- Email: [email protected]
Myelin mapping in the living mouse brain using manganese-enhanced magnetization transfer MRI
NeuroImage 49 (2010) 1200–1204
Contents lists available at ScienceDirect
NeuroImage j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n i m g
Myelin mapping in the living mouse brain using manganese-enhanced magnetization transfer MRI Takashi Watanabe ⁎, Jens Frahm, Thomas Michaelis Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut für biophysikalische Chemie, 37070 Göttingen, Germany
a r t i c l e
i n f o
Article history: Received 19 June 2009 Revised 21 September 2009 Accepted 22 September 2009 Available online 28 September 2009 Keywords: Brain mapping Contrast media Magnetic resonance imaging Manganese chloride White matter
a b s t r a c t This work demonstrates manganese-enhanced magnetization transfer (MT) MRI to improve the contrast of myelinated structures in mouse brain in vivo. Systemic administration of manganese chloride led to a reduction of the MT ratio by 23% in white matter and 35% in gray matter. The effect increased their contrastto-noise ratio by 48% and facilitated a mapping of myelin-rich white matter tissues. Relaxation time measurements revealed the manganese-induced shortening of T1 to be smaller in the corpus callosum (−42%) than in the cortex (−52%) or hippocampus (−60%). These findings are in line with the assumption that a high myelin and correspondingly low water content hinder the free diffusion and uptake of manganese ions. The resulting preferential accumulation of manganese in gray matter structures causes a stronger reduction of the MT saturation in gray matter than in white matter. Extending MRI assessments with conventional MT contrast, manganese-enhanced MT MRI at 76 × 80 × 160 μm3 resolution and 2.35 T field strength allowed for a delineation of small myelinated structures such as the fornix, mammillothalamic tract, and fasciculus retroflexus in the living mouse brain. © 2009 Elsevier Inc. All rights reserved.
Introduction Myelin, which creates the typical glistening appearance of white matter (WM) in freshly cut brain sections, has a lipid content of N40% and a water content of only about 40% (Norton, 1975). The protons of respective lipids are highly immobilized and thus yield T2 relaxation times of b1 ms that are too short to be directly detected by conventional magnetic resonance imaging (MRI). The higher macromolecular content of WM than of gray matter (GM) results in lower signal intensity in magnetization transfer (MT) MRI of murine brain (Natt et al., 2003). The MT method relies upon the indirect saturation of mobile water protons by off-resonance irradiation of highly immobilized protons bound to macromolecules. The saturation transfer is well described by a combination of dipole–dipole interactions, proton exchange, and water exchange between both pools (Henkelman et al., 2001; Sled and Pike, 2001). In general, the saturation of water protons by MT competes with the recovery of the longitudinal magnetization by T1 relaxation. Therefore, the basic idea proposed here is to improve the contrast between myelin-rich WM structures and other tissues with the use of a paramagnetic compound that is preferentially delivered to non-WM tissues. Because hydrophilic substances should become less concentrated in myelin-rich structures with low water content, T1shortening manganese ions (Mn2+) emerge as a potentially useful
⁎ Corresponding author. Fax: +49 551 201 1306. E-mail address: [email protected] (T. Watanabe). 1053-8119/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2009.09.050
contrast agent for this purpose. Their hydrophilic properties are expected to provide little access to the hydrophobic parts of the myelin, so that manganese will be more concentrated in the waterrich GM of the brain. Previous studies have indeed shown that manganese ions can be delivered to brain tissue fluids via systemic circulation of the blood (for a review see Koretsky and Silva, 2004). The purpose of this study was to examine whether the administration of manganese can be exploited for an improved mapping of myelin-rich WM structures using MT MRI of the brain of living mice. Methods Animals Sixteen female mice (NMRI, 8–12 weeks old, 28–38 g) were studied in accordance with German animal protection laws after approval by the responsible governmental authority. Each mouse received manganese chloride (0.5 mmol/kg body weight) dissolved in distilled water via subcutaneous injection. After administration animals were returned to a chamber with unlimited access to food and water. MRI In vivo MRI was performed before as well as 1, 2, 3, and 4 days after manganese injection. The animals were anesthetized by an intraperitoneal injection of ketamine (125 mg/kg body weight) and xylazine (12.5 mg/kg body weight), intubated with a purpose-built polyethylene endotracheal tube (0.58 mm inner diameter, 0.96 mm outer
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
1201
diameter), and artificially ventilated using an animal respirator (TSE, Bad Homberg, Germany) as previously described (Watanabe et al., 2004a,b). The animals were then placed in a prone position on a purpose-built palate holder equipped with an adjustable nose cone. All MRI measurements were carried out at 2.35 T (Magnex Scientific, Abingdon, UK) using 200 mT m−1 gradients (Bruker Biospin MRI GmbH, Ettlingen, Germany). Radiofrequency (RF) excitation and signal reception was accomplished with the use of a Helmholtz coil (inner diameter 100 mm) and an elliptical surface coil (inner diameter 20 × 14 mm), respectively.
volume effect with neighboring tissue, a sagittal MRI section was selected at the midline (for the corpus callosum and thalamus), at 0.5 mm from the midline (for cerebral cortex), and at 0.9 mm from the midline (for the hippocampal formation). The mean MRI signal intensity of a region of interest selected in these structures was fitted for each inversion time, and T1 values were calculated as previously described (Schmitt et al., 2004) using software provided by the manufacturer.
Magnetization transfer MRI
For evaluation of signal intensities, anatomically defined crosssections were obtained from the original 3D MRI data sets by multiplanar reconstructions using software supplied by the manufacturer (Paravision 4.0, Bruker Biospin MRI GmbH, Ettlingen, Germany). The plane of the anterior commissure–posterior commissure (AC-PC) served as a reference for the selection of standardized sections to facilitate comparisons with minimized intra- and interindividual variability. Regions of interest were selected for WM in the corpus callosum, fimbria, and cerebellum, as well as for GM in the cerebral cortex, thalamus, hippocampal formation, and cerebellar cortex. The signal-to-noise ratio (SNR) was defined as the mean MRI signal intensity divided by the standard deviation of the noise. The contrast-to-noise ratio (CNR) was obtained by taking the difference between the SNR values in WM and GM. The analysis followed a strategy previously developed for intraindividual comparisons of MR images obtained after manganese administration (Watanabe et al., 2004a,b).
For MT MRI, an off-resonance Gaussian RF pulse with a frequency offset of 5 kHz, a duration of 12 ms, and a mean amplitude of 200 Hz (flip angle 1045°) was incorporated into a spin density-weighted gradient-echo MRI sequence (RF-spoiled 3D FLASH, TR/TE= 30/7.6 ms, flip angle 5°, field-of-view 15 × 15 × 15 mm3, matrix 128 × 96 × 96, measuring time 84 min) at 117 × 156 × 156 μm3 resolution (Natt et al., 2003). The MT ratio (MTR) was obtained from acquisitions with and without off-resonance irradiation. Data were acquired before (n = 8) and after (n = 9) manganese administration. In addition, T1-weighted gradient-echo images (RF-spoiled 3D FLASH, TR/TE = 17/7.6 ms, flip angle 25°, other parameters as above) were obtained from four animals before and after manganese administration. Finally, high-resolution MT MRI was performed before and 3 days after manganese administration (n = 3). Offresonance RF irradiation in combination with a spin density-weighted gradient-echo MRI sequence (see above) yielded 76 × 80× 160 μm3 resolution (field-of-view 19.6× 23× 20.5 mm3, matrix 256 × 288 × 128) within a measuring time of about 8 h. T1 measurements T1 relaxation times of WM and GM were determined using a segmented inversion-recovery TrueFISP sequence (TR/TE = 7.8/ 3.9 ms, flip angle 50°, field-of-view 19.2 × 19.2 mm2, matrix 128 × 128, in-plane resolution 150 × 150 μm 2 , slice thickness 600 μm) before (n = 6) and after (n = 5) manganese administration. The number of echoes acquired for each segment was four before the injection (or two after injection), the number of segments 32 (64), the segment time 31 ms (15.5 ms), the number of acquired echoes 400 (200), the observation time during recovery 6.2 s (3.1 s), the inversion repetition time 16 s (12 s), the number of averages 6 (3), and the total measuring time 52 min (39 min). In order to minimize the partial
Data evaluation
Results Experimental validation As qualitatively demonstrated in Fig. 1 for MRI of mouse brain in vivo, the delineation of WM with the use of MT contrast is considerably improved after the administration of manganese. This finding is supported by a number of quantitative evaluations summarized in Table 1. For example, 3 days after administration, manganese reduced the mean MTR by 23% and 35% in WM and GM, respectively. Simultaneously, the SNR increased by 24% in WM and 30% in GM, which resulted in a CNR improvement between WM and GM by 48% in manganese-enhanced MT MRI relative to conventional MT MRI. In order to test for the hypothesis of a differential relaxation effect of manganese on tissues with different water content, Fig. 2 compares conventional T1-weighted MRI of the mouse brain with T1-weighted
Fig. 1. The effect of manganese on magnetization transfer (MT) MRI of mouse brain in vivo: horizontal spin density-weighted sections (left) without MT yielding proton density (PD) contrast, (middle) with MT contrast, and (right) with MT contrast 3 days after manganese (Mn) injection. The delineation of white matter improves with MT contrast and even further with manganese-enhanced MT MRI.
1202
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
Table 1 Magnetization transfer ratio (MTR), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) of white and gray matter in MT MRI before (control) and 1–4 days after MnCl2.
MTR White matter Gray matter SNR White matter Gray matter CNR
Control n=8
1 day n=4
2 days n=5
3 days n=9
4 days n=4
0.52 ± 0.01 0.45 ± 0.02
0.38 ± 0.02 0.30 ± 0.01
0.40 ± 0.02 0.31 ± 0.02
0.40 ± 0.01 0.29 ± 0.01
0.38 ± 0.01 0.30 ± 0.02
11.9 ± 1.7 15.5 ± 2.3 3.6 ± 0.7
16.0 ± 1.0 20.9 ± 1.4 5.0 ± 0.5
15.0 ± 0.8 20.0 ± 1.1 5.0 ± 0.4
14.8 ± 0.8 20.1 ± 1.1 5.3 ± 0.4
15.2 ± 0.9 19.6 ± 1.4 4.4 ± 0.5
Table 2 T1 relaxation times (ms) before (control) and 3 days after MnCl2.
Corpus callosum Cortex Thalamus Hippocampus
Control n=6
3 days n=5
Change
876 ± 30 1113 ± 60 1116 ± 28 889 ± 87
505 ± 51 538 ± 22 517 ± 16 355 ± 26
−42% −52% −54% −60%
Values are given as mean ± SD averaged across animals.
High-resolution mouse brain studies
Values are given as mean ± SD averaged across animals.
MRI 3 days after manganese administration as well as with manganese-enhanced MT MRI. While WM in the corpus callosum presents with higher signal intensity than surrounding GM before manganese administration, WM yielded similar signal intensities to GM in manganese-enhanced T1-weighted MRI. In fact, close to the genu of the corpus callosum, fibers of both hemispheres presented with an even lower signal intensity than the surrounding GM. Quantitatively, the underlying T1 shortening due to paramagnetic manganese enhanced the mean SNR in WM by 39% but in GM by 70% and therefore partially equilibrated their native T1 contrast. Of course, bright GM contrast due to the functional accumulation of manganese in specific brain systems is the true reason for the use of manganeseenhanced T1-weighted MRI. This can also be appreciated in Fig. 2, for example, in the hippocampal formation, the habenular nuclei, and the superior and inferior colliculi. In agreement with Fig. 1, however, Fig. 2 again demonstrates that the best CNR between WM and GM is achieved for manganese-enhanced MT MRI both in the cortex and cerebellum. The results of T1 relaxation time measurements in WM and GM structures before and after manganese administration are summarized in Table 2. The paramagnetic effect resulted in a drastic shortening of T1 relaxation times, which turned out to be much more pronounced in GM (−52% to −60%) than in WM (−42% in corpus callosum). As a consequence, the mean T1 relaxation times in the corpus callosum (505 ms) and cerebral cortex (538 ms) approached each other to such a degree that this finding directly explains the reduced contrast in T1-weighted MRI seen in Fig. 2. It is also in line with the assumption that the bulk tissue concentration of manganese is lower in WM than in GM. It should also be noted that all T1 data were well described by a single exponential. This observation is in line with the assumption of a fast exchange of water molecules through the cell membranes (Koenig et al., 1990).
Figs. 3–5 summarize the results of manganese-enhanced MT MRI of mouse brain at high spatial resolution in horizontal, sagittal, and coronal sections. While conventional MT MRI may identify major WM structures such as the corpus callosum, external capsule, fimbria, ventral hippocampal commissure, cerebellar white matter, internal capsule, and cerebral peduncle, these structures become much better distinguishable from surrounding GM after manganese administration. More importantly, several smaller WM structures are much more readily identified using manganese-enhanced MT MRI. This particularly applies to the fornix, optic tract, stria medullaris of the thalamus, and mammillothalamic tract seen in horizontal sections in Fig. 3. Similarly, additional small myelinated structures such as the fasciculus retroflexus, longitudinal fasciculus, and the transverse fibers of the pons become visible in sagittal sections shown in Fig. 4, while the external medullary lamina is depicted in coronal sections in Fig. 5. Discussion This work shows for the first time the use of manganese in MT MRI of the brain. The application of manganese-enhanced MT MRI is advantageous for two reasons. First of all, manganese generally increases the SNR by reducing the conventional degree of MT saturation. The effect is due to a shortening of the water proton T1 relaxation times by the paramagnetic properties of the Mn2+ ion. Secondly, the manganese-induced signal enhancement does not obscure the MT contrast of WM and GM as observed for the T1 contrast, but improves it. This can be explained by a differential T1 shortening effect, which is much more pronounced in GM. The underlying mechanism reflects the preferred accumulation of manganese in the fluid compartments of GM tissues. After administration to the systemic blood circulation and crossing of the capillary endothelium, hydrated Mn2+ ions are less likely to enter the hydrophobic hydrocarbon regions of lipid membranes. In
Fig. 2. The effect of manganese on T1-weighted MRI of mouse brain in vivo: horizontal sections (left) with T1 contrast, (middle) with T1 contrast 3 days after manganese (Mn) injection, and (right) with MT contrast 3 days after manganese injection. The native T1 contrast between white and gray matter is largely abolished after manganese administration. The best contrast is obtained with manganese-enhanced MT MRI.
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
1203
the lipid structures play a central role in MT MRI. They exceed the relevance of other macromolecules such as proteins, carbohydrates, or nucleotides (Clausen, 1969) and define the resulting contrast between WM and GM. Apart from the myelin content, the tissue concentration of manganese also depends on physiological factors as Mn2+ ions can be taken up into cells through Ca2+ channels. However, the cytoplasm of the myelin is expected to accumulate less Mn2+ ions than the cytoplasm of other cell types, because the oligodendrocytes in WM are comparatively inactive in healthy adult animals. Another possible manganese enrichment of WM may be due to axonal transport from the cell bodies of active neurons. However, previous data suggest a lower capacity for manganese accumulation in the axon than in the cell body. For example, after manganese application to the retinal ganglion cells, the axonal signal enhancement in the optic nerve and tract was much weaker and less persistent than that of tissues formed by cell bodies (Watanabe et al., 2004a). Taken together, several factors determine the higher concentration of Mn2+ ions in GM than WM and the more pronounced T1 shortening in GM that favourably supports the MT contrast between both tissues. While the improved WM/GM contrast turns out to be the main advantage of manganese-enhanced MT MRI over manganese-enhanced T1-weighted MRI, a potential drawback may be the observation of a reduced contrast within the cerebral cortex. For example, whereas Fig. 2 readily demonstrates distinct cortical layers in the T1weighted image after manganese enhancement (center), this is no longer the case in the corresponding MT image (right). In conclusion, the administration of Mn2+ ions to neural tissue fluids emerges as a novel tool for MT MRI of animal brain in vivo as it increases the CNR for mapping myelin-rich structures. The approach is expected to contribute to a better neuroanatomical characterization of animal models of human diseases.
References
Fig. 3. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: (top, middle) horizontal sections and (bottom) oblique horizontal-to-coronal section 3 days after manganese injection. AC = anterior commissure, ACI = intrabulbar part of the anterior commissure, ALV = alveus, CC = corpus callosum, CeP = cerebellar peduncle, CeWM = cerebellar white matter, CP = cerebral peduncle, EC = external capsule, Fi = fimbria, Fx = fornix, IC = internal capsule, MTT = mammillothalamic tract, OT = optic tract, SM = stria medullaris of the thalamus, ST = stria terminalis, PC = posterior commissure, VHC = ventral hippocampal commissure.
this respect, myelin sheaths (lipid content N40%, water content 40%) are largely responsible for the fact that WM exhibits a higher lipid content (N15% vs. b6%) and lower water content (70% vs. N80%) than GM (Agranoff and Hajra, 1994; Clausen, 1969; Katzman and Schimmel, 1969). On the other hand, the immobilized protons of
Agranoff, B.W., Hajra, A.K., 1994. Lipids, In: Siegel, G.J., Agranoff, B.W., Albers, R.W., Molinoff, P.B. (Eds.), Basic neurochemistry: molecular, cellular, and medical aspects, 5th ed. Raven Press, New York, pp. 97–116. Clausen, J., 1969. Gray–white matter differences. In: Lajtha, A. (Ed.), Handbook of neurochemistry. Volume 1. Chemical architecture of the nervous system. Plenum Press, New York, pp. 273–300. Henkelman, R.M., Stanisz, G.J., Graham, S.J., 2001. Magnetization transfer in MRI: a review. NMR Biomed. 14, 57–64. Katzman, R., Schimmel, H., 1969. Water movement. In: Lajtha, A. (Ed.), Handbook of Neurochemistry. Volume 2. Structural Neurochemistry. Plenum Press, New York, pp. 11–22. Koenig, S.H., Brown III, R.D., Spiller, M, Lundbom, N., 1990. Relaxometry of brain: why white matter appears bright in MRI. Magn. Reson. Med. 14, 482–495. Koretsky, A.P, Silva, A.C., 2004. Manganese-enhanced magnetic resonance imaging (MEMRI). NMR Biomed. 17, 527–531. Natt, O., Watanabe, T., Boretius, B., Frahm, J., Michaelis, T., 2003. Magnetization transfer MRI of mouse brain reveals areas of high neural density. Magn. Reson. Imaging 21, 1113–1120. Norton, W.T., 1975. Myelin: structure and biochemistry. In: Tower, D.B. (Ed.), The Nervous System. Vol. 1: The Basic Neurosciences. Raven Press, New York, pp. 467–481. Schmitt, P., Griswold, M.A., Jakob, P.M., Kotas, M., Gulani, V., Flentje, M., Haase, A., 2004. Inversion recovery TrueFISP: quantification of T1, T2, and spin density. Magn. Reson. Med. 51, 661–667. Sled, J.G., Pike, G.B., 2001. Quantitative imaging of magnetization transfer exchange and relaxation properties in vivo using MRI. Magn. Reson. Med. 46, 923–931. Watanabe, T., Frahm, J., Michaelis, T., 2004a. Functional mapping of neural pathways in rodent brain in vivo using manganese-enhanced three-dimensional MRI. NMR Biomed. 17, 554–568. Watanabe, T., Radulovic, J., Spiess, J., Natt, O., Boretius, S., Frahm, J., Michaelis, T., 2004b. In vivo 3D MRI staining of the mouse hippocampal system using intracerebral injection of MnCl2. NeuroImage 22, 860–867.
1204
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
Fig. 4. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: sagittal sections 3 days after manganese injection. Cg = cingulum, FR = fasciculus retroflexus, LFP = longitudinal fasciculus of the pons, OC = optic chiasm, TFP = transverse fibers of the pons. For other abbreviations see Fig. 3.
Fig. 5. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: coronal sections 3 days after manganese injection. EML = external medullary lamina. For other abbreviations see Figs. 3 and 4.
Contents lists available at ScienceDirect
NeuroImage j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n i m g
Myelin mapping in the living mouse brain using manganese-enhanced magnetization transfer MRI Takashi Watanabe ⁎, Jens Frahm, Thomas Michaelis Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut für biophysikalische Chemie, 37070 Göttingen, Germany
a r t i c l e
i n f o
Article history: Received 19 June 2009 Revised 21 September 2009 Accepted 22 September 2009 Available online 28 September 2009 Keywords: Brain mapping Contrast media Magnetic resonance imaging Manganese chloride White matter
a b s t r a c t This work demonstrates manganese-enhanced magnetization transfer (MT) MRI to improve the contrast of myelinated structures in mouse brain in vivo. Systemic administration of manganese chloride led to a reduction of the MT ratio by 23% in white matter and 35% in gray matter. The effect increased their contrastto-noise ratio by 48% and facilitated a mapping of myelin-rich white matter tissues. Relaxation time measurements revealed the manganese-induced shortening of T1 to be smaller in the corpus callosum (−42%) than in the cortex (−52%) or hippocampus (−60%). These findings are in line with the assumption that a high myelin and correspondingly low water content hinder the free diffusion and uptake of manganese ions. The resulting preferential accumulation of manganese in gray matter structures causes a stronger reduction of the MT saturation in gray matter than in white matter. Extending MRI assessments with conventional MT contrast, manganese-enhanced MT MRI at 76 × 80 × 160 μm3 resolution and 2.35 T field strength allowed for a delineation of small myelinated structures such as the fornix, mammillothalamic tract, and fasciculus retroflexus in the living mouse brain. © 2009 Elsevier Inc. All rights reserved.
Introduction Myelin, which creates the typical glistening appearance of white matter (WM) in freshly cut brain sections, has a lipid content of N40% and a water content of only about 40% (Norton, 1975). The protons of respective lipids are highly immobilized and thus yield T2 relaxation times of b1 ms that are too short to be directly detected by conventional magnetic resonance imaging (MRI). The higher macromolecular content of WM than of gray matter (GM) results in lower signal intensity in magnetization transfer (MT) MRI of murine brain (Natt et al., 2003). The MT method relies upon the indirect saturation of mobile water protons by off-resonance irradiation of highly immobilized protons bound to macromolecules. The saturation transfer is well described by a combination of dipole–dipole interactions, proton exchange, and water exchange between both pools (Henkelman et al., 2001; Sled and Pike, 2001). In general, the saturation of water protons by MT competes with the recovery of the longitudinal magnetization by T1 relaxation. Therefore, the basic idea proposed here is to improve the contrast between myelin-rich WM structures and other tissues with the use of a paramagnetic compound that is preferentially delivered to non-WM tissues. Because hydrophilic substances should become less concentrated in myelin-rich structures with low water content, T1shortening manganese ions (Mn2+) emerge as a potentially useful
⁎ Corresponding author. Fax: +49 551 201 1306. E-mail address: [email protected] (T. Watanabe). 1053-8119/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2009.09.050
contrast agent for this purpose. Their hydrophilic properties are expected to provide little access to the hydrophobic parts of the myelin, so that manganese will be more concentrated in the waterrich GM of the brain. Previous studies have indeed shown that manganese ions can be delivered to brain tissue fluids via systemic circulation of the blood (for a review see Koretsky and Silva, 2004). The purpose of this study was to examine whether the administration of manganese can be exploited for an improved mapping of myelin-rich WM structures using MT MRI of the brain of living mice. Methods Animals Sixteen female mice (NMRI, 8–12 weeks old, 28–38 g) were studied in accordance with German animal protection laws after approval by the responsible governmental authority. Each mouse received manganese chloride (0.5 mmol/kg body weight) dissolved in distilled water via subcutaneous injection. After administration animals were returned to a chamber with unlimited access to food and water. MRI In vivo MRI was performed before as well as 1, 2, 3, and 4 days after manganese injection. The animals were anesthetized by an intraperitoneal injection of ketamine (125 mg/kg body weight) and xylazine (12.5 mg/kg body weight), intubated with a purpose-built polyethylene endotracheal tube (0.58 mm inner diameter, 0.96 mm outer
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
1201
diameter), and artificially ventilated using an animal respirator (TSE, Bad Homberg, Germany) as previously described (Watanabe et al., 2004a,b). The animals were then placed in a prone position on a purpose-built palate holder equipped with an adjustable nose cone. All MRI measurements were carried out at 2.35 T (Magnex Scientific, Abingdon, UK) using 200 mT m−1 gradients (Bruker Biospin MRI GmbH, Ettlingen, Germany). Radiofrequency (RF) excitation and signal reception was accomplished with the use of a Helmholtz coil (inner diameter 100 mm) and an elliptical surface coil (inner diameter 20 × 14 mm), respectively.
volume effect with neighboring tissue, a sagittal MRI section was selected at the midline (for the corpus callosum and thalamus), at 0.5 mm from the midline (for cerebral cortex), and at 0.9 mm from the midline (for the hippocampal formation). The mean MRI signal intensity of a region of interest selected in these structures was fitted for each inversion time, and T1 values were calculated as previously described (Schmitt et al., 2004) using software provided by the manufacturer.
Magnetization transfer MRI
For evaluation of signal intensities, anatomically defined crosssections were obtained from the original 3D MRI data sets by multiplanar reconstructions using software supplied by the manufacturer (Paravision 4.0, Bruker Biospin MRI GmbH, Ettlingen, Germany). The plane of the anterior commissure–posterior commissure (AC-PC) served as a reference for the selection of standardized sections to facilitate comparisons with minimized intra- and interindividual variability. Regions of interest were selected for WM in the corpus callosum, fimbria, and cerebellum, as well as for GM in the cerebral cortex, thalamus, hippocampal formation, and cerebellar cortex. The signal-to-noise ratio (SNR) was defined as the mean MRI signal intensity divided by the standard deviation of the noise. The contrast-to-noise ratio (CNR) was obtained by taking the difference between the SNR values in WM and GM. The analysis followed a strategy previously developed for intraindividual comparisons of MR images obtained after manganese administration (Watanabe et al., 2004a,b).
For MT MRI, an off-resonance Gaussian RF pulse with a frequency offset of 5 kHz, a duration of 12 ms, and a mean amplitude of 200 Hz (flip angle 1045°) was incorporated into a spin density-weighted gradient-echo MRI sequence (RF-spoiled 3D FLASH, TR/TE= 30/7.6 ms, flip angle 5°, field-of-view 15 × 15 × 15 mm3, matrix 128 × 96 × 96, measuring time 84 min) at 117 × 156 × 156 μm3 resolution (Natt et al., 2003). The MT ratio (MTR) was obtained from acquisitions with and without off-resonance irradiation. Data were acquired before (n = 8) and after (n = 9) manganese administration. In addition, T1-weighted gradient-echo images (RF-spoiled 3D FLASH, TR/TE = 17/7.6 ms, flip angle 25°, other parameters as above) were obtained from four animals before and after manganese administration. Finally, high-resolution MT MRI was performed before and 3 days after manganese administration (n = 3). Offresonance RF irradiation in combination with a spin density-weighted gradient-echo MRI sequence (see above) yielded 76 × 80× 160 μm3 resolution (field-of-view 19.6× 23× 20.5 mm3, matrix 256 × 288 × 128) within a measuring time of about 8 h. T1 measurements T1 relaxation times of WM and GM were determined using a segmented inversion-recovery TrueFISP sequence (TR/TE = 7.8/ 3.9 ms, flip angle 50°, field-of-view 19.2 × 19.2 mm2, matrix 128 × 128, in-plane resolution 150 × 150 μm 2 , slice thickness 600 μm) before (n = 6) and after (n = 5) manganese administration. The number of echoes acquired for each segment was four before the injection (or two after injection), the number of segments 32 (64), the segment time 31 ms (15.5 ms), the number of acquired echoes 400 (200), the observation time during recovery 6.2 s (3.1 s), the inversion repetition time 16 s (12 s), the number of averages 6 (3), and the total measuring time 52 min (39 min). In order to minimize the partial
Data evaluation
Results Experimental validation As qualitatively demonstrated in Fig. 1 for MRI of mouse brain in vivo, the delineation of WM with the use of MT contrast is considerably improved after the administration of manganese. This finding is supported by a number of quantitative evaluations summarized in Table 1. For example, 3 days after administration, manganese reduced the mean MTR by 23% and 35% in WM and GM, respectively. Simultaneously, the SNR increased by 24% in WM and 30% in GM, which resulted in a CNR improvement between WM and GM by 48% in manganese-enhanced MT MRI relative to conventional MT MRI. In order to test for the hypothesis of a differential relaxation effect of manganese on tissues with different water content, Fig. 2 compares conventional T1-weighted MRI of the mouse brain with T1-weighted
Fig. 1. The effect of manganese on magnetization transfer (MT) MRI of mouse brain in vivo: horizontal spin density-weighted sections (left) without MT yielding proton density (PD) contrast, (middle) with MT contrast, and (right) with MT contrast 3 days after manganese (Mn) injection. The delineation of white matter improves with MT contrast and even further with manganese-enhanced MT MRI.
1202
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
Table 1 Magnetization transfer ratio (MTR), signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) of white and gray matter in MT MRI before (control) and 1–4 days after MnCl2.
MTR White matter Gray matter SNR White matter Gray matter CNR
Control n=8
1 day n=4
2 days n=5
3 days n=9
4 days n=4
0.52 ± 0.01 0.45 ± 0.02
0.38 ± 0.02 0.30 ± 0.01
0.40 ± 0.02 0.31 ± 0.02
0.40 ± 0.01 0.29 ± 0.01
0.38 ± 0.01 0.30 ± 0.02
11.9 ± 1.7 15.5 ± 2.3 3.6 ± 0.7
16.0 ± 1.0 20.9 ± 1.4 5.0 ± 0.5
15.0 ± 0.8 20.0 ± 1.1 5.0 ± 0.4
14.8 ± 0.8 20.1 ± 1.1 5.3 ± 0.4
15.2 ± 0.9 19.6 ± 1.4 4.4 ± 0.5
Table 2 T1 relaxation times (ms) before (control) and 3 days after MnCl2.
Corpus callosum Cortex Thalamus Hippocampus
Control n=6
3 days n=5
Change
876 ± 30 1113 ± 60 1116 ± 28 889 ± 87
505 ± 51 538 ± 22 517 ± 16 355 ± 26
−42% −52% −54% −60%
Values are given as mean ± SD averaged across animals.
High-resolution mouse brain studies
Values are given as mean ± SD averaged across animals.
MRI 3 days after manganese administration as well as with manganese-enhanced MT MRI. While WM in the corpus callosum presents with higher signal intensity than surrounding GM before manganese administration, WM yielded similar signal intensities to GM in manganese-enhanced T1-weighted MRI. In fact, close to the genu of the corpus callosum, fibers of both hemispheres presented with an even lower signal intensity than the surrounding GM. Quantitatively, the underlying T1 shortening due to paramagnetic manganese enhanced the mean SNR in WM by 39% but in GM by 70% and therefore partially equilibrated their native T1 contrast. Of course, bright GM contrast due to the functional accumulation of manganese in specific brain systems is the true reason for the use of manganeseenhanced T1-weighted MRI. This can also be appreciated in Fig. 2, for example, in the hippocampal formation, the habenular nuclei, and the superior and inferior colliculi. In agreement with Fig. 1, however, Fig. 2 again demonstrates that the best CNR between WM and GM is achieved for manganese-enhanced MT MRI both in the cortex and cerebellum. The results of T1 relaxation time measurements in WM and GM structures before and after manganese administration are summarized in Table 2. The paramagnetic effect resulted in a drastic shortening of T1 relaxation times, which turned out to be much more pronounced in GM (−52% to −60%) than in WM (−42% in corpus callosum). As a consequence, the mean T1 relaxation times in the corpus callosum (505 ms) and cerebral cortex (538 ms) approached each other to such a degree that this finding directly explains the reduced contrast in T1-weighted MRI seen in Fig. 2. It is also in line with the assumption that the bulk tissue concentration of manganese is lower in WM than in GM. It should also be noted that all T1 data were well described by a single exponential. This observation is in line with the assumption of a fast exchange of water molecules through the cell membranes (Koenig et al., 1990).
Figs. 3–5 summarize the results of manganese-enhanced MT MRI of mouse brain at high spatial resolution in horizontal, sagittal, and coronal sections. While conventional MT MRI may identify major WM structures such as the corpus callosum, external capsule, fimbria, ventral hippocampal commissure, cerebellar white matter, internal capsule, and cerebral peduncle, these structures become much better distinguishable from surrounding GM after manganese administration. More importantly, several smaller WM structures are much more readily identified using manganese-enhanced MT MRI. This particularly applies to the fornix, optic tract, stria medullaris of the thalamus, and mammillothalamic tract seen in horizontal sections in Fig. 3. Similarly, additional small myelinated structures such as the fasciculus retroflexus, longitudinal fasciculus, and the transverse fibers of the pons become visible in sagittal sections shown in Fig. 4, while the external medullary lamina is depicted in coronal sections in Fig. 5. Discussion This work shows for the first time the use of manganese in MT MRI of the brain. The application of manganese-enhanced MT MRI is advantageous for two reasons. First of all, manganese generally increases the SNR by reducing the conventional degree of MT saturation. The effect is due to a shortening of the water proton T1 relaxation times by the paramagnetic properties of the Mn2+ ion. Secondly, the manganese-induced signal enhancement does not obscure the MT contrast of WM and GM as observed for the T1 contrast, but improves it. This can be explained by a differential T1 shortening effect, which is much more pronounced in GM. The underlying mechanism reflects the preferred accumulation of manganese in the fluid compartments of GM tissues. After administration to the systemic blood circulation and crossing of the capillary endothelium, hydrated Mn2+ ions are less likely to enter the hydrophobic hydrocarbon regions of lipid membranes. In
Fig. 2. The effect of manganese on T1-weighted MRI of mouse brain in vivo: horizontal sections (left) with T1 contrast, (middle) with T1 contrast 3 days after manganese (Mn) injection, and (right) with MT contrast 3 days after manganese injection. The native T1 contrast between white and gray matter is largely abolished after manganese administration. The best contrast is obtained with manganese-enhanced MT MRI.
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
1203
the lipid structures play a central role in MT MRI. They exceed the relevance of other macromolecules such as proteins, carbohydrates, or nucleotides (Clausen, 1969) and define the resulting contrast between WM and GM. Apart from the myelin content, the tissue concentration of manganese also depends on physiological factors as Mn2+ ions can be taken up into cells through Ca2+ channels. However, the cytoplasm of the myelin is expected to accumulate less Mn2+ ions than the cytoplasm of other cell types, because the oligodendrocytes in WM are comparatively inactive in healthy adult animals. Another possible manganese enrichment of WM may be due to axonal transport from the cell bodies of active neurons. However, previous data suggest a lower capacity for manganese accumulation in the axon than in the cell body. For example, after manganese application to the retinal ganglion cells, the axonal signal enhancement in the optic nerve and tract was much weaker and less persistent than that of tissues formed by cell bodies (Watanabe et al., 2004a). Taken together, several factors determine the higher concentration of Mn2+ ions in GM than WM and the more pronounced T1 shortening in GM that favourably supports the MT contrast between both tissues. While the improved WM/GM contrast turns out to be the main advantage of manganese-enhanced MT MRI over manganese-enhanced T1-weighted MRI, a potential drawback may be the observation of a reduced contrast within the cerebral cortex. For example, whereas Fig. 2 readily demonstrates distinct cortical layers in the T1weighted image after manganese enhancement (center), this is no longer the case in the corresponding MT image (right). In conclusion, the administration of Mn2+ ions to neural tissue fluids emerges as a novel tool for MT MRI of animal brain in vivo as it increases the CNR for mapping myelin-rich structures. The approach is expected to contribute to a better neuroanatomical characterization of animal models of human diseases.
References
Fig. 3. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: (top, middle) horizontal sections and (bottom) oblique horizontal-to-coronal section 3 days after manganese injection. AC = anterior commissure, ACI = intrabulbar part of the anterior commissure, ALV = alveus, CC = corpus callosum, CeP = cerebellar peduncle, CeWM = cerebellar white matter, CP = cerebral peduncle, EC = external capsule, Fi = fimbria, Fx = fornix, IC = internal capsule, MTT = mammillothalamic tract, OT = optic tract, SM = stria medullaris of the thalamus, ST = stria terminalis, PC = posterior commissure, VHC = ventral hippocampal commissure.
this respect, myelin sheaths (lipid content N40%, water content 40%) are largely responsible for the fact that WM exhibits a higher lipid content (N15% vs. b6%) and lower water content (70% vs. N80%) than GM (Agranoff and Hajra, 1994; Clausen, 1969; Katzman and Schimmel, 1969). On the other hand, the immobilized protons of
Agranoff, B.W., Hajra, A.K., 1994. Lipids, In: Siegel, G.J., Agranoff, B.W., Albers, R.W., Molinoff, P.B. (Eds.), Basic neurochemistry: molecular, cellular, and medical aspects, 5th ed. Raven Press, New York, pp. 97–116. Clausen, J., 1969. Gray–white matter differences. In: Lajtha, A. (Ed.), Handbook of neurochemistry. Volume 1. Chemical architecture of the nervous system. Plenum Press, New York, pp. 273–300. Henkelman, R.M., Stanisz, G.J., Graham, S.J., 2001. Magnetization transfer in MRI: a review. NMR Biomed. 14, 57–64. Katzman, R., Schimmel, H., 1969. Water movement. In: Lajtha, A. (Ed.), Handbook of Neurochemistry. Volume 2. Structural Neurochemistry. Plenum Press, New York, pp. 11–22. Koenig, S.H., Brown III, R.D., Spiller, M, Lundbom, N., 1990. Relaxometry of brain: why white matter appears bright in MRI. Magn. Reson. Med. 14, 482–495. Koretsky, A.P, Silva, A.C., 2004. Manganese-enhanced magnetic resonance imaging (MEMRI). NMR Biomed. 17, 527–531. Natt, O., Watanabe, T., Boretius, B., Frahm, J., Michaelis, T., 2003. Magnetization transfer MRI of mouse brain reveals areas of high neural density. Magn. Reson. Imaging 21, 1113–1120. Norton, W.T., 1975. Myelin: structure and biochemistry. In: Tower, D.B. (Ed.), The Nervous System. Vol. 1: The Basic Neurosciences. Raven Press, New York, pp. 467–481. Schmitt, P., Griswold, M.A., Jakob, P.M., Kotas, M., Gulani, V., Flentje, M., Haase, A., 2004. Inversion recovery TrueFISP: quantification of T1, T2, and spin density. Magn. Reson. Med. 51, 661–667. Sled, J.G., Pike, G.B., 2001. Quantitative imaging of magnetization transfer exchange and relaxation properties in vivo using MRI. Magn. Reson. Med. 46, 923–931. Watanabe, T., Frahm, J., Michaelis, T., 2004a. Functional mapping of neural pathways in rodent brain in vivo using manganese-enhanced three-dimensional MRI. NMR Biomed. 17, 554–568. Watanabe, T., Radulovic, J., Spiess, J., Natt, O., Boretius, S., Frahm, J., Michaelis, T., 2004b. In vivo 3D MRI staining of the mouse hippocampal system using intracerebral injection of MnCl2. NeuroImage 22, 860–867.
1204
T. Watanabe et al. / NeuroImage 49 (2010) 1200–1204
Fig. 4. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: sagittal sections 3 days after manganese injection. Cg = cingulum, FR = fasciculus retroflexus, LFP = longitudinal fasciculus of the pons, OC = optic chiasm, TFP = transverse fibers of the pons. For other abbreviations see Fig. 3.
Fig. 5. High-resolution manganese-enhanced MT MRI of mouse brain in vivo and corresponding drawings indicating white matter structures: coronal sections 3 days after manganese injection. EML = external medullary lamina. For other abbreviations see Figs. 3 and 4.