobe ¹³C MRS to explore possible defects that may persist in the glia of drug-addicted subjects, using as the example, recently AMD subjects. Additional studies, using the neuronal substrate [2-¹³C] glucose, will be necessary to determine the cellular specificity and the frontal localization of these neurochemical injuries. The novel capability to dissect two major cell populations at risk in the intact brain and in patients at varying stages of the abuse–addiction–recovery cycle of methamphetamine abuse promises new insights into an intractable clinical problem.