Alteration of Intracellular Metabolite Diffusion in Rat Brain In Vivo During Ischemia and Reperfusion
Background and Purpose Diffusion-weighted MRI can demonstrate decreases of the apparent diffusion coefficient (ADC) of brain tissue water shortly after the onset of ischemia. To further elucidate underlying mechanisms, this study extended diffusion assessment to intracellular metabolites in rat brain in vivo before, during, and after ischemia.
Methods Changes in molecular mobility were studied in a rat model of global forebrain ischemia (n=8, 20-minute occlusion, 120-minute reperfusion) with the use of diffusion-weighted localized proton MR spectroscopy. During ischemia and early reperfusion the time course of ADC changes was monitored by strongly diffusion-weighted spectra. ADC values of N-acetylaspartate, creatines, cholines, and myo-inositol were evaluated from series of differently diffusion-weighted spectra before ischemia, 90 minutes after reperfusion, and 60 minutes postmortem.
Results Parallel to a rise in diffusion-weighted water signal (133±20%), pertinent intensities of all brain metabolites increased during ischemia. Changes were most pronounced for myo-inositol (46±9%) and smallest for N-acetylaspartate (12±4%). During reperfusion water ADC values returned to basal values, whereas metabolite ADC values were decreased by 22% (after 40 minutes). Postmortem ADC values (after 60 minutes) were reduced by 46% for water and 38% for metabolites.
Conclusions The present findings indicate that water ADC changes during ischemic stroke are accompanied by significant alterations in intracellular mobility in both neuronal and glial cell populations as reflected by N-acetylaspartate and myo-inositol, respectively. Altered metabolite ADC values during reperfusion are consistent with irreversible tissue damage in this model and offer new means to assess circulatory and metabolic compromise.
Diffusion-weighted MRI can detect cerebral ischemia at very early stages by contrasting normal and ischemic tissue with regard to the mean positional displacement of water molecules.1 2 The method relies on changes in the ADC of brain water that develop within the first 10 minutes after the onset of ischemia. The phenomenon may be exploited for early and sensitive imaging of acute human stroke3 preceding mapping by other noninvasive modalities. However, the pathophysiological reasons for the observed reduction in water mobility as well as the diagnostic implications of water ADC changes for tissue viability are still unclear.
Possible mechanisms include the occurrence of restricted diffusion,1 changes in membrane permeability,4 a decrease in brain temperature,1 5 and cessation of blood pulsation.2 Thus far, however, the most plausible explanation stems from the development of cytotoxic edema caused by energy failure.1 Since water has a much higher diffusion coefficient in extracellular than intracellular compartments and since its membrane permeability is sufficient to ensure free exchange for diffusion times of more than 20 ms commonly used in diffusion-weighted MRI, the resulting ADC value represents a weighted mean of extracellular and intracellular diffusion coefficients.6 7 This value decreases for a net volume shift of water from extracellular to intracellular spaces. However, since accurate extracellular and intracellular diffusion coefficients of water are not known, it is not yet clear whether cell swelling alone accounts for the ADC changes in ischemic tissue.
One way of gaining further insight into underlying mechanisms of water ADC changes is to separate intracellular from extracellular contributions by focusing on intracellular metabolites. This can be accomplished by localized proton MRS in vivo, which, for example, allows detection of NAA and myo-inositol exclusively located in neuronal8 and glial9 cell populations, respectively. This study therefore used diffusion-weighted localized proton MRS to probe putative changes in intracellular mobility during ischemia.
Materials and Methods
Ischemia was induced in male Wistar albino rats (n=8; weight, 300 to 400 g) with the use of a four-vessel occlusion model.10 The day before the experiment the vertebral arteries were occluded by electrocoagulation under halothane anesthesia (1% in 3:7 O2/N2O). For MRI and MRS the animals were anesthetized, tracheotomized, and artificially ventilated with the use of an animal respirator (Rhema). The carotid arteries were exposed and nylon threads placed around them. Rats were placed in a supine position with their skull on the surface coil used for signal reception and their snout fixed with a bite bar. During the course of the experiments ischemia was induced in the magnet by pulling the nylon threads tightly. Release of the threads after 20 minutes was followed by at least 120 minutes of reperfusion. Subsequently, the animals were killed (intravenous injection of 5% KCl), and further diffusion measurements were performed postmortem. All procedures were in accordance with German animal protection laws and approved by the responsible governmental authority.
Monitoring of arterial blood pressure (Siemens) and determination of plasma glucose (Beckman) and blood gas and pH (AVL) were accomplished with a femoral artery catheter. Respiratory motion was controlled by a water-filled balloon placed firmly against the chest of the animals. Body temperature (37±1°C) was maintained by a heated water blanket. Table 1⇓ summarizes the physiological parameters during the course of the experiments.
All studies were performed with the use of a 2.35-T magnet (Bruker Biospec) and actively shielded gradients of up to 50 mT · m−1 strength (Oxford Instruments). Diffusion-weighted localized proton MRS involved a STEAM sequence11 (TE, 120 ms; TM, 30 ms) with bipolar diffusion gradients g in both TE/2 intervals to keep diffusion times short (Δ=δ=25 ms) and minimize motion-induced phase errors and associated signal loss.12 Corresponding gradients along the z axis of the magnet resulted in diffusion weightings along the head-to-foot direction of the animals with b factors ranging from 150 to 2960 s · mm−2 (b=γ2g2δ2[Δ−δ/3], γ=42.57 MHz · T−1). Acquisitions were triggered to the respiratory motion of the animals, yielding TR of 3000 ms or greater.
Fig 1⇓ shows coronal and sagittal MRI sections indicating the typical location of a 7.5×5.0×7.5-mm3 (0.28-mL) volume of interest selected to provide a sufficient signal-to-noise ratio for the most strongly diffusion-weighted spectra. Fully automated and user-independent evaluation of resonance areas from NAA, Cr, Cho, and myo-inositol was based on an analysis by LCModel, taking advantage of a linear combination of respective model metabolite spectra.13
ADC values of water and metabolites were determined before ischemia under basal conditions, after approximately 90 minutes of reperfusion, and 60 minutes postmortem. Pertinent measurements comprised the acquisition of 5 spectra (128 accumulations each) with increasing diffusion weightings. Since total measuring time for an ADC determination took approximately 45 minutes, the time course of diffusion effects immediately before, during, and after ischemia was assessed by strongly diffusion-weighted spectra (b=2960 s · mm−2). Interleaved acquisitions of diffusion-weighted metabolite spectra (64 accumulations) and water spectra (1 scan) were obtained, the latter controlling the efficacy of the occlusion in comparison to previously established reductions in water ADC.
Fig 2⇓ shows both mildly and strongly diffusion-weighted metabolite spectra as well as respective analyses by LCModel. Time courses of decreases in ADC values of both water and metabolites, ie, of corresponding signal intensity increases in strongly diffusion-weighted spectra, are demonstrated in Fig 3⇓. Quantitative data are summarized in Table 2⇓, which relates averaged signal intensities during ischemia (period from 5 to 20 minutes only) and at two stages of reperfusion (10 to 25 minutes and 25 to 40 minutes) to mean signal strengths obtained before ischemia. To avoid signal averaging during transition phases, the first 5 minutes after occlusion and the first 10 minutes of reperfusion were omitted (compare Fig 3⇓).
In the acute phase, metabolite signal alterations were strongest for myo-inositol (glial marker) and smallest for NAA (neuronal marker), with intermediate values for Cr and Cho present in both cell types. During reperfusion, water signals nearly normalized in particular when compared with the 133% increase during ischemia. In contrast, however, after an initial recovery metabolite signals tended to develop persistent changes. Although changes were less pronounced than during the acute phase (30%), the mean values for NAA, Cr, Cho, and myo-inositol increased from 6.5% (10 to 25 minutes of reperfusion) to 17.5% (25 to 40 minutes).
This observation was confirmed by water and metabolite ADC determinations before ischemia, during reperfusion, and postmortem, as shown in Table 3⇓. While the water ADC did not differ from the preischemic value after 90 minutes of reperfusion, consistent although statistically not significant ADC decreases by 22% (mean) were observed for all metabolites determined in this study. Postmortem values were reduced by 46% for water and 38% for metabolites (mean). The latter data were obtained 60 minutes after death since a rapid decrease of the water ADC during the first 5 to 10 minutes is followed by a much slower decline as a result of cooling.14
The present study complements previous findings of water ADC changes during early stroke by demonstrating significant changes in intracellular metabolite mobility in both neuronal and glial cells. Since a parallel study using localized proton MRS at short TE revealed constant concentrations of NAA, Cr, Cho, and myo-inositol during 20 minutes of ischemia (M.W., Y.N., F.P., J.F., unpublished data, 1995) in agreement with previous reports of no changes15 16 or slightly decreased NAA levels,17 and since metabolite T2 relaxation times seem to be largely unaffected,15 17 the ischemia-related signal increases of diffusion-weighted spectra (Table 2⇑) may be transformed into ADC changes relative to basal values (Table 3⇑). The calculated ADC decreases were 17% for NAA, 48% for Cr, 44% for Cho, and 68% for myo-inositol during ischemia (mean value, 44%). The 40% decrease obtained for water is in line with numerous MRI studies.1 2 3 5 6 7 14
Putative mechanisms for the intracellular changes in metabolite mobility may be even more complex and potentially different in nature from those proposed for changes in water ADC. A simple increase in intracellular volume (“cell swelling”) due to the development of cytotoxic edema, however, would be expected to increase rather than decrease the ADC of metabolites. It also precludes the occurrence of restricted diffusion since even in normal brain the mean diffusional displacement of 3 μm (25 ms diffusion time, metabolite ADC values as in Table 3⇑) is smaller than the average cell dimension of 10 μm. Moreover, although temperature changes could account for some of the experimental findings, they have been estimated to be unable to quantitatively explain the observed water ADC changes during a short period of ischemia.5 Thus, a working hypothesis may involve an overall change in cytosolic and/or subcellular viscosity caused by as yet unknown derangements at a microstructural level. In fact, although major cell organelles stay intact during short periods of ischemia, clumping of nuclear chromatin has been observed in both neurons and glial cells.18 Swollen mitochondria were frequently observed, and the endoplasmatic reticulum was found to be dilated in astrocytes. The latter observation may be related to the remarkable 68% ADC change obtained for myo-inositol (glial marker) compared with the 17% in NAA (neuronal marker). This indication of distinct glial responses to ischemia clearly deserves further investigation.
A second major observation is that of altered metabolite ADC values during reperfusion. While this finding is consistent with irreversible tissue damage in a 20-minute model of global transient ischemia, its pathophysiological substrate remains to be established. This particularly applies to a potential link to cellular energy failure and distortions of metabolic compartments and fluxes that occur before gross structural deficiencies required for affecting the highly permeable water pool. Moreover, ADC decreases during reperfusion may be due to mechanisms other than those responsible for alterations during the acute phase. One possible explanation stems from postischemic disaggregation of ribosomes in both neurons and glial cells.19 While further studies are warranted, the reperfusion effect may provide a new basis for the assessment of circulatory and metabolic compromise.
Selected Abbreviations and Acronyms
|ADC||=||apparent diffusion coefficient|
|Cr||=||phosphocreatine and creatine|
|MRS||=||magnetic resonance spectroscopy|
|STEAM||=||stimulated echo-acquisition mode|
- Received March 3, 1995.
- Revision received May 19, 1995.
- Accepted June 21, 1995.
- Copyright © 1995 by American Heart Association
Warach S, Chien D, Li W, Ronthal M, Edelman RR. Fast magnetic resonance diffusion-weighted imaging of acute human stroke. Neurology. 1992;42:1717-1723.
Helpern JA, Ordidge RJ, Knight RA. The effect of cell membrane water permeability on the apparent diffusion coefficient of water. Presented at 11th Annual Meeting of the Society of Magnetic Resonance; 1992; Berlin, Germany, Aug 8-14, 1992.
Pulsinelli WA, Brierley JB. A new model of bilateral hemispheric ischemia in the unanesthetized rat. Stroke. 1979;10:267-272.
Frahm J, Michaelis T, Merboldt KD, Bruhn H, Gyngell ML, Hänicke W. Improvements in localized proton NMR spectroscopy of human brain: water suppression, short echo times, and 1 ml resolution. J Magn Reson. 1990;90:464-473.