In Vivo Magnetic Resonance Imaging of Embryonic Neural Grafts in a Rat Model of Striatonigral Degeneration (Multiple System Atrophy)

NeuroImage 12, 209 –218 (2000) doi:10.1006/nimg.2000.0600, available online at http://www.idealibrary.com on

In Vivo Magnetic Resonance Imaging of Embryonic Neural Grafts in a Rat Model of Striatonigral Degeneration (Multiple System Atrophy) Michael F. H. Schocke,* Regina Waldner,† Zoe Puschban,† Christian Kolbitsch,‡ Klaus Seppi,† Christoph Scherfler,† Christian Kremser,* Fritz Zschiegner,‡ Stephan Felber,* Werner Poewe,† and Gregor K. Wenning† *Department of Magnetic Resonance Imaging and Spectroscopy, †Department of Neurology, and ‡Department of Anaesthesia, The Leopold-Franzens University of Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria Received October 22, 1999

The effects of embryonic neural transplantation in experimental models of neurodegenerative disorders are commonly assessed by behavioral tests and postmortem neurochemical or anatomical analysis. The purpose of the present study was to evaluate embryonic neuronal grafts in a novel rat model of multiple system atrophy (MSA) with the help of in vivo magnetic resonance imaging (MRI) and to correlate imaging with histological parameters. Striatonigral double lesions were created in male Wistar rats by unilateral intrastriatal injection of 3-nitropropionic acid (3-NP). Seven weeks following lesion surgery animals were divided into four transplantation groups receiving either pure mesencephalic, pure striatal, mesencephalic-striatal cografts, or sham grafts. In vivo structural imaging was performed 21 weeks after transplantation using a whole body 1.5 Tesla MR scanner. The imaging protocol comprised T2-weighted TSE and T1weighted TIR sequences. Immunohistochemistry using DARPP-32 as striatal marker and tyrosinhydroxylase as marker for nigral neurons was performed for correlation analysis of imaging and histological parameters. The sensitivity of graft detection by in vivo MRI was 100%. The graft tissue was clearly demarcated from the remaining striatal tissue in both T2and T1-weighted sequences. Morphometrically, crosssectional areas of the grafts and spared intact striatum as defined by immunohistochemistry correlated significantly with measurements obtained by in vivo MRI. In conclusion, we were able to evaluate in vivo both lesion-induced damage and graft size in a 3-NP rat model of MSA using a conventional whole body 1.5 Tesla MRI scanner. Additionally, we obtained an excellent correlation between MRI and histological measurements. © 2000 Academic Press

INTRODUCTION Transplantation of embryonic neurons appears to be a promising therapeutic strategy for neurodegenera-

tive disorders such as Parkinson’s disease (PD) or Huntington’s disease (HD) (1). The effects of embryonic neural grafts in PD or HD animal models are commonly assessed by behavioral tests and postmortem neurochemical or anatomical analysis (2,3). This implies that both lesion size as well as graft survival have to be evaluated several weeks or months after surgical intervention. Since 1989 magnetic resonance imaging (MRI) has been applied in a small series of experimental studies to evaluate lesion size (4 –7) and viability of the embryonic neural grafts in PD and HD animal models (8 –11). The first experimental imaging studies in the late eighties and the early nineties which reported on imaging of neurotoxic lesions and neural transplants in animal models were performed on low-field MR scanners (8 –10). Due to limited spatial resolution embryonic grafts were poorly delineated even after administration of contrast agents (9,11). Consequently, the size of neural transplants was indirectly assessed by evaluating the size of the lateral ventricle on the lesioned and subsequently grafted side using T1- and T2weighted images (10). Since the early nineties the MRI technology has improved considerably. Fast MR imaging, new gradient systems and coil design have been established. These improvements have lead to a higher spatial resolution and signal-to-noise ratio as well as reduced scan time (12,13). Importantly, it is now possible to rapidly acquire high quality MRI scans of rodent brains using whole body MR systems (14). However, only one study has characterized embryonic neural grafts using a whole body 1.5 Tesla scanner. The grafts appeared hypointense on the T1-weighted images and hyperintense on the T2-weighted images (11). The present study was performed in a neural transplantation rat model of striatonigral degeneration (SND), the neuropathological substrate of parkinsonism associated with multiple system atrophy (MSA) (15). The purpose of the present MRI study was to

209

1053-8119/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

210

SCHOCKE ET AL.

evaluate presence and size of neural grafts in a MSA model using a whole body 1.5 Tesla MR scanner and a special frame allowing high quality simultaneous MRI acquisition in four animals. MATERIALS AND METHODS

tibody, clone TH-16, Sigma, Product No. T-2928) were used for DARPP-32 and TH immunohistochemistry. The grafts were defined histologically as intrastriatal areas exhibiting TH and/or DARPP-32 immunoactivity. Spared intact striatum was defined by the presence of DARPP-32 immunoactivity.

Surgical Preparation

In Vivo MRI and Histological Morphometry

Under deep anesthesia with halothane 16 male Wistar rats were subjected to unilateral stereotaxic injections of 500 nmol 3-nitropropionic acid (3-NP) dissolved in 0.1 M phosphate-buffered saline into the striatum in order to create a model of MSA-SND (15). The toxin was applied over a period of 2 min using a 5-␮l Hamilton syringe (26S-gauge stainless steel cannula, internal diameter 0.13 mm, external diameter 0.47 mm). Seven weeks following lesioning transplantation was performed. Graft tissue from gestational day 14 embryos was obtained (16,17). The embryonic cell suspensions were grafted by stereotaxic injection via a 5-␮l Hamilton syringe using a stainless steel cannula (26 gauge, internal diameter 0.26 mm, external diameter 0.46 mm) (18). Rats were divided into four transplantation groups, receiving either pure mesencephalic, pure striatal, mesencephalic-striatal cografts, or sham grafts.

The presence of graft tissue in MRI was independently assessed in two measurements by two blinded neuroradiologists (MS, SF) choosing the slice exhibiting optimal graft visibility. Quantification was performed by manual image-segmentation, pixel counting, and multiplication of the number of pixels with the spatial resolution in x- and y-direction. On the MRI and the histological images graft tissue, residual striata, and left lateral ventricles were evaluated. The extent of graft tissue and residual striatum was primarily assessed on the T2-weighted images, while the ventricular size was primarily determined on the T1weighted images. The visual impression and interpretation of the anatomical structures presented on the MR images was increased by consideration of both T1and T2-weighted images. The cross-sectional area of the residual striata excluded the nucleus accumbens. The resulting values of the segmented cross-sectional areas in MRI were correlated with histological measurements derived from a single DARPP-32 and TH immunostained section most closely matching the choosen MR image. Histological measurements were performed blinded by one of the investigators (RW) using a high resolution video camera (SONY 3CCD color video camera) and the software package ImagePro-Plus (Media Cybernetics, U.S.A.).

Magnetic Resonance Imaging Protocol MRI was performed 21 weeks after transplantation using a whole body 1.5 Tesla MR scanner (Magnetom VISION plus, Siemens, Germany). The MR protocol included a T2-weighted turbo spin echo (TSE) sequence (TR, 3500 ms; TE, 96 ms; matrix 256*256 pixels, slice thickness 2 mm; field of view (FoV) 100 mm) and a T1-weighted turbo inversion recovery (TIR) sequence (TR, 5524 ms; TE, 60 ms; IT, 150 ms; matrix 256*256 pixels, slice thickness 2 mm; FoV, 100 mm) with a total acquisition time of about 45 min. For MRI examinations we used a circular polarized (CP) volume coil (CP extremity, internal diameter 20 cm) combined with a home-made animal holder allowing four animals to be examined simultaneously. The rats were anesthetized with ketamine (110 mg/kg) intraperitoneally (i.p.) followed by i.p. injections of midazolam (15 mg/kg) to maintain anesthesia. Histology After MRI animals were transcardially perfused with paraformaldehyde (4%) and brains were processed for haematoxylin-eosin staining and immunohistochemistry. Sections were cut on a freezing microtome with a thickness of 40 ␮m. Mouse monoclonal antibodies to bovine DARPP-32 (dopamine- and cyclic adenosine 3⬘5⬘-monophosphate—regulated phosphoprotein antibody, gift from Prof. Hugh Hemmings, New York) and TH (monoclonal anti-tyrosinhydroxylase an-

Statistical Analysis Data were tabulated and analyzed using SPSS 9.0 for Windows (SPSS Inc., U.S.A.). Since Shapiro–Wilk’s test revealed normal distribution of all parameters, paired t tests and Pearson correlation coefficients were used. Correlation coefficients (r) of 0.35– 0.49 were interpreted empirically as low, those of 0.5– 0.79 as moderate and those of 0.8 or greater as high. Correlation coefficients for grafts, residual striata, and left lateral ventricles were calculated separately for the two measurements of each investigator (intraobserver correlation) and for the average of both investigators’ values (interobserver correlation). Averaged MRI data of both observers were then correlated with histological measurements. Results are presented as mean values (standard deviations, SD). Statistical significance was defined as P ⬍ 0.05 (two-sided). RESULTS In vivo MRI showed in five rats unintended lesion spread with cortical and subcortical damage in the left

DETECTION OF STRIATAL TRANSPLANTS

211

FIG. 1. (a– c) The DARPP-32 immunostained slice (a), the T1-weighted (b), and the T2-weighted (c) MR images of a striatal graft (marked with a white grid in c), which is visible at the floor of the left lateral ventricle. Note the diffuse hypointense areas on the T2-weighted image in the graft tissue. At the lower margin of the graft a slightly hyperintense rim is visible. W, window; C, center; R, right side; A, upper side; SP, slice position. Scale bar, 1 cm.

parietal area. In one of the 12 grafted animals there was no histological evidence of graft survival. In all of the remaining rats with histologically verified graft tissue in vivo MRI clearly visualized graft tissue implanted into lesioned host striatum. Striatal and mixed grafts (Figs. 1a–1c and 2a–2c) showed a similar in vivo MRI appearance. In comparison to undamaged contralateral striatum, the graft tissue was slightly hyperintense in T2-weighted im-

ages containing some hypointense areas, which were shaped irregularly (Figs. 1c and 2c). In the TIR sequence the graft tissue exhibited inhomogeneous signal intensity which ranged between gray and white matter (Figs. 1b and 2b). A differentiation between striatal and mixed graft tissue was not possible by in vivo MRI. The mesencephalic transplants (Figs. 3a–3c) were poorly demarcated from the remaining striatum and

212

SCHOCKE ET AL.

FIG. 2. (a– c) The DARPP-32 immunostained slice (a), the T1-weighted (b), and the T2-weighted (c) MR images of a mesencephalicstriatal cograft (marked with a white grid in c), which is observed in a similar location as the striatal graft and which exhibits a similar signal intensity like the striatal graft. W, window; C, center; R, right side; A, upper side; SP, slice position. Scale bar, 1 cm.

exhibited signal intensities similar to those of the contralateral undamaged striatum. Generally, the mesencephalic transplants appeared homogeneously hypointense on the T2-weighted images. The residual striatal tissue of the lesioned side was identified as hypointense area on T2-weighted MR images (Figs. 1c, 2c, 3c, and 4c). In the T1-weighted TIR sequence the residual striatum appeared hyperintense (Figs. 1b, 2b, 3b, and 4b). In animals with pure striatal and mixed grafts demarcation of residual striatum

from graft was excellent due to the presence of a slightly hyperintense rim which was noted in the periphery of the graft (Fig. 1c). Generally, the signal intensity of the spared intact and contralateral striatum was similar. In contrast, cortex appeared more hyperintense on T2-weighted images and more hypointense on T1-weighted TIR images compared to spared intact striatum allowing accurate delineation of spared intact striatum. Histologically, spared intact striatum was defined by positive DARPP-32 immunohistochem-

DETECTION OF STRIATAL TRANSPLANTS

213

FIG. 3. (a– c) The TH immunostained slice (a), the T1-weighted (b), and the T2-weighted (c) MR images of a mesencephalic graft (marked with a white grid in c), which is poorly demarcated. On the lesioned side, a cortical and subcortical atrophy is detectable reflecting unintended lesion spread in this animal. W, window; C, center; R, right side; A, upper side; SP, slice position. Scale bar, 1 cm.

istry. In animals receiving pure mesencephalic grafts differentiation of graft and residual striatum was impeded by lack of signal changes at the interface of embryonic and adult host tissue. Intraobserver correlations for cross-sectional areas of the transplants, the residual striata and the left lateral ventricles as obtained by MRI were high for both investigators, but correlation of cross-sectional areas of residual striata was moderate as measured by investigator 2. Interobserver correlations for averaged

cross-sectional areas of the transplants and the left lateral ventricles as revealed by MRI were high. No significant interobserver correlation was found for cross-sectional areas of the residual striata as measured by MRI. Intra- and interobserver correlation coefficients with levels of statistical significance are reported in Table 1. The average cross-sectional areas of the transplants, the residual striata and the left lateral ventricles as measured by MRI (averaged values of both observers’

214

SCHOCKE ET AL.

FIG. 4. (a– c) Present a sham graft. The floor of the left lateral ventricle is delineated by the residual striatum which shows a signal-intensity like the undamaged contralateral striatum. There is no visible transplant. W, window; C, center; R, right side; A, upper side; SP, slice position. Scale bar, 1 cm.

measurements) and histology were compared using paired t test (Table 2). In general, MRI measurements yielded greater graft and ventricular cross-sectional areas compared to histological morphometry. The mean cross-sectional areas of transplants, residual striata, and lateral ventricles averaged over all groups amounted to greater values in MRI than in histology. The differences between MRI and histology were significant (Table 2). Comparing MRI and histological measurements, a difference of 41% (SD 32) was received for transplants, of 28% (SD 19) for residual

striata, and of 24% (SD 9) for left lateral ventricles. In the analysis of differences between MRI and histology in the subgroups we did not find significance for grafts of the cotransplant group and for residual striata of the mesencephalic transplant group. Significance was almost reached for lateral ventricles of the cotransplant group. However, as shown in the Diagrams 1, 2, and 3, MRI and histological measurements for grafts (r ⫽ 0.92) and left lateral ventricle (r ⫽ 0.86) were strongly correlated; correlation of MRI and histological measurements for residual striatum was moderate (r ⫽ 0.78).

215

DETECTION OF STRIATAL TRANSPLANTS

TABLE 1 Intra- and Interobserver Correlation Coefficients (r) of MRI Evaluations (Pearson Correlation Coefficient) Intraobserver r Investigator 1

Investigator 2

Interobserver r

* 0.95 * 0.88 * 0.97

* 0.92 * 0.76 * 0.90

* 0.98 0.33 * 0.97

Transplant Residual striatum Left lateral ventricle * P ⬍ 0.001.

DISCUSSION Magnetic resonance imaging (MRI) appears to be a suitable tool in a wide range of basic science. The implementation of improved radio-frequency coils has increased the signal-to-noise ratio and the spatial resolution, respectively, so that high resolution MR studies in rodents have become possible (19). During the last decade small animal MRI of the CNS has been widely employed in a range of experimental paradigms including cerebral ischemia and trauma (20 –24). Most of the experiments were performed on small bore highfield MR systems. For experimental MRI studies, how-

ever, whole body MR scanner can also be used (25). Wolf et al. demonstrated the possibility of noninvasive imaging of living rats using a whole body 1.5 Tesla MR scanner and high resolution coils avoiding time-consuming and expensive adjustments of hard- or software (14). A further advantage of whole body MRI systems was shown by Schmiedl et al., who simultaneously investigated several rats using a special animal holder (26). Moreover, the additional application of fast imaging techniques such as turbo spin-echo dramatically shortens imaging time and provides detailed images (14). We used a whole-body 1.5 Tesla MRI scanner adopted with a home-made head holder to study embryonic neural grafts in a novel rat model of MSASND. Major modifications or implementation of special coils or sequences were not required. We employed a standard CP extremity coil as well as standard fast MRI sequences. The internal diameter of the used coil permits the application of a special home-made frame allowing the simultaneous examination of four rats. In contrast to previous studies we accurately visualized intrastriatal grafts in a time-saving manner with a chosen spatial resolution of 0.39 ⫻ 0.39 ⫻ 2 mm 3 (3,4,8 –11). Previous studies using low- and high-field MRI scanners attempted to visualize striatal transplants in vivo

TABLE 2 The Mean Cross-Sectional Areas of Transplant, Residual Striatal Tissue, and Left Lateral Ventricles Measured by Histology and MRI Transplant Groups

MRI [mm 2] (SD)

Histology [mm 2] (SD)

P value

Striatal transplant Cotransplant Mesencephalic transplant Total

4.5 (2.6) 4.2 (3.7) 2.1 (0.1) 3.6 (2.6)

2.6 (1.8) 3.5 (2.5) 0.6 (0.4) 2.2 (2.0)

0.019 N.S. 0.003 0.001

Residual striatum

Striatal transplant Cotransplant Mesencephalic transplant Sham Total

MRI [mm 2] (SD)

Histology [mm 2] (SD)

4.9 (0.8) 5.4 (0.7) 6.2 (0.5) 5.9 (0.4) 5.6 (0.8)

3.2 (1.0) 3.3 (1.5) 5.9 (1.3) 4.1 (0.3) 4.1 (1.5)

0.004 0.02 N.S. 0.003 ⬍0.001

Left lateral ventricle 2

Striatal transplant Cotransplant Mesencephalic transplant Sham Total

MRI [mm ] (SD)

Histology [mm 2] (SD)

12.6 (2.2) 13.4 (3.9) 12.9 (2.2) 15.8 (3.2) 13.7 (3.0)

8.5 (1.3) 10.8 (2.8) 10.4 (2.0) 11.5 (2.9) 10.3 (2.4)

0.007 0.08 * 0.01 0.005 ⬍0.001

Note. N.S., nonsignificant. SD, standard deviation. *Trend to significance. Differences between MRI and histological values were proved for significance by paired t tests.

216

SCHOCKE ET AL.

DIAGRAM 1. Regression plot showing the correlation of graft cross-sectional area defined by in vivo MRI (averaged data of two measurements of both investigators) versus histology. A Pearson correlation coefficient r of 0.92 (P ⬍ 0.001) indicates a high correlation between MRI and histological measurements for grafts.

(9 –11). Most of these studies were limited to image acquisition in one animal at a time. Some of these studies were performed on low-field MR scanners (9,10) and revealed a limited spatial resolution, so that it was not possible to identify the graft tissue directly. Generally, there is limited information about signal intensity and morphology of embryonic neuronal transplants in in vivo MRI. In a recent experimental MRI study embryonic striatal grafts implanted into a rat model of Huntington‘s disease were visualized using a

DIAGRAM 2. Regression plot showing correlation of residual striatal cross-sectional area defined by in vivo MRI (averaged data of two measurements of both investigators) versus histology. A Pearson correlation coefficient r of 0.78 (P ⬍ 0.001) indicates a moderate correlation between MRI and histological measurements for residual striatum.

DIAGRAM 3. Regression plot showing correlation of left ventricle cross-sectional area defined by in vivo MRI (averaged data of two measurements of both investigators) versus histology. A Pearson correlation coefficient r of 0.86 (P ⬍ 0.001) indicates a high correlation between MRI and histological measurement for left lateral ventricle.

flex coil and whole body MR scanner (11). The investigators scanned single rats and reported a MRI acquisition time in the order of about 46 min for native imaging and could demarcate graft tissue appearing hyperintense in T2 and hypointense in T1. Significant correlations between MRI and histology were reported. In the present study similar findings were observed. Pure striatal and mixed graft tissue appeared slightly hyperintense on T2-weighted and slightly hypointense on the TIR images in comparison to the contralateral undamaged striatum. Additionally, the signal of the grafts was inhomogeneous and comprised diffuse hypointensities in T2. The differences in signal intensity between graft and undamaged contralateral striatum might be explained by the histological morphology of striatal grafts which consist of P and NP zones. The P zones represent the striatal-like compartment of the graft and are characterized by an increase of acetylcholinesterase activity and DARPP-32 positivity, whereas the NP zones contain nonstriatal types of neurons, which develop from other brain regions such as the amygdala, globus pallidus and piriform cortex (27). We were not able to exactly differentiate the internal structure of the grafts with the chosen spatial resolution. The mesencephalic grafts exhibited a similar signal intensity in comparison to the contralateral striatum and were poorly defined, probably due to their minuteness and due to the lack of a hyperintense rim in T2 at the margin of transplant. On MRI, the grafts were usually identified at the base of the left lateral ventricle bulging into its cavity. This typical location helped detect surviving grafts. In contrast to previous studies we could identify and delineate the residual

217

DETECTION OF STRIATAL TRANSPLANTS

striatum appearing similar to undamaged contralateral striatum and extending between graft and cortex. Consistent with Guzman et al. (11) we observed a significant correlation of MRI and histological crosssectional areas for graft and lateral ventricle on the lesioned side. In addition, the MRI morphometric parameters of the residual striatum correlated moderately with histological measurements. We found a high intra- and interobserver correlation for the segmentation results of graft and lateral ventricle consistent with excellent demarcation from the surrounding brain tissue. The intraobserver correlation for residual striatum ranged from moderate to high, while there was no interobserver correlation. These findings might be explained by anatomically distorted ventral borders of the residual striatum. Therefore, dorsal parts of the nucleus accumbens may have been included in some analyses. In general, MRI measurements yielded significant larger cross-sectional areas of grafts and ventricle than histology except for the transplants of the cotransplant group and the residual striata of the mesencephalic transplant group. The different thickness of MRI and histological slices and the limited spatial resolution of the MRI images leading to a partial volume effect (28,29) may have contributed to the observed mismatch of MRI and histological measurements. Partial volume effects might have had a lower impact on well demarcated structures like the cotransplants. Histologically, the cotransplants exhibited the greatest extent resulting in a nonsignificant difference between MRI and histological values. In line with the literature we detected enlarged lateral ventricles on the lesioned side reflecting lesion-induced striatal atrophy and we were also able to identify unintended lesion spread in five rats (4 –7,10). Although the lateral ventricles were the largest structures that we have segmented in the present study, we obtained significantly higher values in MRI compared to histology. Besides the contribution of the partial volume effect the fixation procedure may explain the different crosssectional areas in MRI and histology. In fact, Guzman et al. detected significant differences between MRI and histological measurements which were mainly explained by fixation-induced shrinkage of brain tissue (11). It is well established that the extent of the crosssectional area of histological slices depends on the type and pH of the fixatives. The diminution of the area caused by the fixation procedure may amount up to 30% (30 –32). In conclusion, we demonstrated that fast MRI imaging using a conventional whole body scanner can generate high quality rat brain images in vivo, which allow detection of surviving embryonic grafts and correlation with histology. Our MRI protocol required approximately 45 min for the examination of four rats. It is therefore possible to visualize the presence of neural grafts in vivo in a large numbers of animals. In principle, animals with inappropriate placement of the le-

sion or grafts can be identified at any time point of the protocol. Whole-body MRI systems are widely used for routine diagnostic purposes, therefore experimental transplantation studies using in vivo structural imaging are not restricted to centers with experimental MRI facilities. In our opinion, the application of in vivo MRI to the monitoring of neural grafts should increasingly find access to future neural transplantation protocols and may lead to improved outcome by identification and exclusion of animals with inappropriate lesion placement or graft failure. ACKNOWLEDGMENTS This study was supported by the Austrian Science Foundation (P11748-MED) and the OENB Jubilaeumfonds (No. 8292).

REFERENCES 1. Lindvall, O. 1999. Cerebral implantation in movement disorders: State of the art. Mov. Disord. 14(2): 201–205. 2. Dunnett, S. B., and Bjo¨rklund, A. 1999. Prospects for new restorative and neuroprotective treatments in Parkinson’s disease. Nature 399(6738 Suppl.): A32–39. 3. Watts, C., and Dunnett, S. B. 2000. Towards a protocol for the preparation and delivery of striatal tissue for clinical trials of transplantation in Huntington’s disease. Cell Transplant. 9(2): 223–234. 4. Allegrini, P. R., and Sauer, D. 1992. Application of magnetic resonance imaging to the measurement of neurodegeneration in rat brain: MRI data correlate strongly with histology and enzymatic analysis. Magn. Reson. Imag. 10: 773–778. 5. Ben-Horin, N., Hazvi, S., Bendel, P., and Schul, R. 1996. The ontogeny of a neurotoxic lesion in rat brain revealed by combined MRI and histology. Brain Res. 718: 97–104. 6. Sauer, D., Allegrini, P. R., Thedinga, K. H., Massieu, L., Amacker, H., and Fagg, G. E. 1992. Evaluation of quinolinic acid induced excitotoxic neurodegeneration in rat striatum by quantitative magnetic resonance imaging in vivo. J. Neurosci. Methods 42: 69 –74. 7. Hantraye, P., Leroy-Willig, A., Denys, A., Riche, D., Isacson, O., Maziere, M., and Syrota, A. 1992. Magnetic resonance imaging to monitor pathology of caudate-putamen after excitotoxin-induced neuronal loss in the nonhuman primate brain. Exp. Neurol. 118: 18 –23. 8. Norman, A. B., Thomas, S. R., Pratt, R. G., Samaratunga, R. C., and Sanberg, P. R. 1989. Magnetic resonance imaging of rat brain following kainic acid-induced lesions and fetal striatal tissue transplants. Brain Res. 483: 188 –191. 9. Norman, A. B., Thomas, S. R., Pratt, R. G., Samaratunga, R. C., and Sanberg, P. R. 1989. A magnetic resonance imaging contrast agent differentiates between the vascular properties of fetal striatal tissue transplants and gliomas in rat brain in vivo. Brain Res. 503: 156 –159. 10. Norman, A. B., Thomas, S. R., Pratt, R. G., Samaratunga, R. C., and Sanberg, P. R. 1990. T1 and T2 weighted magnetic resonance imaging of excitotoxin lesions and neural transplants in rat brain in vivo. Exp. Neurol. 109: 164 –170. 11. Guzman, R., Meyer, M., Lo¨vblad, K. O., Ozdoba, C., Schroth, G., Seiler, R. W., and Widmer, H. R. 1999. Striatal grafts in a rat model of Huntington’s disease: Time course comparison of MRI and histology. Exp. Neurol. 156: 180 –190. 12. Marti-Bonamati, L., and Kormano, M. 1997. MR equipment acquisition strategies: Low-field or high-field scanners. Eur. Radiol. 7(Suppl. 5): 263–268.

218

SCHOCKE ET AL.

13. Petersein, J., and Saini, S. 1995. Fast MR imaging: technical strategies. Am. J. Roentgenol. 165(5): 1105–1106. 14. Wolf, R. F. E., Lam, K. H., Mooyaart, E. L., Bleichrodt, R. P., Nieuwenhuis, P., and Schakenraad, J. M. 1992. Magnetic resonance imaging using a clinical whole body system: An introduction to a useful technique in small animal experiments. Lab. Anim. 26: 222–227. 15. Wenning, G. K., Tison, F., Scherfler, C., Puschban, Z., Waldner, R., Granata, R., Ghorayed, I., and Poewe, W. 1999. Towards neurotransplantation in multiple system atrophy: Clinical rationale, pathophysiologic basis and preliminary experimental incidence. Cell Transplant. in press. 16. Wenning, G. K., Granata, R., Laboyrie, P. M., Quinn, N. P., Jenner, P., and Marsden, C. D. 1996. Reversal of behavioural abnormalities by fetal allografts in a novel rat model of striatonigral degeneration. Mov. Disord. 11(5): 522–532. 17. Puschban, Z., Scherfler, C., and Granata, R., et al. 1999. Autoradiographic study of striatal dopamine reuptake sites and dopamine D1 and D2 receptors in a 6-hydroxydopamine and quinolinic acid double-lesion rat model of striatonigral degeneration (multiple system atrophy) and effects of embryonic ventral mesencephalic, striatal or co-grafts. Neuroscience 95: 377–388. 18. Bjo¨rklund, A., Stenevi, U., Schmidt, R. H., Dunnett, S. B., and Gage, F. H. 1983. Intracerebral grafting of neuronal cell suspensions. I. Introduction and general methods of preparation. Acta Physiol. Scand. 522(Suppl): 9 –18. 19. Johnson, G. A., Benveniste, H., Black, R. D., Hedlund, L. W., Maronpot, R. R., and Smith, R. R. 1993. Histology by magnetic resonance microscopy. Magn. Reson. Q. 9(1): 1–30. 20. Baker, L. L., Kucharczyk, J., Sevick, R. J., Mintorovitch, J., and Moseley, M. E. 1991. Recent advances in MR imaging/spectroscopy of cerebral ischemia. Am. J. Roentgenol. 156(6): 1133–1143. 21. Moseley, M. E., Kucharczyk, J., Mintorovitch, et al. 1990. Diffusion-weighted MR imaging of acute stroke: Correlation with T2-weighted and magnetic susceptibility-enhanced MR imaging in cats. Am. J. Neuroradiol. 11: 423– 429. 22. Pierpaoli, C., Righini, A., Linfante, I., Tao-Cheng, J. H., Alger, J. R., and Di Chiro, G. 1993. Histopathologic correlates of abnormal water diffusion in cerebral ischemia: Diffusion-weighted MR

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

imaging and light and electron microscopic study. Radiology 189: 439 – 448. Ishige, N., Pitts, L. H., and Berry, I., et al. 1987. The effect of hypoxia on traumatic head injury in rats: Alterations in neurologic function, brain edema, and cerebral blood flow. J. Cereb. Blood Flow Metab. 7: 759 –767. Smith, D. H., Meaney, D. F., and Lenkinski, R. E., et al. 1995. New magnetic resonance imaging techniques for the evaluation of traumatic brain injury. J. Neurotrauma 12(4): 573–577. Cho, Z. H., Ahn, C. B., and Juh, S. C., et al. 1988. Nuclear magnetic resonance microscopy with 4-microns resolution: Theoretical study and experimental results. Med. Phys. 15(6): 815– 824. Schmiedl, U. P., Kenney, J., and Maravilla, K. R. 1992. Kinetics of pathologic blood– brain barrier permeability in an astrocytic glioma using contrast-enhanced MR. Am. J. Neuroradiol. 13: 5–14. Fricker, R. A., Torres, E. M., and Dunnett, S. B. 1997. The effect of donor stage on the survival and function of embryonic striatal grafts in the adult rat brain. I. Morphological characteristics. Neuroscience 79(3): 695–710. Pakkenberg, B., Boesen, J., Albeck, M., and Gjerris, F. 1989. Unbiased and efficient of total ventricular volume of the brain obtained from CT-scans by a stereological method. Neuroradiology 31(5): 413– 417. Luft, A. R., Skalej, M., Welte, D., Kolb, R., and Klose, U. 1996. Reliability and exactness of MRI-based volumetry: A phantom study. J. Magn. Reson. Imaging 6(4): 700 –704. Baak, J. P., Noteboom, E., and Koevoets, J. J. 1989. The influence of fixatives and other variations in tissue processing on nuclear morphometric features. Anal. Quant. Cytol. Histol. 11(4): 219 –224. Hillman, H., and Deutsch, K. 1978. Area changes in slices of rat brain during preparation for histology or electron microscopy. J. Microsc. 114(1): 77– 84. Doughty, M. J., Bergmanson, J. P., and Blocker, Y. 1997. Shrinkage and distortion of the rabiit corneal endothelial cell mosaic caused by a high osmolality glutaraldehyde-formaldehyde fixative compared to glutaraldehyde. Tissue Cell 29(5): 533–547.